Nonmonetary Effects of the Financial Crisis in the …...Nonmonetary Effects of the Financial Crisis...

Nonmonetary Effects of the Financial Crisis in thePropagation of the Great Depression

By BEN S. BERNANKE*

During 1930-33, the U.S. financial systemexperienced conditions that were among themost difficult and chaotic in its history.Waves of bank failures culminated in theshutdown of the banking system (and of anumber of other intermediaries and markets)in March 1933. On the other side of theledger, exceptionally high rates of defaultand bankruptcy affected every class of bor-rower except the federal government.

An interesting aspect of the general finan-cial crises—most clearly, of the bank failures—was their coincidence in timing with ad-verse developments in the macroeconomy.1

Notably, an apparent attempt at recoveryfrom the 1929-30 recession2 was stalled atthe time of the first banking crisis (Novem-ber-December 1930); the incipient recoverydegenerated into a new slump during themid-1931 panics; and the economy and thefinancial system both reached their respec-tive low points at the time of the bank "holi-day" of March 1933. Only with the NewDeal's rehabilitation of the financial systemin 1933—35 did the economy begin its slowemergence from the Great Depression.

A possible explanation of these synchro-nous movements is that the financial systemsimply responded, without feedback, to thedeclines in aggregate output. This is con-tradicted by the facts that problems of thefinancial system tended to lead output de-

* Stanford Graduate School of Business and HooverInstitution. I received useful comments from too manypeople to list here by name, but I am grateful to each ofthem. The National Science Foundation provided par-tial research support.

1This is documented more carefully in Sections I.Cand IV below.

2 This paper does not address the causes of the initial1929-30 downturn. Milton Friedman and AnnaSchwartz (1963) have stressed the importance of theFederal Reserve's "anti-speculative" monetary tight-ening. Others, such as Peter Temin (1976), have pointedout autonomous expenditure effects.

clines, and that sources of financial panicsunconnected with the fall in U.S. outputhave been documented by many writers. (SeeSection IV below.)

Among explanations that emphasize theopposite direction of causality, the mostprominent is the one due to Friedman andSchwartz. Concentrating on the difficultiesof the banks, they pointed out two ways inwhich these worsened the general economiccontraction: first, by reducing the wealth ofbank shareholders; second, and much moreimportant, by leading to a rapid fall in thesupply of money. There is much support forthe monetary view. However, it is not acomplete explanation of the link between thefinancial sector and aggregate output in the1930's. One problem is that there is no the-ory of monetary effects on the real economythat can explain protracted nonneutrality.Another is that the reductions of the moneysupply in this period seems quantitativelyinsufficient to explain the subsequent falls inoutput. (Again, see Section IV.)

The present paper builds on the Fried-man-Schwartz work by considering a thirdway in which the financial crises (in whichwe include debtor bankruptcies as well as thefailures of banks and other lenders) mayhave affected output. The basic premise isthat, because markets for financial claims areincomplete, intermediation between someclasses of borrowers and lenders requiresnontrivial market-making and information-gathering services. The disruptions of 1930-33 (as I shall try to show) reduced the ef-fectiveness of the financial sector as a wholein performing these services. As the real costsof intermediation increased, some borrowers(especially households, farmers, and smallfirms) found credit to be expensive and dif-ficult to obtain. The effects of this creditsqueeze on aggregate demand helped convertthe severe but not unprecedented downturnof 1929-30 into a protracted depression.

257

Digitized for FRASER http://fraser.stlouisfed.org/ Federal Reserve Bank of St. Louis

258 THE AMERICAN ECONOMIC REVIEW JUNE 1983

It should be stated at the outset that mytheory does not offer a complete explanationof the Great Depression (for example, noth-ing is said about 1929-30). Nor is it neces-sarily inconsistent with some existing ex-planations.3 However, it does have the virtuesthat, first, it seems capable (in a way inwhich existing theories are not) of explainingthe unusual length and depth of the depres-sion; and, second, it can do this withoutassuming markedly irrational behavior byprivate economic agents. Since the reconcili-ation of the obvious inefficiency of the de-pression with the postulate of rational privatebehavior remains a leading unsolved puzzleof macroeconomics, these two virtues aloneprovide motivation for serious considerationof this theory.

There do not seem to be any exact antece-dents of the present paper in the formaleconomics literature.4 The work of LesterChandler (1970, 1971) provides the best his-torical discussions of the general financialcrisis extant; however, he does not developvery far the link to macroeconomic perfor-mance. Beginning with Irving Fisher (1933)and A. G. Hart (1938), there is a literatureon the macroeconomic role of inside debt;an interesting recent example is the paper byFrederic Mishkin (1978), which stresseshousehold balance sheets and liquidity. Ben-jamin Friedman (1981) has written on therelationship of credit and aggregate activity.Hyman Minsky (1977) and Charles Kindle-berger (1978) have in several places arguedfor the inherent instability of the financialsystem, but in doing so have had to departfrom the assumption of rational economicbehavior.5 None of the above authors hasemphasized the effects of financial crisis on

3 See Karl Brunner (1981) for a useful overview ofcontemporary theories of the depression. Also, seeRobert Lucas and Leonard Rapping's article in Lucas(1981).

4 This is especially true of the more recent work,which tends to ignore the nonmonetary effects of thefinancial crisis. Older writers often seemed to take thedisruptive impact of the financial breakdown for granted.

5I do not deny the possible importance of irrational-ity in economic life; however, it seems that the bestresearch strategy is to push the rationality postulate asfar as it will go.

the real costs of credit intermediation, thefocus of the present work.

The paper is organized as follows: SectionI presents some background on the 1930-33financial crisis, its sources, and its correspon-dence with aggregate output movements.Section II begins the principal argument ofthe paper. I explain how the runs on banksand the extensive defaults could have re-duced the efficiency of the financial sector inperforming its intermediary functions. Someevidence of these effects is introduced.

Possible channels by which reduced finan-cial efficiency might have affected output arediscussed in Section III. Reduced-formestimation results, reported in Section IV,suggest that augmenting a purely monetaryapproach by my theory significantly im-proves the explanation of the financial sec-tor-output connection in the short run. Sec-tion V looks at the persistence of theseeffects.

Some international aspects of the financialsector-aggregate output link are briefly dis-cussed in Section VI and Section VII con-cludes.

I. The Financial Collapse: Some Background

The problems faced by the U.S. financialsystem between October 1930 and March1933 have been described in detail by earlierauthors,6 but it will be useful to recapitulatesome principal facts here. Given this back-ground, attention will be turned to the morecentral issues of the paper.

The two major components of the finan-cial collapse were the loss of confidence infinancial institutions, primarily commercialbanks, and the widespread insolvency ofdebtors. I give short discussions of each ofthese components and of their joint relationto aggregate fluctuations.

A. The Failure of Financial Institutions

Most financial institutions (even semipub-lic ones, like the Joint Stock Land Banks)came under pressure in the 1930's. Some,

6See especially Chandler (1970, 1971) and Friedmanand Schwartz.


VOL. 73 NO. 3 BERNANKE: GREAT DEPRESSION 259

such as the insurance companies and themutual savings banks, managed to maintainsomething close to normal operations. Others,like the building-and-loans (which, despitetheir ability to restrict withdrawals by de-positors, failed in significant numbers) weregreatly hampered in their attempts to carryon their business.7 Of most importance, how-ever, were the problems of the commercialbanks. The significance of the banking diffi-culties derived both from their magnitudeand from the central role commercial banksplayed in the financial system.8

The great severity of the banking crises inthe Great Depression is well known to stu-dents of the period. The percentages of oper-ating banks which failed in each year from1930 to 1933 inclusive were 5.6, 10.5, 7.8,and 12.9; because of failures and mergers,the number of banks operating at the end of1933 was only just above half the numberthat existed in 1929.9 Banks that survivedexperienced heavy losses.

The sources of the banking collapse arebest understood in the historical context. Thefirst point to be made is that bank failureswere hardly a novelty at the time of thedepression. The U.S. system, made up as itwas primarily of small, independent banks,had always been particularly vulnerable.(Countries with only a few large banks, suchas Britain, France, and Canada, never hadbanking difficulties on the American scale.)The dominance of small banks in the UnitedStates was due in large part to a regulatoryenvironment which reflected popular fears oflarge banks and "trusts"; for example, therewere numerous laws restricting branch bank-ing at both the state and national level. Com-

7Hart describes the problems of the building-and-loans. An interesting sidelight here is the additionalstrain on housing lenders caused by the existence of thePostal Savings System; see Maureen O'Hara and DavidEasley(1979).

8According to Raymond Goldsmith (1958), commer-cial banks held 39.6 percent of the assets of all financialintermediaries, broadly defined, in 1929. See his Table11.

9Cyril Upham and Edwin Lamke (1934, p. 247).Since smaller banks were more likely to fail, the fractionof deposits represented by suspended banks was some-what less. Eventual recovery by depositors was about 75percent; see Friedman and Schwartz, p. 438.

petition between the state and national bank-ing systems for member banks also tended tokeep the legal barriers to entry in bankingvery low.10 In this sort of environment, asignificant number of failures was to be ex-pected and probably was even desirable.Failures due to "natural causes" (such as theagricultural depression of the 1920's uponwhich many small, rural banks foundered)were common.11

Besides the simple lack of economic viabil-ity of some marginal banks, however, theU.S. system historically suffered also from amore malign source of bank failures; namely,financial panics. The fact that liabilities ofbanks were principally in the form of fixed-price, callable debt (i.e., demand deposits),while many assets were highly illiquid,created the possibility of the perverse expec-tational equilibrium known as a "run" onthe banks. In a run, fear that a bank may failinduces depositors to withdraw their money,which in turn forces liquidation of the bank'sassets. The need to liquidate hastily, or todump assets on the market when other banksare also liquidating, may generate losses thatactually do cause the bank to fail. Thus theexpectation of failure, by the mechanism ofthe run, tends to become self-confirming.12

An interesting question is why banks atthis time relied on fixed-price demand de-posits, when alternative instruments mighthave reduced or prevented the problem ofruns.13 An answer is provided by Friedmanand Schwartz: They pointed out that, beforethe establishment of the Federal Reserve in1913, panics were usually contained by thepractice of suspending convertibility of bankdeposits into currency. This practice, typi-cally initiated by loose organizations of urban

10Benjamin Klebaner (1974) gives a good brief his-tory of U.S. commercial banking.

Upham and Lamke, p. 247, report that approxi-mately 2-3 percent of all banks in operation failed ineach year of the 1920's.

12Douglas Diamond and Philip Dybvig (1981) for-malize this argument. For an alternative analysis of thephenomenon of runs, see Robert Flood and Peter Garber(1981).

13 For example, equity-like instruments, such as thoseused by modern money-market mutual funds, couldhave been used as the transactions medium. See Ken-neth Cone (1982).



banks called clearinghouses, moderated thedangers of runs by making hasty liquidationunnecessary. In conjunction with the suspen-sion of convertibility practice, the use ofdemand deposits created relatively little in-stability.14

However, with the advent of the FederalReserve (according to Friedman-Schwartz),this roughly stable institutional arrangementwas upset. Although the Federal Reserve in-troduced no specific injunctions against thesuspension of convertibility, the clearing-houses apparently felt that the existence ofthe new institution relieved them of the re-sponsibility of fighting runs. Unfortunately,the Federal Reserve turned out to be unableor unwilling to assume this responsibility.

No serious runs occurred between WorldWar I and 1930; but the many pieces of badfinancial news that came in from around theworld in 1930-32 were like sparks aroundtinder. Runs were clearly an important partof the banking problems of this period. Someevidence emerges from contemporary ac-counts, including descriptions of specificevents precipitating runs. Also notable is thefact that bank failures tended to occur inshort spasms, rather than in a steady stream(see Table 1, col. 2, for monthly data on thedeposits of failing banks). The problem wasnot arrested until government interventionbecame important in late 1932 and early1933.

We see, then, that the banking crises of theearly 1930's differed from earlier recordedexperience both in magnitude and in thedegree of danger posed by the phenomenonof runs. The result of this was that the behav-ior of almost the entire system was adverselyaffected, not just that of marginal banks. Thebankers' fear of runs, as I shall argue below,had important macroeconomic effects.

B. Defaults and Bankruptcies

The second major aspect of the financialcrisis (one that is currently neglected byhistorians) was the pervasiveness of debtor

14 Diamond and Dybvig derive this point formally,with some caveats.

insolvency. Given that debt contracts werewritten in nominal terms,15 the protractedfall in prices and money incomes greatlyincreased debt burdens. According to EvansClark (1933), the ratio of debt service tonational income went from 9 percent in 1929to 19.8 percent in 1932-33. The resultinghigh rates of default caused problems forboth borrowers and lenders.

The "debt crisis" touched all sectors. Forexample, about half of all residential proper-ties were mortgaged at the beginning of theGreat Depression; according to the FinancialSurvey of Urban Housing (reported in Hart),as of January 1, 1934,

The proportion of mortgagedowner-occupied houses with some in-terest or principal in default was innone of the twenty-two cities [surveyed]less than 21 percent (the figure forRichmond, Virginia); in half it wasabove 38 percent; in two (Indianapolisand Birmingham, Alabama) between 50percent and 60 percent; and in one(Cleveland), 62 percent. For rentedproperties, percentages in default ranslightly higher. [p. 164]

Because of the long spell of low foodprices, farmers were in more difficulty thanhomeowners. At the beginning of 1933,owners of 45 percent of all U.S. farms, hold-ing 52 percent of the value of farm mortgagedebt, were delinquent in payments (Hart, p.138). State and local governments—many ofwhom tried to provide relief for the unem-ployed—also had problems paying theirdebts: As of March 1934, the governments of37 of the 310 cities with populations over30,000 and of three states had defaulted onobligations (Hart, p. 225).

In the business sector, the incidence offinancial distress was very uneven. Aggregatecorporate profits before tax were negative in1931 and 1932, and after-tax retained earn-ings were negative in each year from 1930 to1933 (Chandler, 1971, p. 102). But the subset

15 Finding an explanation for the lack of indexed debtduring the deflationary 1930's—as in the inflationary1970's—is a point on which I stumble.



of corporations holding more than $50 mil-lion in assets maintained positive profitsthroughout this period, leaving the brunt tobe borne by smaller companies. SolomonFabricant (1935) reported that, in 1932 alone,the losses of corporations with assets of$50,000 or less equalled 33 percent of totalcapitalization; for corporations with assets inthe $50,000-$100,000 range, the comparablefigure was 14 percent. This led to high ratesof failure among small firms.

Although the deflation of the 1930's wasunusually protracted, there had been a simi-lar episode as recently as 1921-22 which hadnot led to mass insolvency. The seriousnessof the problem in the Great Depression wasdue not only to the extent of the deflation,but also to the large and broad-based expan-sion of inside debt in the 1920's. CharlesPersons surveyed the credit expansion of thepredepression decade in a 1930 article: Hereported that outstanding corporate bondsand notes increased from $26.1 billion in1920 to $47.1 billion in 1928, and that non-federal public securities grew from $11.8 bil-lion to $33.6 billion over the same period.(This may be compared with a 1929 nationalincome of $86.8 billion.) Perhaps more sig-nificantly, during the 1920's, small bor-rowers, such as households and unincor-porated businesses, greatly increased theirdebts. For example, the value of urban realestate mortgages outstanding increased from$11 billion in 1920 to $27 billion in 1929,while the growth of consumer installmentdebt reflected the introduction of major con-sumer durables to the mass market.

Like the banking crises, then, the debtcrisis of the 1930's was not qualitatively anew phenomenon; but it represented a breakwith the past in terms of its severity andpervasiveness.

C. Correlation of the Financial Crisiswith Macroeconomic Activity

The close connection of the stages of thefinancial crisis (especially the bank failures)with changes in real output has been notedby Friedman and Schwartz and by others.An informal review of this connection is

facilitated by the monthly data in Table 1.Column 1 is an index of real industrial pro-duction. Columns 2 and 3 are the (nominal)liabilities of failing banks and nonbank com-mercial businesses, respectively.

The industrial production series revealsthat a recession began in the United Statesduring 1929. By late 1930, the downturn,although serious, was still comparable inmagnitude to the recession of 1920-22; asthe decline slowed, it would have been rea-sonable to expect a brisk recovery, just as in1922.

With the first banking crisis, however, therecame what Friedman and Schwartz called a"change in the character of the contraction"(p. 311). The economy first flattened out,then went into a new tailspin just as thebanks began to fail again in June 1931.

A lengthy slide of both the general econ-omy and the financial system followed. Thebanking situation calmed in early 1932, andnonbank failures peaked shortly thereafter.A new recovery attempt began in August,but failed within a few months.16 In March1933, the bottom was reached for both thefinancial system and the economy as a whole.Measures taken after the banking holidayended the bank runs and greatly reduced theburden of debt. Simultaneously aggregateoutput began a recovery that was sustaineduntil 1937.

The leading explanation of the correlationbetween the conditions of the financial sec-tor and of the general economy is that ofFriedman and Schwartz, who stressed theeffects of the banking crises on the supply ofmoney. I agree that money was an importantfactor in 1930-33, but, because of reserva-tions cited in the introduction, I doubt that itcompletely explains the financial sector-aggregate output connection. This motivates

16Judging by Table 1. the failure of this recoveryseems to be unrelated to financial sector difficulties.However, accounts from the time suggest that the bank-ing crisis of late 1932 and early 1933 (which ended inthe banking holiday) was in fact quite severe; see SusanKennedy (1973). The relatively low reported rate ofbank failures at this time may be an artifact of statemoratoria, restrictions on withdrawals, and other in-terventions.



TABLE 1—SELECTED MACROECONOMIC DATA, JULY 1929-MARCH 1933

Month

1929JASOND

1930JF

MAMJJASOND

1931JF

MAMJJASOND

1932JF

MAMJJAS0ND

1933JF

M

IP

11411411211010510010010098989693898685838179787980808077767370686766646362585654535458605958585754

Banks

60.86.79.7

12.522.315.526.532.423.231.919.457.929.822.821.619.7

179.9372.1

75.734.234.341.743.2

190.540.7

180.0233.5471.4

67.9277.1218.9

51.710.931.634.4

132.748.729.513.520.143.370.9

133.162.2

3276.3a

Fails

32.433.734.131.352.062.561.251.356.849.155.563.129.849.246.756.355.383.794.659.660.450.953.451.761.053.047.370.760.773.296.984.993.8

101.183.876.987.277.056.152.953.664.279.165.648.5

L/IP

.163

.007

.079

.177

.121-.214- .228- .102

.076

.058- .028

.085-.055- .027

.008- .010-.067- .144- .187- .144- .043- .104- .133- .120-.013-.103- .050- .310-.101- .120- .117- .138- .183-.225- .154- .170- .219- .130-.091- .095-.133- .039- .139- .059- .767 a

L/DEP

.851

.855

.860

.865

.854

.851

.837

.834

.835

.826

.820

.818

.802

.800

.799

.791

.111

.775

.763

.747

.738

.722

.706

.707

.704

.706

.713

.716

.726

.732

.745

.757

.744

.718

.696

.689

.677

.662

.641

.623

.602

.596

.576

.583

.607 a

DIF

2.312.332.332.502.682.592.492.482.442.332.412.532.522.472.412.733.063.493.213.083.173.453.994.233.934.294.825.415.306.494.874.764.916.787.877.937.214.774.19AM4.795.074.794.094.03

Notes: IP = seasonally adjusted index of industrial production, 1935-39 = 100; Federal Reserve Bulletin.Banks = deposits of failing banks, $millions; Federal Reserve Bulletin.

Fails = liabilities of failing commercial businesses, $millions; Survey of Current Business.L/PI= ratio of net extensions of commercial bank loans to (monthly) personal income; from Banking and

Monetary Statistics and National Income.L/D= ratio of loans outstanding to the sum of demand and time deposits, weekly reporting banks; Banking and

Monetary Statistics.DIF= difference (in percentage points) between yields on Baa corporate bonds and long-term U.S. government

bonds; Banking and Monetary Statistics.aA national bank holiday was declared in March 1933.



my study of a nonmonetary channel throughwhich an additional impact of the financialcrisis may have been felt.

II. The Effect of the Crisis on the Costof Credit Intermediation

This paper posits that, in addition to itseffects via the money supply, the financialcrisis of 1930-33 affected the macroeconomyby reducing the quality of certain financialservices, primarily credit intermediation. Thebasic argument is to be made in two steps.First, it must be shown that the disruption ofthe financial sector by the banking and debtcrises raised the real cost of intermediationbetween lenders and certain classes of bor-rowers. Second, the link between higher in-termediation costs and the decline in aggre-gate output must be established. I presenthere the first step of the argument, leavingthe second to be developed in Sections III-V.

In order to discuss the quality of perfor-mance of the financial sector, I must firstdescribe the real services that the sector issupposed to provide. The specification ofthese services depends on the model of theeconomy one has in mind. We shall clearlynot be interested in economies of the sortdescribed by Eugene Fama (1980), in whichfinancial markets are complete and informa-tion/transactions costs can be neglected. Insuch a world, banks and other intermediariesare merely passive holders of portfolios.Banks' choice of portfolios or the scale of thebanking system can never make any dif-ference in this case, since depositors canoffset any action taken by banks throughprivate portfolio decisions.17

As an alternative to the Fama complete-markets world, consider the following stylizeddescription of the economy. Let us supposethat savers have many ways of transferringresources from present to future, such asholding real assets or buying the liabilities of

governments or corporations on well-orga-nized exchanges. One of the options savershave is to lend resources to a banking sys-tem. The banks also have a menu of differentassets to choose from. Assume, however, thatbanks specialize in making loans to small,idiosyncratic borrowers whose liabilities aretoo few in number to be publicly traded.(Here is where the complete-markets as-sumption is dropped.)

The small borrowers to whom the bankslend will be taken, for simplicity, to be oftwo extreme types, "good" and "bad." Goodborrowers desire loans in order to undertakeindividual-specific investment projects. Theseprojects generate a random return from adistribution whose mean will be assumedalways to exceed the social opportunity costof investment. If this risk is nonsystematic,lending to good borrowers is socially desir-able. Bad borrowers try to look like goodborrowers, but in fact they have no "project."Bad borrowers are assumed to squander anyloan received in profligate consumption, thento default. Loans to bad borrowers are so-cially undesirable.

In this model, the real service performedby the banking system is the differentiationbetween good and bad borrowers.18 For acompetitive banking system, I define the costof credit intermediation (CCI) as being thecost of channeling funds from the ultimatesavers/lenders into the hands of good bor-rowers. The CCI includes screening, moni-toring, and accounting costs, as well as theexpected losses inflicted by bad borrowers.Banks presumably choose operating proce-dures that minimize the CCI. This is done bydeveloping expertise at evaluating potentialborrowers; establishing long-term relation-ships with customers; and offering loan con-ditions that encourage potential borrowers toself-select in a favorable way.19

Given this simple paradigm, I can describethe effects of the two main components of

17It should be noted that the phenomena emphasizedby Friedman and Schwartz—the effects of the contrac-tion of the banking system on the quantity of thetransactions medium and on real output—are also im-possible in a complete-markets world.

18 To concentrate on credit intermediation, I neglectthe transactions and other services performed by banks.

19See Dwight Jaffee and Thomas Russell (1976) andJoseph Stiglitz and Andrew Weiss (1981) on the waybanks induce favorable borrower self-selection.



the financial crisis on the efficiency of thecredit allocation process (i.e., on the CCI).

A. Effect of the Banking Crises on the CCI

The banking problems of 1930-33 dis-rupted the credit allocation process by creat-ing large, unplanned changes in the channelsof credit flow. Fear of runs led to largewithdrawals of deposits, precautionary in-creases in reserve-deposit ratios, and an in-creased desire by banks for very liquid orrediscountable assets. These factors, plus theactual failures, forced a contraction of thebanking system's role in the intermediationof credit.20 Some of the slack was taken upby the growing importance of alternativechannels of credit (see below). However, therapid switch away from the banks (given thebanks' accumulated expertise, information,and customer relationships) no doubt im-paired financial efficiency and raised theCCI.21

It would be useful to have a direct mea-sure of the CCI; unfortunately, no reallysatisfactory empirical representation of thisconcept is available. Reported commercialloan rates reflect loans that are actually made,not the shadow cost of bank funds to arepresentative potential borrower; sincebanks in a period of retrenchment make onlythe safest and highest-quality loans, mea-sured loan rates may well move inversely tothe CCI. I obtained a number of interestingresults using the yield differential betweenBaa corporate bonds and U.S. governmentbonds as a proxy for the CCI; however, theuse of the Baa rate is not consistent with mystory that bank borrowers are those whoseliabilities are too few to be publicly traded.

While we cannot observe directly the ef-fects of the banking troubles on the CCI, wecan see their impact on the extension of bankcredit: Table 1 gives some illustrative data.Column 4 gives, as a measure of the flow of

2 0 For an interesting contemporary account of thisprocess, see the article by Eugene H. Burris in theAmerican Banker, October 15, 1931.

21 Since intermediation resources could have beenshifted out of the beleaguered banking sector (givenenough time), mine is basically a costs-of-adjustmentargument.

bank credit, the monthly change in bankloans outstanding, normalized by monthlypersonal income.22 One might have expectedthe loan-change-to-income ratio to be drivenprimarily by loan demand and thus by therate of production. Comparison with the firsttwo columns of Table 2 shows, however, thatthe banking crises were as important a de-terminant of this variable as output. Forexample, except for a brief period of liquida-tion of speculation loans after the stockmarket crash, credit outstanding declinedvery little before October 1930—this despitea 25 percent fall in industrial production thathad occurred by that time. With the firstbanking crisis of November 1930, however, along period of credit contraction was ini-tiated. The shrinkage of credit shared therhythm of the banking crises; for example, inOctober 1931, the worst month for bankfailure before the bank holiday, net creditreduction was a record 31 percent of per-sonal income.23

The fall in bank loans after November1930 was not simply a balance sheet reflec-tion of the decline in deposits. Column 5 inTable 1 gives the monthly ratio of outstand-ing bank loans to the sum of demand andtime deposits. This ratio declined sharply asbanks switched out of loans and into moreliquid investments.

The perception that the banking crises andthe associated scrambles for liquidity exerteda deflationary force on bank credit wasshared by writers of the time. A 1932 Na-tional Industrial Conference Board survey of

22 In the construction of the bank loans series, datafrom weekly reporting member banks (which held about40 percent of all bank loans) were used to interpolatebetween less frequent aggregate observations. Note that,for our purposes, looking at the change in loans ispreferable to considering the stock of real loans out-standing: In a regime of nominally contracted debt andsharp unanticipated deflation, stability of the stock ofreal debt does not signal a comfortable situation forborrowers.

23 The effect of bank failures on credit outstanding issomewhat exaggerated by the fact that the credit con-traction measure includes the loans of suspending banksthat were not transferred to other banks; however, Iestimate that this accounting convention is responsiblefor less than one-eighth of the total (measured) creditcontraction between October 1930 and February 1933.



credit conditions reported that "During 1930,the shrinkage of commercial loans no morethan reflected business recession. During1931 and the first half of 1932 (the periodstudied), it unquestionably represented pres-sure by banks on customers for repayment ofloans and refusal by banks to grant newloans" (p. 28). Other contemporary sourcestended to agree (see, for example, Chandler,1971, pp. 233-39, for references).

Two other observations about the contrac-tion of bank credit can be made. First, theclass of borrowers most affected by creditreductions were households, farmers, unin-corporated businesses, and small corpora-tions; this group had the highest direct orindirect reliance on bank credit. Second, thecontraction of bank credit was twice as largeas that of other major countries, even thosewhich experienced comparable output de-clines (Klebaner, p. 145).

The fall in bank loans outstanding waspartly offset by the relative expansion ofalternative forms of credit. In the area ofconsumer finance, retail merchants, servicecreditors, and nonbank lending agencies im-proved their position relative to banks andprimarily bank-supported installment fi-nance companies (Rolf Nugent, 1939, pp.114-16). Small firms during this period sig-nificantly reduced their traditional relianceon banks in favor of trade credit (CharlesMerwin, 1942, pp. 5 and 75). But, as arguedabove, in a world with transactions costs andthe need to discriminate among borrowers,these shifts in the loci of credit intermedia-tion must have at least temporarily reducedthe efficiency of the credit allocation process,thereby raising the effective cost of credit topotential borrowers.

B. The Effect of Bankruptcies on the CCI

I turn now to a brief discussion of theimpact of the increase in defaults and bank-ruptcies during this period on the cost ofcredit intermediation.

The very existence of bankruptcy proceed-ings, rather than being an obvious or naturalphenomenon, raises deep questions of eco-nomic theory. Why, for example, do the

creditor and defaulted debtor make the pay-ments to third parties (lawyers, administra-tors) that these proceedings entail, instead ofsomehow agreeing to divide those paymentsbetween themselves? In a complete-marketsworld, bankruptcy would never be observed;this is because complete state-contingent loanagreements would uniquely define eachparty's obligations in all possible circum-stances, rendering third-party arbitration un-necessary. That we do observe bankruptcies,in our incomplete-markets world, suggeststhat creditors and debtors have found thecombination of simple loan arrangementsand ex post adjudication by bankruptcy(when necessary) to be cheaper than attempt-ing to write and enforce complete state-con-tingent contracts.

To be more concrete, let us use the "goodborrower-bad borrower" example. In writinga loan contract with a potential borrower,the bank has two polar options. First, itmight try to approximate the completestate-contingent contract by making the bor-rower's actions part of the agreement and byallowing repayment to depend on the out-come of the borrower's project. This con-tract, if properly written and enforced, wouldcompletely eliminate the possibility of eitherside not being able to meet its obligations;its obvious drawback is the cost of monitor-ing which it involves. The bank's other op-tion is to write a very simple agreement("payment of such-amount to be made onsuch-date"), then to make the loan only if itbelieves that the borrower is likely to repay.The second approach usually dominates thefirst, of course, especially for small bor-rowers.

A device which makes the cost advantageof the simpler approach even greater is theuse of collateral. If the borrower has wealththat can be attached by the bank in the eventof nonpayment, the bank's risk is low. More-over, the threat of loss of collateral providesthe right incentives for borrowers to use loansonly for profitable projects. Thus, the combi-nation of collateral and simple loan contractshelps to create a low effective CCI.

A useful way to think of the 1930-33 debtcrisis is as the progressive erosion of bor-rowers' collateral relative to debt burdens.



As the representative borrower became moreand more insolvent, banks (and other lendersas well) faced a dilemma. Simple, noncontin-gent loans faced increasingly higher risks ofdefault; yet a return to the more complextype of contract involved many other costs.Either way, debtor insolvency necessarilyraised the CCI for banks.

One way for banks to adjust to a higherCCI is to increase the rate that they chargeborrowers. This may be counterproductive,however, if higher interest charges increasethe risk of default. The more usual responseis for banks just not to make loans to somepeople that they might have lent to in bettertimes. This was certainly the pattern in the1930's. For example, it was reported that theextraordinary rate of default on residentialmortgages forced banks and life insurancecompanies to "practically stop making mort-gage loans, except for renewals" (Hart, p.163). This situation precluded many bor-rowers, even with good projects, from gettingfunds, while lenders rushed to compete forexisting high-grade assets. As one writer ofthe time, D. M. Frederiksen, put it:

We see money accumulating at thecenters, with difficulty of finding safeinvestment for it; interest rates drop-ping down lower than ever before;money available in great plenty forthings that are obviously safe, but notavailable at all for things that are infact safe, and which under normal con-ditions would be entirely safe (and thereare a great many such), but which arenow viewed with suspicion by lenders.

[1931, p. 139]

As this quote suggests, the idea that the lowyields on Treasury or blue-chip corporationliabilities during this time signalled a generalstate of "easy money" is mistaken; moneywas easy for a few safe borrowers, but dif-ficult for everyone else.

An indicator of the strength of lenderpreferences for safe, liquid assets (and henceof the difficulty of risky borrowers in ob-taining funds) is the yield differential be-tween Baa corporate bonds and Treasury

bonds (Table 1, column 6). Because this vari-able contains no adjustment for the reclassi-fication of firms into higher risk categories, ittends to understate the true difference inyields between representative risky and safeassets. Nevertheless, this indicator showedsome impressive shifts, going from 2.5 per-cent during 1929-30 to nearly 8 percent inmid-1932. (The differential never exceeded3.5 percent in the sharp 1920-22 recession.)The yield differential reflected changing per-ceptions of default risk, of course; but notealso the close relationship of the differentialand the banking crises (a fact first pointedout by Friedman and Schwartz). Bank crisesdepressed the prices of lower-quality invest-ments as the fear of runs drove banks intoassets that could be used as reserves or forrediscounting. This effect of bank portfoliochoices on an asset price could not happen ina Fama-type, complete-markets world.

Finally, it is instructive to consider theexperience of a country that had a debt crisiswithout a banking crisis. Canada entered theGreat Depression with a large external debt,much of it payable in foreign currencies. Thecombination of deflation and the devalua-tion of the Canadian dollar led to manydefaults. Internally, debt problems in agri-culture and in mortgage markets were assevere as in the United States, while majorindustries (notably pulp and paper) expe-rienced many bankruptcies (A. E. Safarian,1959, ch. 7). Although Canadian bankers didnot face serious danger of runs, they shiftedaway from loans to safer assets. This shifttoward safety and liquidity, though less pro-nounced than in the U.S. case, drew criticismfrom all facets of Canadian society. TheAmerican Banker of December 6, 1932, re-ported the following complaint from a non-populist Canadian politician:

The chief criticism of our presentsystem appears to be that in good timescredit is expanded to great extremes...but, when the pinch of hard times isfirst being felt, credit is suddenly anddrastically restricted by the banks... Atthe present time, loans are only beingmade when the banks have a very wide



margin of security and every effort isbeing made to collect outstanding loans.All our banks are reaching out in anendeavor to liquefy their assets....

[p.l]

Canadian lenders other than banks also triedto retrench: According to the Financial Post,May 14, 1932, "Insurance, trust, and loancompanies were increasingly unwilling tolend funds with real estate and rental valuesfalling, a growing number of defaults ofinterest and principal, the increasing burdenof property taxes, and legislation whichadversely affected creditors" (quoted inSafarian, p. 130).

More careful study of the Canadian expe-rience in the Great Depression would beuseful. However, on first appraisal, that ex-perience does not seem to be inconsistentwith the point that even good borrowers mayfind it more difficult or costly to obtaincredit when there is extensive insolvency.The debt crisis should be added to the bank-ing crises as a potential source of disruptionof the credit system.

III. Credit Markets and MacroeconomicPerformance

If it is taken as given that the financialcrises during the depression did interfere withthe normal flows of credit, it still must beshown how this might have had an effect onthe course of the aggregate economy.

There are many ways in which problems incredit markets might potentially affect themacroeconomy. Several of these could begrouped under the heading of "effects onaggregate supply." For example, if creditflows are dammed up, potential borrowers inthe economy may not be able to secure fundsto undertake worthwhile activities or invest-ments; at the same time, savers may have todevote their funds to inferior uses. Otherpossible problems resulting from poorlyfunctioning credit markets include a reducedfeasibility of effective risk sharing and greaterdifficulties in funding large, indivisible proj-ects. Each of these might limit the economy'sproductive capacity.

These arguments are reminiscent of someideas advanced by John Gurley and E. S.Shaw (1955), Ronald McKinnon (1973), andothers in an economic development context.The claim of this literature is that immatureor repressed financial sectors cause the"fragmentation" of less developed econo-mies, reducing the effective set of productionpossibilities available to the society.

Did the financial crisis of the 1930's turnthe United States into a " temporarily under-developed economy" (to use Bob Hall's felic-itous phrase)? Although this possibility isintriguing, the answer to the question isprobably no. While many businesses didsuffer drains of working capital and invest-ment funds, most larger corporations enteredthe decade with sufficient cash and liquidreserves to finance operations and any de-sired expansion (see, for example, FriedrichLutz, 1945). Unless it is believed that theoutputs of large and of small businesses arenot potentially substitutes, the aggregatesupply effect must be regarded as not ofgreat quantitative importance.

The reluctance of even cash-rich corpora-tions to expand production during the de-pression suggests that consideration of theaggregate demand channel for credit marketeffects on output may be more fruitful. Theaggregate demand argument is in fact easy tomake: A higher cost of credit intermediationfor some borrowers (for example, householdsand smaller firms) implies that, for a givensafe interest rate, these borrowers must facea higher effective cost of credit. (Indeed, theymay not be able to borrow at all.) If thishigher rate applies to household and smallfirm borrowing but not to their saving (theymay only earn the safe rate on their savings),then the effect of higher borrowing costs isunambiguously to reduce their demands forcurrent-period goods and services. This puresubstitution effect (of future for present con-sumption) is easily derived from the classicaltwo-period model of savings.24

24 The classical model may be augmented, if thereader desires, by considerations of liquidity constraints,bankruptcy costs, or risk aversion; see my 1981 paper.



Assume that the behavior of borrowersunaffected by credit market problems is un-changed. Then the paragraph above impliesthat, for a given safe rate, an increase in thecost of credit intermediation reduces the totalquantity of goods and services currently de-manded. That is, the aggregate demand curve,drawn as a function of the safe rate, isshifted downward by a financial crisis. Inany macroeconomic model one cares to use,this implies lower output and lower safe in-terest rates. Both of these outcomes char-acterized 1930-33, of course.

Some evidence on the magnitude of theeffect of the financial market problems onaggregate output is now presented.

IV. Short-Run Macroeconomic Impactsof the Financial Crisis

This section studies the short-run or "im-pact" effects of the financial crisis. For thispurpose, I use only monthly data on therelevant variables. In addition, rather thanconsider the 1929-33 episode outside of itscontext, I have widened the sample to in-clude the entire interwar period (January1919-December 1941).

Section I.C above has already given someevidence of the relationship between thetroubles of the financial sector and those ofthe economy as a whole. However, supportfor the thesis of this paper requires thatnonmonetary effects of the financial crisis onoutput be distinguished from the monetaryeffects studied by Friedman and Schwartz.My approach will be to fit output equationsusing monetary variables, then to show thatadding proxies for the financial crisis sub-stantially improves the performance of theseequations. Comparison of financial to totallynonfinancial sources of the Great Depres-sion, such as those suggested by Temin, isleft to future research.

To isolate the purely monetary influenceson the economy, one needs a structural ex-planation of the money-income relationship.Lucas (1972) has presented a formal modelin which monetary shocks affect productiondecisions by causing confusion about theprice level. Influenced by this work, mostrecent empirical studies of the role of money

have related national income to measuresof "unanticipated" changes in money orprices.25

The most familiar way of constructing aproxy for unanticipated components of avariable is the two-step method of RobertBarro (1978), in which the residuals from afirst-stage prediction equation for (say) mon-ey are employed as the independent variablesin a second-stage regression. I experimentedwith both the Barro approach and somealternatives.26 Since my conclusions wereunaffected by choice of technique, I reporthere only the Barro-type results.

In the spirit of the Lucas-Barro analysis, Iconsidered the effects of both "moneyshocks" and "price shocks" on output. Mon-ey shocks (M—Me) were defined as theresiduals from a regression of the rate ofgrowth of Ml on four lags of the growthrates of industrial production, wholesaleprices, and Ml itself; price shocks (P - Pe)were defined symmetrically.27I used ordinaryleast squares to estimate the effects of moneyand price shocks on the rate of growth ofindustrial production, relative to trend.

The basic regression results for the inter-war sample period are given as equations (1)and (2) in Table 2. These two equations areof interest, independently of the other resultsof this paper. The estimated "Lucas supplycurve," equation (2), shows an effect of priceshocks on output that is statistically andeconomically significant. As such, it comple-ments the results of Thomas Sargent (1976),who found a similar relationship for thepostwar. The relationship of output to mon-ey surprises, equation (1), is a bit weaker.The fact that we discover a smaller role formoney in the monthly data than does PaulEvans (1981) is primarily the result of ourinclusion of lagged values of production onthe right-hand side. This inclusion seemsjustified both on statistical grounds and for

25A notable exception is Mishkin (1982).26 Principal alternatives tried were 1) the use of antic-

ipated as well as unanticipated quantities as explanatoryvariables; and 2) reestimation of some equations by themore efficient but computationally more complexmethod of Andrew Abel and Mishkin (1981).

27The first-stage regressions were unsurprising and,for the sake of space, are not reported.


VOL. 73 NO. 3 BERNANKE: GREAT DEPRESSION

TABLE 2—ESTIMATED OUTPUT EQUATIONS

269

Notes: Yt= rate of growth of industrial production (Federal Reserve Bulletin), relative to exponential trend.( M - Me)t = rate of growth of M1, nominal and seasonally adjusted (Friedman and Schwartz, Table 4-1), less

predicted rate of growth.(P - Pe)t= rate of growth of wholesale price index (Federal Reserve Bulletin), less predicted rate of growth.DBANKSt= first difference of deposits of failing banks (deflated by wholesale price index).DFAILSt= first difference of liabilities of failing businesses (deflated by wholesale price index).

Data are monthly; t-statistics are shown in parentheses.

the economic reason that costs of adjustingproduction can be presumed to create a serialdependence in output. Like Evans, I was notable to find effects of money (or prices)lagged more than three months.

While these regression results exhibit sta-tistical significance and the expected signsfor coefficients, they are disappointing in thefollowing sense: When equations (1) and (2)

are used to perform dynamic simulations ofthe path of output between mid-1930 and thebank holiday of March 1933, they capture nomore than half of the total decline of outputduring the period. This is the basis of thecomment in the introduction that the de-clines in money seem "quantitatively insuffi-cient" to explain what happened to output in1930-33.



Given the basic regressions (1) and (2), thenext step was to examine the effects of in-cluding proxies for the nonmonetary finan-cial impact as explanators of output. Basedon the earlier analysis of this paper, the mostobvious such proxies are the deposits of fail-ing banks and the liabilities of failing busi-nesses.

A preliminary problem with the bank de-posits series that needs to be discussed is thevalue for March 1933, the month of the bankholiday. As can be seen in Table 1, thedeposits of banks suspended in March 1933is seven times that of the next worse month.The question arises if any adjustment shouldbe made to that figure before running theregressions.

We believe that it would be a mistake toeliminate totally the bank holiday episodefrom the sample. According to contemporaryaccounts, rather than being an orderly andplanned-in-advance policy, the imposition ofthe holiday was a forced response to themost panicky and chaotic financial condi-tions of the period. The deposits of sus-pended banks figure for March, as large as itis, reflects not all closed banks but onlythose not licensed to reopen by June 30,1933. Of these banks, most were liquidatedor placed in receivership; less than 25 per-cent had been licensed to reopen as of De-cember 31, 1936.28 Qualitatively, then, theMarch 1933 episode resembled the earliercrises; it would be throwing away informa-tion not to include in some way the effects ofthis crisis and of its resolution on the econ-omy.

On the other hand, the mass closing ofbanks by government action probably createdless confusion and fear of future crises thanwould have a similar number of suspensionsoccurring without government intervention.As a conservative compromise, I assumedthat the "supervised" bank closings of March1933 had the same effect as an "unsuper-vised" bank crisis involving 15 percent asmuch in frozen deposits. This scales downthe March 1933 episode to about the size ofthe events of October 1931. The sensitivity ofthe results to this assumption is as follows:

28Federal Reserve Bulletin, 1937, pp. 866-67.

increasing the amount of importance attrib-uted to the March 1933 crisis raises themagnitude and statistical significance of themeasured effects of the financial crises onoutput. (It is in this sense that the 15 percentfigure is conservative.) However, the bankfailure coefficients in the regressions retainedhigh significance even when less weight wasgiven to March 1933.

I turn now to the results of adding (real)deposits of failing banks and liabilities offailing businesses to the output equations(see equations (3) and (4) in Table 2). Thesample period begins in 1921 because of theunavailability of data on monthly bankfailures before then. In both regressions, cur-rent and lagged first differences of the addedvariables enter the explanation of the growthrate of industrial production (relative totrend) with the expected sign and, takenjointly, with a high level of statistical signifi-cance. The magnitudes and significance ofthe coefficients of money and price shocksare not much changed. This provides at leasta tentative confirmation that nonmonetaryeffects of the financial crisis augmentedmonetary effects in the short-run determina-tion of output.

Some alternative proxies for the nonmone-tary component of the financial crisis werealso tried. For the sake of space, only asummary of these results is given. 1) Toexamine the direct effects of the contractionof bank credit on the economy, I began byregressing the rate of growth of bank loanson current and lagged values of suspendedbank deposits and of failing business liabili-ties. (This regression indicated a powerfulnegative effect of financial crisis on bankloans.) The fitted series from this regressionwas used as a proxy for the portion of thecredit contraction induced by the financialcrisis. In the presence of money or priceshocks, the effect of a decline in this variableon output was found to be negative for twomonths, positive for the next two months,then strongly negative for the fifth and sixthmonths after the decline. For the period from1921 until the bank holiday, and with mone-tary variables included, the total effect ofcredit contraction on output (as measured bythe sum of lag coefficients in a polynomial



distributed lag) was large (comparable to themonetary effect), negative, and significant atthe 95 percent level. For the entire interwarsample, however, the statistical significanceof this variable was much reduced. This lastresult is due to the fact that the recovery of1933-41 was financed by nonbank sources,with bank loans remaining at a low level.

2) Another proxy for the financial crisisthat was tried was the differential betweenBaa corporate bond yields and the yields onU.S. bonds. As described in Section I.C, thisvariable responded strongly to both bankcrises and the problems of debtors, and assuch was a sensitive indicator of financialmarket conditions. The yield differentialvariable turned out to enter very strongly asan explanator of current and future outputgrowth, overall and in every subsample. Asmuch of this predictive power was no doubtdue to pure financial market anticipations offuture output declines, I also put the dif-ferential variable through a first-stage regres-sion on the liabilities of bank and businessfailures. Assuming that these latter variablesthemselves were not determined by anticipa-tions of future output declines (see below),the use of the fitted series from this regres-sion "purged" the differential variable of itspure anticipatory component. The fittedseries entered the output equations lessstrongly than the raw series, but it retainedthe right sign and statistical significance atthe 95 percent confidence level.

In almost every case, then, the addition ofproxies for the general financial crisis im-proved the purely monetary explanation ofshort-run (monthly) output movements. Thisfinding was robust to the obvious experi-ments. For example, with the above-notedexception of the credit variable in 1933-41,coefficients remained roughly stable oversubsamples. Another experiment was to in-clude free dummy variables for each quarterfrom 1931:1 to 1932:IV in the above re-gressions. The purpose of this was to test thesuggestion that our results are only a reflec-tion of the fact that both the output andfinancial crisis variables "moved a lot" dur-ing 1930-33. The rather surprising discoverywas that the inclusion of the dummiesincreased the magnitude and statistical sig-

nificance of the coefficients on bank andbusiness failures. Finally, the economic sig-nificance of the results was tested by usingthe various estimated equations to run dy-namic simulations of monthly levels of in-dustrial production (relative to trend) formid-1930 to March 1933. Relative to thepure money-shock and price-shock simula-tions described above, the equations includ-ing financial crisis proxies did well. Equa-tions (3) and (4) reduced the mean squaredsimulation error over (1) and (2) by about 50percent. The other (nonreported) equationsdid better; for example, those using the yielddifferential variable reduced the MSE ofsimulation from 90 to 95 percent.

These results are promising. However, acaveat must be added: To conclude that theobserved correlations support the theory out-lined in this paper requires an additionalassumption, that failures of banks and com-mercial firms are not caused by anticipationsof (future) changes in output. To the extentthat, say, bank runs are caused by the receiptof bad news about next month's industrialproduction, the fact that bank failures tendto lead production declines does not provethat the bank problems are helping to causethe declines.29

While it may not be possible to convincethe determined skeptic that bank and busi-ness failures are not purely anticipatory phe-nomena, a good case can be made againstthat position. For example, while in somecases a bad sales forecast may induce a firmto declare bankruptcy, more often that op-tion is forced by insolvency (a result of pastbusiness conditions). For banks, it might wellbe argued that not only are failures relativelyindependent of anticipations about output,but that they are not simply the product ofcurrent and past output performance either:First, banking crises had never previous tothis time been a necessary result of declinesin output.30 Second, Friedman and Schwartz,as well as other writers, have identified

29Actually, a similar criticism might be made ofBarro's work and my own money and price regressions.

30Philip Cagan (1965) makes this point; see pp. 216,227-28. The 1920-22 recession, for example, did notgenerate any banking problems.



specific events that were important sourcesof bank runs during 1930-33. These includethe revelation of scandal at the Bank of theUnited States (a private bank, which in De-cember 1930 became the largest bank to failup to that time); the collapse of the Kredit-anstalt in Austria and the ensuing financialpanics in central Europe; Britain's going offgold; the exposure of huge pyramidingschemes in the United States and Europe;and others, all connected very indirectly (ifat all) with the path of industrial productionin the United States.

If it is accepted that bank suspensions andbusiness bankruptcies were the product offactors beyond pure anticipations of outputdecline, then the evidence of this sectionsupports the view that nonmonetary aspectsof the financial crisis were at least part of thepropagatory mechanism of the Great De-pression. If it is further accepted that thefinancial crisis contained large exogenouscomponents (there is evidence for this in thecase of the banking panics), then there areelements of causality in the story as well.

V. Persistence of the Financial Crisis

The claim was made in the introductionthat my theory seems capable, unlike themajor alternatives, of explaining the unusuallength and depth of the Great Depression. Inthe previous section, I attempted to deal withthe issue of depth; simulations of the esti-mated regressions suggested that the com-bined monetary and nonmonetary effects ofthe financial crisis can explain much of theseverity of the decline in output. In thissection, the question of the length of theGreat Depression is addressed.

As a matter of theory, the duration of thecredit effects described in Section II abovedepends on the amount of time it takes to 1)establish new or revive old channels of creditflow after a major disruption, and 2) re-habilitate insolvent debtors. Since theseprocesses may be difficult and slow, the per-sistence of nonmonetary effects of financialcrisis has a plausible basis. (In contrast, per-sistence of purely monetary effects relies onthe slow diffusion of information or unex-plained stickiness of wages and prices.) Of

course, plausibility is not enough; some evi-dence on the speed of financial recoveryshould be adduced.

After struggling through 1931 and 1932,the financial system hit its low point in March1933, when the newly elected PresidentRoosevelt's "bank holiday" closed downmost financial intermediaries and markets.March 1933 was a watershed month in severalways: It marked not only the beginning ofeconomic and financial recovery but also theintroduction of truly extensive governmentinvolvement in all aspects of the financialsystem.31 It might be argued that the federallydirected financial rehabilitation—which tookstrong measures against the problems of bothcreditors and debtors—was the only majorNew Deal program that successfully pro-moted economic recovery.32 In any case, thelarge government intervention is prima facieevidence that by this time the public had lostconfidence in the self-correcting powers ofthe financial structure.

Although the government's actions set thefinancial system on its way back to health,recovery was neither rapid nor complete.Many banks did not reopen after the holi-day, and many that did open did so on arestricted basis or with marginally solventbalance sheets. Deposits did not flow backinto the banks in great quantities until 1934,and the government (through the Recon-struction Finance Corporation and otheragencies) had to continue to pump largesums into banks and other intermediaries.Most important, however, was a noticeablechange in attitude among lenders; theyemerged from the 1930-33 episode chas-tened and conservative. Friedman andSchwartz (pp. 449-62) have documented theshift of banks during this time away frommaking loans toward holding safe and liquidinvestments. The growing level of bankliquidity created an illusion (as Friedmanand Schwartz pointed out) of easy money;

31See Chandler (1970), ch. 15, and Friedman andSchwartz, ch. 8.

32 E. Carey Brown (1956) has argued that New Dealfiscal policy was not very constructive. A paper byMichael Weinstein in Brunner (1981) points out coun-terproductive aspects of the N.R.A.



however, the combination of lender reluc-tance and continued debtor insolvency in-terfered with credit flows for several yearsafter 1933.

Evidence of postholiday credit problems isnot hard to find. For example, small busi-nesses, which (as I have noted) suffered dis-proportionately during the Contraction, hadcontinuing difficulties with credit during re-covery. Lewis Kimmel (1939) carried out asurvey of credit availability during 1933-38as a companion to the National IndustrialConference Board's 1932 survey. His conclu-sions are generally sanguine (this may reflectthe fact that the work was commissioned bythe American Bankers Association). How-ever, his survey results (p. 65) show that, ofresponding manufacturing firms normallydependent on banks, refusal or restriction ofbank credit was reported by 30.2 percent ofvery small firms (capitalization less than$50,000); 14.3 percent of small firms($50,001-$500,000); 10.3 percent of mediumfirms ($500,001-$1,000,000); and 3.2 percentof the largest companies (capital over $1million). (The corresponding results from the1932 NICB survey were 41.3, 22.2, 12.5, and9.7 percent.)

Two well-known economists, Hardy andViner, conducted a credit survey in the Sev-enth Federal Reserve District in 1934-35.Based on "intensive coverage of 2600 indi-vidual cases," they found "a genuineunsatisfied demand for credit by solvent bor-rowers, many of whom could make economi-cally sound use of working capital.... Thetotal amount of this unsatisfied demand forcredit is a significant factor, among manyothers, in retarding business recovery." Theyadded, "So far as small business is con-cerned, the difficulty in getting bank credithas increased more, as compared with a fewyears ago, than has the difficulty of gettingtrade credit." (These passages are quoted inW. L. Stoddard, 1940.)

Finally, another credit survey for the1933-38 period was done by the Small Busi-ness Review Committee for the U.S. Depart-ment of Commerce. This study surveyed6,000 firms with between 21 and 150 em-ployees. From these they chose a specialsample of 600 companies "selected because

of their high ratings by a standard commer-cial rating agency." Even within the elitesample, 45 percent of the firms reporteddifficulty in securing funds for workingcapital purposes during this period; and 75percent could not obtain capital or long-termloan requirements through regular markets.(See Stoddard.)

The reader may wish to view the AmericanBankers Association and Small Business Re-view Committee surveys as lower and upperbounds, with the Hardy-Viner study in themiddle. In any case, the consensus fromsurveys, as well as the opinion of carefulstudents such as Chandler, is that credit dif-ficulties for small business persisted for atleast two years after the bank holiday.33

Home mortgage lending was another im-portant area of credit activity. In this sphere,private lenders were even more cautious after1933 than in business lending. They had areason for conservatism; while businessfailures fell quite a bit during the recovery,real estate defaults and foreclosures con-tinued high through 1935.34 As has beennoted, some traditional mortgage lendersnearly left the market: life insurance compa-nies, which made $525 million in mortgageloans in 1929, made $10 million in new loansin 1933 and $16 million in 1934.35 Duringthis period, mortgage loans that were madeby private institutions went only to the verybest potential borrowers. Evidence for this isthe sharp drop in default rates of loans madein the early 1930's as compared to loansmade in earlier years (see Carl Behrens, 1952,p. 11); this decline was too large to be ex-plained by the improvement in business con-ditions alone.

To the extent that the home mortgagemarket did function in the years immediatelyfollowing 1933, it was largely due to thedirect involvement of the federal govern-ment. Besides establishing some importantnew institutions (such as the FSLIC and thesystem of federally chartered savings andloans), the government "readjusted" existingdebts, made investments in the shares of

33See Chandler (1970), pp. 150-51.34U.S. Department of Commerce (1975), series N301.35U.S. Department of Commerce (1975), N282.



thrift institutions, and substituted for re-calcitrant private institutions in the provisionof direct credit. In 1934, the government-sponsored Home Owners' Loan Corporationmade 71 percent of all mortgage loans ex-tended.36

Similar conditions obtained for farm creditand in other markets, but space does notpermit this to be pursued here. Summarizingthe reading of all of the evidence byeconomists and by other students of theperiod, it seems safe to say that the return ofthe private financial system to normal condi-tions after March 1933 was not rapid; andthat the financial recovery would have beenmore difficult without extensive governmentintervention and assistance. A moderateestimate is that the U.S. financial systemoperated under handicap for about five years(from the beginning of 1931 to the end of1935), a period which covers most of thetime between the recessions of 1929-30 and1937-38. This is consistent with the claimthat the effects of financial crisis can helpexplain the persistence of the depression.

VI. International Aspects

The Great Depression was a worldwidephenomenon; banking crises, though occur-ring in a number of important countriesbesides the United States, were not soubiquitous. A number of large countries hadno serious domestic banking problems, yetexperienced severe drops in real income inthe early 1930's. Can this be made consistentwith the important role we have ascribed tothe financial crisis in the United States? Acomplete answer would require anotherpaper; but I offer some observations:

1) The experience of different countriesand the mix of depressive forces each facedvaried significantly. For example, Britain,suffering from an overvalued pound, hadhigh unemployment throughout the 1920's;after leaving gold in 1931, it was one of thefirst countries to recover. The biggest prob-lems of food and raw materials exporterswere falling prices and the drying up of

36U.S. Department of Commerce (1975), N278 andN283.

overseas markets. Thus we need not look tothe domestic financial system as an im-portant cause in every case.

2) The countries in which banking crisesoccurred (the United States, Germany,Austria, Hungary, and others) were amongthe worst hit by the depression. Moreover,these countries held a large share of worldtrade and output. The United States aloneaccounted for almost half of world industrialoutput in 1925-29, and its imports of basicraw materials and foodstuffs in 1927-28made up almost 40 percent of the trade inthese commodities.37 The reduction of im-ports as these economies weakened exerteddownward pressure on trading partners.

3) There were interesting parallels be-tween the troubles of the domestic financialsystem and those of the international system.One of the Federal Reserve's proudestaccomplishments had been the establish-ment, during the 1920's, of an internationalgold-exchange standard. Unfortunately, likedomestic banking, the gold-exchange stan-dard had the instability of a fractional-reserve system. International reservesincluded not only gold but also foreign cur-rencies, notably the dollar and the pound;for countries other than the United Statesand the United Kingdom, foreign exchangewas 35 percent of total reserves.

In 1931, the expectations that the interna-tional financial system would collapse be-came self-fulfilling. A general attempt toconvert currencies into gold drove one cur-rency after another off the gold-exchangestandard. Restrictions on the movement ofcapital or gold were widely imposed. By 1932,only the United States and a small numberof other countries remained on gold.

As the fall of the gold standard parallelleddomestic bank failures, the domestic in-solvency problem had an international ana-logue as well. Largely due to fixed exchangerates, the deflation of prices was worldwide.Countries with large nominal debts, notablyagricultural exporters (the case of Canadahas been mentioned), became unable to pay.Foreign bond values in the United Stateswere extremely depressed.

37U.S. Department of Commerce (1947), pp. 29-31.



As in the domestic economy, these prob-lems disrupted the worldwide mechanism ofcredit. International capital flows were re-duced to a trickle. This represented a seriousproblem for many countries.

To summarize these observations: the factthat the Great Depression hit countries whichdid not have banking crises does not pre-clude the possibility that banking and debtproblems were important in the United States(or, for that matter, that countries with strongbanks had problems with debtor insolvency).Moreover, my analysis of the domestic finan-cial system may be able to shed light onsome of the international financial difficul-ties of the period.

VII. Conclusion

Did the financial collapse of the early1930's have real effects on the macroecon-omy, other than through monetary channels?The evidence is at least not inconsistent withthis proposition. However, a stronger reasonfor giving this view consideration is the onestated in the introduction: this theory hashope of achieving a reconciliation of theobvious suboptimality of this period with thepostulate of reasonably rational, market-con-strained agents. The solution to this paradoxlies in recognizing that economic institutions,rather than being a "veil," can affect costs oftransactions and thus market opportunitiesand allocations. Institutions which evolve andperform well in normal times may becomecounterproductive during periods when ex-ogenous shocks or policy mistakes drive theeconomy off course. The malfunctioning offinancial institutions during the early 1930'sexemplifies this point.

REFERENCES

Abel, Andrew and Mishkin, Frederic, "An In-tegrated View of Tests of Rationality,Market Efficiency, and the Short-RunNeutrality of Monetary Policy," WorkingPaper 726, National Bureau of EconomicResearch, 1981.

Barro, Robert, "Unanticipated Money, Out-put, and the Price Level in the UnitedStates," Journal of Political Economy,

August 1978, 86, 549-80.Behrens, Carl, Commercial Bank Activities in

Urban Mortgage Financing, New York:National Bureau of Economic Research,1952.

Bernanke, Ben, "Bankruptcy, Liquidity, andRecession," American Economic ReviewProceedings, May 1981, 71, 155-59.

Brown, E. Carey, "Fiscal Policy in the Thir-ties: A Reappraisal," American EconomicReview, December 1956, 46, 857-79.

Brunner, Karl, The Great Depression Revisited,Boston: Martinus Nijhoff, 1981.

Cagan, Philip, Determinants and Effects ofChanges in the Stock of Money, 1875-1960,New York: National Bureau of EconomicResearch, 1965.

Chandler, Lester, America's Greatest Depres-sion, New York: Harper & Row, 1970.

, American Monetary Policy, 1928-1941, New York: Harper & Row, 1971.

Clark, Evans, The Internal Debts of the UnitedStates, New York: Macmillan Co., 1933.

Cone, Kenneth, "Regulation of DepositoryFinancial Intermediaries," unpublisheddoctoral dissertation, Stanford University,1982.

Diamond, Douglas and Dybvig, Philip, "BankRuns, Deposit Insurance, and Liquidity,"mimeo., University of Chicago, 1981.

Evans, Paul, "An Econometric Analysis of theCauses of the Great Depression in theU.S.," mimeo., Stanford University, 1981.

Fabricant, Solomon, Profits, Losses, and Busi-ness Assets, 1929-1934, Bulletin 55, Na-tional Bureau of Economic Research, 1935.

Fama, Eugene, "Banking in the Theory ofFinance," Journal of Monetary Economics,January 1980, 6, 39-57.

Fisher, Irving, "The Debt-Deflation Theory ofGreat Depressions," Econometrica, Oc-tober 1933, 7, 337-57.

Flood, Robert and Garber, Peter, "A SystematicBanking Collapse in a Perfect ForesightWorld," Working Paper No. 691, NationalBureau of Economic Research, 1981.

Frederiksen, D. M., "Two Financial RoadsLeading Out of Depression," HarvardBusiness Review, October 1931, 10,131-48.

Friedman, Benjamin, "Debt and Economic Ac-tivity in the United States," Working Paper



No. 704, National Bureau of EconomicResearch, 1981.

Friedman, Milton, and Schwartz, Anna J., AMonetary History of the United States,1867-1960, Princeton: Princeton Univer-sity Press, 1963.

Goldsmith, Raymond, Financial Institutions inthe American Economy Since 1900, Prince-ton: Princeton University Press, 1958.

Gurley, John G. and Shaw, E. S., "FinancialAspects of Economic Development,"American Economic Review, September1955, 45, 515-38.

Hart, A. G., Debts and Recovery, 1929-1937,New York: Twentieth Century Fund, 1938.

Jaffee, Dwight, and Russell, Thomas, "Imper-fect Information and Credit Rationing,"Quarterly Journal of Economics, November1976, 90, 651-66.

Kennedy, Susan E., The Banking Crisis of 1933,Lexington: University Press of Kentucky,1973.

Kimmel, Lewis H., The Availability of BankCredit, 1933-1938, New York: NationalIndustrial Conference Board, 1939.

Kindleberger, Charles P., Manias, Panics, andCrashes, New York: Basic Books, 1978.

Klebaner, Benjamin, Commercial Banking inthe United States: A History, Hinsdale:Dryden Press, 1974.

Lucas, Robert E., Jr., "Expectations and theNeutrality of Money," Journal of Eco-nomic Theory, April 1972, 4, 103-24.

, Studies in Business Cycle Theory,Cambridge: Massachusetts Institute ofTechnology Press, 1981.

Lutz, Friedrich, Corporate Cash Balances,1914-43, New York: National Bureau ofEconomic Research, 1945.

McKinnon, Ronald J., Money and Capital inEconomic Development, Washington: TheBrookings Institution, 1973.

Merwin, Charles L., Financing Small Corpora-tions, New York: National Bureau of Eco-nomic Research, 1942.

Minsky, Hyman P., "A Theory of SystematicFragility," in E. I. Altman and A. W.Sametz, eds., Financial Crises, New York:Wiley-Interscience, 1977.

Mishkin, Frederic, "The Household BalanceSheet and the Great Depression," Journal

of Economic History, December 1978, 38,918-37.

., "Does Anticipated Money Matter?An Econometric Investigation," Journal ofPolitical Economy, February 1982, 90,22-51.

Nugent, Rolf, Consumer Credit and EconomicStability, New York: Russell Sage Foun-dation, 1939.

O'Hara, Maureen and Easley, David, "The PostalSavings System in Depression," Journal ofEconomic History, September 1979, 39,741-53.

Persons, Charles E., "Credit Expansion, 1920to 1929 and Its Lessons," Quarterly Jour-nal of Economics, November 1930, 45,94-130.

Safarian, A. E., The Canadian Economy in theGreat Depression, Toronto: University ofToronto Press, 1959.

Sargent, Thomas J., "A Classical Macroecono-metric Model for the United States," Jour-nal of Political Economy, April 1976, 84,207-38.

Stiglitz, Joseph E. and Weiss, Andrew, "CreditRationing in Markets with Imperfect In-formation," American Economic Review,June 1981, 71, 393-410.

Stoddard, W. L., "Small Business WantsCapital," Harvard Business Review, Spring1940, 18, 265-74.

Temin, Peter, Did Monetary Forces Cause theGreat Depression!, New York: W. W. Nor-ton, 1976.

Upham, Cyril B. and Lamke, Edwin, Closed andDistressed Banks: A Study in Public Ad-ministration, Washington: The BrookingsInstitution, 1934.

Board of Governors of the Federal Reserve Sys-tem, Banking and Monetary Statistics, 1943.

Federal Reserve Bulletin, various issues.National Industrial Conference Board, The Avail-

ability of Bank Credit, New York, 1932.Survey of Current Business, various issues.U.S. Department of Commerce, Historical Sta-

tistics of the United States, Washington:USGPO, 1975.

, National Income, Washington:USGPO, 1954.

_, The United States in the World Econ-omy, Washington: USGPO, 1947.


Nonmonetary Effects of the Financial Crisis in the …...Nonmonetary Effects of the Financial Crisis...

Documents

Transcript of Nonmonetary Effects of the Financial Crisis in the …...Nonmonetary Effects of the Financial Crisis...