The misuse of economic history: flawed analogies with Japan's "liquidity trap" and the great depression.

• Author: Reynolds, Alan

Paul Samuelson (2005: 242) advises that, "The sage economist must muster best available knowledge about history and theory in giving plausible pragmatic advice." Economists frequently use historical anecdotes to justify theories they prefer and policies they advocate. Unfortunately, policy-motivated history often depends on "stylized" facts and hazy metaphors--such as bubbles bursting, central banks pushing on strings, and budget deficits jump-starting the economy.

By early 2009, the global recession and financial crisis were being widely compared to the Great Depression or Japan's "Lost Decade." Writing about the U.S. economy in mid-2009, Krugman (2009a) notes this is "the third time in history that a major economy has found itself in a liquidity trap." The other two traps were the United States in 1929-39 and Japan in the 1990s. In all three cases, he argues, the alleged impotence of monetary policy justifies very large debt-financed government spending plans. On the basis of theory, Krugman (1998) once proclaimed, "When the economy is in a liquidity trap, government spending should expand up to the point at which full employment is restored" (emphasis in the original).

The Federal Reserve more than doubled the monetary base in five months, between August 2008 and January 2009. In response, Krugman (2009a) wrote that, "A rising monetary base isn't inflationary when you're in a liquidity trap. America's monetary base doubled between 1929 and 1939; prices fell 19 percent. Japan's monetary base rose 85 percent between 1997 and 2003; deflation continued apace."

There is no question that the demand for base money increases substantially during any period of widespread bank runs and failures. Banks naturally want more reserves as a cushion against possible bank runs, and the public wants to keep less cash in banks and more in currency. A liquidity trap, however, suggests the demand for reserves and currency is insatiable, so central banks cannot possibly finance inflation or even resist deflation (Boianovsky 2004). As Krugman (2009b) explains, "My definition of a liquidity trap is, purely and simply, a situation in which conventional monetary policy--open-market purchases of short-term government debt--has lost effectiveness. Period. End of story." Taken literally, that definition implies the Federal Reserve could buy up outstanding Treasury bills and commercial paper without the slightest risk of inflation (Hamilton 2008, Grier 2008).

This article questions the data Krugman uses to suggest that the Federal Reserve's recent doubling of the monetary base in five months was in any sense comparable to (1) what the Fed did in 1929-39 and (2) what the Bank of Japan did during the Lost Decade. I find that monetary policy was not ineffective (for good or ill) in the United States during the 1930s, or in Japan since 1991.

Krugman uses the alleged impotence of monetary policy as an argument for aggressive use of debt-financed government purchases and transfers. But his argument is problematic. Ineffectiveness of monetary stimulus would not demonstrate the effectiveness of fiscal stimulus. Indeed, I find no evidence that traditional fiscal policy stimulated real or nominal GDP growth during the Great Depression or Japan's Lost Decade. These historical case studies are consistent with other evidence casting doubt on the empirical validity of the view that budget deficits are an effective way to accelerate growth of domestic demand.

Liquidity Trap in the Great Depression?

Keynes (1937: 207-8) described something similar to a hypothetical liquidity trap. But he said, "I know of no example of it" other than a "very abnormal" episode "in the United States at certain dates in 1932."

By writing that "America's monetary base doubled between 1929 and 1939; prices fell 19 percent," by contrast, Krugman implies that monetary policy was ineffective against an entire decade of continuous deflation. Yet two years are not an adequate time series to determine what happened between 1929 and 1939, when there were two deep recessions and two brisk recoveries.

Table 1 shows that real GDP grew, on average, by 10.9 percent a year during 1934-36 with consumer prices rising, on average, by 2.3 percent a year (and by 3.6 percent in 1937). There was another deep recession from May 1937 to June 1938, followed by another vigorous recovery.

Krugman claims the U.S. economy grew rapidly from March 1933 to May 1937 because of "New Deal policies." That suggests rapid increases in federal spending or large budget deficits. Yet the last two columns of Table 1 show federal budgets grew much more slowly in 1933-46 (1.8 percent a year) than in 1930-42 (14.7 percent a year). If big government spending was the solution, 1932 should have been a terrific year.

Contemporary estimates of budget deficits as a share of GDP, in the last column, go back to 1934. Even within a Keynesian model, those deficits were much too small to explain the strength of economic growth before and after the 1937-38 contraction.

Describing the May 1937 downturn, Krngman (2009a) says, "The Federal Reserve tightened monetary policy, while FDR tried to balance the federal budget. Sure enough, the economy slumped again, and full recovery had to wait for World War II." The Fed doubled reserve requirements from May 1, 1937, to April 15, 1938 (Orphanides 2004). The recession began in May 1937 and ended in June 1938 (two months after the Fed reversed that policy). Since there was no discernable impact on short-term interest rates, however, Krugman's liquidity trap hypothesis precludes him from blaming the ensuing recession on monetary policy. If changes in Fed policy were that powerful (and they were), monetary policy was not ineffective.

The allusion to World War II looks like a red herring designed to put a fiscal spin on a recession that began and ended with changes in monetary policy. Drafting 11 million soldiers certainly reduced civilian unemployment, but the draft is an in-kind tax rather than a "fiscal stimulus." Investment and per capita consumption fell during World War II, Higgs (1992) finds, once we account for the statistical illusions created by price controls and rationing.

Krugman contrasts the gradual doubling of the U.S. monetary base from 1929 to 1939 with falling prices (during half of those years) to suggest the Fed was trying to stop an inexorable deflation for a decade but was thwarted by a liquidity trap. Those conjectures are inconsistent with all relevant facts.

Table 2 shows four key measures of monetary policy from 1928 to 1940. The first column reflects access to the Fed's discount window, which was tightly limited after 1929. Federal Reserve credit in the second column ("Fed credit") includes discounts but adds the effect of open market operations on the monetary base. Aside from April to June of 1932, the Fed made no significant effort to buy securities to expand bank reserves (Epstein and Ferguson 1984). On the contrary, the Fed sold securities from 1929 through early 1931, and again in 1937, to sterilize gold inflows. How could the shrinkage of Federal Reserve assets from 1928 to 1931, or the unchanged level of Fed credit from 1933 to 1940, be compared to the rapid doubling of the Fed's portfolio after August 2008?

Krugman (2009c) argues "a Friedman-style focus on a broad monetary aggregate gives the false impression that Fed policy wasn't very expansionary. But it was; the problem was that since banks weren't lending out...