an empirical analysis of the term auction facility

42
8/13/2019 An Empirical Analysis of the Term Auction Facility http://slidepdf.com/reader/full/an-empirical-analysis-of-the-term-auction-facility 1/42 1 An Empirical Analysis of the Fed’s Term Auction Facility Efraim Benmelech ABSTRACT The U.S. Federal Reserve used the Term Auction Facility to provide term funding to eligible depository institutions from December 2007 to March 2010. According to the Fed, the purpose of the TAF was to inject term funds through a broader range of counterparties and against a broader range of collateral than open market operations. The overall goal of the TAF was to ensure that liquidity provisions could be dis- seminated efficiently even when the unsecured interbank markets were under stress. In this paper I use the TAF micro-level loan data and find that about 60 percent of TAF loans went to foreign banks that pledged asset-backed securities as collateral for these loans. The data and analysis illustrate the major role that foreign—in particular, European—banks currently play in the U.S. financial system and the resultant cur- rency mismatch in their balance sheets. The data suggest that foreign banks had to  borrow from the Federal Reserve Bank to meet their dollar-denominated liabilities. Efraim Benmelech is an associate professor of finance in the Kellogg School of Management at Northwestern University and a research associate of the National Bureau of Economic Research. I thank Ian Dew-Becker, Charlie Calomiris, Simon Gilchrist, James Hamilton, Laura Jones Dooley, Ross Levine, Jeff Miron, and seminar participants at the Cato Papers on Public Policy Conference for useful comments. I acknowledge financial support from the Cato Institute. All errors are my own. Cato Papers on Public Policy  , Vol. 2 © 2012, Cato Institute

Upload: shchoksy

Post on 04-Jun-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 142

1

An Empirical Analysis of theFedrsquos Term Auction Facility

Efraim Benmelech

ABSTRACT

The US Federal Reserve used the Term Auction Facility to provide term fundingto eligible depository institutions from December 2007 to March 2010 According tothe Fed the purpose of the TAF was to inject term funds through a broader range ofcounterparties and against a broader range of collateral than open market operationsThe overall goal of the TAF was to ensure that liquidity provisions could be dis-seminated efficiently even when the unsecured interbank markets were under stressIn this paper I use the TAF micro-level loan data and find that about 60 percent ofTAF loans went to foreign banks that pledged asset-backed securities as collateral forthese loans The data and analysis illustrate the major role that foreignmdashin particularEuropeanmdashbanks currently play in the US financial system and the resultant cur-rency mismatch in their balance sheets The data suggest that foreign banks had to

borrow from the Federal Reserve Bank to meet their dollar-denominated liabilities

Efraim Benmelech is an associate professor of finance in the Kellogg School ofManagement at Northwestern University and a research associate of the NationalBureau of Economic Research

I thank Ian Dew-Becker Charlie Calomiris Simon Gilchrist James Hamilton Laura JonesDooley Ross Levine Jeff Miron and seminar participants at theCato Papers on Public Policy

Conference for useful comments I acknowledge financial support from the Cato InstituteAll errors are my own

Cato Papers on Public Policy Vol 2 copy 2012 Cato Institute

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 342

3

An Empirical Analysis of the Fedrsquos Term Auction Facility

3

An Empirical Analysis of theFedrsquos Term Auction Facility

1 INTRODUCTION

The Term Auction Facility program was one of the main toolsused by the Federal Reserve and US fiscal authorities during therecent financial crisis The goal of this program as described by theFederal Reserve was to intervene in the interbank money marketsin response to the difficulties experienced by banks in the UnitedStates and Europe Initially the Federal Reserve used open marketoperations to maintain the effective federal funds rate near its targetrate and enacted several measures to encourage borrowing at the dis-count window1 However these moves failed to stimulate the marketas much as the Fed had expected On December 12 2007 therefore

the Federal Reserve introduced the TAF The TAF provided longer-term financing to eligible depository institutions through auctions atpredetermined dates At its peak the TAF amounted to more than$500 billion and was the largest expansion on the Federal Reserversquos

balance sheet Lending through the TAF gradually faded away andthe final TAF auction was conducted on March 8 2010

One of the reasons for the introduction of the TAF during the earlystages of the financial crisis was to provide banks with Federal Reserve

liquidity without forcing them to face the stigma of borrowing fromthe discount window Indeed according to Federal Reserve BoardChairman Ben Bernanke the associated stigma made banks reluctantto use the discount window

In August 2007 banks were reluctant to rely on discountwindow credit to address their funding needs The banksrsquoconcern was that their recourse to the discount window ifit became known might lead market participants to inferweaknessmdashthe so-called stigma problem (Bernanke 2009)

1 To encourage banks to borrow at the discount window the Federal Reserve reducedthe discount window penalty rate from 100 basis points to 50 basis points on August 172007 and extended the term of financing from overnight to as long as 30 days

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 442

CATO PAPERS ON PUBLIC POLICY

4

However even borrowing from the TAF had a stigma attached toit As a result data on the loans that were made in the TAF as wellas the identity of the banks that participated in the auctions were notdisclosed initially Later the Federal Reserve disclosed data on theloans made under the TAF as well as information on the other creditand liquidity programs it used during the crisis

While the effectiveness of the TAF in reducing rates in the in-terbank market has been debated by both academic economistsand policymakers (see eg McAndrews Sarkar and Wang 2008and Taylor and Williams 2009) little is known about the identity

of the banks that participated in the auctions the nature of the col-lateral used or the terms on the individual loans This paper fillsin that gap by using the micro-level loan dataset released by theFederal Reserve The TAF data which contain detailed informationon the loans and the participating financial institutions provide arare glimpse into the injection of emergency liquidity by the FederalReserve as well as the identity of the banks obtaining credit and inparticular the type of assets they pledged as collateral

I found that foreign banks accounted for 58 percent of TAF lendingwith a total amount of $22 trillion compared to $16 trillion forUS banks During the auction of December 2007 and through mostof 2008 foreign banks accounted for the vast majority of the lendingwith amounts that ranged between twofold and fourfold the totallending to US banks United Kingdomndashbased Barclays was thelargest borrower in the TAF followed by Bank of America RoyalBank of Scotland Wells Fargo and Wachovia Out of the 10 largest

borrowers 5 are foreign banks and out of the 50 largest borrowers

more than 30 are from foreign countriesNext I compared the collateral structure of domestic and foreign

banks I found that most of the banks and financial institutions thatpledged asset-backed securities (ABSs) as collateral were foreignmdashprimarily Europeanmdashbanks For example the bank that pledgedthe largest amount of ABSs for a given loan is Socieacuteteacute Geacuteneacuterale(France) followed by Norinchukin Bank (Japan) Dexia (Belgium)Barclays (UK) and UBS (Switzerland) Among the 10 banks that

pledged the largest amounts of collateral only two are Ameri-can banks (State Street and US Central Federal Credit Union)Why did the Federal Reserve allocate the majority of TAF loansto foreign banks Why were foreign banks more likely to pledge

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 542

5

An Empirical Analysis of the Fedrsquos Term Auction Facility

the riskier ABSs and collateralized debt obligations (CDOs) ascollateral

One potential explanation is that the meltdown of the structuredfinance market and the severe deterioration in the credit ratings ofABSs necessitated liquidity injections to institutions that sufferedmajor losses because of their exposure to the structured financemarket However US banks that borrowed from the TAF and hadlarge exposures to ABSs such as Citibank and Bank of Americadid not pledge ABSs at the same level as European banks Thuswhile some of the Federal Reserve lending was probably aimed at

in jecting liquidity into financial institutions that held securities thatwere illiquid at the time this is unlikely to be the only reason for thedominance of European banks in the TAF

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch intheir balance sheets Many foreign banks were active players in thecreation and issuance of structured finance products As moneymarkets ground to a halt those banks required financing to rollover their short-term liabilities Furthermore foreign banks weresubject to a currency mismatch in their assets and liabilities Themain source of funding for some of the banks was demand depositsand other forms of credit in their home countries and these weredenominated in their home currencies (mostly the British poundand the euro) However many European banks issued liabilitiesin US money markets that were denominated in the US dollarThus not only were foreign banks subject to roll-over risk but theyalso suffered from a currency mismatch and had to rely on special

facilities such as the currency swap lines between central banks(eg the European Central Bank Bank of England Swiss NationalBank and Federal Reserve) as well as on special lending programssuch as the TAF European banks were more likely to bid for TAFmoney because they were more severely affected by the financialcrisis given their exposure to a currency mismatch between assetsand liabilities

The rest of this paper is organized as follows Section 2 provides

the institutional details of the TAF Section 3 describes the datasetand provides summary statistics on the evolution of the TAF overtime Section 4 displays the empirical analysis Section 5 discussesthe Federal Reserversquos lending to foreign banks Section 6 concludes

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 642

CATO PAPERS ON PUBLIC POLICY

6

2 THE TERM AUCTION FACILITY

Global money markets suffered serious disruptions in the summerof 2007 when the rates of interbank term loans rose to unusually highlevels2 The TED spreadmdashthe difference between the three-monthLondon interbank offered rate (LIBOR) and the three-month USTreasury billmdashrose from its typical level of 30 basis points to about50 basis points and then to 200 basis points by the summer of 2007This widening was a reason for major concern because the TED spreadis an indicator of perceived credit risk in the general economy More-over according to a New York Federal Reserve Bank research paper

[T]he volume of transactions in the inter-bank market de-clined and borrowers reportedly often could not obtainfunds at the posted rates Since the LIBOR aects interestrates on a wide variety of loans and securities (eg homemortgages and corporate loans) unusually high term ratescan have disruptive eects on the economy (McAndrewsSarkar and Wang 2008 1)

The Federal Reserve responded to the disruptions in the moneymarkets with the traditional tool of monetary policy open marketoperations to maintain the effective federal funds rate near its targetrate However despite the Federal Reserversquos efforts in the overnightfunding market the rates on term loans in the interbank market keptrising In an attempt to ease the strains in the money markets theFederal Reserve resorted to nontraditional tools of monetary policyPerhaps the most important tool used for this purpose was the TAF

The TAF was introduced on December 2007 in the early stages of

the financial crisis to provide Federal Reserve liquidity funding byauctioning off short-term funding without forcing banks to face thestigma of borrowing from the Federal Reserversquos discount windowUnder the TAF the Federal Reserve auctioned term funds to deposi-tory institutions All depository institutions that were eligible to bor-row under the primary credit program of the Federal Reserve wereeligible to participate in TAF auctions All loans extended under theTAF were fully collateralized The funds were allocated through an

auction in which participating depository institutions placed bidsspecifying an amount of funds up to a pre-specified limit and an

2 Term funding is typically made with maturity terms of one month or longer

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 742

7

An Empirical Analysis of the Fedrsquos Term Auction Facility

interest rate that they would be willing to pay for such funds Thefunds were allocated beginning with the highest interest rate offereduntil either all funds were allocated or all bids were satisfied All

borrowing institutions paid the same interest rate either the rateassociated with the bid that would fully subscribe the auction or inthe case that total bids were less than the amount of funds offeredthe lowest rate that was bid The TAF was created under the FederalReserversquos standard discount window lending authority grantedunder Section 10B of the Federal Reserve Act The auctions wereadministered by the Federal Reserve Bank of New York with loans

granted through the 12 Federal Reserve banksTAF funding supplemented the US dollar funding received by

global banks around the world under the central bank swap facilities between the Federal Reserve banks and the Banco Central do BrasilBank of Canada Denmarkrsquos Nationalbank Bank of England Euro-pean Central Bank Bank of Japan Bank of Korea Banco de MexicoReserve Bank of New Zealand Norges Bank Monetary Authority ofSingapore Sveriges Riksbank and Swiss National Bank

From the first TAF auction on December 17 2007 to the last onMarch 8 2010 the Federal Reserve conducted 60 auctions Theamount of term loans auctioned was initially between $20 billionand $30 billion but was later increased to between $50 billion and$75 billion The size increased to $150 billion in October 2008 andremained at that level until June 2009 During the second half of 2009and the first three months of 2010 the amount auctioned graduallydeclined and by the final auction in March 2010 only $34 billionwas loaned out

Whether the TAF was effective in reducing rates in the interbankmarket has been debated by both academic economists and policy-makers McAndrews Sarkar and Wang (2008) provide empirical evi-dence that the TAF has helped to ease strains in the interbank marketIn contrast according to Taylor and Williams (2009) the TAF had noimpact on interest rate spreads According to McAndrews Sarkarand Wang the major problem in the money markets in 2007ndash2008was lack of liquidity hence the TAF was effective because it provided

central bank liquidity to the banking system when the interbankingsystem collapsed In contrast Taylor and Williams (2009) argue thatthe main problem in the market was not liquidity but rather counter-party risk which TAF funding could not have solved

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 2: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 342

3

An Empirical Analysis of the Fedrsquos Term Auction Facility

3

An Empirical Analysis of theFedrsquos Term Auction Facility

1 INTRODUCTION

The Term Auction Facility program was one of the main toolsused by the Federal Reserve and US fiscal authorities during therecent financial crisis The goal of this program as described by theFederal Reserve was to intervene in the interbank money marketsin response to the difficulties experienced by banks in the UnitedStates and Europe Initially the Federal Reserve used open marketoperations to maintain the effective federal funds rate near its targetrate and enacted several measures to encourage borrowing at the dis-count window1 However these moves failed to stimulate the marketas much as the Fed had expected On December 12 2007 therefore

the Federal Reserve introduced the TAF The TAF provided longer-term financing to eligible depository institutions through auctions atpredetermined dates At its peak the TAF amounted to more than$500 billion and was the largest expansion on the Federal Reserversquos

balance sheet Lending through the TAF gradually faded away andthe final TAF auction was conducted on March 8 2010

One of the reasons for the introduction of the TAF during the earlystages of the financial crisis was to provide banks with Federal Reserve

liquidity without forcing them to face the stigma of borrowing fromthe discount window Indeed according to Federal Reserve BoardChairman Ben Bernanke the associated stigma made banks reluctantto use the discount window

In August 2007 banks were reluctant to rely on discountwindow credit to address their funding needs The banksrsquoconcern was that their recourse to the discount window ifit became known might lead market participants to inferweaknessmdashthe so-called stigma problem (Bernanke 2009)

1 To encourage banks to borrow at the discount window the Federal Reserve reducedthe discount window penalty rate from 100 basis points to 50 basis points on August 172007 and extended the term of financing from overnight to as long as 30 days

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 442

CATO PAPERS ON PUBLIC POLICY

4

However even borrowing from the TAF had a stigma attached toit As a result data on the loans that were made in the TAF as wellas the identity of the banks that participated in the auctions were notdisclosed initially Later the Federal Reserve disclosed data on theloans made under the TAF as well as information on the other creditand liquidity programs it used during the crisis

While the effectiveness of the TAF in reducing rates in the in-terbank market has been debated by both academic economistsand policymakers (see eg McAndrews Sarkar and Wang 2008and Taylor and Williams 2009) little is known about the identity

of the banks that participated in the auctions the nature of the col-lateral used or the terms on the individual loans This paper fillsin that gap by using the micro-level loan dataset released by theFederal Reserve The TAF data which contain detailed informationon the loans and the participating financial institutions provide arare glimpse into the injection of emergency liquidity by the FederalReserve as well as the identity of the banks obtaining credit and inparticular the type of assets they pledged as collateral

I found that foreign banks accounted for 58 percent of TAF lendingwith a total amount of $22 trillion compared to $16 trillion forUS banks During the auction of December 2007 and through mostof 2008 foreign banks accounted for the vast majority of the lendingwith amounts that ranged between twofold and fourfold the totallending to US banks United Kingdomndashbased Barclays was thelargest borrower in the TAF followed by Bank of America RoyalBank of Scotland Wells Fargo and Wachovia Out of the 10 largest

borrowers 5 are foreign banks and out of the 50 largest borrowers

more than 30 are from foreign countriesNext I compared the collateral structure of domestic and foreign

banks I found that most of the banks and financial institutions thatpledged asset-backed securities (ABSs) as collateral were foreignmdashprimarily Europeanmdashbanks For example the bank that pledgedthe largest amount of ABSs for a given loan is Socieacuteteacute Geacuteneacuterale(France) followed by Norinchukin Bank (Japan) Dexia (Belgium)Barclays (UK) and UBS (Switzerland) Among the 10 banks that

pledged the largest amounts of collateral only two are Ameri-can banks (State Street and US Central Federal Credit Union)Why did the Federal Reserve allocate the majority of TAF loansto foreign banks Why were foreign banks more likely to pledge

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 542

5

An Empirical Analysis of the Fedrsquos Term Auction Facility

the riskier ABSs and collateralized debt obligations (CDOs) ascollateral

One potential explanation is that the meltdown of the structuredfinance market and the severe deterioration in the credit ratings ofABSs necessitated liquidity injections to institutions that sufferedmajor losses because of their exposure to the structured financemarket However US banks that borrowed from the TAF and hadlarge exposures to ABSs such as Citibank and Bank of Americadid not pledge ABSs at the same level as European banks Thuswhile some of the Federal Reserve lending was probably aimed at

in jecting liquidity into financial institutions that held securities thatwere illiquid at the time this is unlikely to be the only reason for thedominance of European banks in the TAF

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch intheir balance sheets Many foreign banks were active players in thecreation and issuance of structured finance products As moneymarkets ground to a halt those banks required financing to rollover their short-term liabilities Furthermore foreign banks weresubject to a currency mismatch in their assets and liabilities Themain source of funding for some of the banks was demand depositsand other forms of credit in their home countries and these weredenominated in their home currencies (mostly the British poundand the euro) However many European banks issued liabilitiesin US money markets that were denominated in the US dollarThus not only were foreign banks subject to roll-over risk but theyalso suffered from a currency mismatch and had to rely on special

facilities such as the currency swap lines between central banks(eg the European Central Bank Bank of England Swiss NationalBank and Federal Reserve) as well as on special lending programssuch as the TAF European banks were more likely to bid for TAFmoney because they were more severely affected by the financialcrisis given their exposure to a currency mismatch between assetsand liabilities

The rest of this paper is organized as follows Section 2 provides

the institutional details of the TAF Section 3 describes the datasetand provides summary statistics on the evolution of the TAF overtime Section 4 displays the empirical analysis Section 5 discussesthe Federal Reserversquos lending to foreign banks Section 6 concludes

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 642

CATO PAPERS ON PUBLIC POLICY

6

2 THE TERM AUCTION FACILITY

Global money markets suffered serious disruptions in the summerof 2007 when the rates of interbank term loans rose to unusually highlevels2 The TED spreadmdashthe difference between the three-monthLondon interbank offered rate (LIBOR) and the three-month USTreasury billmdashrose from its typical level of 30 basis points to about50 basis points and then to 200 basis points by the summer of 2007This widening was a reason for major concern because the TED spreadis an indicator of perceived credit risk in the general economy More-over according to a New York Federal Reserve Bank research paper

[T]he volume of transactions in the inter-bank market de-clined and borrowers reportedly often could not obtainfunds at the posted rates Since the LIBOR aects interestrates on a wide variety of loans and securities (eg homemortgages and corporate loans) unusually high term ratescan have disruptive eects on the economy (McAndrewsSarkar and Wang 2008 1)

The Federal Reserve responded to the disruptions in the moneymarkets with the traditional tool of monetary policy open marketoperations to maintain the effective federal funds rate near its targetrate However despite the Federal Reserversquos efforts in the overnightfunding market the rates on term loans in the interbank market keptrising In an attempt to ease the strains in the money markets theFederal Reserve resorted to nontraditional tools of monetary policyPerhaps the most important tool used for this purpose was the TAF

The TAF was introduced on December 2007 in the early stages of

the financial crisis to provide Federal Reserve liquidity funding byauctioning off short-term funding without forcing banks to face thestigma of borrowing from the Federal Reserversquos discount windowUnder the TAF the Federal Reserve auctioned term funds to deposi-tory institutions All depository institutions that were eligible to bor-row under the primary credit program of the Federal Reserve wereeligible to participate in TAF auctions All loans extended under theTAF were fully collateralized The funds were allocated through an

auction in which participating depository institutions placed bidsspecifying an amount of funds up to a pre-specified limit and an

2 Term funding is typically made with maturity terms of one month or longer

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 742

7

An Empirical Analysis of the Fedrsquos Term Auction Facility

interest rate that they would be willing to pay for such funds Thefunds were allocated beginning with the highest interest rate offereduntil either all funds were allocated or all bids were satisfied All

borrowing institutions paid the same interest rate either the rateassociated with the bid that would fully subscribe the auction or inthe case that total bids were less than the amount of funds offeredthe lowest rate that was bid The TAF was created under the FederalReserversquos standard discount window lending authority grantedunder Section 10B of the Federal Reserve Act The auctions wereadministered by the Federal Reserve Bank of New York with loans

granted through the 12 Federal Reserve banksTAF funding supplemented the US dollar funding received by

global banks around the world under the central bank swap facilities between the Federal Reserve banks and the Banco Central do BrasilBank of Canada Denmarkrsquos Nationalbank Bank of England Euro-pean Central Bank Bank of Japan Bank of Korea Banco de MexicoReserve Bank of New Zealand Norges Bank Monetary Authority ofSingapore Sveriges Riksbank and Swiss National Bank

From the first TAF auction on December 17 2007 to the last onMarch 8 2010 the Federal Reserve conducted 60 auctions Theamount of term loans auctioned was initially between $20 billionand $30 billion but was later increased to between $50 billion and$75 billion The size increased to $150 billion in October 2008 andremained at that level until June 2009 During the second half of 2009and the first three months of 2010 the amount auctioned graduallydeclined and by the final auction in March 2010 only $34 billionwas loaned out

Whether the TAF was effective in reducing rates in the interbankmarket has been debated by both academic economists and policy-makers McAndrews Sarkar and Wang (2008) provide empirical evi-dence that the TAF has helped to ease strains in the interbank marketIn contrast according to Taylor and Williams (2009) the TAF had noimpact on interest rate spreads According to McAndrews Sarkarand Wang the major problem in the money markets in 2007ndash2008was lack of liquidity hence the TAF was effective because it provided

central bank liquidity to the banking system when the interbankingsystem collapsed In contrast Taylor and Williams (2009) argue thatthe main problem in the market was not liquidity but rather counter-party risk which TAF funding could not have solved

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 3: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 342

3

An Empirical Analysis of the Fedrsquos Term Auction Facility

3

An Empirical Analysis of theFedrsquos Term Auction Facility

1 INTRODUCTION

The Term Auction Facility program was one of the main toolsused by the Federal Reserve and US fiscal authorities during therecent financial crisis The goal of this program as described by theFederal Reserve was to intervene in the interbank money marketsin response to the difficulties experienced by banks in the UnitedStates and Europe Initially the Federal Reserve used open marketoperations to maintain the effective federal funds rate near its targetrate and enacted several measures to encourage borrowing at the dis-count window1 However these moves failed to stimulate the marketas much as the Fed had expected On December 12 2007 therefore

the Federal Reserve introduced the TAF The TAF provided longer-term financing to eligible depository institutions through auctions atpredetermined dates At its peak the TAF amounted to more than$500 billion and was the largest expansion on the Federal Reserversquos

balance sheet Lending through the TAF gradually faded away andthe final TAF auction was conducted on March 8 2010

One of the reasons for the introduction of the TAF during the earlystages of the financial crisis was to provide banks with Federal Reserve

liquidity without forcing them to face the stigma of borrowing fromthe discount window Indeed according to Federal Reserve BoardChairman Ben Bernanke the associated stigma made banks reluctantto use the discount window

In August 2007 banks were reluctant to rely on discountwindow credit to address their funding needs The banksrsquoconcern was that their recourse to the discount window ifit became known might lead market participants to inferweaknessmdashthe so-called stigma problem (Bernanke 2009)

1 To encourage banks to borrow at the discount window the Federal Reserve reducedthe discount window penalty rate from 100 basis points to 50 basis points on August 172007 and extended the term of financing from overnight to as long as 30 days

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 442

CATO PAPERS ON PUBLIC POLICY

4

However even borrowing from the TAF had a stigma attached toit As a result data on the loans that were made in the TAF as wellas the identity of the banks that participated in the auctions were notdisclosed initially Later the Federal Reserve disclosed data on theloans made under the TAF as well as information on the other creditand liquidity programs it used during the crisis

While the effectiveness of the TAF in reducing rates in the in-terbank market has been debated by both academic economistsand policymakers (see eg McAndrews Sarkar and Wang 2008and Taylor and Williams 2009) little is known about the identity

of the banks that participated in the auctions the nature of the col-lateral used or the terms on the individual loans This paper fillsin that gap by using the micro-level loan dataset released by theFederal Reserve The TAF data which contain detailed informationon the loans and the participating financial institutions provide arare glimpse into the injection of emergency liquidity by the FederalReserve as well as the identity of the banks obtaining credit and inparticular the type of assets they pledged as collateral

I found that foreign banks accounted for 58 percent of TAF lendingwith a total amount of $22 trillion compared to $16 trillion forUS banks During the auction of December 2007 and through mostof 2008 foreign banks accounted for the vast majority of the lendingwith amounts that ranged between twofold and fourfold the totallending to US banks United Kingdomndashbased Barclays was thelargest borrower in the TAF followed by Bank of America RoyalBank of Scotland Wells Fargo and Wachovia Out of the 10 largest

borrowers 5 are foreign banks and out of the 50 largest borrowers

more than 30 are from foreign countriesNext I compared the collateral structure of domestic and foreign

banks I found that most of the banks and financial institutions thatpledged asset-backed securities (ABSs) as collateral were foreignmdashprimarily Europeanmdashbanks For example the bank that pledgedthe largest amount of ABSs for a given loan is Socieacuteteacute Geacuteneacuterale(France) followed by Norinchukin Bank (Japan) Dexia (Belgium)Barclays (UK) and UBS (Switzerland) Among the 10 banks that

pledged the largest amounts of collateral only two are Ameri-can banks (State Street and US Central Federal Credit Union)Why did the Federal Reserve allocate the majority of TAF loansto foreign banks Why were foreign banks more likely to pledge

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 542

5

An Empirical Analysis of the Fedrsquos Term Auction Facility

the riskier ABSs and collateralized debt obligations (CDOs) ascollateral

One potential explanation is that the meltdown of the structuredfinance market and the severe deterioration in the credit ratings ofABSs necessitated liquidity injections to institutions that sufferedmajor losses because of their exposure to the structured financemarket However US banks that borrowed from the TAF and hadlarge exposures to ABSs such as Citibank and Bank of Americadid not pledge ABSs at the same level as European banks Thuswhile some of the Federal Reserve lending was probably aimed at

in jecting liquidity into financial institutions that held securities thatwere illiquid at the time this is unlikely to be the only reason for thedominance of European banks in the TAF

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch intheir balance sheets Many foreign banks were active players in thecreation and issuance of structured finance products As moneymarkets ground to a halt those banks required financing to rollover their short-term liabilities Furthermore foreign banks weresubject to a currency mismatch in their assets and liabilities Themain source of funding for some of the banks was demand depositsand other forms of credit in their home countries and these weredenominated in their home currencies (mostly the British poundand the euro) However many European banks issued liabilitiesin US money markets that were denominated in the US dollarThus not only were foreign banks subject to roll-over risk but theyalso suffered from a currency mismatch and had to rely on special

facilities such as the currency swap lines between central banks(eg the European Central Bank Bank of England Swiss NationalBank and Federal Reserve) as well as on special lending programssuch as the TAF European banks were more likely to bid for TAFmoney because they were more severely affected by the financialcrisis given their exposure to a currency mismatch between assetsand liabilities

The rest of this paper is organized as follows Section 2 provides

the institutional details of the TAF Section 3 describes the datasetand provides summary statistics on the evolution of the TAF overtime Section 4 displays the empirical analysis Section 5 discussesthe Federal Reserversquos lending to foreign banks Section 6 concludes

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 642

CATO PAPERS ON PUBLIC POLICY

6

2 THE TERM AUCTION FACILITY

Global money markets suffered serious disruptions in the summerof 2007 when the rates of interbank term loans rose to unusually highlevels2 The TED spreadmdashthe difference between the three-monthLondon interbank offered rate (LIBOR) and the three-month USTreasury billmdashrose from its typical level of 30 basis points to about50 basis points and then to 200 basis points by the summer of 2007This widening was a reason for major concern because the TED spreadis an indicator of perceived credit risk in the general economy More-over according to a New York Federal Reserve Bank research paper

[T]he volume of transactions in the inter-bank market de-clined and borrowers reportedly often could not obtainfunds at the posted rates Since the LIBOR aects interestrates on a wide variety of loans and securities (eg homemortgages and corporate loans) unusually high term ratescan have disruptive eects on the economy (McAndrewsSarkar and Wang 2008 1)

The Federal Reserve responded to the disruptions in the moneymarkets with the traditional tool of monetary policy open marketoperations to maintain the effective federal funds rate near its targetrate However despite the Federal Reserversquos efforts in the overnightfunding market the rates on term loans in the interbank market keptrising In an attempt to ease the strains in the money markets theFederal Reserve resorted to nontraditional tools of monetary policyPerhaps the most important tool used for this purpose was the TAF

The TAF was introduced on December 2007 in the early stages of

the financial crisis to provide Federal Reserve liquidity funding byauctioning off short-term funding without forcing banks to face thestigma of borrowing from the Federal Reserversquos discount windowUnder the TAF the Federal Reserve auctioned term funds to deposi-tory institutions All depository institutions that were eligible to bor-row under the primary credit program of the Federal Reserve wereeligible to participate in TAF auctions All loans extended under theTAF were fully collateralized The funds were allocated through an

auction in which participating depository institutions placed bidsspecifying an amount of funds up to a pre-specified limit and an

2 Term funding is typically made with maturity terms of one month or longer

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 742

7

An Empirical Analysis of the Fedrsquos Term Auction Facility

interest rate that they would be willing to pay for such funds Thefunds were allocated beginning with the highest interest rate offereduntil either all funds were allocated or all bids were satisfied All

borrowing institutions paid the same interest rate either the rateassociated with the bid that would fully subscribe the auction or inthe case that total bids were less than the amount of funds offeredthe lowest rate that was bid The TAF was created under the FederalReserversquos standard discount window lending authority grantedunder Section 10B of the Federal Reserve Act The auctions wereadministered by the Federal Reserve Bank of New York with loans

granted through the 12 Federal Reserve banksTAF funding supplemented the US dollar funding received by

global banks around the world under the central bank swap facilities between the Federal Reserve banks and the Banco Central do BrasilBank of Canada Denmarkrsquos Nationalbank Bank of England Euro-pean Central Bank Bank of Japan Bank of Korea Banco de MexicoReserve Bank of New Zealand Norges Bank Monetary Authority ofSingapore Sveriges Riksbank and Swiss National Bank

From the first TAF auction on December 17 2007 to the last onMarch 8 2010 the Federal Reserve conducted 60 auctions Theamount of term loans auctioned was initially between $20 billionand $30 billion but was later increased to between $50 billion and$75 billion The size increased to $150 billion in October 2008 andremained at that level until June 2009 During the second half of 2009and the first three months of 2010 the amount auctioned graduallydeclined and by the final auction in March 2010 only $34 billionwas loaned out

Whether the TAF was effective in reducing rates in the interbankmarket has been debated by both academic economists and policy-makers McAndrews Sarkar and Wang (2008) provide empirical evi-dence that the TAF has helped to ease strains in the interbank marketIn contrast according to Taylor and Williams (2009) the TAF had noimpact on interest rate spreads According to McAndrews Sarkarand Wang the major problem in the money markets in 2007ndash2008was lack of liquidity hence the TAF was effective because it provided

central bank liquidity to the banking system when the interbankingsystem collapsed In contrast Taylor and Williams (2009) argue thatthe main problem in the market was not liquidity but rather counter-party risk which TAF funding could not have solved

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 4: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 442

CATO PAPERS ON PUBLIC POLICY

4

However even borrowing from the TAF had a stigma attached toit As a result data on the loans that were made in the TAF as wellas the identity of the banks that participated in the auctions were notdisclosed initially Later the Federal Reserve disclosed data on theloans made under the TAF as well as information on the other creditand liquidity programs it used during the crisis

While the effectiveness of the TAF in reducing rates in the in-terbank market has been debated by both academic economistsand policymakers (see eg McAndrews Sarkar and Wang 2008and Taylor and Williams 2009) little is known about the identity

of the banks that participated in the auctions the nature of the col-lateral used or the terms on the individual loans This paper fillsin that gap by using the micro-level loan dataset released by theFederal Reserve The TAF data which contain detailed informationon the loans and the participating financial institutions provide arare glimpse into the injection of emergency liquidity by the FederalReserve as well as the identity of the banks obtaining credit and inparticular the type of assets they pledged as collateral

I found that foreign banks accounted for 58 percent of TAF lendingwith a total amount of $22 trillion compared to $16 trillion forUS banks During the auction of December 2007 and through mostof 2008 foreign banks accounted for the vast majority of the lendingwith amounts that ranged between twofold and fourfold the totallending to US banks United Kingdomndashbased Barclays was thelargest borrower in the TAF followed by Bank of America RoyalBank of Scotland Wells Fargo and Wachovia Out of the 10 largest

borrowers 5 are foreign banks and out of the 50 largest borrowers

more than 30 are from foreign countriesNext I compared the collateral structure of domestic and foreign

banks I found that most of the banks and financial institutions thatpledged asset-backed securities (ABSs) as collateral were foreignmdashprimarily Europeanmdashbanks For example the bank that pledgedthe largest amount of ABSs for a given loan is Socieacuteteacute Geacuteneacuterale(France) followed by Norinchukin Bank (Japan) Dexia (Belgium)Barclays (UK) and UBS (Switzerland) Among the 10 banks that

pledged the largest amounts of collateral only two are Ameri-can banks (State Street and US Central Federal Credit Union)Why did the Federal Reserve allocate the majority of TAF loansto foreign banks Why were foreign banks more likely to pledge

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 542

5

An Empirical Analysis of the Fedrsquos Term Auction Facility

the riskier ABSs and collateralized debt obligations (CDOs) ascollateral

One potential explanation is that the meltdown of the structuredfinance market and the severe deterioration in the credit ratings ofABSs necessitated liquidity injections to institutions that sufferedmajor losses because of their exposure to the structured financemarket However US banks that borrowed from the TAF and hadlarge exposures to ABSs such as Citibank and Bank of Americadid not pledge ABSs at the same level as European banks Thuswhile some of the Federal Reserve lending was probably aimed at

in jecting liquidity into financial institutions that held securities thatwere illiquid at the time this is unlikely to be the only reason for thedominance of European banks in the TAF

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch intheir balance sheets Many foreign banks were active players in thecreation and issuance of structured finance products As moneymarkets ground to a halt those banks required financing to rollover their short-term liabilities Furthermore foreign banks weresubject to a currency mismatch in their assets and liabilities Themain source of funding for some of the banks was demand depositsand other forms of credit in their home countries and these weredenominated in their home currencies (mostly the British poundand the euro) However many European banks issued liabilitiesin US money markets that were denominated in the US dollarThus not only were foreign banks subject to roll-over risk but theyalso suffered from a currency mismatch and had to rely on special

facilities such as the currency swap lines between central banks(eg the European Central Bank Bank of England Swiss NationalBank and Federal Reserve) as well as on special lending programssuch as the TAF European banks were more likely to bid for TAFmoney because they were more severely affected by the financialcrisis given their exposure to a currency mismatch between assetsand liabilities

The rest of this paper is organized as follows Section 2 provides

the institutional details of the TAF Section 3 describes the datasetand provides summary statistics on the evolution of the TAF overtime Section 4 displays the empirical analysis Section 5 discussesthe Federal Reserversquos lending to foreign banks Section 6 concludes

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 642

CATO PAPERS ON PUBLIC POLICY

6

2 THE TERM AUCTION FACILITY

Global money markets suffered serious disruptions in the summerof 2007 when the rates of interbank term loans rose to unusually highlevels2 The TED spreadmdashthe difference between the three-monthLondon interbank offered rate (LIBOR) and the three-month USTreasury billmdashrose from its typical level of 30 basis points to about50 basis points and then to 200 basis points by the summer of 2007This widening was a reason for major concern because the TED spreadis an indicator of perceived credit risk in the general economy More-over according to a New York Federal Reserve Bank research paper

[T]he volume of transactions in the inter-bank market de-clined and borrowers reportedly often could not obtainfunds at the posted rates Since the LIBOR aects interestrates on a wide variety of loans and securities (eg homemortgages and corporate loans) unusually high term ratescan have disruptive eects on the economy (McAndrewsSarkar and Wang 2008 1)

The Federal Reserve responded to the disruptions in the moneymarkets with the traditional tool of monetary policy open marketoperations to maintain the effective federal funds rate near its targetrate However despite the Federal Reserversquos efforts in the overnightfunding market the rates on term loans in the interbank market keptrising In an attempt to ease the strains in the money markets theFederal Reserve resorted to nontraditional tools of monetary policyPerhaps the most important tool used for this purpose was the TAF

The TAF was introduced on December 2007 in the early stages of

the financial crisis to provide Federal Reserve liquidity funding byauctioning off short-term funding without forcing banks to face thestigma of borrowing from the Federal Reserversquos discount windowUnder the TAF the Federal Reserve auctioned term funds to deposi-tory institutions All depository institutions that were eligible to bor-row under the primary credit program of the Federal Reserve wereeligible to participate in TAF auctions All loans extended under theTAF were fully collateralized The funds were allocated through an

auction in which participating depository institutions placed bidsspecifying an amount of funds up to a pre-specified limit and an

2 Term funding is typically made with maturity terms of one month or longer

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 742

7

An Empirical Analysis of the Fedrsquos Term Auction Facility

interest rate that they would be willing to pay for such funds Thefunds were allocated beginning with the highest interest rate offereduntil either all funds were allocated or all bids were satisfied All

borrowing institutions paid the same interest rate either the rateassociated with the bid that would fully subscribe the auction or inthe case that total bids were less than the amount of funds offeredthe lowest rate that was bid The TAF was created under the FederalReserversquos standard discount window lending authority grantedunder Section 10B of the Federal Reserve Act The auctions wereadministered by the Federal Reserve Bank of New York with loans

granted through the 12 Federal Reserve banksTAF funding supplemented the US dollar funding received by

global banks around the world under the central bank swap facilities between the Federal Reserve banks and the Banco Central do BrasilBank of Canada Denmarkrsquos Nationalbank Bank of England Euro-pean Central Bank Bank of Japan Bank of Korea Banco de MexicoReserve Bank of New Zealand Norges Bank Monetary Authority ofSingapore Sveriges Riksbank and Swiss National Bank

From the first TAF auction on December 17 2007 to the last onMarch 8 2010 the Federal Reserve conducted 60 auctions Theamount of term loans auctioned was initially between $20 billionand $30 billion but was later increased to between $50 billion and$75 billion The size increased to $150 billion in October 2008 andremained at that level until June 2009 During the second half of 2009and the first three months of 2010 the amount auctioned graduallydeclined and by the final auction in March 2010 only $34 billionwas loaned out

Whether the TAF was effective in reducing rates in the interbankmarket has been debated by both academic economists and policy-makers McAndrews Sarkar and Wang (2008) provide empirical evi-dence that the TAF has helped to ease strains in the interbank marketIn contrast according to Taylor and Williams (2009) the TAF had noimpact on interest rate spreads According to McAndrews Sarkarand Wang the major problem in the money markets in 2007ndash2008was lack of liquidity hence the TAF was effective because it provided

central bank liquidity to the banking system when the interbankingsystem collapsed In contrast Taylor and Williams (2009) argue thatthe main problem in the market was not liquidity but rather counter-party risk which TAF funding could not have solved

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 5: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 542

5

An Empirical Analysis of the Fedrsquos Term Auction Facility

the riskier ABSs and collateralized debt obligations (CDOs) ascollateral

One potential explanation is that the meltdown of the structuredfinance market and the severe deterioration in the credit ratings ofABSs necessitated liquidity injections to institutions that sufferedmajor losses because of their exposure to the structured financemarket However US banks that borrowed from the TAF and hadlarge exposures to ABSs such as Citibank and Bank of Americadid not pledge ABSs at the same level as European banks Thuswhile some of the Federal Reserve lending was probably aimed at

in jecting liquidity into financial institutions that held securities thatwere illiquid at the time this is unlikely to be the only reason for thedominance of European banks in the TAF

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch intheir balance sheets Many foreign banks were active players in thecreation and issuance of structured finance products As moneymarkets ground to a halt those banks required financing to rollover their short-term liabilities Furthermore foreign banks weresubject to a currency mismatch in their assets and liabilities Themain source of funding for some of the banks was demand depositsand other forms of credit in their home countries and these weredenominated in their home currencies (mostly the British poundand the euro) However many European banks issued liabilitiesin US money markets that were denominated in the US dollarThus not only were foreign banks subject to roll-over risk but theyalso suffered from a currency mismatch and had to rely on special

facilities such as the currency swap lines between central banks(eg the European Central Bank Bank of England Swiss NationalBank and Federal Reserve) as well as on special lending programssuch as the TAF European banks were more likely to bid for TAFmoney because they were more severely affected by the financialcrisis given their exposure to a currency mismatch between assetsand liabilities

The rest of this paper is organized as follows Section 2 provides

the institutional details of the TAF Section 3 describes the datasetand provides summary statistics on the evolution of the TAF overtime Section 4 displays the empirical analysis Section 5 discussesthe Federal Reserversquos lending to foreign banks Section 6 concludes

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 642

CATO PAPERS ON PUBLIC POLICY

6

2 THE TERM AUCTION FACILITY

Global money markets suffered serious disruptions in the summerof 2007 when the rates of interbank term loans rose to unusually highlevels2 The TED spreadmdashthe difference between the three-monthLondon interbank offered rate (LIBOR) and the three-month USTreasury billmdashrose from its typical level of 30 basis points to about50 basis points and then to 200 basis points by the summer of 2007This widening was a reason for major concern because the TED spreadis an indicator of perceived credit risk in the general economy More-over according to a New York Federal Reserve Bank research paper

[T]he volume of transactions in the inter-bank market de-clined and borrowers reportedly often could not obtainfunds at the posted rates Since the LIBOR aects interestrates on a wide variety of loans and securities (eg homemortgages and corporate loans) unusually high term ratescan have disruptive eects on the economy (McAndrewsSarkar and Wang 2008 1)

The Federal Reserve responded to the disruptions in the moneymarkets with the traditional tool of monetary policy open marketoperations to maintain the effective federal funds rate near its targetrate However despite the Federal Reserversquos efforts in the overnightfunding market the rates on term loans in the interbank market keptrising In an attempt to ease the strains in the money markets theFederal Reserve resorted to nontraditional tools of monetary policyPerhaps the most important tool used for this purpose was the TAF

The TAF was introduced on December 2007 in the early stages of

the financial crisis to provide Federal Reserve liquidity funding byauctioning off short-term funding without forcing banks to face thestigma of borrowing from the Federal Reserversquos discount windowUnder the TAF the Federal Reserve auctioned term funds to deposi-tory institutions All depository institutions that were eligible to bor-row under the primary credit program of the Federal Reserve wereeligible to participate in TAF auctions All loans extended under theTAF were fully collateralized The funds were allocated through an

auction in which participating depository institutions placed bidsspecifying an amount of funds up to a pre-specified limit and an

2 Term funding is typically made with maturity terms of one month or longer

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 742

7

An Empirical Analysis of the Fedrsquos Term Auction Facility

interest rate that they would be willing to pay for such funds Thefunds were allocated beginning with the highest interest rate offereduntil either all funds were allocated or all bids were satisfied All

borrowing institutions paid the same interest rate either the rateassociated with the bid that would fully subscribe the auction or inthe case that total bids were less than the amount of funds offeredthe lowest rate that was bid The TAF was created under the FederalReserversquos standard discount window lending authority grantedunder Section 10B of the Federal Reserve Act The auctions wereadministered by the Federal Reserve Bank of New York with loans

granted through the 12 Federal Reserve banksTAF funding supplemented the US dollar funding received by

global banks around the world under the central bank swap facilities between the Federal Reserve banks and the Banco Central do BrasilBank of Canada Denmarkrsquos Nationalbank Bank of England Euro-pean Central Bank Bank of Japan Bank of Korea Banco de MexicoReserve Bank of New Zealand Norges Bank Monetary Authority ofSingapore Sveriges Riksbank and Swiss National Bank

From the first TAF auction on December 17 2007 to the last onMarch 8 2010 the Federal Reserve conducted 60 auctions Theamount of term loans auctioned was initially between $20 billionand $30 billion but was later increased to between $50 billion and$75 billion The size increased to $150 billion in October 2008 andremained at that level until June 2009 During the second half of 2009and the first three months of 2010 the amount auctioned graduallydeclined and by the final auction in March 2010 only $34 billionwas loaned out

Whether the TAF was effective in reducing rates in the interbankmarket has been debated by both academic economists and policy-makers McAndrews Sarkar and Wang (2008) provide empirical evi-dence that the TAF has helped to ease strains in the interbank marketIn contrast according to Taylor and Williams (2009) the TAF had noimpact on interest rate spreads According to McAndrews Sarkarand Wang the major problem in the money markets in 2007ndash2008was lack of liquidity hence the TAF was effective because it provided

central bank liquidity to the banking system when the interbankingsystem collapsed In contrast Taylor and Williams (2009) argue thatthe main problem in the market was not liquidity but rather counter-party risk which TAF funding could not have solved

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 6: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 642

CATO PAPERS ON PUBLIC POLICY

6

2 THE TERM AUCTION FACILITY

Global money markets suffered serious disruptions in the summerof 2007 when the rates of interbank term loans rose to unusually highlevels2 The TED spreadmdashthe difference between the three-monthLondon interbank offered rate (LIBOR) and the three-month USTreasury billmdashrose from its typical level of 30 basis points to about50 basis points and then to 200 basis points by the summer of 2007This widening was a reason for major concern because the TED spreadis an indicator of perceived credit risk in the general economy More-over according to a New York Federal Reserve Bank research paper

[T]he volume of transactions in the inter-bank market de-clined and borrowers reportedly often could not obtainfunds at the posted rates Since the LIBOR aects interestrates on a wide variety of loans and securities (eg homemortgages and corporate loans) unusually high term ratescan have disruptive eects on the economy (McAndrewsSarkar and Wang 2008 1)

The Federal Reserve responded to the disruptions in the moneymarkets with the traditional tool of monetary policy open marketoperations to maintain the effective federal funds rate near its targetrate However despite the Federal Reserversquos efforts in the overnightfunding market the rates on term loans in the interbank market keptrising In an attempt to ease the strains in the money markets theFederal Reserve resorted to nontraditional tools of monetary policyPerhaps the most important tool used for this purpose was the TAF

The TAF was introduced on December 2007 in the early stages of

the financial crisis to provide Federal Reserve liquidity funding byauctioning off short-term funding without forcing banks to face thestigma of borrowing from the Federal Reserversquos discount windowUnder the TAF the Federal Reserve auctioned term funds to deposi-tory institutions All depository institutions that were eligible to bor-row under the primary credit program of the Federal Reserve wereeligible to participate in TAF auctions All loans extended under theTAF were fully collateralized The funds were allocated through an

auction in which participating depository institutions placed bidsspecifying an amount of funds up to a pre-specified limit and an

2 Term funding is typically made with maturity terms of one month or longer

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 742

7

An Empirical Analysis of the Fedrsquos Term Auction Facility

interest rate that they would be willing to pay for such funds Thefunds were allocated beginning with the highest interest rate offereduntil either all funds were allocated or all bids were satisfied All

borrowing institutions paid the same interest rate either the rateassociated with the bid that would fully subscribe the auction or inthe case that total bids were less than the amount of funds offeredthe lowest rate that was bid The TAF was created under the FederalReserversquos standard discount window lending authority grantedunder Section 10B of the Federal Reserve Act The auctions wereadministered by the Federal Reserve Bank of New York with loans

granted through the 12 Federal Reserve banksTAF funding supplemented the US dollar funding received by

global banks around the world under the central bank swap facilities between the Federal Reserve banks and the Banco Central do BrasilBank of Canada Denmarkrsquos Nationalbank Bank of England Euro-pean Central Bank Bank of Japan Bank of Korea Banco de MexicoReserve Bank of New Zealand Norges Bank Monetary Authority ofSingapore Sveriges Riksbank and Swiss National Bank

From the first TAF auction on December 17 2007 to the last onMarch 8 2010 the Federal Reserve conducted 60 auctions Theamount of term loans auctioned was initially between $20 billionand $30 billion but was later increased to between $50 billion and$75 billion The size increased to $150 billion in October 2008 andremained at that level until June 2009 During the second half of 2009and the first three months of 2010 the amount auctioned graduallydeclined and by the final auction in March 2010 only $34 billionwas loaned out

Whether the TAF was effective in reducing rates in the interbankmarket has been debated by both academic economists and policy-makers McAndrews Sarkar and Wang (2008) provide empirical evi-dence that the TAF has helped to ease strains in the interbank marketIn contrast according to Taylor and Williams (2009) the TAF had noimpact on interest rate spreads According to McAndrews Sarkarand Wang the major problem in the money markets in 2007ndash2008was lack of liquidity hence the TAF was effective because it provided

central bank liquidity to the banking system when the interbankingsystem collapsed In contrast Taylor and Williams (2009) argue thatthe main problem in the market was not liquidity but rather counter-party risk which TAF funding could not have solved

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 7: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 742

7

An Empirical Analysis of the Fedrsquos Term Auction Facility

interest rate that they would be willing to pay for such funds Thefunds were allocated beginning with the highest interest rate offereduntil either all funds were allocated or all bids were satisfied All

borrowing institutions paid the same interest rate either the rateassociated with the bid that would fully subscribe the auction or inthe case that total bids were less than the amount of funds offeredthe lowest rate that was bid The TAF was created under the FederalReserversquos standard discount window lending authority grantedunder Section 10B of the Federal Reserve Act The auctions wereadministered by the Federal Reserve Bank of New York with loans

granted through the 12 Federal Reserve banksTAF funding supplemented the US dollar funding received by

global banks around the world under the central bank swap facilities between the Federal Reserve banks and the Banco Central do BrasilBank of Canada Denmarkrsquos Nationalbank Bank of England Euro-pean Central Bank Bank of Japan Bank of Korea Banco de MexicoReserve Bank of New Zealand Norges Bank Monetary Authority ofSingapore Sveriges Riksbank and Swiss National Bank

From the first TAF auction on December 17 2007 to the last onMarch 8 2010 the Federal Reserve conducted 60 auctions Theamount of term loans auctioned was initially between $20 billionand $30 billion but was later increased to between $50 billion and$75 billion The size increased to $150 billion in October 2008 andremained at that level until June 2009 During the second half of 2009and the first three months of 2010 the amount auctioned graduallydeclined and by the final auction in March 2010 only $34 billionwas loaned out

Whether the TAF was effective in reducing rates in the interbankmarket has been debated by both academic economists and policy-makers McAndrews Sarkar and Wang (2008) provide empirical evi-dence that the TAF has helped to ease strains in the interbank marketIn contrast according to Taylor and Williams (2009) the TAF had noimpact on interest rate spreads According to McAndrews Sarkarand Wang the major problem in the money markets in 2007ndash2008was lack of liquidity hence the TAF was effective because it provided

central bank liquidity to the banking system when the interbankingsystem collapsed In contrast Taylor and Williams (2009) argue thatthe main problem in the market was not liquidity but rather counter-party risk which TAF funding could not have solved

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 8: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 842

CATO PAPERS ON PUBLIC POLICY

8

3 DATA AND SUMMARY STATISTICS

The data analyzed here come from Federal Reserve disclosure ofeach of the individual term loans provided under the TAF3 Thedataset lists 4214 individual loans spanning the auctions fromDecember 12 2007 to March 8 2010

The dataset includes micro-level detailed information for eachloan contract on the contract terms the borrowerrsquos identity andthe broad categories of the securities against which the loans weremade The loan contract terms include the interest rate on the loan(in percent) the loan maturity (in days) and the loan amount (in

millions of dollars) The dataset also provides information on the bor-rower that includes the borrowerrsquos name city and state4 In additionthe Federal Reserve discloses information on the underlying col-lateral against which the loan was granted In particular it reportsthe amount of unencumbered collateral (defined as the lendablevalue of the borrowerrsquos collateral) as well as the broad categoriesof the assets used as collateral The data comprise 12 asset-typecategories commercial loans residential mortgages commercial

real estate loans consumer loans US TreasuryAgency securitiesmunicipal securities corporate securities mortgage-backed securi-ties (MBSs) and collateralized mortgage obligations (CMOs) issued

by government-sponsored enterprises MBSs and CMOs issued byprivate corporations ABSs international securities and other col-lateral Finally the dataset breaks down the dollar value of collateral

by broad credit rating categories

31 Loan Characteristics

Table 1 displays descriptive statistics for the main loan charac-teristics The average loan amount (in millions) is $9061 millionand the median is $1250 million The dispersion in loan amountranges widely from a minimum of $14 million (First Merchant Bankof Indiana) to the largest loans of $15 billion (to Bank of AmericaBarclays Citibank JP Morgan Chase Wachovia and Wells Fargo)The average loan term is 456 days and ranges from 13 days to 85 daysThe average annualized interest rate is 0900 percent and ranges from

3 The data can be downloaded at httpwwwfederalreservegovnewseventsreform_tafhtmdatadesc4 For foreign borrowers the dataset lists the city and state of their US branch whichin most cases is New York City

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 9: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 942

9

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

1

L o a n C h a r a c t e r i s t i c s

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h

P e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u

m

M a x i m u m

O b s e r v a t i o n s

L o a n a m o u n t ( m i l l i o n s )

$ 9 0 6 1

$ 2 2 0

$ 1 2 5 0

$ 1 0 0

0 0

$ 1 9 2 2 5

$ 1 4

$ 1 5 0 0 0 0

4 2 1 4

L o a n t e r m

( d a y s )

4 5 6

2 8

2 8

8 4

2 5 6

1 3

8 5

4 2 1 4

I n t e r e s t r a t e

0 9 0 0

0 2 5 0

0 2 5 0

1 3 9 0

1 0 9 3

0 2 0 0

4 6 7 0

4 2 1 4

C o l l a t e r a l ( m i l l i o n s )

$ 4 2 8 4 5

$ 7 9 3

$ 5 7 1 0

$ 4 1 5

7 9

$ 1 0 5 4 4 7

$ 5 1

$ 1 8 5 4 1 0 1

4 2 1 4

L o a n - t o - c o

l l a t e r a l r a t i o

0 3 3 4

0 1 5 0

0 2 8 6

0 4 7

7

0 2 2 7

0 0 0 4

1 0 0 1

4 2 1 4

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 10: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1042

CATO PAPERS ON PUBLIC POLICY

10

0200 percent to 4670 percent As I explained in Section 2 the TAFwas conducted through auctions in which all successful bids weresubject to the same interest rate and loan terms Thus although loanamounts varied across banks and over time all banks borrowing inthe same auction obtained loans with the same interest rates and loanmaturities

In addition to the loan amount there is strong heterogeneity inthe amounts and types of collateral posted by the borrowing banksBorrowers pledged unencumbered collateral with an averagevalue of $42854 million Collateral values range from $51 million

(Timberwood Bank) to $1854100 million (Bank of America) themedian collateral value is $5710 million I also calculated the ratioof the face amount of the loan to the value of the unencumberedcollateral and report it in the last row of Table 1 As the table showsthe average loan-to-collateral ratio is 0334 and the median is 0286Loan-to-collateral rates increased after the peak of the crisis as col-lateral values increased and haircuts on collateral declined Forexample the average loan-to-value in 2008 was 0255 compared to0370 and 0460 in 2009 and 2010 respectively Although the loan-to-collateral ratios appear to be low and conservative it is not clearwhether these numbers are based on market values or on face valuesof the underlying collateral

32 The Collateral Structure of TAF Loans

Next I analyzed the composition of collateral in TAF loans Giventhat loan terms as well as loan rates were determined at the auctionlevel the only sources of interbank variation were the amount of

the loan and the amount and type of the collateral Indeed bankspledged different types of assets as collateral for their loans andmost TAF loans were secured by numerous securities from differentasset types Table 2 provides a detailed analysis of collateral struc-ture for the 4214 TAF loans The table reports summary statistics forthe dollar amount (in millions) as well as the number of loans forwhich collateral was pledged in each asset category

The largest collateral category (based on the dollar amount of the

assets pledged) is residential mortgages The mean amount of resi-dential mortgages used as collateral is $37863 million and it wasused as collateral in 465 individual loans The next largest categoryis ABSs which according to the Federal Reserve definitions include

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 11: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1142

11

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

2

C

o l l a t e r a l C o m p o s i t i o n b y S e c u r i t y T y p e

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P e r c e n t i l e

M e d i a n

7 5 t h P

e r c e n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

R e s i d e n t i a l

m o r t g a g e s

$ 3 7 8 6 3

$ 7 9

$ 2 7 5

$

4 0 2 9

$ 1 1 0 7 0 0

$ 0 0

$ 7 6 8 4 7 5

4 6

5

A B S s

$ 2 5 6 2 8

$ 9 1 5

$ 7 8 0 6

$ 2 5 1 3 4

$ 4 4 8 6 9

$ 0 4

$ 2 5 9 5 3 7

1 3 0

1

C o m m e r c i a

l l o a n s

$ 2 2 3 2 1

$ 4 0 1

$ 3 3 8 0

$ 1 5 4 4 4

$ 6 9 1 1 8

$ 0 1

$ 7 6 7 8 4 0

2 2 9

1

C o n s u m e r l o a n s

$ 1 4 6 2 5

$ 8 8

$ 8 6 1

$ 1 1 9 4 1

$ 3 8 7 5 5

$ 0 0

$ 3 2 6 7 9 2

1 0 8

7

P r i v a t e M B S s C M O s

$ 1 1 5 4 9

$ 6 9 6

$ 2 4 1 7

$

9 2 2 5

$ 2 3 7 8 8

$ 0 0

$ 1 4 5 9 9 2

1 0 4

5

C o m m e r c i a

l r e a l

e s t a t e

$ 1 0 9 1 0

$ 3 5 9

$ 1 0 4 3

$

7 1 8 9

$ 2 4 7 1 5

$ 0 0

$ 3 0 4 6 9 6

1 6 2

4

C o r p o r a t e

$ 7 4 7 6

$ 4 3 6

$ 2 0 9 3

$ 1 1 1 5 9

$ 1 0 9 0 0

$ 0 1

$ 6 8 4 0 6

1 5 0

7

I n t e r n a t i o n a l

s e c u r i t i e s

$ 7 0 3 1

$ 2 8 7

$ 1 2 9 7

$

5 8 0 4

$ 1 4 9 2 0

$ 0 1

$ 1 1 3 0 2 5

1 1 3

8

A g e n c y M B

S s

C M O s

$ 5 6 7 9

$ 1 2 2

$ 8 0 2

$

4 9 8 7

$ 1 6 8 1 2

$ 0 0

$ 2 6 6 7 9 8

1 1 5

1

M u n i c i p a l s

$ 3 7 0 3

$ 7 0

$ 2 0 1

$ 9 3 6

$ 1 1 8 0 6

$ 0 0

$ 8 9 1 1 9

1 0 9

9

U S T r e a s u

r y

A g e n c y

$ 3 4 8 8

$ 1 1 3

$ 4 8 5

$

2 4 1 6

$ 9 6 8 7

$ 0 0

$ 8 7 6 2 4

8 3

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 12: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1242

CATO PAPERS ON PUBLIC POLICY

12

securities collateralized by assets other than first-lien mortgagesincluding CDOs More than 1301 loans were backed by ABSs andthe mean collateral pledged in this category is $25628 million andranges from $04 million to $259537 million

The most popular asset class based on the number of loans thatused it as collateral is commercial loans which were used in 2291loans followed by commercial real estate and corporate securitieswhich were used in 1624 and 1507 loans respectively FinallyUS TreasuryAgency securities were used in 833 loans with a meancollateral value of $3488 million

The dataset also breaks down the collateral pool by credit ratingcategories5 Table 3 reports summary statistics for the collateral assets

by the major credit rating classifications AAA-rated US TreasuryAgency securities (including agency MBSs and CMOs) amounted onaverage to $6503 million The amount of other AAA-rated securitiespledged as collateral was on average $18458 million per loan andthese were used in 1859 loans AA-rated and A-rated securities wereused in 1681 and 1817 loans respectively and accounted for about$380 million each of the collateral pool Other rating categories in-clude BBB-rated (mean $2380 million) and ldquoother investment graderdquosecurities (mean $12326 million)

4 EMPIRICAL ANALYSIS

41 Determinants of Loan Characteristics

I began the empirical analysis of TAF loans by analyzing the char-acteristics of the loans The eight ordinary least-squares regressionsreported in Table 4 use different specifications to predict the determi-

nants of the loan terms For each of the four loan determinants I reportresults from regressions that do not include bank fixed-effects (betweenanalysis) and regressions that utilize variation over time using bankfixed-effects (within analysis) As explanatory variables I used collateraldummy variables that take the value of ldquo1rdquo if a particular asset is in-cluded in the collateral pool and ldquo0rdquo otherwise All regressions includeyear983091 month fixed-effects to account for time-varying effects6

5 The dataset reports asset types and credit ratings separately and hence does notenable classification that is based on both credit ratings and asset class6 Although I used collateral dummy variables the analysis yields similar results whenusing the actual share of collateral in each asset category

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 13: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1342

13

An Empirical Analysis of the Fedrsquos Term Auction Facility

T a b l e

3

C o l l a t e r a l C o m p o s i t i o n b y C r e d i t R a t i n g

( M i l l i o n s e x c e p t o b s e r v a t i o n s )

M e a n

2 5 t h

P

e r c e n t i l e

M e d i a n

7 5

t h

P e r c e

n t i l e

S t a n d a r d

D e v i a t i o n

M i n i m u m

M a x i m u m

O b s e r v a t i o n s

A A A - r a t e d

U S T r e a s u r y A g e n c y

a n d A g e n c y M B S - C M O s

$ 6 5 0 3

$ 1 3 4

$ 1 0 6 1

$ 5

5 5 8

$ 1 7 0 9 2

$ 0 0

$ 2 6 6 7 9 8

1 3 7 5

O t h e r A A A - r a t e d

$ 1 8 4 5 8

$ 2 2 2

$ 3 5 2 9

$ 1 8

0 1 7

$ 3 5 3 2 8

$ 0 0

$ 2 2 3 6 4 7

1 8 5 9

A A - r a t e d

$ 3 8 1 6

$ 7 9

$ 7 4 2

$ 3

7 0 9

$ 8 5 7 8

$ 0 0

$ 8 5 0 5 2

1 6 8 1

A - r a t e d

$ 3 8 6 5

$ 1 3 3

$ 9 6 3

$ 4

2 0 8

$ 7 8 3 3

$ 0 1

$ 7 7 7 5 7

1 8 1 7

B B B - r a t e d

$ 2 3 8 0

$ 1 2 2

$ 6 8 5

$ 2

9 6 8

$ 4 2 2 9

$ 0 1

$ 4 8 8 1 4

1 6 9 4

O t h e r i n v e s t m e n t - g r a d e

$ 1 2 3 2 6

$ 4 2 3

$ 2 9 5 7

$ 1 0

8 8 1

$ 2 7 1 4 3

$ 0 0

$ 2 2 7 2 6 2

1 4 1 7

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 14: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1442

CATO PAPERS ON PUBLIC POLICY

14

T a b l e 4

D e t e r m i n a n t s o f L o a n T e r m s

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t

h i n

A B S d u m

m y

1 5 7 0

983090 0 0 0 1

0 0 0 8

983090 0 0 1 6

0 4 4 4

0 5 4 4

983090 0 1 5 0

983090 0 0 5 2

( 0 0 8 4 )

( 0 0 8 8 )

( 0 0 0 6 )

( 0 0 2 0 )

( 1 1 5 2 )

( 2 7 0 8 )

( 0 0 1 0 )

( 0 0

1 9 )

C o m m e r c

i a l r e a l e s t a t e d u m m y

0 4 4 6

0 2 2 0

983090 0 0 0 4

983090 0 0 1 0

983090 1 0 7 9

983090 0 6 8 0

983090 0 0 6 9

983090 0 0 3 4

( 0 0 6 1 )

( 0 0 6 8 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 8 1 3 )

( 1 8 5 9 )

( 0 0 0 7 )

( 0 0

1 4 )

I n t e r n a t i o

n a l d u m m y

0 8 5 0

0 0 4 7

983090 0 0 0 7

983090 0 0 0 8

2 4 1 7

983090 0 7 7 5

0 0 0 4

983090 0 0 0 5

( 0 0 8 2 )

( 0 0 6 9 )

( 0 0 0 7 )

( 0 0 1 4 )

( 1 1 2 2 )

( 1 9 6 7 )

( 0 0 1 1 )

( 0 0

1 6 )

M u n i c i p a

l d u m m y

983090 0 2 4 2

983090 0 0 1 7

0 0 0 2

0 0 1 7

983090 1 3 9 6

983090

2 9 5 1

0 0 1 1

983090 0 0 2 2

( 0 0 5 9 )

( 0 0 5 9 )

( 0 0 0 4 )

( 0 0 1 1 )

( 0 8 9 9 )

( 1 7 6 2 )

( 0 0 0 8 )

( 0 0

1 4 )

T r e a s u r y d u m m y

0 2 7 9

0 2 1 7

983090 0 0 0 4

983090 0 0 0 5

983090 1 1 7 0

983090 2 7 6 2

0 0 5 1

0 0 3

0

983090 0

0 6 7

( 0 0 6 2 )

( 0 0 0 5 )

( 0 0 1 0 )

( 0 9 9 8 )

( 1 8 9 9 )

( 0 0 0 9 )

( 0 0

1 4 )

P r i v a t e M

B S d u m m y

0 8 8 1

0 2 7 5

0 0 0 1

983090 0 0 1 7

983090 0 2 1 4

983090 1 8 9 9

983090 0 4 5

983090 0 0 1 9

( 0 0 8 7 )

( 0 0 8 6 )

( 0 0 1 0 )

( 0 0 1 6 )

( 1 1 5 2 )

( 2 2 6 8 )

( 0 0 1 0 )

( 0 0

1 6 )

A g e n c y d

u m m y

0 0 4 6

0 0 4 8

983090 0 0 0 3

983090 0 0 0 8

0 0 8 8

983090

3 3 9 8

0 0 0 4

0 0

0 2

( 0 0 6 2 )

( 0 0 5 7 )

( 0 0 0 4 )

( 0 0 0 9 )

( 0 9 6 0 )

( 1 6 7 3 )

( 0 0 0 9 )

( 0 0

1 3 )

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 15: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1542

15

An Empirical Analysis of the Fedrsquos Term Auction Facility

R e s i d e n t i a l m o r t g a g e d u m m y

0 0 3 5

0 3 2 4

983090 0 0 0 0

0 0 2 4

983090 6 2 6 7

983090

7 2 2 3

0 0 5 2

983090 0 0 3 7

( 0 0 9 5 )

( 0 1 0 8 )

( 0 0 0 5 )

( 0 0 2 0 )

( 1 1 0 5 )

( 2 5 4 9 )

( 0 0 1 2 )

( 0 0

2 2 )

C o n s u m e r l o a n s d u m m y

0 4 5 4

0 2 0 3

983090 0 0 0 3

983090 0 0 0 5

0 9 5 9

983090

4 0 3 7

983090 0 0 7 3

983090 0 0 6 3

( 0 0 7 0 )

( 0 1 0 3 )

( 0 0 0 4 )

( 0 0 1 7 )

( 0 8 6 5 )

( 2 4 0 1 )

( 0 0 0 8 )

( 0 0

2 0 )

L

o g ( l o a n a m o u n t )

I n t e r e s t R a t e

L o a n T e r m

L o a n t o C o l l a t e r a l

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

B e t w e e n

W i t h i n

C o r p o r a t e s e c u r i t i e s d u m m y

0 1 1 1

0 3 4 0

0 0 0 7

0 0 0 7

983090 1 6 1 3

1 7 4 0

0 0 4 7

983090 0 0

3 4

( 0 0 7 7 )

( 0 0 8 2 )

( 0 0 0 6 )

( 0 0 1 5 )

( 1 0 9 1 )

( 2 1 9 2 )

( 0 0 1 1 )

( 0 0

2 0 )

F i x e d - e f f e c t s

y e a r 983091 m o n t h

Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

B a n k

N

o

Y e s

N o

Y e s

N o

Y e s

N o

Y e s

P s e u d o A

d j u s t e d R 2

0 4 1

0 9 3

0 9 9

0 9 9

0 3 9

0 3 9

0 1 9

0 5 9

O b s e r v a t i o n s

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2 1 4

4 2

1 4

N o t e O m

i t t e d c o l l a t e r a l c a t e g o r y i s c o m

m e r c i a l l o a n s 983093 1 0 s i g n i f i c a n c e l e v e l 983093 5 s i g n i f i c a n

c e l e v e l 983093 1 s i g n i f i c a n c e

l e v e l S t a n

d a r d e r r o r s i n p a r e n t h e s e s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 16: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1642

CATO PAPERS ON PUBLIC POLICY

16

The table reports results for the following loan characteristics loanamount (in logs) interest rate loan term and loan-to-collateral ratioHowever it should be noted that because TAF loans were grantedat auction the same interest rate and loan term applied to all banksparticipating in each auction In contrast the loan amount the ratioof loan to collateral (the inverse of the loan ldquohaircutrdquo) and the na-ture of the assets pledged as collateral varied across banks within anauction As Table 4 shows the composition of the collateral has littleexplanatory power in bearing on loan outcomes First few if any ofthe explanatory variables turn out to be significant in regressions

that use the interest rate or loan term as dependent variables Secondthe R-squared in the regressions shows that the addition of bankfixed-effects does not change the adjusted R-squared in the interestrate and loan term regressions indicating that bank-specific effectshad no impact on the loan rate and maturity7

In contrast collateral composition significantly affected both loanamount and loan-to-collateral ratio As the first column shows ABSscommercial real estate international securities Treasuries privateMBSs and consumer loans are associated with larger loans whereasmunicipal securities are correlated with smaller loans However giventhat the regressions do not control for bank characteristics it is likelythat some of the collateral results are driven by omitted variables Forexample if larger banks are more likely to hold ABSs or international

bonds then the positive coefficient in column one might be captur-ing the simple correlation between bank size and loan amount In anattempt to address this concern the regression specification reportedin column two adds bank fixed-effects to the analysis and hence uses

variation within a bank from repeated loans in several TAF auctionsover time Indeed as the second column of the table demonstratesonly commercial real estate Treasury and private MBSs survive theaddition of fixed-effects and are still positive and significant In addi-tion residential mortgages and corporate bonds turn out to be posi-tive and significant when fixed-effects are added

Turning to the last two columns of the table I find that loanssecured by ABSs obtained loan-to-collateral ratios between 9830900150

7 The high R-squared in the interest rate is completely driven by the year 983091 monthfixed-effects since there was an overall trend of declining interest rates throughoutthe TAF time period

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 17: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1742

17

An Empirical Analysis of the Fedrsquos Term Auction Facility

and 9830900052 lower Likewise consumer loans led to lower loan-to-value ratios while loans secured by Treasuries had loan-to-collateralratios that were higher by 0030 The results are consistent with thenotion that haircuts on collateral are an important tool for monetarypolicy This is important especially when nontraditional monetarypolicy is conducted through auctions in which the interest rate andloan terms do not vary across borrowers

42 The Evolution of TAF over Time

Figure 1 displays the evolution of the TAF lending facility size

over time As described in Section 2 the Federal Reserve announcedthe offering amount in advance of each auction As Figure 1 showsthe initial auctions were smaller with amounts between $20 and$30 billion The offering amount was raised to $50 billion in theMarch 10 2008 auction and was further increased to $75 billion onMay 5 2008 While the amounts fluctuated between $25 billion and$75 billion in August and September 2008 the lending facilities in-creased dramatically to $150 billion on October 6 2008 during thepeak of the financial crisis and remained at that level until the end

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

12607 31508 62308 10108 1909 41909 72809 11509 21310

M i l l i o n s

Figure 1Facility Size

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 18: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1842

CATO PAPERS ON PUBLIC POLICY

18

of June 2009 The offering amount gradually declined to $125 billionand then $100 billion in July and August 2009 and later fluctuated

between $75 billion and $25 billion The final auction was held onMarch 8 2010 for $25 billion

Although Figure 1 plots the offering amounts in each of the auc-tions and hence the potential (the supply of funds) it does not show

the amount demanded by banks that submitted bids for TAF moneyor the amount that was actually loaned Figure 2 supplements theinformation in Figure 1 by plotting both the total amount of propo-sition submitted by banks (the demand for loans) and the amountthat was actually awarded8 As the figure illustrates the demand forfunds exceeded the supply from the first auction in December 2007until the auction of September 22 2008 For example on December12 2007 the offering amount by the Federal Reserve was $20 bil-lion but the amount demanded by the 93 banks that submitted bids

8 I use the notion of demand and supply here fairly loosely Of course given theauction structure there was no excess demand at a given rate

$0

$20000

$40000

$60000

$80000

$100000

$120000

$140000

$160000

10107 11508 42408 8208 111008 21809 52909 9609 121509 32510 7310

A m o u n t ( m i l l i o n s )

total proposion submied

amount awarded

Figure 2

Total Proposition Submitted and Amount Awarded

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 19: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 1942

19

An Empirical Analysis of the Fedrsquos Term Auction Facility

was $616 billion In the auction of September 22 2008 the facilitywas increased to $75 billion but 85 banks submitted bids totaling$1336 billion

Following the Federal Reserversquos increase of the facility size to$150 billion in October 2008 the amount of propositions submitted

by banks dropped below the amount offered by the Federal Reserveuntil the end of the TAF In the October 8 2008 auctionmdashthe firstauction with a facility size of $150 billionmdash71 banks submitted

bids totaling $1381 billion The largest amount requested by bankswas $1425 billion when 117 banks participated in the auction of

February 9 2009 The largest number of banks participating in asingle auction was 124 (May 4 2009) compared to only 16 banks onNovember 24 2008

43 TAF Lending to Foreign Banks

Table 5 lists the number of loans average loan size and totalamount loaned in each month from the first auction in December2007 through the final auction in March 2010 The table further breaksdown monthly lending by whether the borrowing bank is a US de-pository institution or a foreign bank9 Overall foreign banks received58 percent of the total amount lent over time with a total amount of$2214688 million compared to only $1603723 for US banks FromDecember 2007 through most of 2008 foreign banks accounted forthe vast majority of the lending with amounts that were two to fourtimes the total lending to US banks However during the peak of thecrisis and following the collapse of Lehman Brothers and especiallyin October and November 2008 lending to US banks exceeded bor-

rowing by foreign banks By April 2008 and until the end of the TAFforeign banks again accounted for the majority of TAF lending

Table 6 and Figure 3 present the 50 largest borrowers (measured bythe total amount borrowed) For each of the largest borrowers Table 6lists the total loan amount the average loan size the number of loansobtained under the TAF and the home country of the bank LikewiseFigure 3 displays the largest 50 borrowers in a bar chart As bothTable 6 and Figure 3 show UK-based Barclays is the largest borrowerwith a total amount of $232283 million in 49 loans followed by Bank

9 Foreign banks were eligible to participate in the TAF through their agencies or branches in the United States

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 20: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2042

CATO PAPERS ON PUBLIC POLICY

20

T a b l e

5

E v o l u t i o n o f t h e T A F o v e r T i m e D o

m e s t i c B a n k s v s F o r e

i g n B a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

N u m b e r o f

L o a n s

A v e r a g e L o a n

S i z e

T o t a l A m

o u n t

N u m b e r o f

L o a n s

A

v e r a g e L o a n S i z e

T o t a

l

A m o u

n t

D e c e m b e r 2

0 0 7

1 3

$ 3 7 4 9

$ 4 8 7 3

4 2

$ 8 3 6 4

$ 3 5 1 2 7

J a n u a r y 2 0 0

8

4 6

$ 4 6 0 3

$ 2 1 1 7 2

3 5

$ 1 1 0 9 4

$ 3 8 8 2 8

F e b r u a r y 2 0

0 8

4 1

$ 4 2 3 6

$ 1 7 3 6 8

3 4

$ 1 2 5 3 9

$ 4 2 6 3 2

M a r c h 2 0 0 8

4 5

$ 6 1 5 2

$ 2 7 6 8 3

4 1

$ 1 7 6 3 8

$ 7 2 3 1 8

A p r i l 2 0 0 8

3 2

$ 4 8 7 5

$ 1 5 5 6 0

4 4

$ 1 9 1 8 2

$ 8 4 4 0 0

M a y 2 0 0 8

7 6

$ 7 2 9 2

$ 5 5 4 1 8

5 6

$ 1 6 6 9 0

$ 9 4 5 8 1

J u n e 2 0 0 8

5 7

$ 8 4 8 0

$ 4 8 3 3 5

5 0

$ 2 0 3 3 3

$ 1 0 1 6 6 5

J u l y 2 0 0 8

1 0 5

$ 8 2 1 2

$ 8 6 2 3 0

7 9

$ 1 7 5 6 6

$ 1 3 8 7 7 0

A u g u s t 2 0 0 8

7 3

$ 9 1 2 7

$ 6 6 6 2 7

5 5

$ 1 5 1 5 9

$ 8 3 3 7 3

S e p t e m b e r 2 0 0 8

4 0

$ 1 0 6 4 9

$ 4 2 5 9 5

4 6

$ 1 7 9 1 4

$ 8 2 4 0 5

O c t o b e r 2 0 0

8

8 6

$ 1 7 3 0 4

$ 1 4 8 8 1 8

6 1

$ 1 6 8 1 1

$ 1 0 2 5 4 5

N o v e m b e r 2 0 0 8

1 1 3

$ 1 3 9 4 4

$ 1 5 7 5 6 1

6 0

$ 1 4 3 0 9

$ 8 5 8 5 6

D e c e m b e r 2

0 0 8

1 2 4

$ 6 0 8 2

$ 7 5 4 1 2

6 1

$ 1 6 0 2 9

$ 9 7 7 7 6

J a n u a r y 2 0 0

9

1 8 5

$ 1 0 6 2 8

$ 1 9 6 6 1 7

8 9

$ 1 6 8 7 2

$ 1 5 0 1 6 0

F e b r u a r y 2 0

0 9

1 5 0

$ 1 0 2 5 6

$ 1 5 3 8 3 8

6 6

$ 1 5 1 9 7

$ 1 0 0 3 0 2

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 21: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2142

21

An Empirical Analysis of the Fedrsquos Term Auction Facility

M a r c h 2 0 0 9

1 4 9

$ 6 8 9 6

$ 1 0 2 7 4 3

7 1

$ 1 6 4 3 3

$ 1 1 6 6 7 2

A p r i l 2 0 0 9

1 3 6

$ 6 0 8 4

$ 8 2 7 4 3

6 7

$ 1 6 0 2 8

$ 1 0 7 3 8 6

M a y 2 0 0 9

1 5 9

$ 5 6 8 8

$ 9 0 4 4 6

6 1

$ 1 5 8 4 1

$ 9 6 6 3 3

J u n e 2 0 0 9

1 4 0

$ 2 5 8 5

$ 3 6 1 9 5

6 2

$ 1 7 2 3 5

$ 1 0 7 4 1 6

J u l y 2 0 0 9

2 1 0

$ 2 4 0 1

$ 5 0 4 1 7

8 7

$ 1 9 0 8 8

$ 1 6 6 0 6 2

A u g u s t 2 0 0 9

1 3 2

$ 1 9 2 2

$ 2 5 3 7 1

5 6

$ 1 6 2 4 5

$ 9 0 9 7 4

S e p t e m b e r 2 0 0 9

1 0 9

$ 2 1 0 2

$ 2 2 9 1 2

4 9

$ 1 3 2 1 6

$ 6 4 7 5 9

O c t o b e r 2 0 0

9

1 1 6

$ 1 9 2 5

$ 2 2 3 2 6

4 2

$ 1 0 0 1 7

$ 4 2 0 7 0

N o v e m b e r 2 0 0 9

1 0 4

$ 1 6 0 7

$ 1 6 7 1 3

3 1

$ 8 8 9 0

$ 2 7 5 5 8

D e c e m b e r 2

0 0 9

1 1 0

$ 1 5 5 8

$ 1 7 1 4 0

4 1

$ 1 1 1 2 8

$ 4 5 6 2 5

J a n u a r y 2 0 1

0

9 8

$ 1 1 1 2

$ 1 0 8 9 3

2 3

$ 1 2 0 1 7

$ 2 7 6 3 8

F e b r u a r y 2 0

1 0

8 9

$ 6 4 8

$ 5 7 6 3

1 4

$ 6 9 0 2

$ 9 6 6 3

M a r c h 2 0 1 0

4 9

$ 3 9 1

$ 1 9 1 5

4

$ 3 7 3 8

$ 1 4 9 5

D e c e m b e r 2 0 0 7 ndash

M a r c h 2 0 1 0

2 7 8 7

$ 5 7 5 4

$ 1 6 0 3 7 2 3

1 4 2 7

$ 1 5 5 2 0

$ 2 2 1 4 6 8 8

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 22: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2242

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 23: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2342

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 24: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2442

CATO PAPERS ON PUBLIC POLICY

24

$ 0

$ 5 0

0 0 0

$ 1 0 0

0 0 0

$ 1 5 0

0 0 0

$ 2 0 0

0 0 0

B A R C L A Y S B K P L C N Y B R

B A N K O F A M E R N A

B A N K O F S C O T L A N D P L C N Y B R

W E L L S F A R G O B K N A

W A C H O V I A B K N A

S O C I E T E G E N E R A L E N Y B R

D R E S D N E R B K A G N Y B R

R B S C I T I Z E N S N A C I T I B A N K N A

B A Y E R I S C H E L A N D E S B A N K N Y B R

D E X I A C R E D I T L O C A L N Y B R

N O R I N C H U K I N B K N Y B R

J P M O R G A N C H A S E B K N A

W E S T L B A G N Y B R

D E U T S C H E B K A G N Y B R

R E G I O N S B K

U N I C R E D I T N Y B R

F O R T I S B K S A N V N Y B R

S U M I T O M O M I T S U I B K G N Y B R

U B S A G P A R K A V E B R

R O Y A L B K O F S C O T L A N D P L C N Y B R

H S H N O R D B K A G N Y B R M I Z U H O C O R P O R A T E N Y B R

C O M M E R Z B A N K A G N Y B R

F I A C A R D S V C N A

D E P F A B K P L C N Y B R

F I R S T T N B K N A

F I F T H T H I R D B K

S T A T E S T R E E T B amp T C

K E Y B A N K N A

D Z B K A G D E U T S C H E Z E N T R A N Y B R

C I T I Z E N S B K O F P A

B A N K T O K - - M I T U F J N Y B R

R O Y A L B K O F C A N A D A N Y B R

A L L I E D I R I S H B K S N Y B R

B A Y E R I S C H E H Y P O V E R E I N S N Y B R

N A T I X I S N Y B R

B N P P A R I B A S S F B R

T O R O N T O - - D O M I N I O N B K N Y B R

B A N K O F N O V A S C O T I A N Y A G Y

A R A B B K G C O R P N Y B R

S T A N D A R D C H A R T E R E D B K N Y B R

M I T S U B I S H I U F J T R amp B K G N Y B R

C R E D I T I N D U S E T C M R L N Y B R

R A B O B A N K N E D E R L A N D N Y B R

B R A N C H B K G amp T C

L A N D E S B K B A D E N W U E R T T E M B N Y B R

A L L Y B K

M amp I M A R S H A L L amp I L S L E Y B K

C O U N T R Y W I D E B K F S B

A m o u n t b o r r o w e d ( m i l l i o n s )

F i g u r e

3

L a r g e s t T A F B

o r r o w e r s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 25: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2542

25

An Empirical Analysis of the Fedrsquos Term Auction Facility

of America with a total amount of $212617 million in 15 loans Thenext largest borrowers are Royal Bank of Scotland ($180920 million)Wells Fargo ($153953 million) and Wachovia ($147025 million) Fur-thermore out of the 10 largest borrowers five are foreign banks andout of the 50 largest borrowers 33 are from foreign countries

44 The Collateral Structure of Foreign Banks

As I argued previously the loan term and interest rate weredetermined at the auction level regardless of the identity of the bor-rowing bank participating in the auction In contrast the size of the

loan and the collateral pledged by the bank were the only marginsthat both the bank and the Federal Reserve could adjust at the loanlevel Given the importance of collateral in general and in particu-lar given the unique setup of the TAF I compared the collateralstructure of domestic banks to the collateral used by foreign banksTable 7 presents summary statistics on the use of collateral by do-mestic and foreign banks For each asset category reported by theFederal Reserve the table lists the mean share of the asset category

in the collateral pool the standard deviation of the share and thenumber of loans pledging that asset as part of their collateralThe summary statistics are reported separately for domestic and

foreign banks and a two-sample T-test for equal means is also pre-sented As Table 7 shows foreign banks rarely used residential mort-gages as collateral only 5 loans made to foreign banks were secured

by residential mortgages compared to 460 loans to domestic banksConversely ABSs were used in 983 loans to foreign banks com-pared to 318 loans to domestic banks Furthermore ABSs account

for a larger share of the overall collateral pool in foreign banks (9323versus 0151 significant at the 1 percent level) As in the case of resi-dential mortgages foreign banks rarely used consumer loans (only44) as collateral while US-based banks used consumer loans in1043 loans Private MBSs and CMOs were more prevalent amongforeign banks (although their share is slightly lower than in US

banks) and commercial real estate loans were used in only 222 for-eign loans Other significant differences between foreign and domes-

tic banks are that foreign banks were less likely to use Agency MBSsand CMOs US TreasuryAgency securities and US municipal

bonds and were much more likely to pledge international securitiesas collateral

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 26: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2642

CATO PAPERS ON PUBLIC POLICY

26

T a b l e 7

C o

l l a t e r a l S h a r e D o m e s t i c B a n k s v s F o r e i g n B

a n k s

D o m e s t i c B a n k s

F o r e i g n B a n k s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e r o f

L o a n s

M e a n

S t a n d a r d

D e v i a t i o n

N u m b e

r o f

L o a n

s

D i f f e r e n c e

T w o - S a m p l e

T - t e s

t

R e

s i d e n t i a l m o r t g a g e s

0 2 8 0

0 0 1

4 6 0

0 7 3 3

0 0 1

5

983090 0 4 5 4

983090 3 3

7

A s s e t - b a c k e d

s e c u r i t i e s

0 1 5 1

0 0 1

3 1 8

0 3 2 3

0 0 1

9 8 3

983090 0 1 7 2

983090 1 1 3

4

C o

m m e r c i a l l o a n s

0 4 7 0

0 0 1

1 6 5 6

0 4 2 7

0 0 2

6 3 5

0 0 4 3

2 8

3

C o

n s u m e r l o a n s

0 3 5 8

0 0 1

1 0 4 3

0 4 7 7

0 0 7

4 4

983090 0 1 1 9

983090 2 1

5

P r i v a t e M B S s C M O s

0 1 8 3

0 0 1

3 4 9

0 1 4 1

0 0 1

6 9 6

0 0 4 2

3 1

9

C o

m m e r c i a l r e a l

e s t a t e

0 5 6 3

0 0 1

1 4 0 2

0 1 5 9

0 0 1

2 2 2

0 4 0 4

2 0 3

7

C o

r p o r a t e s e c u r i t i e s

0 1 9 2

0 0 1

4 2 8

0 2 6 3

0 0 1

1 0 7 9

983090 0 0 7 1

983090 4 7

4

I n t e r n a t i o n a l

s e c u r i t i e s

0 1 3 3

0 0 2

1 9 4

0 2 0 8

0 0 1

9 4 4

983090 0 0 7 3

983090 3 8

3

A g e n c y M B S s C M O s

0 3 9 2

0 0 1

6 7 3

0 2 2 5

0 0 1

4 7 8

0 1 6 7

8 8

2

M u n i c i p a l s

0 2 3 3

0 0 1

6 1 1

0 0 5 5

0 0 1

4 8 8

0 1 7 7

1 2 2

6

U S T r e a s u r y A g e n c y

0 2 3 8

0 0 1

3 8 4

0 1 4 9

0 0 1

4 4 9

0 0 8 9

5 4

3

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 27: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2742

27

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 8 supplements the results in Table 7 using regression analy-sis of the collateral composition of foreign banks For each of the mainasset categories I use two dependent variables The first dependentvariable is a dummy variable for whether a security type is pledgedas collateral for a particular loan This variable captures the averagetendency to use an asset as collateral The second dependent variableis the actual share of the collateral in each asset group conditional onthe asset being used as collateral That is while the first variable usesinformation on all loans the second variable captures only the cross-sectional variation within an asset category conditional on its use All

regressions include year 983091 month fixed-effects as well as a controlfor the loan amount (in logs) and a dummy variable that takes thevalue of ldquo1rdquo for foreign banks and ldquo0rdquo otherwise Regressions forwhich the dependent variable is a dummy variable are estimatedusing probit where marginal effects are reported Table 8 confirmsthe univariate findings Foreign banks are more likely to use ABSsinternational assets and Treasuries and are less likely to use com-mercial real-estate

5 WHY FOREIGN BANKS

Given that more than 58 percent of TAF lending went to foreign banks it is important to understand why the Federal Reserveallocated its lending to foreign banks that are not under its directsupervision In addition the information contained in the collateralstructure of these banks suggests that the collateral pledged bythe foreign banks consisted of harder-to-value riskier assets suchas ABSs In particular the ABSs held by the foreign banks aremdash

according to the data definitions provided by the Federal ReserveBoardmdashcollateralized debt obligations secured by ABSs whichwere the securitized assets that declined the most during the crisis(Benmelech and Dlugosz 2009)

51 Exposure to Asset-Backed Securities

One potential explanation for both the elevated lending to foreign banks and their use of ABSs as collateral is that foreign banks were

hit harder than US banks and hence required m ore liquidity Giventhat many foreign banks had exposure to assets that deterioratedin value (mostly ABSs and CDOs) these banks had weaker balancesheets

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 28: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2842

CATO PAPERS ON PUBLIC POLICY

28

T a b l e

8

C

o l l a t e r a l C o m p o s i t i o n o f F o r e i g n B a n k s

A s s e t - B a c k e d

S e c u r i t i e s

C o m m e r c i a l R

e a l E s t a t e

I n t e r n a t i o n a l S e c u r i t i e s

T r e a s u r y S e c u r i t i e s

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

E x t e n s i v e

I n t e n s i v e

L o g ( l o a n a m o u n t )

0 0 9 0

0 0 2 5

0 0 3 4

983090 0 0 7 2

0 0 1 8

983090 0 0 0 8

0 0 0 6

983090 0 0

6 4

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 3 )

( 0 0 0 4 )

( 0 0 0 5 )

( 0 0 0 4 )

( 0 0 0 5 )

F o r e i g n d u m m y

0 3 9 3

0 1 4 3

983090 0 4 0 3

983090 0 2 3 1

0 6 1 1

0 1 3 6

0 1 7 8

0 0 8

2

( 0 0 1 8 )

( 0 0 1 5 )

( 0 0 1 5 )

( 0 0 1 8 )

( 0 0 1 7 )

( 0 0 2 1 )

( 0 0 1 6 )

( 0 0 1 5 )

F i x e d - e f f e c t s

y e a r 983091 m o n

t h

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y

e s

P s e u d o A d

j u s t e d R 2

0 3 9

0 1 3

0 1 1

0 4 1

0 3 9

0 0 9

0 0 5

0 2 6

O b s e r v a t i o n s

4 2 1 4

1 3 0 1

4 2 1 4

1 6 2 4

4 2 1 4

1 1 3 8

4 2 1 4

8 3 3

N o t e 983093 1 0 s i g n i f i c a n c e l e v e l 983093

5 s i g n i f i c a n c e l e v e l 983093 1 s i g n i f i c a n c e l e v e l S t a

n d a r d e r r o r s i n p a r e n t h e s e

s

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 29: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 2942

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 30: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3042

CATO PAPERS ON PUBLIC POLICY

30

Table 9

ABS CDOs and Write-DownsPanel A Crisis-Related Write-Downs for Selected Banks (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American Banks

Bank of America $9089 $932 ndash $2834 $12855Bear Stearns $2300 ndash ndash ndash $2300Citigroup $34106 $4053 $1319 $15904 $55382Goldman Sachs ndash $4100 $1700 $1400 $7200

JP MorganChase

$1300 $5467 $5305 ndash $12072

Lehman Brothers $200 $1300 $4100 $3400 $9000Merrill Lynch $26100 $2845 $12998 $13125 $55068Morgan Stanley $7800 $3810 $3781 $1992 $17383European Banks

Credit Suisse $3427 $3057 $530 $2523 $9357Deutsche Bank $2092 $5820 $3386 $3677 $14974Fortis Bank $4359 $3660 $144 ndash $8163ING $565 ndash $8028 $25 $8617

Royal Bank ofScotland

$3609 $1849 $2566 $4122 $12146

UBS $21870 $348 $1716 $13871 $37805

Asian and Emerging Market Banks

Aozora Bank $510 ndash ndash ndash $510Mitsubishi UFJ $360 $2348 $921 $11 $3640Mizuho $3898 $629 $2539 $584 $7650NationalAustralia Bank

$670 ndash ndash ndash $670

Sumitomo Mitsui $562 ndash ndash ndash $562Panel B Aggregate Crisis-Related Write-Downs (Millions)

ABS CDOs

Corporate

Credit RMBS Other Total

North American $84319 $23702 $42272 $59011 $209305European $63464 $18579 $26423 $62634 $171100Asian ampEmergingMarkets

$9358 $4724 $5728 $3743 $23553

TOTAL $218216 $53324 $84810 $163735 $520084

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 31: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3142

31

An Empirical Analysis of the Fedrsquos Term Auction Facility

Table 10

Banks Pledging Most Asset-Backed SecuritiesRank Bank ABS Amount (Millions) Country

1 Socieacuteteacute Geacuteneacuterale $165320 France2 Norinchukin Bank $146079 Japan3 Dexia $114297 Belgium4 Barclays $98051 UK5 UBS $94190 Switzerland6 State Street $91256 US7 Royal Bank of Scotland $82278 UK

8 Bank of Scotland $65185 UK9 US Central Federal Credit

Union$52932 US

10 Bank of Tokyo Mitsubishi $46505 Japan11 Depfa Bank $34050 Ireland12 Abbey National Treasury $31433 UK13 Bayerische Landesbank $26054 Germany14 Deutsche Bank $25900 Germany15 Landesbank Baden $25054 Germany16 WestLB $20963 UK

17 HSH Nordbank $20288 Germany18 Calyon $19047 France19 Shinkin Central Bank $18240 Japan20 DZ Bank $14965 Germany21 Skandinaviska Enskilda $14443 Sweden22 Dresdner Bank $14363 Germany23 PNC Bank $13909 US24 Natixis $13083 France25 Sumitomo $9590 Japan26 Washington Mutual $9202 US

27 Erste Bank $8841 Austria28 Standard Chartered $8693 UK29 Fortis Bank $8384 Belgium30 Royal Bank of Canada $8020 Canada31 Allied Irish $7705 Ireland32 HSBC $7610 UK33 Citibank $7608 US34 Fifth Third Bank $7368 US35 Bank of Montreal $6677 Canada36 Commerzbank $5650 Germany

37 Mizuho $5103 Japan38 Metlife $5043 US39 Sallie Mae $5030 US

(continued)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 32: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3242

CATO PAPERS ON PUBLIC POLICY

32

40 Zions First National Bank $4263 US41 RBC Bank $4175 US42 Advanta $2366 US43 Creacutedit Industriel et Commercial $2263 France44 Ally Bank $1948 US45 Mitsubishi UFJ $1921 Japan46 First Hawaiian Bank $1550 US47 Bank Hapoalim $1493 Israel48 California National Bank $1130 US49 Norddeutsche Landesbank $922 Germany50 MampT Bank $897 US

Table 10

(continued)

example as Table 9 demonstrates Citibank had the largest write-downs due to ABS CDOs borrowed against $7608 million of ABSs

compared to Socieacuteteacute Geacuteneacuterale with $163520 million and UBS with$94190 million Thus despite their exposure to ABSs and struc-tured finance assets American banks were less likely to obtain termfunding through the TAF or to pledge ABSs as collateral

52 The European Banksrsquo Dollar Crisis

Another explanation for the large number of loans made to foreign banks is that these banks suffered from a currency mismatch in their balance sheets Many foreign banks were active players in the creation

and issuance of structured finance products As money markets cameto a halt these banks required financing to meet the needs of rollingover their short-term liabilities Foreign banks were also subject to acurrency mismatch in managing their assets and liabilities Althoughthe main source of funding for some of these banks was demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies (mostly the British pound andthe euro) many European banks issued liabilities in US money mar-

kets that were denominated in the US dollar Thus not only wereforeign banks subject to a roll-over risk but they also suffered froma currency mismatch and had to rely on special facilities such as thecurrency swap lines between central banks including the European

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 33: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3342

33

An Empirical Analysis of the Fedrsquos Term Auction Facility

Central Bank Bank of England Swiss National Bank and FederalReserve as well as special lending programs such as the TAF

Foreign banks played an important role in American financial mar-kets during the years leading up to the financial crisis According toShin (2011 p 3) ldquoThe US dollarndashdenominated assets of banks outsidethe United States are comparable in size to the total assets of the UScommercial banking sector peaking at over $10 trillion prior to thecrisis The [Bank for International Settlements] banking statistics revealthat a substantial portion of external US dollar claims are the claims ofEuropean banks against US counterpartiesrdquo Likewise studies from

the Bank for International Settlements by Baba McCauley and Ramas-wamy (2009) and McGuire and von Peter (2009) show that US dollarwholesale deposits and money market funds were an important sourceof funding for European global banks in the years leading to the crisis

Moreover Shin (2011) provides evidence that European global banks raised their assets in the United States in the years leading tothe crisis increasing their claims against US borrowers by almost40 percent from 2005 to 2007 Although European banks had accessto US credit markets they still had their core funding in their homecountries in European currencies This currency mismatch betweentheir assetsmdashmany in the form of private-label ABSs and CDOsmdashand their liabilities is what made them vulnerable to the halt in USshort-term lending markets

According to this view European banks were more likely to bidfor TAF money because they were affected more severely by thefinancial crisis given their exposure to a currency mismatch betweenassets and liabilities Shin draws similar conclusions from the fact that

a large fraction of TAF lending went to European banks He writes

Two features stand out from the charts in Figure 11 The rstis that the non-US banksrsquo total borrowing is large relative toUS banksrsquo borrowing The relative magnitudes are roughlycomparable at the peak The second feature that stands outis the preponderance of European banks in the list of non-US recipients of TAF funding The UK banks are especiallyprominent led by Barclays RBS and Bank of Scotland Thelist also reveals some unlikely names such as Norinchukin

(the Agricultural Savings Bank of Japan) and the Germanlandesbanks who are likely to have ventured into US dollarlending in their search for higher yielding assets to deploytheir large domestic deposit bases (Shin 2011 17ndash18)

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 34: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3442

CATO PAPERS ON PUBLIC POLICY

34

Thus the elevated lending to foreign banks and in particularto European banks likely reflects their prominent role in the USfinancial system their involvement in the structured finance markets(especially the private-name ABSs and CDOs) and the currencymismatch in their balance sheets

6 CONCLUSION

This paper provides detailed analysis of the TAF plan usingmicro-level data on the individual loans the assets posted ascollateral and the identity of the borrowing banks I found that

foreign banks accounted for about 60 percent of TAF lending andthat the largest borrowers in the program were mostly European

banks Moreover most of the banks that pledged ABSs as collat-eral were European banks

I argue that the main reason for the large number of loans madeto foreign banks was the currency mismatch in European banksrsquo bal-ance sheets Many European banks were active players in the creationand issuance of structured finance products and as money markets

came to a halt these banks required financing to roll over their short-term liabilities These European banks also faced a currency mis-match in managing their assets and liabilities Although the mainsource of funding for some of these banks was based on demanddeposits and other forms of credit in their home countries that weredenominated in their home currencies they issued liabilities in USmoney markets that were denominated in the US dollar Thus for-eign banks not only were subject to a roll-over risk but also sufferedfrom a currency mismatch and had to rely on special facilities such

as the TAFThe data illustrate the scale of the operation of foreignmdashin

particular Europeanmdashbanks in US financial markets What preciserole do European banks play in the American economy What led totheir involvement in the US financial system These questions areleft for future research

REFERENCES

Baba Naohiko Robert N McCauley and Srichander Ramaswamy 2009 ldquoUS DollarMoney Market Funds and Non-US Banksrdquo BIS Quarterly Review March 65ndash81

Benmelech Efraim and Jennifer Dlugosz 2009 ldquoThe Credit Rating Crisisrdquo NBER Macroeconomics Annual 26 161ndash207

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 35: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3542

35

An Empirical Analysis of the Fedrsquos Term Auction Facility

Bernanke Ben 2009 ldquoThe Federal Reserversquos Balance Sheet An Updaterdquo Speech atthe Federal Reserve Board Conference on Key Developments in Monetary Policy

Washington DC October 8McAndrews James Asani Sarkar and Zhenyu Wang 2008 ldquoThe Effect of the TermAuction Facility on the London Inter-Bank Offered Raterdquo Staff Report no 335Federal Reserve Bank of New York

McGuire Patrick and Goetz von Peter 2009 ldquoThe US Dollar Shortage in GlobalBankingrdquo BIS Quarterly Review March 47ndash63

Shin Hyun Song 2011 ldquoGlobal Banking Glut and Loan Risk Premiumrdquo 2011Mundell-Fleming Lecture Conference Draft Paper presented at the 12th JacquesPolak Annual Research Conference Washington DC International Monetary Fund

Taylor John B and John C Williams 2009 ldquoA Black Swan in the Money MarketrdquoAmerican Economic Journal Macroeconomic 1 (1) 58ndash83

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 36: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3642

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 37: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3742

37

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentSimon Gilchrist

Efraim Benmelechrsquos paper presents a welcome overview andanalysis of the Federal Reserversquos Term Auction Facility (TAF) based

on the micro lending data recently made available by the FederalReserve The paper documents both the size of loans received andthe forms of collateral that were pledged It also highlights the factthat foreign banks were large recipients of TAF funds throughout thefinancial crisis

The TAF was one of many policy initiatives put in place by theFederal Reserve to combat financial market turmoil since the onset ofthe financial crisis Some of these programs represent conventional

monetary policy measures aimed directly at increasing overall mar-ket liquidity and reducing interest rates via the purchase of safe-asset securities such as Treasury bonds mortgage-backed securitiesand agency debt In the early stages of the crisis interest rates werereduced via standard open market operations combined with reduc-tions in the primary credit rate obtained through borrowing at thediscount window In later stages of the crisis with the effective Fedfunds rate at the zero lower bound the Federal Reserve conductedlarge-scale asset purchases (quantitative easing) to reduce long-term

interest rates relative to short-term interest rates It also providedextensive guidance on the future path of short-term interest rates

In contrast other programs pursued by the Federal Reserve are better viewed as unconventional in the sense that they are moreclosely linked to credit than to monetary policy These include theextension of liquidity to primary dealers (through the Term Securi-ties Lending Facility and the Primary Dealer Credit Facility) theprovision of liquidity to the private sector (through direct lending

Simon Gilchrist is professor of economics at Boston University and a research associ-ate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 38: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3842

CATO PAPERS ON PUBLIC POLICY

38

programs such as the asset-backed commercial papermoney mar-ket funds liquidity facility the commercial paper funding facility themoney market investor funding facility and the term asset-backedsecurities loan facility) and the direct bailout of financial institutionssuch as Bear-Stearns and AIG

The TAF is a straightforward extension of discount window lend-ing and therefore should be understood as part of the traditionaltool kit of the monetary authority It differed from discount windowlending in the anonymous nature of the auction which was designedto reduce any perceived stigma associated with such borrowing

It also potentially differed from discount window lending in thatthe primary credit rate was the minimum bid rate but not neces-sarily the effective borrowing rate Thus the TAF gave the FederalReserve the potential to ldquoprice-discriminaterdquo between the two formsof borrowing Indeed one of the interesting facts documented in thepaper is that the auctions were over-subscribed prior to September2008 but under-subscribed thereafter In effect prior to September2008 the Fed controlled the aggregate quantity of borrowing andlet the price adjust to market conditions Because the borrowing ratewas higher than the primary (minimum bid) rate during this period

banks effectively viewed discount window borrowing as carrying astigma Indeed the difference between the bid rate and the primaryrate provides a lower bound on the premium placed on anonymityduring this time period PostndashSeptember 2008 the Fed set the priceat the primary rate and effectively let banks borrow freely at that ratevia the TAF program

As emphasized by Benmelechrsquos paper another important aspect

of the TAF program is the amount and type of collateral that wasposted The paper documents that the amount and type is bank-specific with foreign banks more likely to post what appears to beriskier collateral Because one cannot observe the haircuts applied todifferent asset classes when determining eligible collateral it is notreally feasible to infer the risk structure of such loans based on postedcollateral The paper also suggests that the collateral requirementsmay be an important tool that the Fed could use in its conduct of

monetary policy Two reasons for caution emerge First the availablecollateral and therefore haircuts are determined in the same manneras borrowing at the discount window and both are therefore set bythe individual Reserve Banks within whose district a bank would

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 39: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 3942

39

An Empirical Analysis of the Fedrsquos Term Auction Facility

borrow It is unlikely that the Federal Reserve actively manipulatedthese requirements from one auction to the next or sought to applyhaircuts differentially across foreign versus domestic banks Secondin nearly all cases the amount of collateral posted greatly exceeds theamount borrowed Thus it is difficult to view the collateral postingsas representative of the marginal value of an additional unit of col-lateral that may vary systematically across banks Indeed differencesin the type of collateral posted across foreign versus domestic bankslikely reflect the type of assets held by the banks Thus for examplethe relatively low usage of residential mortgages as collateral by

foreign banks simply reflects the fact that these banks are not activelyengaged in the residential mortgage business in the United States

The main thrust of the paper is to highlight the importance of bor-rowing by foreign banks under the TAF program Any foreign bankthat is regulated as a foreign branch or US subsidiary is eligible to

borrow at the discount window and therefore through the TAF pro-gram As emphasized in the paper the fact that foreign banks foundit desirable to do so must to a great extent reflect the overall need fordollar funding in international money markets Such funding becameincreasingly scarce as the interbank markets ceased to function duringthe depth of the financial crisis The extent of borrowing by foreign

banks may also indicate that relative to the alternative the TAF pro-gram offered a good deal in terms of lending against collateral thatwould otherwise not be accepted in the marketplace or for that mat-ter as collateral by the European Central Bank during this time period

From this perspective it is useful to ask what are the consequencesand policy tradeoffs associated with the TAF program The Federal

Reserve clearly decided that providing dollar funding was an impor-tant tool in its tool kit during the crisis Indeed the provision of dollarfunding through swap lines to the European Central Bank and othercentral banks accounted for the largest share of the Federal Reserversquos

balance sheet at the height of the crisis Are swap lines the desirablealternative To the extent that haircuts were not sufficient the heavyuse of the TAF program by foreign banks likely exposed the FederalReserve to additional risk from European banks Swap lines reduce

the Federal Reserversquos exposure to foreign-bank risks but increase itsexposure to sovereign risk The costs and benefits of such a tradeoffare highly relevant to policymakers in todayrsquos environment wheremarkets perceive little difference between bank and sovereign risk

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 40: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4042

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 41: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4142

41

An Empirical Analysis of the Fedrsquos Term Auction Facility

CommentRoss Levine

During an extraordinary period in the economic and financialhistory of the United States a period when the familiar tools of

monetary policy did not function the Federal Reserve developedand implemented new procedures for addressing the liquidity prob-lems plaguing banks In particular by the end of 2007 banks had

become worried about the creditworthiness of other banks uncertainabout their own ability to borrow in the future and consequentlyexceedingly reluctant to lend to other banks This breakdown ininterbank lending disrupted the normal functioning of open marketoperations which relies on banks that sell securities to the Federal

Reserve to then lend excess funds to other banks instead of simplyaccumulating excess reserves at the FedWith one monetary policy tool malfunctioning the Fed attempted to

employ another traditional tool lending funds directly to banks throughthe discount window But the stigma associated with borrowing fromthe Fedmdashthe view that only weak banks use the discount windowmdashmeant that banks were disinclined to use the discount window hin-dering the efficacy of this monetary policy tool as well Thus the Fedfaced a challenge it viewed the burgeoning financial crisis as emanat-

ing from the liquidity problems plaguing banks but its traditional toolsfor addressing the problem did not work So it created new tools

One of the first new tools that the Fed developed to ease liquid-ity problems was the Term Auction Facility (TAF) which started inDecember of 2007 and ceased operations in March of 2010 At its peakthe TAF was almost a $500 billion item on the Fedrsquos balance sheetUnder the TAF the Fed would choose a quantity of money to auction to

banks All banks that were eligible to borrow under the Fedrsquos traditional

Ross Levine is the Willis H Booth Chair in Banking and Finance in the Haas School ofBusiness at the University of California Berkeley a senior fellow at the Milken Insti-tute and a research associate of the National Bureau of Economic Research

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242

Page 42: An Empirical Analysis of the Term Auction Facility

8132019 An Empirical Analysis of the Term Auction Facility

httpslidepdfcomreaderfullan-empirical-analysis-of-the-term-auction-facility 4242