recent arguments against the gold standard

7/28/2019 Recent Arguments against the Gold Standard

1/28

Executive Summary

The presidential primary contests of 201112 brought renewed attention to the idea ofreinstituting a gold standard. The 2012 Re-publican Party platform ultimately included aplank calling for the creation of a commissionto study the issue.

The favorable attention given to the idea ofreinstituting a gold standard has attracted criti-cism of the idea from a variety of sources. Con-

sidered here are the most important argumentsagainst the gold standard that have been madeby economists and economic journalists in re-cent years.

A few recent arguments are novel to some ex-tent, but not all add weight to the case againsta gold standard. Several authors identify genu-ine historical problems that they blame on thegold standard when they should instead blamecentral banks for having contravened the goldstandard.

Gold standards, being real-world human in-

stitutions, fall short of perfection. No doubt a

well-trained academic economist can describeon the whiteboard an idealmonetary system thatproduces greater stability in the purchasing pow-er of money than a gold standard doesor scoreshigher on whatever one criterion the economistfavorswhile sparing us a gold standards re-source costs by employing fiat money. But otherwell-trained economists have proposed differentcriteria, and even a flawless central bank cannot

pursue all criteria with one policy.More important, fiat standards in practice

have been far from perfect monetary systems.We need to examine historical evidence if wewant to come to an informed judgment aboutwhether actualgold-based systems or actualfiat-based systems display the smaller set of flaws.I find that the most automatic and least man-aged kind of gold-based systema gold stan-dard with free bankingcan be expected to out-perform a gold standard with central bankingand to outperform the kind of fiat monetary

systems that currently prevail.

Recent Arguments against the Gold Standard

by Lawrence H. White

No. 728 June 20, 2013

Lawrence H. White is a professor of economics at George Mason University and an adjunct scholar with theCato Institute.


2/28

2

We needto examine

historicalevidence if we

want to cometo an informed

judgment aboutwhetheractual

gold-basedsystems or

actualfiat-basedsystems display

the smaller set offlaws.

Introduction

The presidential primary contests of201112 brought renewed attention to theidea of reinstituting a gold standard. At least

four candidates spoke favorably about thegold standard. One suggested a commissionon gold to look at the whole concept of howdo we get back to hard money. The 2012 Re-publican Party platform ultimately includeda plank calling for the creation of just such acommission, explicitly viewing it as a sequelto the U.S. Gold Commission of 1981: Now,three decades later . . . , we propose a similarcommission to investigate possible ways toset a fixed value for the dollar.1

The favorable attention given to the idea

of reinstituting a gold standard has attractedcriticism of the idea from a variety of sources.In the popular press,Atlanticwriter MatthewOBrien has expounded on Why the GoldStandard Is the Worlds Worst EconomicIdea,2 while Washington Post columnist EzraKlein has declared that The problems withthe gold standard are legion.3 On the morescholarly side, Federal Reserve Chairman andformer Princeton economics professor BenBernanke, guest lecturing at George Wash-ington University on the history of monetary

policy in the United States, in the words ofthe New York Times account, framed muchof this history as a critique of the gold stan-dard, which was dropped in the early 1930sin a decision that mainstream economistsregard as obviously correct, hugely beneficialand essentially irreversible.4 Well-knownUniversity of CaliforniaBerkeley economistBarry Eichengreen has offered A Critique ofPure Gold.5

In a Briefing Paper published by the CatoInstitute, I addressed a number of then-

common theoretical and historical objec-tions to a gold standard, sorting those thathave some substance from those that are ill-founded.6 Here I consider the most impor-tant arguments against the gold standardthat have been made by economists and eco-nomic journalists since then. Some of theless-substantial arguments that I criticized

in 2008 reappear in the recent literatureOther arguments are novel to some extent,but not all add weight to the case against agold standard. Several authors identify gen-uine historical problems that they blame on

the gold standard, when they should insteadblame central banks for having contravenedthe gold standard.

Bernanke told the students at GeorgeWashington University, Unfortunately goldstandards are far from perfect monetary sys-tems.7 We can all agree that gold standards,being real-world human institutions, fallshort of perfection. There is no doubt that awell-trained academic economist can describeon the whiteboard an ideal monetary systemthat, through the flawlessly timed and flaw-

lessly calibrated policy actions of a centralbank, produces greater stability in the pur-chasing power of money than a gold standarddoesor scores higher on whatever one crite-rion the economist favorswhile sparing us agold standards resource costs by employingfiat (noncommodity) money.8 But other well-trained economists have proposed differentcriteria, and even a flawless central bank can-not pursue all criteria with one policy.

More important, fiat standards in prac-tice have been far from perfect monetary sys-

tems. We need to examine historical evidenceif we want to come to an informed judgmentabout whether actualgold-based systems oractualfiat-based systems display the smallerset of flaws. We need to recognize the varietyof institutional arrangements that the worldhas seen under gold standards and likewiseunder fiat standards. In particular, we needto distinguish an automatic gold-stan-dard systemlike the classical gold standardin countries without central banksfromthe interwar gold-exchange system that was

managed or mismanaged by the discretionof central bankers. I find that the most auto-matic and least managed kind of gold-basedsystema gold standard with free bank-ingcan be expected to outperform a goldstandard with central banking, and to out-perform the kind of fiat monetary systemsthat currently prevail.


3/28

3

Gold standardshave historicallyrequired onlyfractionalreservesthat is,the holding ofenough gold toback only a smalportion of the

money supply.

What follows are critical analyses of theleading recent arguments against a goldstandard. I spell out each argument as crit-ics have made it, and evaluate its logicaland historical merits. I begin with the least

substantial arguments, and proceed to theweightier.

Claim 1: There Isnt Enough Gold toOperate a Gold Standard Today

Personal finance columnist John Wag-goner recently claimed in USA Today thattheres not enough gold in the world to re-turn to a gold standard.9 He explained:

In the gold standard, the amount ofcurrency issued is tied to the govern-

ments gold holdings. The price ofgold would have to soar to accommo-date U.S. trade in goods and services.. . . Total gold owned by the [UnitedStates] governmentincluding theFederal Reserve and the U.S. Mintis 248 million ounces. Thats about$405 billion dollars at todays prices,hardly enough to support a $15 tril-lion economy.

The government could use a kind

of semi-gold standard, limiting theamount of money printed to a per-centage of its gold reserves. For exam-ple, it could say that at least 40% ofall currency outstanding be backedby gold. This would limit the moneysupply, but be vulnerable to govern-ment manipulationrevising the lim-it downward to 5%, for example.

Waggoners figures of 248 million ouncesand $405 billion are approximately correct,

but his claim that the price of gold wouldhave to soar to make that an adequate stockof gold reserves is not. The August 31st Sta-tus Report of U.S. Treasury-Owned Goldputs theU.S. governments total holdings at 261.5million ounces.10 (The source of Waggonerslower figure is unclear.) At a market priceof $1,700 per fine troy ounces (to choose

a recently realized round number), thoseholdings are worth $444.6 billion. Currentrequired bank reserves (as of October 2012)are less than one fourth as large, $107.3 bil-lion. Looked at another way, $444.6 billion

is 18.4 percent of the current money supplymeasure M1 ($2,417.2 billion as of Oc-tober 22), which is the sum of currency incirculation and checking-account balances.That is a more than healthy reserve ratio byhistorical standards.11

Waggoner labors under several miscon-ceptions. First, gold standards have histori-cally required only fractional reservesthatis, the holding of enough gold to back onlya small portion of the money supply. So longas banks or the government can satisfy the

actual demand of conversion of money togold, fractional reserves do not make a goldstandard into a kind of semi-gold stan-dard. Second, it is not generally true thatthe amount of currency issued is tied to thegovernments gold holdings. It is true onlyif the government monopolizes the issue ofgold-redeemable currency and the holdingof gold reserves, but history offers 60-plusexamples of competitive private-note issueunder historical gold and silver standards.12Third, the vulnerability of the average reserve

ratio to government manipulation is not in-evitable. It can be avoided by leaving com-mercial banks to determine their own reserveratios, as in historical free banking systems.

Claim 2: The Gold Standard Is anExample of Price-fixing by Government

Barry Eichengreen writes that countriesusing gold as money fix its price in domes-tic-currency terms (in the U.S. case, in dol-lars). He finds this perplexing:

But the idea that government shouldlegislate the price of a particularcommodity, be it gold, milk or gaso-line, sits uneasily with conservativeRepublicanisms commitment to let-ting market forces work, much lesswith Tea Partyesque libertarianism.Surely a believer in the free market


4/28

4

The worldpurchasing

power of goldwas better-

behaved underthe classical

internationalgold standard

than thepurchasing powerof fiat money has

been since 1971.

would argue that if there is an increasein the demand for gold, whateverthe reason, then the price should beallowed to rise, giving the gold-min-ing industry an incentive to produce

more, eventually bringing that priceback down. Thus, the notion that theU.S. government should peg the price,as in gold standards past, is curious atthe least.13

To describe a gold standard as fixing goldsprice in terms of a distinct good, domesticcurrency, is to begin with a confusion. Agold standard means that a standard massof gold (so many troy ounces of 24-karatgold) defines the domestic currency unit. The

currency unit (dollar) is nothing other thana unit of gold, not a separate good with apotentially fluctuating market price againstgold. That $1, defined as so many ouncesof gold, continues to be worth the specifiedamount of goldor, in other words, that xunits of gold continue to be worth x unitsof golddoes not involve the pegging ofany relative price. Domestic currency notes(and checking-account balances) are denomi-natedin and redeemable for gold, not pricedin gold. They dont have aprice in gold any

more than checking account balances in ourcurrent system, denominated in fiat dol-lars, have a price in fiat dollars. PresumablyEichengreen does not find it curious or ob-jectionable that his bank maintains a fixeddollar-for-dollar redemption rate, cash forchecking balances, when he withdraws cashat its automatic teller machine.

As to what a believer in the free marketwould argue, surely Eichengreen under-stands that if there is an increase in the de-mand for gold under a gold standard, what-

ever the reason, then the relative price of gold(the purchasing power per unit of gold overother goods and services) will in fact rise,that this rise will in fact give the gold-min-ing industry an incentive to produce more,and that the increase in gold output will infact eventually bring the relative price backdown.14

Claim 3: The Volatility of the Price ofGold Since 1971 Shows that Gold WouldBe an Unstable Monetary Standard

Eichengreen argues that golds inherentprice volatility makes it unsuitable to pro-

vide a basis for international commercialand financial transactions on a twenty-first-century scale.15

Klein declares, The problems with thegold standard are legion, but the most obvi-ous is that our currency fluctuates with theglobal price of gold as opposed to the needsof our economy.16 It is not entirely clearwhat our currency fluctuates with the glob-al price of gold means in this declaration. Ifit means that, for a country that is part of aninternational gold standard, the purchasing

power of domestic currency moves with theworld purchasing power of gold, then it istrue, but it fails to identify a problem. Theworld purchasing power of gold was better-behaved under the classical internationalgold standard than the purchasing power offiat money has been since 1971. If it meansto invoke the volatility of the real or dollarprice of gold since gold was fully demone-tized in 1971, it identifies a problem, but itis a problem experienced under a fiat stan-dard and notunder a gold standard. Today

demonetized gold rises and falls in price assavers and investors rush into and out ofgold as a hedge against fiat-money inflation

The respected University of CaliforniaSan Diego economist and blogger James DHamilton makes an argument that is lessambiguous, but puzzling nonetheless. Ham-ilton charts how much the average dollarwage would have varied if it was initially fixedin ounces of gold but instead was paid in thedollar equivalent as the price of gold variedbetween January 2000 and July 2012.17 He

observes that if the real value of gold hadchanged as much as it has since then, thedollar wage that an average worker receivedwould need to have fallen from $13.75/hourin 2000 to $3.45/hour in 2012. That soundsalarming, but in fact it is of very little signifi-cance. It is relevant only if the behavior of thereal value (purchasing power) of gold is in-


5/28

5

It makes littlesense to attributvolatility in thereal price of gold

to the growth indemand fromsteadily risingincomes.

dependent of the monetary regime so that thepurchasing power of gold-backed currencywould fluctuate on the world market. Such acalculation would be relevant if a small openeconomy (say, the Bahamas) should unilat-

erally adopt the gold standard today. Thatwould indeed be a bad idea.18 But thoughtfuladvocates of the gold standard propose thatit should again be an international standard.Hamiltons calculation is completely irrel-evant to that proposal. A Lucas critique ap-plies: observations drawn from a world of fiatregimes are not informative about the behav-ior of the purchasing power of money underan international gold standard.

Hamilton anticipates such an objectionand has a reply ready:

[G]old advocates respond with theclaim that if the U.S. had been on agold standard since 2000, then thehuge change in the real value of goldthat we observed over the last decadenever would have happened in the firstplace. The first strange thing about thisclaim is its supposition that events andpolicies within the U.S. are the mostimportant determinants of the realvalue of gold. According to the World

Gold Council, North America accountsfor only 8% of global demand.19

This, too, is irrelevant to the evaluation ofproposals for an international gold stan-dard. By the way, Hamiltons 8 percent fig-ure is North Americas share of global pur-chases of new gold jewelry, a nonmonetaryand flow measure, rather than its shareof the stock transactions demand to holdmonetary gold, which under an interna-tional gold standard would presumably be

closer to North Americas 30 percent shareof world output.

The purchasing power of money wasmore stable under the classical interna-tional gold standard (18791914) than ithas been under fiat money standards since1971. In a blog entry a few days after theone just quoted, Hamilton recognizes this

fact: It is true that the biggest concern Ihave about going back on a gold standardtodaythat it would tie the monetary unitof account to an object whose real value canbe quite volatilewas not the core problem

associated with the system of the 19th cen-tury. He then continues: But the fact thatthis wasnt the core problem with the goldstandard in the nineteenth century does notmean that it wouldnt be a big problem if wetried to go back to the system in the twenty-first century.20

But its unlikely that purchasing-power in-stability would be any more of a problem fora present-day international gold standard.Hamilton attributes recent movements inthe real value of gold to the surge in income

from the emerging economies rather thanU.S. monetary policy, citing data showingglobal gold jewelry sales up strongly in 2010over 2009, led by large increases in sales to In-dia, Hong Kong, and mainland China.21 It isreasonable to suppose that demand for goldjewelry rises with income. But real incomein India and China is rising fairly steadily. Itmakes little sense to attribute volatility in thereal price of gold to the growth in demandfrom steadily rising incomes.

Hamiltons drawing of a trend from two

data points, moreover, is not a careful read-ing of the data source he cites. Even if wefocus exclusively on 2010 over 2009, only asmall fraction of the extraordinary increaseof 69 percent in gold jewelry sales to Indiacan possibly be attributed to Indias real in-come growth, which was 10 percent that yearaccording to the International MonetaryFund. The income-elasticity of demand forgold jewelry is nothing like 6.9 if we observelonger-run trends. The text of the article con-taining the data provides a clue to the lions

share of that one years increase: Histori-cally savvy gold buyers, Indias influx of buy-ing implies an expectation that gold pricesstill have much higher to go. The [WorldGold Council] says that Indian consumersappeared almost universally to expect thatthe local gold price was likely to continue ris-ing.22 That is, Indians did not buy so much


6/28

6

Inflation-hedging demand

is volatilebecause inflation

expectations arevolatile under

unanchoredmonetary

systems.

gold jewelry in 2010 just for ornamentation,but also as an investment or inflation hedge.Likewise, the article notes, many in Chinasmiddle class are looking to gold as a meansfor long-term savings and a possible hedge

against inflation.If we look at additional years of the data, wesee that global gold jewelry sales in 2010 weredown from the levels of 2007 or 2008, whichis hardly consistent with the hypothesis thatgold demand is rising mainly due to risingemerging-economy income. If we look at thearticles entire 200410 range of sales data forgold in allforms, we see as much or more vola-tility in investment sales of gold (bars, coins,medallions, exchange-traded funds) as in jew-elry sales. Absent fiat inflation hedging, there

is little cause for concern about the volatilityof demand for gold or golds real price.

Like Hamilton, the respected George Ma-son University economist and blogger TylerCowen23 also expresses concern about vola-tility in the real price of gold:

Why put your economy at the mercyof these essentially random forces? Ibelieve the 19th century was a rela-tively good time to have had a goldstandard, but the last twenty years,

with their rising commodity prices,would have been an especially badtime. When it comes to the next twen-ty years, who knows?

In a later blog entry, Cowen adds, I think agold standard today would be much worsethan the 19th century gold standard, in partbecause commodity prices are currentlymore volatile and may be for some time.24

Cowen does not directly address the pos-sibility that the current volatility of several

commodity price series, most importantlythat of gold, is principally caused by theinflation-hedging prompted by our currentfiat monetary systems. Inflation-hedgingdemand is volatile because inflation expec-tations are volatile under unanchored mon-etary systems. Inflation-hedging involvesother commodities in addition to gold and

silver. Under a reliably anchored monetarysystem this source of commodity price vola-tility would disappear.

The answer to Cowens first questionwhy put your economy at the mercy of es-

sentially random supply and demandshocks for gold?is that, to judge by thehistorical evidence, doing so engenders lessvolatility than the alternative of putting youreconomy at the mercy of a central banksmonetary policy committee. Monetary sup-ply and demand shocks under fiat moneysystems have been much larger. Underthe classical gold standard, changes in thegrowth rate of the base money stock wererelatively smallperhaps surprisingly smallto those who havent looked at the numbers

The largestsupply shock, the California GoldRush, caused a cumulative world price levelrise of 26 percent (as measured by the UnitedKingdoms Retail Price Index) stretched over18 years (184967), which works out to aninflation rate of only 1.3 percent per annumAs Cowen recognizes, gold discoveries thesize of Californias are hardly likely today.25

Barry Eichengreen also worries that vola-tility in the demand for gold would persisteven in an international gold standard:

There could be violent fluctuationsin the price of gold were it to againbecome the principal means of pay-ment and store of value, since thedemand for it might change dramati-cally, whether owing to shifts in thestate of confidence or general econom-ic conditions. Alternatively, if the priceof gold were fixed by law, as undergold standards past, its purchasingpower (that is, the general price level)would fluctuate violently.26

The concern that Eichengreen expresses inhis first sentence seems baseless. It wouldrequire a separation of monetary functionssuch that gold serves as the commonly ac-cepted medium of exchange, but a unit ofsomething else (what?) serves as the unitof account. Only under such a peculiar ar-


7/28

7

The classicalgold standardconstrained

inflation in amore credibleway, betterpinning downinflationaryexpectations.

rangement could one ounce of monetarygold have a fluctuating price. In every his-torically known system where gold or gold-redeemable claims were the principal meansof payment, a specified amount of gold also

defined the pricing unit.The concern Eichengreen expresses in

his second sentence, that under a gold stan-dard dramatic shifts in the demand for goldwould result in violently fluctuating pricelevels, seems also to lack merit. The histori-cal evidence shows that price levels duringthe classical gold standard of 18211914 didnot fluctuate any more violently than thefiat money era post-1971. Figure 1 showsprice index movements in the United King-dom over 253 years under gold and paper

sterling standards.There is a good reason why the demand

for monetary gold did not change dramati-cally under the classical gold standard. AsRobert Barro noted 30 years ago, the clas-sical gold standard constrained inflation ina more credible way, thereby better pinningdown inflationary expectations and better

stabilizing the demand to hold money rela-tive to income (or stated inversely, it betterstabilized velocity) than the fiat money sys-tem that followed it.27 He explained:

Since the move in 1971 toward flex-ible exchange rates and the completedivorce of United States monetarymanagement from the objective ofa pegged gold price, it is clear thatthe nominal anchor for the mone-tary systemweak as it was earlier[under Bretton Woods]is now entire-ly absent. Future monetary growthand long-run inflation appear nowto depend entirely on the year-to-yeardiscretion of the monetary author-

ity, that is, the Federal Reserve. Notsurprisingly, inflationary expectationsand their reflection in nominal interestrates and hence in short-run inflationrates have all become more volatile.

Volatility of inflation and expectations ofvolatility of inflation did diminish during

Figure 1Composite Price Index 1750 to 2003, January 1974 = 100 (logarithmic scale)

Source: Jim ODonoghue, Louise Goulding, and Grahame Allen, Consumer Price Inflation since 1750, Officefor National Statistics [UK]Economic Trends 604 (March 2004): 3846.

LogarithmicScale


8/28

8

Whenproductivity

growth allowsparticular goods

to be produced atlower cost, those

goods becomecheaper in both

real and nominalterms.

the Great Moderation after the 1980s,but since 2006 they have returned. In the14 years between August 1991 and August2005, the annual U.S. Consumer Price In-dex inflation rate (year-over-year, observed

monthly) stayed between 1 and 4 percent,a band of just 3 percentage points. But be-tween July 2008 and July 2009, the year-over-year inflation rate went from a high of 5.5percent to a low ofminus 2.0 percent, a swingof 7.5 percentage points in a single year. Ithas since risen as high as 3.9 percent. Aslong as the Fed retains discretion, inflationexpectations will remain variable.

Claim 4: A Gold Standard Would Be aSource of Harmful Secular Deflation

The most fundamental argumentagainst a gold standard, writes Cowen, isthat when the relative price of gold is go[ing]up, that creates deflationary pressures onthe general price level, thereby harming out-put and employment.28 Eichengreen offersa similar criticism:

As the economy grows, the price levelwill have to fall. The same amount ofgold-backed currency has to supporta growing volume of transactions,

something it can do only if the pricesare lower, unless the supply of newgold by the mining industry magicallyrises at the same rate as the outputof other goods and services. If not,prices go down, and real interest ratesbecome higher. Investment becomesmore expensive, rendering job cre-ation more difficult all over again.29

Eichengreen concludes: The robust invest-ment and job creation prized by the gold

standards champions and the deflationthey foresee are not easily reconciled, in oth-er words. In a nutshell, he maintains thatvigorous economic growth is at war with it-self under a gold standard because the mon-ey stock wont keep up.

Eichengreens argument here is theo-retically incorrect andsurprisingly from a

leading economic historianinconsistentwith the historical record of the gold stan-dard. First, as Eichengreen surely under-stands, the condition for the price level notfalling isnt an unlikely or magical exact

equality (=) between the rate of growth inthe stock of monetary gold and the rate ofgrowth in the output of other goods andservices (which proxies for demand to holdmonetary gold for transactions), but ratherthat the rate of growth in the stock of mone-tary gold is as at least as great() as that of therate of growth of output. How rare was that?Not very. During the period of the classicalgold standard, given that the long-run aver-age inflation rate was close to zero, this con-dition was met about half of the time. The

index numbers compiled by ODonoghue,Goulding, and Allen in fact show a fewmoreyears of a rising, rather than a falling, priceindex during the 93 years from the UnitedKingdoms resumption of the gold standardin 1821 to its departure in 1914.30 Over theperiod as a whole, the compound inflationrate was one-tenth of 1 percent per annum.

It is true that if the output of goods andservices grows too fast for the stock of mon-etary gold to keep up, the price level fallsIn such an environment, when productivity

growth allows particular goods to be pro-duced at lower cost, those goods becomecheaper in both real and nominal terms. 31

Such deflation, which results from rapidgrowth in real output, can hardly be a causefor regret.

Eichengreens case for fearing deflationunder a gold standard overlooks the im-portant historical findings of Atkeson andKehoe.32 Examining inflation rates and realoutput growth rates for 17 countries overmore than 100 years, they found that there

is no link between deflation (falling prices)and depression (falling real output) outsideof one extraordinary episode, the Great De-pression period of 192934. Their evidencesuggests to them that the Great Depressionshould be considered a special experiencewith little to offer policymakers consider-ing a deflationary policy today. Outside of


9/28

9

We need torecognize thebasic distinctionwhich appliesunder anymonetarystandard,between agooddeflation and a

baddeflation.

the Great Depression, in their database 65of 73 deflation episodes had no depression(and most of these deflations without de-pression occurred under a gold standard),while 21 of 29 depressions occurred without

deflation. We consider the Great Depressionin more detail below, but the Atkeson-Kehoeevidence makes it clear that the combinationof rapid deflation and rapid output shrink-age of 193033, which occurred under theinterwar system managed (or mismanaged)by central banks, was unlike experience un-der the much milder deflations of the classi-cal gold standard.

We need to recognize the basic distinction,which applies under any monetary standard,between agooddeflation and abaddeflation.

Selgin,33

Atkeson and Kehoe,34

and Bordo,Landon-Lane, and Redish35 have made thisdistinction conspicuously clear, but Eichen-green neglects it, as does Bernanke routinely.In brief, a good deflation is a situation wherethe price level falls because output growsmore rapidly than the money stock. It is asituation of ongoing approximate monetaryequilibrium, involving no significant excessdemand for money and therefore no sig-nificant excess supply of goods at any datesprice level. Prices fall one by one as the selling

prices of particular goods follow their costsof production downward. Real living stan-dards rise as goods become cheaper. A defla-tion driven by real growth does not make realgrowth more difficult to sustain.

A bad deflation, in a world with some de-gree of downward price and wage stickiness,is a situation where prices fall as a laggedresponse to an unexpected shrinkage in themoney stock or a spike in money demand.(The degree of price and wage stickiness islower in a system where the expected infla-

tion rate is lower, but stickiness was not zeroeven under the classical gold standard whenthe long-run expected inflation rate wasnear zero.) Such shocks create a monetarydisequilibrium, an unsatisfied demand tohold money at the existing price level. Con-sumers and businesses cut their spendingfor the sake of adding to money balances,

creating unsold inventories of goods, lead-ing to recessionary cutbacks in productionand employment until prices and wagesdecline sufficiently to clear the markets forgoods, labor, and money balances (a classic

discussion is provided by Yeager 1956.)

36

A good deflation involves no such un-planned inventory accumulation, so it doesnot depress output. In terms of the standardequation of exchange, MV = Py, a good defla-tion has the price levelPfalling contempora-neously with real income y rising. A bad de-flation hasPfalling with a lag (andy fallingin the interim) behind a shrinking moneystock M or shrinking velocity of money V.Bad deflation was a major problem in theearly 1930s, as a series of banking panics

led to the hoarding of currency by the pub-lic and the stockpiling of reserves by banks(events that can be described either as a fallin the velocity of base money or a fall in thequantity of broader money). It was briefly aproblem during the pre-Fed banking panicsin the United States. But banking panics arenot caused by being on a gold standard (seeClaim 6 below).

The nonconflict between deflation androbust growth is evident during the mostextended deflationary period under the clas-

sical gold standard in the United States,the 15 years from 1882 to 1897. The GrossDomestic Product deflator (as constructedby Romer 1989), which is a measure of theprice level, fell from 8.267 to 6.383, a com-pound inflation rate of approximately 1.7percent per annum.37 Over the same period,real GDP grew at the healthy rate of ap-proximately 3.0 percent per annum. Robustinvestment and real income growth wereeasily reconciled with deflation. The similarexperience in Britain during the same period

has sometimes been called a great depres-sion, but use of that label confuses defla-tion, which did happen, with falling output,which did not.38

The same confusion is evident when po-litical commentator Bruce Bartlett writesthat while a gold standard provided sta-ble purchasing power over long periods of


10/28

10

Deflations under

the classical goldstandard were notdebilitating. That

is, they were notassociated withfalling output.

time, that was only because inflations weresubsequently offset with debilitating de-flations.39 In fact, as the 188297 periodshows, and as Atkeson and Kehoe showmore generally, deflations under the clas-

sical gold standard were not debilitating.

40

That is, they were not associated with fallingoutput. Bartlett is mistaken in thinking that,as a consequence of deflation, there weregreater economic instabilities, higher unem-ployment and longer recessions during thegold-standard era. Despite a weak bankingsystem, the record of the gold-standard erabefore 1914 in the United States does not infact show greater economic instabilities orlonger recessions than the postWorld WarII era.41

Atkeson and Kehoe also address specifi-cally the case of slow-growing Japan in re-cent decades, which has often been cited asevidence of the depressing effect of falling ornegative inflation.42 They show that Japansgrowth rate began falling around 1960,while its inflation rate began falling around1970, suggesting that the former is a seculartrend independent of the latter. They aptlycomment: Attributing this 40-year slow-down to monetary forces is a stretch.43

Returning to the quotation from Eichen-

green, let us consider his claim that whenprices go down real interest rates becomehigher with the result that [i]nvestmentbecomes more expensive, rendering jobcreation more difficult.44 The statementunfortunately fails to keep straight the stan-dard distinction between two kinds of realinterest rates, ex ante (anticipated) and expost (retrospective). The identity that de-fines a real interest rate is: (1 + real interestrate) = (1 + nominal interest rate) (1 + infla-tion rate). The inflation rate in question can

either be an anticipated rate or a rate mea-sured retrospectively. Correspondingly, thederived real interest rate can either be antici-pated or retrospective. The standard theoryof the Fisher Effect tells us that when (say)a drop to minus 1 percent from 0 percentannual inflation is anticipated, the nominalinterest rate also drops by approximately 1

percent to keep the anticipated real interestrate constant. Therefore an anticipated defla-tion has no effect on the cost of investment. A de-cline in the price level greater than anticipat-ed over the period of a loan does raise the ex

post real interest rate paid on the loan. Butsuch an unanticipated decline, occurring af-ter an investment loan was taken out, doesnot raise the interest rate at the time of theloan contract, and thus cannot make invest-ment more expensive.

To be fair, Eichengreen may have had inmind (and simply neglected to specify) theone atypical set of conditions where his ar-gument would apply. Namely, if the nomi-nal interest rate is already near or at the zerolower bound, then the nominal rate cannot fall

enough in response to a large downward shiftin the anticipated inflation rate to keep theex ante real interest constant. The ex ante realinterest rate then does rise. This was a prob-lem during the extreme deflation of 193032; three-month Treasury rates fell close tozero at the end of 1932. Below I argue thatthis deflationunder the Federal Reserveswatchwas not due to the gold standard, butdue to its contravention. The zero low boundmay be a problem today under the FederalReserves deliberate policy of ultralow short-

term interest rates. During the period of theclassical gold standard, there were no cases ofan anticipated deflation so great as to bringthe nominal interest rate close to zero or cre-ate a lower-bound problem.

Claim 5: A Gold Standard too RigidlyTies the Governments Hands

One of the slides for Ben Bernankes lec-ture at GWU reads as follows:45

The strength of a gold standard is its

greatest weakness too: Because themoney supply is determined by thesupply of gold, it cannot be adjustedin response to changing economicconditions.

Note the passive wording: be adjusted. Adjust-ed by whom or by what? On a previous slide


11/28

11

The historicalrecord does notshow the FederaReserve carrying

its own weight,successfullyadjusting themoney supply toconditions.

Bernanke indicated that he was assuming anautomatic gold standard, without a centralbank able to do any significant adjusting ofthe money supply. But under a gold stan-dard, a change in the money supply can also

be brought about by market forces. Under agold standard, market forces in gold mining,minting, and banking do adjust the moneysupply in response to changing economicconditions, that is, in response to changesin the demand to hold monetary gold orto hold bank-issued money. The supply ofbank-issued money is not determined by thesupply of gold alone. If such a market-drivenchange counts as the supply being adjust-edand why shouldnt it?then Bernankesstatement is false. The money supply does

adjust in response to changing economicconditions.46

But perhaps the Bernanke slides phrasecannot be adjusted only intends to saythat under a fully decentralized and auto-matic gold standard there is no central mon-etary policy committee or other small groupof people who can deliberately adjust the ag-gregate money supply. Under that readingthe statement is true. But read that way thestatement does not deny that market forceswill adjust the money supply appropriately.

Bernanke neglects to provide a compara-tive analysis here. One might, with equal orgreater justice, invert his statement and say,The strength of a fiat standard is its great-est weakness too: because the money supplyis not automatically determined by marketforces but by the discretion of a committee,it can change in ways that are inappropri-ate to changing economic conditions. Thecomparative historical question remains: un-der which systemautomatic adjustment bymarket forces under a gold standard or de-

liberate adjustment by central bankers on afiat standardis the money supply better ad-justed to economic conditions? Those whounderstand why central economic planninggenerally fails should presume that marketguidance works better, absent a persuasiverebuttal showing that money is an excep-tion. The historical record does not show the

Federal Reserve carrying its own weight, suc-cessfully adjusting the money supply to con-ditions.47 That is, the Fed has not reducedcyclical volatility in the economy.

Bernanke apparently thinks that mar-

ket determination of the money supply is aweakness because it eliminates the option touse monetary policy to reduce the unemploy-ment rate (or in economists jargon, rules outexploiting the short-run Phillips Curve). Ac-cording to theNew York Times account of hisGWU lecture, Bernanke told the class thatbeing on the gold standard means swearingthat no matter how bad unemployment getsyou are not going to do anything about it.True, an automatic gold standard does elimi-nate the option to respond to the unemploy-

ment rate. But that is a feature, not a bug.Any economist who takes to heart the casethat Kydland and Prescott have made for thebenefit of rules over discretion in monetarypolicy will recognize that such a restraint is astrength rather than a weakness.48

When job seekers recognize the centralbanks intention to use monetary expansionto reduce unemployment, they will raisetheir inflation-rate expectations and thustheir reservation wage demands. Monetaryexpansion will then only ratify their ex-

pectations, not surprise them, and therebywill achieve only higher inflation and noreduction in the unemployment rate. Justas Ulysses strengthened his ability to sailhome, past the island of the Sirens, by tyinghimself to the mast and plugging his helms-mans ears with wax, so too a monetary sys-tem strengthens its ability to achieve thegood outcome it can achieve by foreswearingother goals. Kydland and Prescott identifythe goal as zero inflation, but more gener-ally the goal is to facilitate tradeincluding

intertemporal trademost efficiently.

Claim 6: A Gold Standard AmplifiesBusiness Cycles (or Fails to Dampenthem as a Well-managed Fiat MoneySystem Does)

In response to my 2008 piece, Tyler Cow-en wrote:49


12/28

12

The U.S. bankingpanics, both

under the pre-Fedsystem and in the1930s, came fromlegal restrictions

that weakenedthe banking

system, not from

the United Statesbeing on the gold

standard.

My main worry with the gold stan-dard is simply the pro-cyclicality ofthe money supply. . . . For instancewould you really want a contract-ing money supply in todays envi-

ronment? And yes credit crunches ofthis kind happen in market settingstoo so you cant blame it all on AlanGreenspan.

Cowens worry here does not appear to beabout the pro-cyclicality of the gold sup-ply. Gold mining is actually countercyclicalwith respect to the price level: that is, a fall-ing price level denominated in gold unitsraises the purchasing power of gold andso increases global mining output. For any

single economic region, the price-specie-flowmechanism is likewise countercyclical withrespect to the price level, meaning a fallinglocal price level attracts gold from the rest ofthe world. Cowen instead appears to worryabout the supposed pro-cyclicality of bank-issued money (deposits and banknotes) as aresult of bank runs and credit crunches. Heworries that the banking system is prone tocontract its liabilities in a downturn, andthereby to amplify the economys contrac-tion.

The inside money supply does fall ina banking panic if there are runs for basemoney, whether that base money is metallicor fiat.50 But it is not true that a gold stan-dard or free banking makes the banking sys-tem prone to bank runs and credit crunches.

The U.S. banking panics, both under thepre-Fed system and in the 1930s, came fromlegal restrictions that weakened the bankingsystem, not from the United States being onthe gold standard. Comparing the UnitedStates to Canada illustrates this strikingly.

Canada was equally on the gold standard,and had a similar agricultural economy, butexperienced no panics. Its banking systemwas far less restricted and consequently farstronger. The most important legal restric-tions on U.S. banks were the prohibition ofinterstate branching, which would have al-lowed better diversification of assets and lia-

bilities (Canada allowed nationwide branch-ing), and the rules (originally imposed tohelp finance federal expenditures in the CivilWar) requiring note-issuing banks to holdfederal bonds as collateral (no such rules op-

erated in Canada). The banknote restrictionprevented banks from issuing more notesduring seasons of peak currency demand,which in turn led to reserve drains everyautumn (not seen in Canada). Because pan-ics are not inherent to a gold standard, butrather to a banking system weakened by legalrestrictions, the pre-1933 panics do not in-dict the gold standard, but rather indict legalrestrictions that weaken banks. While Ber-nanke was correct to say in his lecture thatThe gold standard did not prevent frequent

financial panics, neither did it cause them.51

Financial Times columnist Martin Wolfexpresses a worry similar to Cowens, thata gold standard with fractional-reservebanking is inherently pro-cyclical: In goodtimes, credit, deposit money and the ratioof deposit money to the monetary base ex-pands. In bad times, this pyramid collapsesThe result is financial crises, as happened re-peatedly in the 19th century.52 In fact, freebanks did not exhibit exuberant swings intheir reserve ratios. 53 Less-regulated bank-

ing systems were more robust than Wolfsuspects, as seen not only in Canada butalso in Scotland, Sweden, Switzerland, andother systems without central banks underthe gold standard. Repeated financial criseswere a feature of the 19th-century bankingsystems in the United States and England,weakened as they were by legal restrictionsbut not of the less restricted systems else-where.54

Claim 7: The Gold Standard Was

Responsible for the Deflation thatUshered in the Great Depression in theUnited States

The most prominent set of criticisms ofthe gold standard among academic econo-mists in recent years blames the gold stan-dard for creating the Great Depression inthe United States and for then spreading


13/28

13

To understand

the deflation of193032, we needto review thedeflation of theinterwar periodas a whole.

it internationally. Douglas Irwin summa-rizes the case and identifies its most citedsource:55

Modern scholarship regards the De-

pression as an international phenom-enon, rather than as something thataffected different countries in isola-tion. The thread that bound countriestogether in the economic collapse wasthe gold standard. Barry Eichengreens1992 book Golden Fetters is most com-monly associated with the view thatthe gold standard was the key factorin the origins and transmission of theGreat Depression around the world.56

The piece of evidence most often cited forthis view is [t]he fact that countries noton the gold standard managed to avoid theGreat Depression, while countries on thegold standard did not begin to recover untilthey left it.57

This section addresses the factor in theorigins charge. The next section addressesthe transmission charge.

James D. Hamilton argues that between1929 and 1933, the U.S. and much of the restof the world were on a gold standard. That

did not prevent (indeed, I have argued it wasan important cause of) a big increase in thereal value of gold over that period. Becausethe price of gold was fixed at a dollar priceof $20/ounce, the increase in the real valueof gold required a huge drop in U.S. nomi-nal wages over those years.58 Because wageswere sticky downward, the drop in nominaldemand for labor created a massive loss ofemployment.

To understand the deflation of 193032,we need to review the deflation of the inter-

war period as a whole. And to understandthe interwar deflation as a whole, we needto review the monetary events of World WarI. During the war, the major combatant na-tions suspended the gold standard in orderto print copious amounts of money to fi-nance war expenditures. At wars end theywere left with price levels in local currency

units much higher than before the war, andmuch higher than postwar price levels mea-sured in gold units. As Robert Mundellnoted in his Nobel lecture, large volumes ofEuropean gold flowed to the United States,

which continuously remained on gold (al-though the federal government embargoedgold exports in 191719).59 The gold inflowsubstantially raised the U.S. dollar price levelduring the war. Despite a major correction in192021, the dollar (and gold) price levelremained 40 percent above the prewar equi-librium, a level at which the Federal Reservekept it until 1929.60 For the United States,this meant that the price level would eventu-ally have to fall.

Meanwhile in Europe, wartime money

printing had pushed the price levels in theUnited Kingdom, France, and other coun-tries much higher than 40 percent abovetheir prewar levels. For the United Kingdomand France to return to the gold standard(that is, to reinstitute convertibility at a de-fined parity between the domestic monetaryunit and gold), even without further U.S. de-flation, would require some combination ofdevaluation and deflation. Mundell pointsout that some notable staunch defenders ofthe gold standard, such as Charles Rist and

Ludwig von Mises, saw devaluation as a moreprudent option than a painfully large defla-tion. Mises is reported to have criticized therecommendation that a deflation should beundertaken to reverse the effects of wartimeinflation by remarking that, once you haverun a man over with a truck, you do himno favor by putting the truck in reverse anddriving over him in the other direction.

France chose to adjust the francs goldcontent downward (to devalue) fully in pro-portion to its lost purchasing power, which

enabled them to keep the postwar francprice level. The United Kingdom and mostother countries chose to restore the prewargold content to the monetary unit, whichforced a major downward adjustment in theprice level to reverse most of the wartime in-flation. As Mundell put it, The deflation ofthe 1930s was the mirror image of the war-


14/28

14

The globaldeflation of theinterwar period

was not due

to the worldsbeing on the gold

standard.

time rise in the price level that had not beenreversed in the 192021 recession.61 Ma-zumder and Wood detail the economic logicof this reversal in an important recent pa-per, and show how the movement of prices

parallels the pattern seen in resumptions ofthe gold standard at the old parity followingprevious wartime inflations.62

The global deflation of the interwarperiod, in other words, was not due to theworlds being on the gold standard. It wasdue to many countries leaving the gold stan-dard, inflating massively while off the goldstandard, and then resuming the gold stan-dard at the old parity (not devaluing to accom-modate the inflated price level).

Attempts to reduce the demand for mon-

etary gold through international coordina-tion among central banks came to naught.The Federal Reserve System, and especiallythe Bank of France, absorbed large amountsof gold by sterilizing inflows to block the risein prices that otherwise makes a regions in-flow self-limiting.63 They were not acting inaccordance with the gold standard. Rather, asBen Bernanke puts it, in defiance of the so-called rules of the game of the internationalgold standard, neither country allowed thehigher gold reserves to feed through to their

domestic money supplies and price levels.64The U.S. recession that became the Great

Depression, according to the National Bu-reau of Economic Research business-cyclechronology, began once the previous busi-ness expansion ended in August 1929. Pric-es began to fall three months later. Monthlydata show the consumer price index risingup until November 1929, with Decemberthe first month of decline. The arrival of de-flation cannot then have been the initiatingcause for the expansion turning into reces-

sion. Better explanations for why the boomdid not continue are beyond our subjectmatter here, but some contemporary observ-ers, such as F. A. Hayek, argued that the Fedhad amplified the boom to an unsustain-able degree by deliberately expanding creditto keep wholesale prices from falling.65 InHayeks view, a milder downturn would have

occurred sooner had the Fed not increasedits expansionary efforts from June 1927 toDecember 1928. The Fed finally tightenedcredit in early 1929 to moderate the rapidrise in stock market share prices.

In the view famously spelled out by Mil-ton Friedman and Anna J. Schwartz in theirA Monetary History of the United States,66 whatmight have been a garden-variety recessionthough perhaps a fairly severe one, becamethe Great Depression when bank runs wereallowed to shrink the broader money supplydramatically.67 The Fed stood idly by, nottrying to counter the shrinkage, while thestock of money fell by over a third betweenAugust 1929 and March 1933.68 The result-ing inflation rates in 1930, 1931, and 1932

were deeply negative: 6.4, 9.3, and 10.3percent, respectively.

In Golden Fetters, Eichengreen chargesthat the gold standard was responsiblefor the failure of monetary and fiscal au-thorities to take offsetting action once theDepression was underway.69 More specifi-cally, he claims that the gold standard wasthe binding constraint preventing policy-makers from averting the failures of banksand containing the spread of financialpanic.70 Friedman and Schwartz, however

had already provided some evidence to thecontrary. They showed that the Fed duringthis period was not obeying the dictates ofthe gold standard, but was in fact violatingthem by sterilizing gold inflows.71 The U.Sgold stock rose in 1931 and again in 1932,but the Fed prevented bank reserves and themoney supply from expanding and therebyprevented a moderation of the downwardpressure on prices and output. If not the goldstandard, what stopped the Fed from ex-panding? Most plausibly, to judge by its own

pronouncements at the time, we can blamethe Federal Reserve Boards adherence to anow-discarded credit policy doctrine knownas the Real Bills Doctrine, which held thatthe issuance of short-term, self-liquidatingloans would ensure that the created moneywould go to real goods, and thus the lendingwould be non-inflationary.72


15/28

15

The interwarperiod showsus a case wherecentral banksnot the goldstandardranthe show.

Eichengreen acknowledges that the Fedhad extensive gold reserves, but none-theless maintains that it had very limitedroom to maneuver.73 A more recent studycoauthored by Anna J. Schwartz, Michael D.

Bordo, and Ehsan U. Choudhri provides ad-ditional evidence that, in fact, the Fed hadmore than enough spare gold reserves (inexcess of its legally mandated gold cover re-quirements) to offset the contraction of thebroad money supply and thereby offset thedownward pressure on real output.74 Theysummarize their findings as follows:75

[T]he United States, . . . holding mas-sive gold reserves . . . , was not con-strained from using expansionary

policy to offset banking panics, defla-tion, and declining economic activ-ity. Simulations, based on a model ofa large open economy, indicate thatexpansionary open market operationsby the Federal Reserve at two criticaljunctures (October 1930 to February1931; September 1931 through January1932) would have been successfulin averting the banking panics thatoccurred, without endangering convert-ibility [through losses of gold reserves].

Indeed had expansionary open marketpurchases been conducted in 1930, thecontraction would not have led to theinternational crises that followed.

Specifically they find that, under a simulatedprogram of large open-market purchases tooffset the contraction of the broader moneysupply, U.S. gold reserves would have de-clined significantly but not sufficiently toreduce the gold ratio below the statutoryminimum requirement.

Claim 8: The Gold Standard WasResponsible for Spreading the GreatDepression from the United States to theRest of the World

The second part of the Golden Fettersindictment, to quote a recent statement ofit by Michael Bordo, is that The Great De-

pression spread across the world via the fixedexchange rate gold standard.76 In Eichen-greens earlier words, the international goldstandard transmitted the destabilizing im-pulse from the United States to the rest of

the world.

77

This description of events hassome truth to it, but is misleadingly incom-plete. The destabilizing impulse, as empha-sized in the previous section, came from theFederal Reserve and Bank of France steriliz-ing gold inflows and thereby absorbing ever-greater amounts of gold. These policies, asBernanke has noted, and not the gold stan-dard as such, created deflationary pressuresin deficit countries that were losing gold.78Even more important, as discussed above,counties such as the United Kingdom were

already headed for deflation once they decid-ed to return to the gold standard at their pre-war parities while their price levels were wellabove their prewar (and equilibrium) levels.

The interwar period shows us a case wherecentral banksnot the gold standardranthe show. To put it mildly, they failed to runit as well as the classical gold standard. AsRichard H. Timberlake has emphasized, itis illogical to blame the international goldstandard for the interwar disaster.79 Theinternational gold standard worked well in

the prewar period, when central banks wereless active in trying to manage gold flows(and in many countries, such as the UnitedStates and Canada, did not yet exist). Blamefor the unfortunate results of the interwarsystem rests instead on decisions to resumethe gold standard at the old parity and onthe discretionary policies of central bankers.The illogic is compounded when the failureof the discretionary interwar central bank-ing system is taken to provide evidence insupport of giving central banks more discre-

tion than they have under an automatic in-ternational gold standard.

The interwar experience does carry a les-son for advocates of reinstating an interna-tional gold standard. It indicates that the in-ternational gold standard works best whenit works most automatically. A valid pointis therefore made by Bernankes lecture


16/28

16

Free bankingdecentralizescurrency issue

and goldreserve holding,

subjecting itto competitive

interbankclearing

discipline, andthereby all buteliminates the

risk of large orpersistent money-

supply errors.

slide that reads, The effects of bad poli-cies in one country can be transmitted toother countries if both are on the gold stan-dard.80 Bad monetary policies can comefrom discretionary central banks in other

countries. It would therefore be better for allif a treaty reinstating an international goldstandard could also institute enforceableconstraints against central banks disturbingthe peace. The most thorough constraint isto eliminate central banking in favor of freebanking. Among other reforms, free bank-ing decentralizes currency issue and goldreserve holding, subjecting it to competitiveinterbank clearing discipline, and thereby allbut eliminates the risk of large or persistentmoney-supply errors.

Claim 9: A Gold Standard,Like any Fixed Exchange-rate System,Is Vulnerable to Speculative Attacks

George Selgin finds it more doubtful [to-day] than ever before that any government-sponsored and administered gold standardwill be sufficiently credible to either bespared from or to withstand redemptionruns.81 He quotes Hamilton to similar ef-fect: given that central banks and treasurieson the gold standard can, and often have, left

it, and given that speculators know this, itfollows that any currency adhering to a goldstandard will . . . be subject to a speculativeattack.82 Selgin adds, The breakdown inthe credibility of central bank exchange ratecommitments since World War I cannot beeasily repaired, if it can be repaired at all.

Hamiltons any currency is too sweep-ing, but the lesson Selgin draws is persua-sive. As he notes, the noncredibility of a gov-ernment central banks promises to stay onthe gold standard is not a case against the

gold standard but a case against weakeningcommitment to the gold standard by com-bining it with central banking. Because atypical central bank has a legal monopolyof currency notes denominated in the localmonetary unit, it has the power to devalueor to take the economy entirely off the goldstandard by ending gold redemption of its

liabilities. The devaluation or departurefrom gold can be coordinated with the trea-sury, which has a legal monopoly on coins.

A more durable and credible approach tosustaining the gold standard is to let the pri-

vate sector competitively issue currency. Pri-vate firms in a competitive market are morestrongly committed to gold redemption fortwo reasons: they can be legally held to theirpromises (unlike central banks, which enjoysovereign immunity from lawsuits over de-valuation or nonredemption), and they needto compete for customers who can go else-where by avoiding practices that raise theirrisk of not being able to redeem. In the eventthat any single bank among dozens fails orsuspends payment as a result of its poor

management, the gold standard survivesFree banking thus delivers a more robustgold standard,83 and the combination ofgold and free banking is even an antifragilemonetary system.84

In an attack on a fixed exchange rate, sayon the pound sterling when it is pegged tothe deutsche mark, speculators borrow inpounds, redeem them for marks, and holdmarks until the Bank of England runs outof marks and must devalue the pound. Theymake a profit if and when devaluation oc-

curs, because they now get more poundsfor each mark they hold and can repay theirpound-denominated loans with plenty ofmarks left over. A similar path to profit ex-ists under a gold-dollar standard in whichthe Federal Reserve is empowered to devaluethe dollar against gold. There was, in fact, arun on the dollar in anticipation of Frank-lin D. Roosevelts devaluation in 1933. Butno such path is available with decentralizedprivate issue of gold-redeemable currencyentirely by commercial banks, because no

commercial bank can devalue the dollar. If acommercial bank fails, whether because of arun or otherwise, those who have borrowedfrom it must still pay back their loans in un-diminished dollars. Hence there is no profitin borrowing, running for reserve money,and repaying later, even if the run bringsdown the bank.


17/28

17

The UnitedStates was onthe classical goldstandard withou

a central bankfrom 1879 to1914.

Claim 10: Fiat Money Is Necessary toHave a Lender of Last Resort Able toMeet the Liquidity Needs of the BankingSystem

Barry Eichengreen writes:

Under a true gold standard, moreover,the Fed would have little ability to actas a lender of last resort to the bank-ing and financial system. The kind ofliquidity injections it made to preventthe financial system from collapsingin the autumn of 2008 would becomeimpossible because it could provideadditional credit only if it somehowcame into possession of additionalgold. Given the fragility of banks and

financial markets, this would seema recipe for disaster. Its proponentspaint the gold standard as a guaran-tee of financial stability; in practice, itwould be precisely the opposite.85

The classical conception of the lender oflast resort, as spelled out by the Englishjournalist and banking historian WalterBagehot during the classical internationalgold standard era, is an institution thatlends reserves to illiquid-but-solvent com-

mercial banks in a period of peak demandfor currency or bank reserves, in the extremeduring a period of bank runs.86 Its aims areto prevent regrettable bank insolvencies thatresult from hasty asset liquidations, and tosatisfy the publics demand for currency orreserve money so that the runs cease and themarket calms. This appears to be the notionthat Eichengreen has in mind.

Assuming that a central bank such as theFederal Reserve is assigned the role of lenderof last resort, Eichengreen takes a true gold

standard to imply that the central bankcould provide additional credit only if itsomehow came into possession of addition-al gold. That is, the gold standard is nottrue unless it imposes a 100 percent goldmarginal reserve requirement on centralbank liabilities. This is a highly idiosyncrat-ic understanding of a true gold standard.

Peels Act of 1844 did impose a 100 percentmarginal gold reserve requirement on ex-pansion of the Bank of Englands banknotecirculation, but the Bank could still provideadditional credit by expanding its deposit

liabilities. Indeed, the Bank is generally un-derstood to have first acted as a lender oflast resort during the Baring Crisis in 1890,while Peels Act was still in place.

It is true that a 100 percent gold marginalreserve requirement on all central bank li-abilities would constrain last-resort lending.But imposing such a rule on the central bankis not required in order to have a true goldstandard, and indeed having a central bankis not even required. A gold standard, again,is generically defined by gold serving as the

medium of redemption and medium of ac-count, not by any reserve requirement im-posed on a central bank. The United Stateswas on the classical gold standard without acentral bank from 1879 to 1914. During thatperiod, private clearinghouse associationsacted as lenders of last resort to their mem-ber banks.87 So a central bank is not evennecessary to have a lender of last resort.

Eichengreen argues that confidenceproblems are intrinsic to fractional-reservebanking and why an economy with a mod-

ern banking system needs a lender of last re-sort.88 But as noted under Claim 6 above,historical evidence indicates that confidenceproblems are minimal if no legal restrictionsprevent banks from adequately capitalizingand diversifying themselves.

Claim 11: Setting the New Gold Parity IsToo Hard

The danger of setting the new gold paritytoo low (too few dollars per ounce of gold) isexemplified, as Selgin notes, by Great Brit-

ains choice in 1925 to restore the old par-ity to the pound sterling.89 Because the pricelevel had risen sharply, a return to the oldparity required a sharp deflation to return tothe old price level. The danger of setting theparity too high is, conversely, a transitionalinflation to reach the new equilibrium pricelevel. Eichengreen summarizes the problem


18/28

18

To avoidtransitionalinflation or

deflation, thenew parity must

equate monetarygold supply and

demand at thecurrentprice level.

this way:

Envisioning a statute requiring theFederal Reserve to redeem its notesfor fixed amounts of specie is easy,

but deciding what that fixed amountshould be is hard. Set the price toohigh and there will be large amountsof gold-backed currency chasing limit-ed supplies of goods and services. Thenew gold standard will then becomean engine of precisely the inflationthat its proponents abhor. But setthe price too low, and the result willbe deflation, which is not exactly ahealthy state for an economy.90

To avoid transitional inflation or defla-tion, the new parity must equate monetarygold supply and demand at the currentpricelevel. If we could assume that the supply anddemand for monetary gold were unaffectedby the reinstatement of the gold standard,the solution would be easy: choose the cur-rent price of gold. But that is unlikely towork in todays financial world. The demandfor gold stocks today includes an inflation-hedging demand that would be absent un-der a gold standard. On the other hand, be-

cause a gold standard lowers the mean andmedium-term variance of the inflation rate,the demand to hold currency and demanddeposits for transaction purposes, againstwhich banks would hold gold reserves,would rise. As Selgin notes:

The problem here is, not that thereis no new gold parity such as wouldallow for a smooth transition, butthat the correct parity cannot be deter-mined with any precision, but must

instead be discovered by trial and error.Consequently the transition couldinvolve either costly inflation or itsopposite. . . .91

Tyler Cowen cites the same problem: Onefive or ten percent deflation is enough tocrush the economy and indeed the whole

gold standard idea.92 Given the socialist cal-culation debate, can we really know the righttransition price?

Choosing a new parity is indeed a prob-lem. There are at least two approaches to

estimating the new parity that would avoidtransitional inflation or deflation. Note thatnew parities need to be chosen simultane-ously by all participating currency areas inorder to agree to return to the gold standardsimultaneously so as to create the broadestpossible international gold standard. Themore conventional approach is to use econo-metric studies of recent inflation-hedgingdemand for gold, and of transactions de-mand for zero-yielding bank reserves atgold-standard-type expected inflation rates

The less conventional approach, which callsfor further study, is to derive guidance frommarket signals, in particular from the goldfutures market or some new kinds of predic-tion markets. Under such a regime, marketplayers would put money on their own esti-mates of what the real purchasing power ofgold will be following a return to the inter-national gold standard.

In the current world where prices andwages exhibit greater downward than up-ward stickiness, playing it safe in the choice

of a new parity means erring on the side ofa small transitional inflation rather than adeflation.

So as not to overstate the relative size ofthe problem, however, we should note thatthe same problem attends any significantchange in the inflation path, or significantchange in other policy (such as the rate ofinterest on reserves) under a fiat standardThe switch to a lower inflation rate targetfor example, will cause the path of transac-tions demand to hold money relative to the

volume of spending to jump upward (shift-ing the velocity-of-money path downward)Underestimating the increased demandand failing to offset it with a one-time in-crease in the stock of money, will cause thepolicy to create an excess demand for moneyand will thus create a recession with unsoldinventories of goods and unemployed labor


19/28

19

The Feds trackrecord for

real economicstability underfiat money doenot weigh infavor of fiatmoney.

services. The Bernanke Feds switch fromzero to positive interest on bank reserves inOctober 2008 sharply increased the bankingsystems demand to hold reserves, swamp-ing the money-supply-expanding effect of

the accompanying Quantitative Easing Iexpansion of reserves. The result was sevenmonths in 2009 (March through Septem-ber) in which the year-over-year inflationrate was negative. The downturn in real out-put already underway was amplified. Curi-ously, this bad deflationand the firstdeflation of either kind in more than fivedecadesoccurred on the watch of an ex-pressly deflation-averse Fed chairman.

Claim 12: Inflation Is so Low Today that

We Dont Need a Gold StandardEzra Klein comments:

In 1981, the country really was facingan inflation problem. It made sensethat people would be looking forradical alternatives that would helpcontrol inflation. Today, inflation isabout as low as its ever been, and ifyou look at market expectationsyoudo believe in the market, dont you?its expected to stay low.93

It is, of course, true that the urgency ofadopting a gold standard to fight inflationis lower when the inflation rate is lower. Ifinflation were our exclusive concern, and wecould trust the central bank to keep infla-tion as low under a fiat standard as it wasunder the classical gold standard, then itwould be foolish to bear any cost to rein-stitute a gold standard. Inflation today iscertainly lower than it was in the 1970s and1980s, but it is not true that inflation is as

low today as it was under the classical goldstandard. Recall that the inflation rate wasonly 0.1 percent over Britains 93 years onthe classical gold standard. Over the most re-cent 10 years (August 2002 to August 2012)in the United States, the CPI for urban con-sumers rose 27.5 percent, for an annualizedinflation rate of 2.5 percent. Over the last 40

years (since August 1972, shortly after Presi-dent Nixon closed the gold window), the risehas been 449.2 percent, for an annualizedrate of 4.4 percent. There remains a case forthe gold standard based on inflation alone.

How low are market expectations of theinflation rate to come? According to theFinancial Times (September 17, 2012), theannouncement of the Feds latest round ofquantitative easing, QE3, pushed the mar-kets expectation of the U.S. inflation rateover the next 10 years (derived from priceson the inflation-indexed bond market) to2.73 percent per annum. Inflation expecta-tions are not as low today as they were un-der the classical gold standard, and they arecertainly more volatile. There is no tangible

institutional assurance that the U.S. infla-tion rate will never again return north of 5percent or even 10 percent.

Of course, consumer price inflation isnot our exclusive concern. The past decadehas reminded us that, even with consumerinflation rates around 2.5 percent or lower,we face the serious danger of asset price bub-bles and unsustainable credit booms undera central bank policy of artificially low inter-est rates. The ultralow Fed Funds rate policyof 1.25 percent or less from November 2002

through June 2004 helped fuel the hous-ing bubble.94 Todays rate policy has beenholding the Fed Funds rate at 0.25 percentor less since December 2008, with the an-nounced prospect of another three years ofultralow rates. Time will tell where a newbubble is now forming. More generally, theFeds track record for real economic stabilityunder fiat money does not weigh in favor offiat money.95

Claim 13: A Gold Standard Needs to Be

International, and the Rest of the WorldWont Come Along

Selgin makes an important point whenhe notes that

the historical gold standard that . . .performed so well was an internationalgold standard, and [its] advantages . . .


20/28

20

Compared toa fiat money

standard, a goldstandard is a

source of stabilityin the purchasingpower of money.

were to a large extent advantages due tobelonging to a very large monetary net-work. Consequently, a gold standardthat is limited to a single country, andeven to a very large country, cannot be

expected to offer the same advantagesas a multi-country gold standard or setof gold standards.96

The strongest case for reinstating the goldstandard is for an international gold stan-dard. Getting other nations to join in thereinstatement is therefore a genuine prob-lem.97 But this is not a reason for rejectingthe case for an international gold standard.It is, rather, a reason for taking the case forreinstating the international gold standard

to other countries while developing it athome. China and much of Latin America al-ready link to or shadow the U.S. dollar. Sothe most important places to take the argu-ment are the Eurozone, Japan, and GreatBritain.

Representatives of the leading nationscame together to reconstruct the internation-al monetary system in 1944, at the famousconference in Bretton Woods, New Hamp-shire. Such a gathering can happen again oncedissatisfaction with the postBretton Woods

system of completely unanchored currenciesbecomes deep and widespread enough. Theinfluential leader of the United Kingdomdelegation at Bretton Woods was John May-nard Keynes, who famously considered thegold standard a barbarous relic and wasdetermined to minimize its role to widen thescope for discretionary central bank policy-making.98 The challenge for those who favorrestoration of an international gold standardwill be to insure that the delegates to the newconference have a better understanding.

Conclusion

Assuming that the federal governmenthas the gold it says it has, there is enoughgold in the United States to operate a goldstandard today with a free banking system,

without requiring a transitional inflationor deflation if the reentry dollar-gold par-ity is set near the current market price. Thegold standard is not an example of price fix-ing by government, but a system in which a

unit of gold defines the unit of account, andpieces of gold serve as the ultimate mediumof redemption. The volatility of the dollarprice of gold since gold was demonetized in1971 does not show that gold is an unstablemonetary standard. The dollar price of goldrises and falls these days largely because ofswings in the demand for gold as an infla-tion hedgeswings driven by the instabilityof fiat currencies.

Compared to a fiat money standard, agold standard is a source of stability in the

purchasing power of money. It is a source ofmild secular deflation if the output of goodsgrows more rapidly than the gradually grow-ing stock of gold, but that is a benign kindof deflation. A gold standard does tie thegovernments hands against printing moneyto cover its expenses, but that is a desirablefeature of the system and not a flaw. It doesnot prevent a government from borrow-ing in the international financial market,provided that it credibly commits to repaywhich means that it credibly commits to

balancing its budget in present-value termsThe lack of a constraint on printing-pressfinance under a fiat standard is one of itsgreatest weaknesses. Because a fiat moneysupply is not automatically determined bymarket forces, but instead by the discretionof a committee, it can change in ways thatare inappropriate to changing economicconditions.

An automatic gold standard does notamplify business cycles as compared with amanaged fiat money system. If free banking

on a gold standard were to render the bank-ing system prone to bank runs and panics,creating unanswerable spikes in the demandfor monetary gold, which would, of course,be a serious problem. But such is not the his-torical record. Runs and panics are not inher-ent to free banking on a gold standard, butonly to a banking system weakened by legal


21/28

21

restrictions. The pre-1933 banking panicsin the United States therefore do not indictthe gold standard, but rather indict the legalrestrictions that weakened banks. The mon-etary instability of the interwar period that

ushered in the global Great Depression wasnot due to what remained of the gold stan-dardnothing of the sort happened underthe classical gold standardbut ultimatelycan be traced to the inflationary policies ofcentral banks during the First World Warwhile they were off the gold standard andto their subsequent decisions to return togold only intermittently, insincerely, and atparities inconsistent with the high domesticprice levels they had created.

There are at least two genuine problems

to be faced in planning a transition from adiscretionary fiat standard to an automaticgold standard. The first is choosing the newgold-dollar parity so as to minimize disrup-tive inflation or deflation in the transition.Prediction markets could help to estimatethe sustainable parity. Staying with the sta-tus quo fiat standard does not avoid theproblem of transitional changes in the de-mand for base money, it should be noted, be-cause such changes accompany every majorswing in projected inflation. The Feds track

record for real economic stability under fiatmoney does not weigh in favor of fiat moneybeing the path of least disruption. The sec-ond problem is getting as much of the restof world as possible to opt into the transi-tion at the same time, so that the benefitsof an international gold standard are maxi-mized. This is not a reason for embracingthe status quo, but for reviving appreciationfor the international gold standard aroundthe globe as well as at home.

NotesI thank Vipin Veetil for research assistance.

1. See Ralph Benko, The Gold Standard: ALitmus Test for GOP Candidates, Forbes.com,

July 5, 2011, http://www.forbes.com/sites/ral-phbenko/2011/07/05/gold-standard-litmus-test-gop-candidates/; and Chris Isidore, Gin-

grich: U.S. Should Reconsider Gold Standard,January 18, 2012, http://money.cnn.com/2012/01/18/news/economy/gingrich_gold_standard/index.htm; Republican Platform (2012), p. 4,http://www.gop.com/wp-content/uploads/2012/08/2012GOPPlatform.pdf.

2. Matthew OBrien, Why the Gold StandardIs the Worlds Worst Economic Idea, In 2 Charts,The Atlantic, August 26, 2012, http://www.theatlantic.com/business/archive/2012/08/why-the-gold-standard-is-the-worlds-worst-economic-idea-in-2-charts/261552/.

3. Ezra Klein, The GOP has Picked the WrongTime to Rediscover Gold, Wonkblog, August 26,2012, http://www.washingtonpost.com/blogs/ezra-klein/wp/2012/08/24/the-gop-has-picked-the-wrong-time-to-rediscover-gold/.

4. Binyamin Appelbaum, Bernanke, as Profes-sor, Tries to Buff Feds Image, New York Times,

March 20, 2012, p. B3.

5. Barry Eichengreen, A Critique of PureGold, The National Interest(SeptemberOctober),http://nationalinterest.org/article/critique-pure-gold-5741.

6. Lawrence H. White, Is the Gold StandardStill the Gold Standard Among Monetary Sys-tems? Cato Institute Briefing Paper no. 100,February 8, 2008, http://www.cato.org/pubs/bp/bp100.pdf.

7. Ben Bernanke, The Federal Reserve and

the Financial Crisis [slides for lecture at GeorgeWashington University], March 20, 2012, www.federalreserve.gov/newsevents/files/bernanke-lecture-one-20120320.pdf.

8. As humorist Dave Barry jokingly puts it,Over the years, all the governments in the world,having discovered that gold is, like, rare, decidedthat it would be more convenient to back theirmoney with something that is easier to come by,namely: nothing.Dave Barrys Money Secrets: Like:Why Is There a Giant Eyeball on the Dollar? (New

York: Three Rivers Press, 2006), p. 10.

9. John Waggoner, Should We Return to

the Gold Standard? USA Today, May 23, 2012,http://www.usatoday.com/money/markets/story/2012-04-23/return-to-the-gold-standard

/54493710/1.

10. United States Department of the Treasury,Status Report of U.S. Treasury-Owned Gold,(August 31, 2012), https://www.fms.treas.gov/gold/backissues.html.

11. At $1,600 per ounce, the ratio of government


22/28

22

gold to current M1 is 17.3 percent. Numbers arefrom the Federal Reserve Economic Data, FederalReserve Bank of St. Louis, data series RequiredReserves, Not Adjusted for Changes in ReserveRequirements (REQRESNS), http://research.stlouisfed.org/fred2/series/REQRESNS?cid=123;and series M1 Money Stock (M1), http://research.

stlouisfed.org/fred2/series/M1?cid=25. The ratiosreported here update, but are very close to, thosein Lawrence H. White, Making the Transition toa New Gold Standard, Cato Journal 32 (Spring/Summer 2012), pp. 41121, http://www.cato.org/pubs/journal/cj32n2/v32n2-14.pdf.

12. Kurt Schuler, The World History of FreeBanking: An Overview, in The Experience of Free

Banking, Kevin Dowd, ed. (London: Routledge,1992), pp. 447.

13. Eichengreen, A Critique of Pure Gold.

14. I have also made these arguments against

Eichengreen in Lawrence H. White, Making theTransition to a New Gold Standard, Cato Journal32 (Spring/Summer 2012): 41920.

15. Barry Eichengreen, When Currencies Col-lapse, Foreign Affairs 91 (January/February): 128,http://www.foreignaffairs.com/articles/136779/barry-eichengreen/when-currencies-collapse.

16. Klein, The GOP has Picked the Wrong Timeto Rediscover Gold.

17. James D. Hamilton, Return to the Gold Stan-dard, Econobrowser (blog), September 1, 2012,

http://www.econbrowser.com/archives/2012/09/return_to_the_g.html.

18. Because it has so much trade with the Unit-ed States, I would recommend that the Bahamasadopt official dollarization (with private-note is-sue) in place of its current exchange-rate peg tothe dollar.

19. Hamilton, Return to the Gold Standard.

20. James D. Hamilton, The Gold Standardand Economic Growth,Econobrowser(blog), Sep-tember 5, 2012.

21. Hamilton, Return to the Gold Standard.

22. Frank Holmes, Jewelry Drives the Gold LoveTrade,Advisory Analyst(blog), February 19, 2011,http://advisoranalyst.com/glablog/2011/02/19/

jewelry-drives-the-gold-love-trade/.

23. Tyler Cowen, What Exactly Is the Argu-ment against Gold? Marginal Revolution (blog),December 29, 2011, http://marginalrevolution.com/marginalrevolution/2011/12/what-exactly-

is-the-argument-against-gold.html.

24. Tyler Cowen, A Short Note on the GoldStandard, Marginal Revolution (blog), Septem-ber 3, 2012, http://marginalrevolution.com/marginalrevolution/2012/09/a-short-note-on-the-gold-standard.html.

25. Hugh Rockoff, Some Evidence on the RealPrice of Gold, Its Costs of Production, and Com-modity Prices, in A Retrospective on the ClassicaGold Standard, 18211931, Michael D. Bordo and

Anna J. Schwartz, eds. (Chicago: University ofChicago Press, 1982): 61350, http://www.nberorg/chapters/c11139.

26. Barry, A Critique of Pure Gold.

27. See Robert J. Barro, United States Inflationand the Choice of Monetary Standard, in Inflation: Causes and Effects, Robert Hall, ed. (ChicagoUniversity of Chicago Press, 1982), p. 105.

28. Cowen, What Exactly Is the Argumentagainst Gold?

29. Eichengreen, A Critique of Pure Gold.

30. Jim ODonoghue, Louise Goulding, andGrahame Allen, Consumer Price Inflation since1750, Office for National Statistics [UK] Economic Trends 604 (March 2004): 3846.

31. George Selgin,Less Than Zero: The Case for Falling Prices in a Growing Economy (London: Instituteof Economic Affairs, 1997), http://www.iea.orguk/publications/research/less-zero.

32. Andrew Atkeson and Patrick J. Kehoe, De-flation and Depression: Is There an EmpiricalLink? American Economic Review 94 (May 2004)100.

33. Selgin,Less Than Zero.

34. Atkeson et al., Deflation and Depression.

35. Michael D. Bordo, John Landon-Lane, andAngela Redish, Good versus Bad Deflation: Lessons from the Gold Standard Era, in Monetary

Policy in Low-Inflation Countries, David E. Altig and

Ed Nosal, eds. (Cambridge: Cambridge Univer-sity Press, 2009), pp. 12774.

36. Leland B. Yeager, A Cash-Balances Interpre-tation of Depression, Southern Economic Journa22 (April 1956), 43847.

37. Christina D. Romer, The Prewar BusinessCycle Reconsidered: New Estimates of GrossNational Product, 18691908,Journal of Politica

Economy 97 (February 1989): 137.


23/28

23

38. S. B. Saul, The Myth of the Great Depression,18731896, 2nd ed. (London: Macmillan, 1985).

39. Bruce Bartlett, Republicans Are Wrongon Call for Gold Standard, NY Times Economix(blog), September 4, 2012, http://economix.blogs.nytimes.com/2012/09/04/the-gold-standard-is-

not-ready-for-prime-time/.

40. Atkeson et al., Deflation and Depression.

41. George Selgin, William D. Lastrapes, andLawrence H. White, Has the Fed Been a Failure?

Journal of Macroeconomics 34 (September 2012):56996.

42. Atkeson et al., Deflation and Depression: IsThere an Empirical Link? p. 102.

43. Ibid., p. 99.

44. Barry, A Critique of Pure Gold.

45. Bernanke, The Federal Reserve and the Fi-nancial Crisis.

46. On the interaction of gold su

recent arguments against the gold standard

Documents