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

Bernanke accepts that “money was an important factor in 1930-33” but argues the monetary account is incomplete on two counts: “there is no theory of monetary effects on the real economy that can explain protracted non-neutrality,” and “the reductions of the money supply in this period seem[] quantitatively insufficient to explain the subsequent falls in output” (Introduction, pp. 2-3, 25). He proposes a third channel, additional to the wealth and money-supply effects Friedman and Schwartz already identified: financial disruption reduced “the effectiveness of the financial sector as a whole” in performing credit-allocation services, raising the “real costs of intermediation” facing “especially…households, farmers, and small firms” (p. 3). He is explicit that the theory does not fully explain the Depression (it says nothing about the initial 1929-30 downturn) but claims two virtues: it can help explain the unusual length and depth of the Depression in a way existing theories could not, and it does so “without assuming markedly irrational behavior by private economic agents” (pp. 3-4) – addressing what he calls “a leading unsolved puzzle of macroeconomics,” reconciling the Depression’s apparent inefficiency with rational private behavior.

The two components are “the loss in confidence in financial institutions, primarily commercial banks, and the widespread insolvency of debtors” (Section II, p. 5). On banking, he documents that 5.6%, 10.5%, 7.8%, and 12.9% of operating banks failed in 1930-33 respectively, roughly halving the number of banks by 1933, and traces the danger of self-confirming bank runs to banks’ reliance on fixed-price, callable demand deposits against illiquid assets. On debt, he cites a rise in the debt-service-to-national-income ratio from 9% in 1929 to 19.8% in 1932-33, with default rates as high as 50-62% of mortgaged owner-occupied housing in some surveyed cities and delinquency on mortgage debt covering 45% of U.S. farms by early 1933 (Section II.2, pp. 8-9).

Departing from Fama’s (1980) complete-markets world “in which banks and other intermediaries are merely passive holders of portfolios,” Bernanke posits a stylized economy with “good” borrowers (whose projects have positive expected social return) and “bad” borrowers (who squander loans and default), where the bank’s real service is differentiating between the two (Section III, pp. 12-13). The CCI is defined as the cost of channeling funds from savers to good borrowers, “includ[ing] screening, monitoring, and accounting costs, as well as the expected losses inflicted by bad borrowers”; competitive banks minimize this cost by building borrower-evaluation expertise, long-term customer relationships, and self-selecting loan terms (p. 13). This is the theoretical scaffolding that lets Bernanke argue institutions “rather than being a ‘veil,’ can affect costs of transactions and thus market opportunities and allocations” (Conclusion, p. 38).

“Fear of runs led to large withdrawals of deposits, precautionary increases in reserve-deposit ratios, and an increased desire by banks for very liquid or re-discountable assets,” forcing “a contraction of the banking system’s role in the intermediation of credit,” and “the rapid switch away from the banks (given the banks’ accumulated expertise, information, and customer relationship[s]) no doubt impaired financial efficiency and raised the CCI” (Section III.1, pp. 13-14). Because a direct measure of the CCI is unavailable, Bernanke uses the flow of bank credit (monthly change in outstanding loans, normalized by personal income) as an indicator: credit outstanding “declined very little before October 1930” despite a 25% fall in industrial production, but the first banking crisis of November 1930 “initiated” a long credit contraction whose rhythm tracked the banking crises, reaching a “record 31% of personal income” net reduction in October 1931 (p. 14). A 1932 National Industrial Conference Board survey corroborated that by 1931 the credit shrinkage “unquestionably represented pressure by banks on customers for repayment of loans and refusal by banks to grant new loans,” not merely reduced loan demand (p. 15).

Bernanke frames the value of collateral and simple, non-contingent loan contracts as a device that “helps to create a low effective CCI” by giving borrowers the right incentive and limiting the bank’s risk; “a useful way to think of the 1930-33 debt crisis is as the progressive erosion of borrowers’ collateral relative to debt burdens,” which forced banks into “a dilemma” between increasingly risky simple loans and costlier contingent contracts – “either way, debtor insolvency necessarily raised the CCI for banks” (Section III.2, pp. 17-18). Rather than raising interest rates (which could increase default risk), “the more usual response is for banks just not to make loans to some people that they might have lent to in better times,” a pattern Bernanke documents with contemporary reports of lenders “practically stop[ping] making mortgage loans, except for renewals” (p. 18) and with the Baa-Treasury yield spread rising from 2.5% in 1929-30 to nearly 8% by mid-1932 – a shift Bernanke notes “never exceeded 3.5%” even in the comparably sharp 1920-22 recession (Section II.3, p. 19).

He considers and largely dismisses an aggregate-supply channel – most larger corporations “entered the decade with sufficient cash and liquid reserves to finance operations,” so a “temporarily underdeveloped economy” story is judged “probably no” (Section IV, p. 21). He instead builds an aggregate-demand argument: “a higher cost of credit intermediation for some borrowers…implies that, for a given safe interest rate, these borrowers must face a higher effective cost of credit,” and if that higher borrowing cost does not apply symmetrically to what those same households and firms earn on saving, “the effect of higher borrowing costs is unambiguously to reduce their demands for current-period goods and services” – a substitution effect derivable from the standard two-period savings model, which shifts the aggregate demand curve down “for a given safe rate,” implying “in any macroeconomic model one cares to use…lower output and lower safe interest rates,” exactly the pattern observed in 1930-33 (Section IV, pp. 21-22).

Following the Lucas-Barro tradition of using unanticipated money and price “shocks” (residuals from an autoregression) as the monetary explanatory variables, Bernanke first fits baseline “Lucas supply curve” regressions of industrial-production growth on these monetary/price surprises for the full interwar period (1919-1941), then tests whether adding proxies for the financial crisis – deposits of failing banks and liabilities of failing businesses – improves the fit (Section V, pp. 23-24). The baseline monetary-surprise equations are “disappointing” on their own: dynamic simulations from mid-1930 to the March 1933 bank holiday “capture no more than half of the total decline of output during the period,” which Bernanke identifies as the basis for his Introduction’s claim that money declines were “quantitatively insufficient” to explain the output collapse (p. 24).

Adding the failing-bank-deposits and failing-business-liabilities variables to the output equations, both enter “with the expected sign and, taken jointly, with a high level of statistical significance,” while “the magnitudes and significance of the coefficients of money and price shocks are not much changed” (Section V, p. 25), evidence Bernanke reads as “at least tentative confirmation that non-monetary effects of the financial crisis augmented monetary effects.” Dynamic simulations using these augmented equations “reduced the mean squared simulation error…by about fifty per cent” relative to the pure monetary/price-shock equations, and unreported equations using the Baa-Treasury yield differential “reduced the MSE of simulation from ninety to ninety-five per cent” (p. 29). Results were robust to alternative proxies (a fitted bank-credit-contraction series, quarterly dummy variables for 1931-32) and, with the noted exception of the 1933-41 recovery period (financed mainly through non-bank channels), to subsample splits.

The results require “an additional assumption, that failures of banks and commercial firms are not caused by anticipations of (future) changes in output” – otherwise, bank failures merely forecasting rather than causing output declines would produce the same correlation (Section V, p. 29). Bernanke argues this reverse-causation story is not fully convincing: business bankruptcy is “more often…forced by insolvency (a result of past business conditions)” than by forward-looking sales forecasts, banking crises had “never previous to this time been a necessary result of declines in output,” and Friedman and Schwartz and others identified specific triggering events for the 1930-33 runs – the Bank of United States failure, the Kreditanstalt collapse, Britain’s departure from gold, exposed pyramid schemes – that were “connected very indirectly (if at all) with the path of industrial production in the United States” (pp. 29-30), supporting a causal rather than purely anticipatory reading of the bank-failure evidence.

Unlike purely monetary theories, which must rely on “the slow diffusion of information or unexplained stickiness of wages and prices” to generate persistence, credit-market effects have “a plausible basis” for persistence because rebuilding disrupted lending channels and rehabilitating insolvent debtors “may be difficult and slow” (Section VI, p. 30). Bernanke documents that even after the March 1933 bank holiday and subsequent New Deal financial rehabilitation, “recovery was neither rapid nor complete”: bank deposits did not recover in volume until 1934, and survey evidence (Kimmel 1939; Hardy and Viner’s Seventh Federal Reserve District study; a U.S. Department of Commerce Small Business survey) documents continued credit rationing of small firms and mortgage borrowers as late as 1935-38, including a report that “5% of the firms” in an elite sample of otherwise highly-rated companies “reported difficulty in securing funds for working capital purposes” and 75% “could not obtain capital or long-term loan requirements through regular markets” during 1933-38 (Section VI, pp. 32-34). He estimates “the U.S. financial system operated under handicap for about five years (from the beginning of 1931 to the end of 1935),” spanning most of the gap between the 1929-30 and 1937-38 recessions (p. 34).

He offers three observations rather than a complete answer: first, “the experience of different countries and the mix of depressive forces each faced varied significantly” (e.g., Britain’s overvalued pound and early departure from gold); second, the countries that did have banking crises (the U.S., Germany, Austria, Hungary) accounted for a disproportionate share of world trade and output, so their credit disruptions could transmit internationally through reduced imports; and third, the international gold-exchange standard “had the instability of a fractional-reserve system” analogous to domestic banking, and its 1931 collapse, together with worldwide debt-deflation among countries with large nominal (often foreign-currency) debts, “disrupted the worldwide mechanism of credit” much as domestic bank and debt crises did (Section VII, pp. 35-37). He concludes that the absence of a domestic banking crisis elsewhere “does not preclude the possibility that banking and debt problems were important in the U.S.”