Will Reducing Government Spending Derail The Recovery?
Lee E. Ohanian, 07.08.09, 12:01 AM EDT Lessons from the 1937-38 Depression.
Recently Professor Alan Blinder, former vice chair of the Federal Reserve Board, and Dr. Christina Romer, chair of President Obama's Council of Economic Advisors, both strongly cautioned the Federal Reserve and Congress against reducing either monetary or fiscal stimulus until economic recovery is well underway.
Both scholars support this view with a comparison to the Great Depression, arguing that the recession of 1937 to 1938 was caused by policy miscues, which tightened monetary policy through higher bank reserve requirements, and which tightened fiscal policy by reducing government spending in an effort to achieve budget balance. Both Blinder and Romer argue that it is necessary to maintain expansionary policy, and Romer adds that more fiscal stimulus may be necessary in the future as the American Recovery and Reinvestment Act winds down in 2011.
My research suggests that Federal Reserve and government spending changes are not the major factors behind the 1937-38 recession. In terms of fiscal policy, the decline in government spending is not plausibly large enough to be the main culprit. The economy declined considerably at this time, with industrial production, which is available on a monthly basis, falling by almost 30% between mid-1937 and mid-1938, and with real gross domestic product, which is available on an annual basis, falling by about 3.5%, or by $3.2 billion between 1937 and 1938.
Now, government spending actually rose between 1936 and 1938, reflecting a $600 million decline in 1937, followed by a $1 billion increase in 1938. But even if one ignores the $1 billion increase in government spending in 1938, the 1937 $600 million decline doesn't translate into a $3.2 billion GDP decline a year later, much less a 30% drop in industrial output.
Regarding monetary policy, half of the increase in bank reserve requirements occurred more than a year before the economy began to decline. Moreover, several of the usual telltale signs of a recession that is created by tighter credit were notably absent during this decline: Commercial lending rates didn't increase, lending relative to GDP didn't fall, and interest rate spreads between risky securities and government bonds didn't increase until economic decline was well underway.
But perhaps the most important evidence against both of these factors is that the 1937-38 recession hit industry much harder than it hit agriculture. Given that agriculture was particularly sensitive to credit and demand conditions at this time, the factors cited by Blinder and Romer imply that farmers should have been hit just as hard, if not harder, than manufacturers by these changes in policy. Understanding the 1937-38 contraction requires understanding why industrial employment was so much more depressed than agriculture.
My work with the University of Pennsylvania's Harold Cole suggests a very different causal factor, which is consistent with both the timing and the size of the 1937-38 downturn, and which is consistent with the fact that industry was impacted more than agriculture. What was it? A large change in labor policies that substantially increased unionization rates and that led to a large increase in industrial wages relative to worker productivity.
During the 1930s, many policymakers, including President Roosevelt, subscribed to what was known as the "high wage doctrine," which held that large wage hikes were necessary for achieving economic recovery.
Now it is true that high wages tend to accompany economic prosperity, but many economists today interpret this correlation differently, as high worker productivity generates high wages and prosperity, not that raising wages in and of itself generates high output. And this means that raising wages above worker productivity growth will lead to job loss and lower output, not higher employment and output as hoped for by 1930s policymakers.
Policymakers were indeed able to raise real wages above productivity during the New Deal, first through the National Industrial Recovery Act of 1933, which gave industry the right to collude provided that they raised wages substantially. In fact, government approval of each industry's "code of fair competition" required explicit labor provisions, including specific wage schedules for workers in the industry. Industrial prices and wages jumped after the NIRA was passed, and Harold Cole and I have argued that these policies were important for understanding why hours worked did not recover very much during the New Deal.
While the NIRA was struck down by the Supreme Court in 1935, the high wage component of the New Deal continued with the passage of the National Labor Relations Act in 1935, which not only provided workers with the right to collectively bargain, but also placed few explicit restrictions on workers during strikes. This latter feature of the NLRA (before it was significantly amended by the Taft-Hartley Act in 1947) gave labor much more bargaining power than it has today, and the NLRA had an enormous impact on labor markets. Workers used the "sit-down strike," in which workers forcibly occupy factories and prevent production successfully against General Motors ( GMGMQ.PK - news - people ) in late 1936 and early 1937, and the threat of a sit-down strike was used to successfully organize U.S. Steel ( X - news - people ) in 1937. Unionization rates doubled in just two years, rising from about 13 per cent in 1936 to nearly 27% in 1938.
And with this increase in unionization came significantly higher wages.
All told, real average hourly earnings in manufacturing rose nearly 10% above trend between 1936 and 1938, and rose significantly around May 1937 in several industries, which interestingly enough was immediately after the Supreme Court's decision that upheld the constitutionality of the NLRA. These wage increases clearly were much higher than productivity growth, which means that industrial labor costs rose significantly.
Standard economics--the economics that tells us that demand curves slope down--implies that industrial employment and output should have fallen considerably. So what happened to the economy after these industrial wage increases? The industrial sector began to fall in June 1937, with manufacturing hours declining by about 13% between 1936 and 1938. In contrast, agricultural hours rise about 2% over the same period, and there are no increases in agricultural wages above productivity.
The manufacturing sector ultimately did recover during World War II, reflecting both large increases in productivity that significantly reduced labor costs, and the fact that the National War Labor Board significantly limited wage increases. And even though unionization rates remained high in the 1950s, the gap between manufacturing wages and productivity was much smaller than in the 1930s, which means that postwar labor and unionization policies did not impact the economy as much as in the 1930s.
Where does this leave us in terms of drawing lessons from the 1930s for our current crisis? Whether or not you support the idea that fiscal stimulus is a useful tool for addressing our current crisis, the 1937-38 contraction doesn't tell us that reducing government spending will derail economic recovery.
Lee E. Ohanian is a professor of economics and director of the Ettinger Family Program in Macroeconomic Research at UCLA. (Forbes columnist Thomas F. Cooley is away for two weeks.)
forbes.com |
