What ended the Great Depression? It was not World War II.

• Author: Steindl, Frank G.

There are two prominent views about what ended the Great Depression. The most widely accepted one by far emphasizes U.S. entry into World War II, with its attendant government spending for armaments. (1) According to this view, the U.S. economy in 1941, though approaching its potential output, remained well below it, so the war's fiscal stimulus moved the economy completely out of the Depression. (2) That the war ended the Depression is a view held not only by economists, but also by the public in general. The second view, in contrast, prominent particularly among economists, sees monetary expansion, which began in spring 1933, as the principal reason for its end.

In this article, I offer another interpretation, arguing that neither of the conventional views is fundamentally correct. I maintain that forces inherent in the economy that promote productivity growth drove the recovery following the 1933 trough and moved the economy back to its trend. This view is not a radically new one, but it is important to recognize that productivity-raising forces promoted recovery specifically during the 1930s--a time when aggregate-demand measures are held to have been fundamental to moving the economy back to its trend.

One of the difficulties of macroeconomic interpretation, indeed of scientific inquiry in general, is that of isolating a dominant mechanism from the complex of operative forces. Among the likely forces promoting the recovery was the return to trend attributable principally to monetary changes, fiscal measures, credit-channel influences, and inherent "mean reversion" influences. Or did it result from purely random developments? The years following the 1937-38 depression facilitate a natural experiment in which some of these influences can be identified and disentangled, and therefore are of special importance in any attempt to identify what ended the Great Depression.

The Current Canonical Chronicle

The trough of the 1929-33 debacle came in early 1933. (3) The ensuing recovery ran to late spring of 1937, when an extremely sharp, year-long depression began, reducing industrial production by 33 percent. (4) Its nadir was in May 1938. The ensuing revival carried into early 1942. Figure 1 depicts the evolution of industrial production in the Great Depression decade, showing the respective 52 and 33 percent declines in the two 1930s contractions. Industrial production in September 1939 was at the same level as a decade earlier. The economy was, however, still well below its trend. Using the Cole and Ohanian (1999) framework, updated to take account of National Income and Product Accounts (NIPA) revisions of annual gross domestic product (GDP), we find that in 1939 the economy was 22 percent below its trend. In 1940, it was 18 percent below, and in 1941, 8 percent below. Only in 1942 did it reach the trend and rise 6 percent above it. (5)

For understandable reasons, economists have concentrated on the behavior of output relative to capacity. Their interest in the course of prices has been secondary to their interest in the course of output. Figure 2 shows the behavior of wholesale prices in the Depression decade (this index was the most widely used one in the 1930s). Prices fell 37 percent in the 1929-33 contraction. They subsequently rose 46 percent, until the May 1937 onset of the later real-output contraction, during which they fell 11 percent.

In contrast to the normal, expected course during a recovery, prices continued to decline for twenty-seven months after the May 1938 trough, until August 1940. This deflation occurred despite an expanding stock of money (monetary base up 53 percent, M1 up 36 percent, M2 up 25 percent) and an expanding output (index of industrial production up 58 percent). The deflation was interrupted by a price spike in September-October 1939 associated with the outbreak of war in Europe, when wholesale prices rose at an annual rate of 89 percent in September alone. Prices then resumed their decline for almost another year.

The Depression was a worldwide phenomenon, affecting all countries except those on a monetary standard other than gold (Choudri and Kochin 1980). The importance of the gold standard as fundamental to the international transmission of the Depression and of its abandonment as basic to recovery is Eichengreen's (1992, 2002) thesis: the transmission of depression reflected national monetary authorities' failure to play by the rules. According to this view, national economies with international-payments deficits contracted as their money stocks shrank, but economies with surpluses did not expand; rather, their monetary authorities sterilized the incoming gold flows. Hence, there was a net deflationary bias. Recovery came as countries abandoned the gold standard because then their monetary authorities could pursue independent expansionary policies. Long before Eichengreen's contributions, the Brookings Institution, in a comprehensive review of developments in the early stages of recovery, linked the recovery to abandonment of the gold standard. One of its conclusions was that "©ountries which remained on the gold standard show as a rule a considerably smaller degree of recovery than do those which abandoned or modified that standard" (1936, 109). This interpretation has become the theoretical context in which much research has proceeded, particularly in regard to monetary actions. These analyses generally conclude that the largely exogenous expansion of the money stock was primarily responsible for the recovery.

The earliest, most cogent, generally accepted analysis of the role of money in the recovery, Friedman and Schwartz's (1963, 493-545), appeared two decades after the economy had returned to its trend. Their preferred method of analysis was narrative, with heavy stress on timing relations based on charts. The data were generally monthly observations. The analysis carried into 1941, but the bulk of the investigation pertains to events through the onset of the late-1930s contraction.

Friedman and Schwartz advance two themes integral to the current understanding of the monetary interpretation of the recovery. First, the recovery was driven by increases in the stock of money: "the broad movements in the stock of money correspond[ed] with those in income" (1963, 497), and "the rapid rate of rise in the money stock certainly promoted and facilitated the concurrent economic expansion" (1963, 544). Second, the rapid monetary growth reflected a growing monetary base, which itself reflected increases in the stock of gold, initially because of the 1933-34 devaluation and then because of gold inflows associated with the growing uncertainties in Europe. (6) Changes in the money stock thus resulted largely from exogenous developments.

A complement to Friedman and Schwartz's narrative is Christina Romer's 1992 study, in which simulations allow her to assess the importance of monetary and fiscal actions. Her conclusion with regard to monetary policy is unambiguously strong: changes in the money stock were "crucial to the recovery" (768). The reason for the "large estimated effect of monetary developments is not hard to find": "the extraordinarily high rates of money growth in the mid- and late 1930s" (769). Further, the source of "the huge increases in the U.S. money supply during the recovery was the tremendous gold inflow that began in 1933" (773). In sum, fiscal policy "contributed almost nothing to the recovery" (767), and "monetary developments were crucial" (782).

In a comparative study involving at times twenty-six countries, Ben S. Bernanke (1995) investigated changes in the money stock as they were linked to gold. He did not study the forces of recovery throughout the entire decade, but concentrated on the period from 1929 to 1936 because the world economy spun down early in these years and then began to recover, with different countries initiating their respective returns to trend at different times. He focused on changes in the quantity of money as determinants of the decline and recovery, with an emphasis on the gold standard as the mechanism underlying the action. (7) In this view, the gold standard acted as a restraint on independent monetary measures: "In particular, the evidence that monetary shocks played a major role in the Great Contraction, and that these shocks were transmitted around the world primarily through the workings of the gold standard, is quite compelling" (2). As countries severed their ties to gold, recovery commenced: "the evidence is that countries leaving the gold standard recovered substantially more rapidly and vigorously than those who [sic] did not" (12). "[T]his divergence arose because countries leaving the gold standard had greater freedom to initiate expansionary monetary policies" (15). (8)

The most recent investigation is Allan Meltzer's (2003). He similarly finds that the recovery was driven by increases in the money stock, based on an expanding monetary base attributable principally to gold inflows. (9) "The main policy stimulus to output came from the rise in money, an unplanned and for the most part undesired consequence of the 1934 devaluation of the dollar against gold. Later in the decade, German mobilization and annexation of the Rhineland, Austria, and Czechoslovakia, the rising threat of war, and war itself supplemented the $35 gold price as a cause of the rise in gold and money" (573). The important effect of the growth of the money stock is underscored in his assessment of the revival from the 1937-38 depression: "The principal force for recovery from the 1937-38 recession came from the decline in prices that raised the real value of the money stock and, later from the rise in the nominal money stock" (574).

In sum, the widely accepted present view that gold-flow-induced growth of the money stock drove the recovery after the spring 1933 trough appears to be a robust conclusion. This view, I might note...