美联储的退出时机应以史为鉴

The recession is now over and the US economy is recovering. To deal with the crisis, the Federal Reserve followed an unusually aggressive monetary policy expansion – cutting the Federal funds rate from 5.25 per cent in August 2007 to close to zero by the end of 2008, and introducing uNPRecedented quantitative easing with its purchases of mortgage-backed securities and Treasury securities.

The question of the moment is the Fed's exit strategy: when will it shift from its highly expansionary policy to a neutral one consistent with long-run growth and low inflation? Ben Bernanke, the Fed chairman, has recently announced steps to disengage from expansionary policy, including the termination of MBS purchases at the end of March and measures to drain reserves from the banking system.

The key question, however, is: when will the Fed start raising the Fed funds rate? The timing is fraught with peril. Tightening too soon risks pushing the economy again into recession. Waiting too long risks firing up hard-to-reverse inflationary expectations.

We have examined the record of exit strategies since the Fed was established in 1913. Our analysis, a National Bureau of Economic Research working paper, points to a pattern where monetary policy tightens close to when unemployment peaks and after inflation troughs – with some differences across cycles. We compared the timing of the turning points in the Fed's policy instruments (the nominal and real discount rate and funds rate) and nominal and real monetary aggregates (the monetary base and broad money) with the turning points of key macroeconomic variables (such as inflation and unemployment) across 14 business cycles.

From the 1920s to the 1950s, the Fed tightened when the price level began to increase and it paid less attention to the real economy. Since the mid-1960s, it has generally waited for unemployment to peak before tightening and has allowed inflation to rise. Even after inflation fell during the Great Moderation, from the mid-1980s, the timing of tightening has focused primarily on unemployment levels.

Several disasters provide cautionary tales. The Fed's decision to double the reserve requirements of banks in 1936-37 to reduce excess reserves in the banking system and prevent future inflation is a classic case of tightening too soon. The result was the serious recession of 1937-38. Banks cut lending and sold assets to restore the cushion of excess reserves.

The classic examples of tightening too late (and not enough) occurred in the 1960s and 1970s, creating the Great Inflation. In a number of cycles the Fed held back from tightening because of concern over high unemployment; and when it did tighten, it stopped too soon to stem an inflationary spiral because of political pressure both from Congress and the Johnson and Nixon administrations. In the last two recessions (1990-91 and 2000-01) the Fed delayed tightening until several quarters after the recession ended. Many believe these delays fuelled the technology boom of the 1990s and the later housing boom.

These varied patterns reflect political and economic frameworks. Before the second world war, the Fed followed gold standard orthodoxy, placing primary importance on price stability. After the Employment Act of 1945, it followed a dual mandate for price stability and high employment. Then, from the 1960s the Fed adhered to Keynesian theories and was pressured by the White House and Congress to prioritise low unemployment. Even since the Fed's emphasis on low inflation and credibility during the mid-1980s, the timing of exits has favoured a focus on unemployment.

From this historical perspective, how will the exit strategy play out now? Our evidence suggests that if we follow the patterns of postwar business cycles, and if unemployment has peaked in the fourth quarter of 2009, we may see a tightening in the first quarter of 2010 but more likely in the second quarter. However, if unemployment falls slowly – and if the last two recessions are any guide – the Fed may delay longer. This raises concerns that prolonged maintenance of low rates will fuel future inflation. Indeed, the public's rush into inflation-protected government bonds may be a harbinger of a future rise in inflation expectations.

Michael Bordo is director of the Centre for Monetary and Financial History at Rutgers University. John Landon Lane is an associate professor of economics at Rutgers