The 1937-1938 Recession
With the economy wobbling around along a slow recovery path, a lot of commentators (Paul Krugman here and here, John Taylor here and Christina Romer here) have made connections to 1937/38 – the so-called “recession within the depression.” Economic historians have written a fair amount about that recession for some time and I thought it would be interesting to revisit the major findings. Below is a summary of some of the arguments, along with a few comments.
(1) Contractionary fiscal policy. According to NBER, the U.S. economy fell back into recession in May 1937, where it remained until June 1938. The fiscal policy argument is a straightforward Keynesian story: The recovery that started in 1933 was rapid – so rapid, in fact, that fiscal policy turned contractionary too early. Government spending was cut (by 3.6 percent from 1936 to 1937) and taxes were increased (June 1936 Revenue Act and the start of Social Security taxation in 1937) in a premature tightening of fiscal policy. The budget went from a deficit of around 5.5 percent of GDP in 1936 to nearly balanced in 1938.
One problem with the argument (as Scott Sumner has pointed out here) is that fiscal policy was not all that contractionary in absolute terms; moreover, the deficit was partly due to a one-time World War One veterans’ bonus payments in 1936. Sumner also pointed out that the timing of the government spending cuts does not match the timing of the recession all that well: ”Government output does decline slightly in 1937, but is still far higher than 1934 and 1935…but RGDP rose more than 5% in 1937. Then in 1938 government output rises by twice as much as it fell in 1937, and RGDP plunges by almost 4%.”
Quarterly growth rates shed a bit more light on the timing issue. Government purchases (based on the decomposition of GNP growth rates by Francois Velde of the Chicago Fed) grew at annualized rates of -2.0 and -19.9 percent in 1936Q4 and 1937Q1, respectively. The recession had already begun by the time we started seeing reductions in residential and nonresidential structures and durable goods purchases (all of which fell sharply by the end of 1937)
(2) Contractionary monetary policy, specifically the increase in reserve requirements. Starting in 1935, the FOMC had become concerned with excess reserve buildup in the banking system and incipient inflation (sound familiar?), so raised required reserve ratios shaprly.
The theory that tight monetary policy created the 1937/38 downturn originated with Milton Friedman and Anna Schwarz in their Monetary History of the United States, but more recent empirical studies have generally failed to confirm it. For example, a recent paper by Calomiris, Mason and Wheelock suggests that the reserve requirements were not binding on Fed member banks, so even though they were doubled they could not have produced significant money multiplier reductions. Cargill and Mayer (2005) found that while banks that were members of the Federal Reserve system (and thus subject to the changes in reserve requirements) increased their cash reserve ratios relative to non-member banks, the magnitude of the adjustment was not large enough to explain the recession. What this empirical evidence may be missing is the increased concern by banks about what this signaled about future monetary policy; nonetheless, the direct impact of the reserve requirement change does not seem to explain the downturn.
(3) Unionization and Increased Labor Costs. Wage rate rose sharply in 1937 partly because of the Wagner Act of that year. Labor union membership rose from 3.8 million in 1935 to at least 8.7 and as much as 10.2 million in 1941 (up to 28 percent of the nonagricultural workforce). Nominal wages rose by about 14.5 percent in 1937 (but, they rose by an even higher percentage in 1934 without disrupting the strong recovery going on in that year – these sorts of changes may well have been important, but not in isolation, as Sumner pointed out).
While labor costs were rising, it is not clear that those higher costs alone can account for much of the downturn. See Francois Velde’s paper here, for one argument.
(4) Robert Higgs focused on the possibility of regime uncertainty related to the anti-business policies of FDR. Higgs argued that a number of legislative measures of starting in 1933 attenuated private property rights and reduced the certainty required to make long-term business investments. FDR may have become more vocal in his criticisms and anti-business policies as the recovery progressed and as he became more confident in its ability to self-sustain. Higgs makes, I think, some convincing arguments; FDR’s challenges from his political left probably did push him further left and give louder voice to his anti-business rhetoric. His efforts to pack the court, among other things, must have increased uncertainty. But it is not clear why Higgs so quickly dismissed contractionary policy.
If regime uncertainty were the only factor, it is not clear why long-term investment collapsed in 1937/38 and then rapidly rebounded in 1939. In fact, as Higgs himself noted, Congress actually reversed some of FDR’s policies in 1938; the perceived threat from FDR’s second term were undoubtedly still present, but how significant were the actual measures imposed? Some of the specific policies cited by Higgs don’t always work out in terms of the timing; for example, the Temporary National Economic Committee started Dec 31, 1938 and may not have been a significant factor in the decline in private investment that had begun earlier (Miron and Romer’s monthly industrial production index shows sharp declines starting in January 1938).
It could be argued that the policy effects were cumulative and the various bits of legislation reinforced each othe (this was Schumpeter’s view). Private investors certainly had plenty of sources of uncertainty, as Higgs ably demonstrated; however, the data supporting the causal role they played has never been that strong.