drivers of credit losses in australasian banking slides prepared by kurt hess university of waikato...

Drivers of Credit Losses in Australasian Banking

Slides prepared by Kurt Hess

University of Waikato Management School, Department of Finance

Hamilton, New Zealand

Apr 19, 2023 Kurt Hess, WMS [email protected]

2

MotivationLiterature reviewCredit loss data AustralasiaMethodological issuesResultsConclusions

Topics


3

Motivation Stability and integrity of banking

systems are of utmost importance to national economies

Credit losses, or more generally, asset quality problems have repeatedly been identified as the ultimate trigger of bank failures [e.g. in Graham & Horner (1988), Caprio & Klingebiel (1996)]


4

Motivation Prudential supervisory agencies need to

understand drivers of credit losses in banking system– Validation of proprietary credit risk models

& parameters under Basel II This is the first specific research of long

term drivers of credit losses for Australian banking system


5

Literature review

1. Literature with regulatory focus looks at macro & micro factors

2. Literature looks discretionary nature of loan loss provisions and behavioural factors which affect them

Two main streams of research that analyse drivers of banks’ credit losses (or more specifically loan losses):


6

Literature review

Behavioural hypotheses in the literature on the discretionary nature of loan loss provisions

– Income smoothing:Greenawalt & Sinkey (1988)

– Capital management: Moyer (1990)– Signalling: Akerlof (1970), Spence (1973)– Taxation Management


7

Literature review Studies with global samples (using

commercial data providers):– Cavallo & Majnoni (2001),

Bikker & Metzemakers (2003)

Country-specific samples– Austria: Arpa et al. – (2001)– Italy: Quagliarello (2004)– Australia: Esho & Liaw (2002)

(in this APRA report the authors study level of impaired assets for loans in Basel I risk buckets for 16 Australian banks 1991 to 2001)


8

Literature review Research based on original published

financial accounts is rare (very large effort to collect data).– Pain (2003): 7 UK commercial banks &

4 mortgage banks 1978-2000– Kearns (2004):

14 Irish banks, early 1990s to 2003– Salas & Saurina (2002): Spain


9

Credit Loss Data Australasia

The database includes extensive financial and in particular credit loss data for

– 23 Australian + 10 New Zealand banks– Time period from 1980 to 2005– Approximately raw 55 data elements per

institution, of which 12 specifically related to the credit loss experience (CLE) of the bank


10


Sample selection criteria Registered banks Must have substantial retail and/or

rural banking business Exclude pure wholesale and/or

merchant banking institutions


11

Credit Losses and GDP Growth (New Zealand Banks)

Total Stock of Loan Loss Provisions as % of Loan Assets(excluding BNZ and Rural Bank)

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

Net Write offs/ Avg Loans(excl BNZ, Rural Bk)

Charge to P&L/ Avg Loans(excl BNZ, Rural Bk)

-5.0%

0.0%

5.0%

10.0%

GDP YoY% Real

Annual Debt Write-offs and Charges to P&L as % of Loan Assets(excluding BNZ and Rural Bank)

0.0%0.2%0.4%0.6%0.8%1.0%1.2%1.4%1.6%

19

80

19

81

19

82

19

83

19

84

19

85

19

86

19

87

19

88

19

89

19

90

19

91

19

92

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

Net Write offs/ Avg Loans(excl BNZ, Rural Bk)

Charge to P&L/ Avg Loans(excl BNZ, Rural Bk)

Provisioning/write-off behaviour correlated to macro factors

Note: chart for NZ Bank sub-sample only


12


AU ANZ, 1993, 2.6%

AU CoWthBk, 1993, 2.5%

AU Westpac, 1993, 3.7%

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

1980 1985 1990 1995 2000

AU ANZ

AU CoWthBk

AU NAB

AU Westpac

(on average loans, annualized)


Methodology


14

Principal Model

1,...., ;1,.....,

;q

1s)(

1k 0s)(

Tqtni

uCLExConstCLE itstis

K z

stkiksit

k

CLEit Credit loss experience for bank i in period txkit Observations of the potential explanatory variable k for bank i

and period tuit Random error term with distribution N(0,), Variance-covariance matrix of it error termsn Number of banks in sampleT Years in observation periodK Number of explanatory variableszk Maximum lag of the explanatory variable k of the modelq Maximum lag of the dependent variable of the model


15

Principal Model

Principal model on previous slide allows for many potential functional forms.

There are choices with regard to– Dependent CLE proxy– Suitable drivers of credit losses and lags

for these drivers– Estimation techniques


16

Determinants of Credit LossesMacro Factors (1)

Real GDP growth -ve Ability of borrowers to service debt determined by the economic cycle.

Unemployment rate

+ve Unemployment rate not only reflects the business cycle (like GDP growth) but also longer term and structural imbalances in economy.

Liabilities of households/firms as % of disp. income

+ve The more households and firms in the system are indebted, the more financially vulnerable they will be.


17

Determinants of Credit LossesMacro Factors (2)

Asset prices / interest rates

Disturbances in the asset markets can impair the value of banks’ assets both directly and indirectly (i.e. through reduced collateral values). Experience shows that especially the property sector and the share markets may play a critical role in triggering losses in the banking system. Similar effects are expected in a volatile interest rate environment.

Housing price index (changes)

-ve

Return leading share indices

-ve

Change real/nominal interest ratesonly CPI growth

+ve


18

Determinants of Credit LossesBank Specific Factors (1)

Past credit expansion

+veor-ve

Fast growth of the loan portfolio is often associated with subsequent loan losses. Alternatively, a slow growing loan portfolio may be caused by a weak economy and thus increase CLE.


19


Pricing of risks( net interest margins) +ve/

(-ve)

A bank’s deliberate choice to lend to more risky borrowers is likely reflected in higher interest margins. Lower past margins might induce greater risk-taking by bank

Characteristic of lending portfolio(share of housing loans)

-veThe share of comparably lower risk housing loans as % of loans proxies the risk characteristic of the bank’s loan portfolio.


20

Determinants of Credit LossesBank Specific Factors (3)Diversification

Market power

-ve

+ve/(-ve)

A bank’s assets in proportion to the overall banking system assets provides a crude proxy for loan portfolio diversification or market power

Cost efficiency(cost-income ratio)

+ve/(-ve)

Inefficient banks can be expected to suffer greater credit losses. Alternatively, such banks could maintain an expensive credit monitoring procedure and will thus exhibit lower credit losses.


21


Income smoothing(Earnings before provisions & taxes as % of assets)

+veSome literature finds evidence of banks using discretionary provisions to smooth earnings for a variety of motivations.

Capital management(Capital measured as tier 1 or tier 1+2 capital as % of risk weighted assets)

-ve General provisions count towards Basel I minimum capital and weaker banks might thus be tempted to engage in capital management through provisioning.


22

Pooled regression model as per equation 1 in paper Dependent

– Impaired asset expense as CLE proxy Determinants (as per table next slide)

– Alternative macro factors: GDP growth, unemployment rate

– Alternative asset shock proxies: share index, house prices

– Misc. bank-specific proxies – Bank past growth


23

Dependent variables in model

Variable Description Lags (yrs.)

GDPGRWUNEMP

Macro state proxies: GDP growth or level/change Unemployment rate

0 to -2

RET_SHINDX HPGRW

Asset price shock proxies: Return share index or change house prices

0 to -2

CPIGRW Change CPI 0 to -2

SH_SYSLNS Share of system loans (size proxy) 0

NIM Net interest margin 0 to -2

CIR Cost-income ratio 0 to -2

EBTP_AS Pre-provision/tax earnings / assets 0 to -2

ASGRW Bank past asset growth 0 to -4

Agg

rega

teB

anks

peci

fic


Empirical results


25

Results macro state factors

GDP growth (GDPPGRW), change and level of the unemployment rate (UNEMP, UNEMP) have expected effect (not all lags significant)

Unemployment with best explanatory power for overall sample

see Table 8, 9,10 in papersee Table 8, 9,10 in paper


26

Results macro state factors (2)

Country-specific differences between Australia and New Zealand– Australia’s results show much greater

sensitivities to GDP growth (see Table 9)– New Zealand results are less significant

and effects of GDP and UNEMP seem more delayed



27

Results asset price factors

Contemporaneous coefficient of share index return negative & significant for overall and Australia. Less significant for NZ.

Housing price index has less sigificanceIntuition: early 90s crises not rooted in particular problems of the housing sector



28

Results CPI growth

Positive, but not significant coefficients for most regressions, i.e. inflationary pressure tends to lift credit losses

Contemporaneous term negative and significant for Australian sub-sample, in line with evidence elsewhere that inflation may lead to temporary improvement of borrower quality (Tommasi, 1994)



29

Results size proxy

Higher level of provisioning for larger banks – no significance of coefficients, however

Intuition: portfolios of smaller institutions often dominated by (comparably) lower risk housing loans



30

Results net interest margin

Generally negative, contemporaneous and 2yr lagged term significant, i.e. – Lower past margins lead to higher subsequent

losses (induce risk taking)– Difficult to explain contemporaneous negative

term Inconclusive results also in comparable

studies, e.g. Salas & Saurina (2002) for Spain



31

Results net interest margin (2)

Endogenous nature of net interest margins as postulated by Ho & Saunders (1981) dealership model. Spread increases with …– Market power (inelastic demand)– Bank risk aversion– Larger size of transactions (loans/deposits)– Interest rate volatility

Net interest margins may thus control for other bank specific & market characteristics



32

Results cost efficiency (CIR)

High and increasing cost income ratios are associated with higher credit losses

Results reject alternative hypothesis that banks are inefficient because they spend to much resources on borrower monitoring

Not surprising as “gut feel” would tell that excessive monitoring might not pay



33

Results earnings proxy

Very clear evidence of income smoothing activities, i.e. banks increase provisions in good years, withhold them in weak years.

Confirms similar results found in many other studies



34

Results past bank growth

Clear evidence of the fast growing banks faced with higher credit losses in future (lags beyond 2 years)

Managers seem unable (or unwilling) to assess true risks of expansive lending

Much clearer results than in other studies. Possibly due to test design with longer lags considered.



35

Conclusions Model presented here is very suitable

for assessing general / global effects on impaired assets in the banking sector

The dynamics of this transmission seems to differ among systems

A study of particular effects might thus call for alternative models


36

Conclusions (2) Income smoothing is a reality, possibly

also with new tighter IFRS provisioning rules as this ultimately remains a discretionary managerial decision


37

Conclusions (3) Use data base for comparative studies

of alternative CLE dependent variables

First results show that they (in part) correlate rather poorly which means there must be caution comparing results of studies unless CLE is defined in exactly the same way

Credit Loss Experience of Australasian Banks

Back-up Slides


39

Basel II PillarsPillar 1:

– Minimum capital requirementsPillar 2:

– A supervisory review processPillar 3:

– Market discipline (risk disclosure)


40

Pages in New Basel Capital Accord (issued June 2004)

Basel II Pillars

Pillar 3 Market Discipline:16 of 216

pages

General: 6 of 216 pages

Pillar 2 Supervisory

Review Process: 15 of 216

pages

Pillar 1 Minimum

Capital Requirements

179 of 216 pages


41

Pro Memoria: Calculation Capital Requirements under Basel II

Total CapitalCredit Risk + Market Risk + Operational Risk 8%

SignificantlyRefined

New

Unchanged

(Could be set higher under pillar 2)

RelativelyUnchanged

Source: slide inspired by PWC presentation slide retrieved 27/7/2005 from http://asp.amcham.org.sg/downloads/Basel%20II%20Update%20-%20ACC.ppt ,


42

Basel II – IRB Approach

Two approaches developed for calculating capital minimums for credit risk:

Standardized Approach (essentially a slightly modified version of the current Accord)

Internal Ratings-Based Approach (IRB)– foundation IRB - supervisors provide some inputs– advanced IRB (A-IRB) - institution provides inputs


43

Basel II – IRB Approach

Internal Ratings-Based Approach (IRB)– Under both the foundation and advanced

IRB banks are required to provide estimates for probability of default (PD)

– It is commonly known that macro factor are the main determinants of PD


44

Primer Loan Loss AccountingBeginning of period Transactions during period End of period

Profit & loss statement (P&L) - Bad debt charge

Provision accountLoan balance Provisons initial balance Loan balanceGross loan amount + New provisions made Gross loan amount

- Debt write-offs + Recovery of debt previously written off

Net loan amount Provisons final balance Net loan amount

Gross loan accountOpening balance -/+ Loans issued/repaid - Debt write-offs + Recovery of debt previously written offEnding balance

- Provisions initial balance

- Provisions final balance


45

Primer Loan Loss Accounting

Potential loan loss identified

Initiation of loan

Loan account

Cash account

Bad debt provision expense

1,000 +50950

1,000

General provision recognized

50

1,000 50 +350600

Additional specif ic provisons

Loan account

Bad debt provision expense

350

Loan write-off(derecognition)

1,000 400 - 400 - 400

600

Loan recovery

600 + 700 + 100

-

700

Cash account

Bad debt provisionrecovery income

- 100

Loan account Loan account


46


1984

1986

1988

1990

1993

18m

1995

1997

1999

2001

(200)0

200400600800

1,0001,2001,400

NZ$ million

Bad debt charge to P&L

Net write-offs

BNZ books bad debt credits 1994-1997

BNZ 1984 - 2002


47

Credit Loss Data AustralasiaBanks in sample

AUSTRALIA: Adelaide Bank, Advance Bank, ANZ, Bendigo Bank, Bank of Melbourne, Bank West, Bank of Queensland, Commercial Banking Company of Sydney, Challenge Bank, Colonial State Bank, Commercial Bank of Australia, Commonwealth Bank, Elders Rural Bank, NAB, Primary Industry Bank of Australia, State Bank of NSW, State Bank of SA, State Bank of VIC, St. George Bank, Suncorp-Metway, Tasmania Bank, Trust Bank Tasmania, WestpacNEW ZEALAND: ANZ National Bank, ASB, BNZ, Countrywide Bank, NBNZ, Rural Bank, Trust Bank NZ, TSB Bank, United Bank, Westpac (NZ)


48


Data issues Macro level statistics

– Differing formats between NZ and Australiae.g. indebtedness of households / firms

– House price series back to 1986 only for Australia

– Balance sheets of M3 institutions only back to 1988 for New Zealand (use private sector credit statistics instead)


49


Data issues (2) Micro / bank specific data

– Lack of reporting limits choice of proxies(particularly through the very important crisis time early 1990)

– Comparability due to inconsistent reporting(e.g. segment credit exposures)


50

Measuring CLE

Dedicated nature of database allows tests for many proxies for a bank’s credit loss experience (CLE)

– Level of bad debt provisions, impaired assets, past due assets

– Impaired asset expense (=provisions charge to P&L)

– Write-offs (either gross or net of recoveries)– Components of above proxies, e.g. general or

specific component of provisions (stock or expense)


51

Med

ian

-0.2

5 S

tDan

d le

ss

Med

ian

-0.2

5 S

td to

Med

ian

Med

ian

to

Med

ian

+0.

25 S

tD

Med

ian

+0.

25 S

tD to

Med

ian

+0.

5 S

tD

Med

ian

+0.

5 S

tD to

Med

ian

+0.

75 S

tD

Med

ian

+0.

75 S

tD to

Med

ian

+1

StD

Med

ian

+1

StD

toM

edia

n +

1.25

StD

Med

ian

+1.

25 S

tD to

Med

ian

+1.

5 S

tD

Med

ian

+1.

5 S

tD to

Med

ian

+1.

75 S

tD

Med

ian

+1.

75 S

tD to

Med

ian

+2

StD

Med

ian

+2

StD

and

mor

e

Impaired assets / assets

Stock of provisions / loans

Net write-offs / loans

Imp. asset exp. / gross interest

Imp. asset expense / net interest

Impaired asset expense / loans

0

50

100

150

200

250

Measuring CLE

Med

ian

Histogram of selected CLE proxies

Pooled observations of Australian and NZ Banks 1980 - 2005

Credit Loss Experience of Australasian Banks

Selected References


53

Selected ReferencesBikker, J. A., & Metzemakers, P. A. J. (2003).

Bank Provisioning Behaviour and Procyclicality, De Nederlandsche Bank Staff Reports, No. 111.

Caprio, G., & Klingebiel, D. (1996). Bank insolvencies : cross-country experience. Worldbank Working Paper WPS1620.

Cavallo, M., & Majnoni, G. (2001). Do Banks Provision for Bad Loans in Good Times? Empirical Evidence and Policy Implications, World Bank, Working Paper 2691.


54

Selected ReferencesEsho, N., & Liaw, A. (2002). Should the

Capital Requirement on Housing Lending be Reduced? Evidence From Australian Banks. APRA Working Paper(02, June).

Graham, F., & Horner, J. (1988). Bank Failure: An Evaluation of the Factors Contributing to the Failure of National Banks, Federal Reserve Bank of Chicago.


55

Selected ReferencesKearns, A. (2004). Loan Losses and the

Macroeconomy: A Framework for Stress Testing Credit Institutions’ Financial Well-Being, Financial Stability Report 2004. Dublin: The Central Bank & Financial Services Authority of Ireland.

Pain, D. (2003). The provisioning experience of the major UK banks: a small panel investigation. Bank of England Working Paper No 177, 1-45.


56

Selected References

Salas, V., & Saurina, J. (2002). Credit Risk in Two Institutional Regimes: Spanish Commercial and Savings Banks. Journal of Financial Services Research, 22(3), 203 - 224.

drivers of credit losses in australasian banking slides prepared by kurt hess university of waikato...

Documents

drivers of banks credit

particular credit loss

bank slide

spain slide

new zealand slide

australian banks

literature review research

credit loss experience