Comment on "The Great Depression and the Friedman-Schwartz Hypothesis"

1. Overview

The Great Depression has been the litmus test of macroeconomic theories. Several generations of models have tried to explain the downturn, prolonged slump, and slow recovery. Each explanation was then rejected (all or in part) by the next generation of economists for something better.

The original explanation, put forth in the 1930s was the Austrian view of Hayek and Robbins who blamed the depression on the imbalances created by a credit boom and consequent overinvestment in the 1920s. The severe downturn was viewed as a necessary unwinding of the excesses of the previous boom (Laidler, 2003, Eichengreen and Mitchener, 2003).

The next explanation was by the Keynesians with emphasis on the collapse of investment. Gordon (1952) viewed the collapse of the housing market as the trigger; Hansen (1939) attributed the long slump and slow recovery to the end of the U.S. frontier and the completion of capital extensive investment. Models by Tinbergen (1937) and Samuelson (1939 [1966]) emphasized the role of the multiplier and accelerator. The Keynesian approach did not incorporate monetary factors since Keynes had ruled out an active role for monetary policy with his liquidity trap hypothesis (Friedman 1970). Indeed, Temin (1976) dismissed the role of money in favor of an unexplained collapse of the consumption function in 1930 as the key cause of the depression. Irving Fisher's emphasis on debt deflation and financial factors was also ignored by these models.

Friedman and Schwartz (1963), in reaction to the prevailing Keynesian view, attributed the contraction to a collapse in money supply, initially triggered by tight Fed policy in 1928-29 to stem the stock market boom (and according to Hamilton (1987) to prevent gold outflows set off by tight French monetary policy). The key cause of monetary collapse in 1930-33 according to Friedman and Schwartz was a series of banking panics, evident in a decline in the deposit currency and deposit reserve ratios. They then blamed the Fed for not offsetting the banking panics by open market purchases or by using other tools of monetary policy—pursuing a lender of last resort function which was well known since the publication of Walter Bagehot's Lombard Street in 1870. Indeed, they argue that counterfactual open market purchases at key episodes of the contraction could have stopped the monetary collapse. Later econometric work by McCallum (1990) and Bordo, Choudhri, and Schwartz (1995) confirm this insight. Also in their monetarist approach, a key propagation mechanism of the depression in the U.S. was via nominal wage stickiness (as well as the effects of deflation on real interest rates according to Schwartz, 1981).

The next generation emphasized international and financial factors. Building on Fisher (1936) and Friedman and Schwartz (1963), who viewed the gold standard as the mechanism by which the depression was transmitted abroad, Eichengreen (1992) as well as Temin (1989) posited that adherence to the gold standard constrained monetary authorities across the world from adopting expansionary policy. Bernanke (1983), following Fisher (1933), emphasized the role of financial factors in propagating the contraction via the collapse of financial intermediation, the effect of collapsing asset values on firm and household net worths, and debt deflation.

The most recent approach to understanding the depression is by the application of modern dynamic general equilibrium (DGE) models. Kehoe and Prescott (2002), Cole and Ohanian (1999) stress on factors such as the role of negative productivity shocks in the downturn and adverse labor policies such as the NIRA and the Wagner Act, in preventing the recovery. Bordo, Erceg, and Evans (2000) using a DGE model with sticky money wages, demonstrate that the monetary collapse explains most of the contraction. Like Cole and Ohanian (1999), they attribute the sluggish recovery to the effects of the NIRA labor market codes on the supply of labor.

Christiano et al. (2003, this issue of JMCB) also use a DGE model but they go much farther than the earlier literature by synthesizing it with the financial frictions stressed by Bernanke. Their elaborate model includes a household sector, firm sector, financial intermediaries, and...