Why Fed Policy Worked So Well

A study updates the Phillips curve

If Alan Greenspan & Co. deserve any credit for the long expansion of the 1990s, it is for not hitting the monetary brakes when unemployment sank to new lows in the second half of the decade. In so doing, the Federal Reserve ignored orthodox theory, which posited that allowing joblessness to fall below its so-called natural rate would inevitably result in accelerating inflation.

As it happens, inflation--already modest by the mid-1990s--continued to subside, even as unemployment dropped far below its then estimated 6% natural level. And unburdened by a restrictive monetary policy, the economy grew faster than anyone anticipated.

Why didn't inflation take off, as economists expected? Experts have offered many explanations, from the impact of global competition and the strong dollar to the Asian crisis and productivity gains from the new high-tech economy.

But the most intriguing may be one suggested by a recent study by George A. Akerlof of the University of California at Berkeley and William Dickens and George Perry of the Brookings Institution. In the study, the three economists find that unemployment can be reduced below its normal natural rate without sparking a rise in inflation--if the reduction occurs in a climate of already moderate inflation.

To see why, you have to first understand natural rate theory. Most economists once bought the Phillips curve notion that accepting higher inflation would allow you to achieve lower unemployment. But economists Milton Friedman and Edmund Phelps in the 1970s suggested that such gains don't last. Over the long run, they argued, employers and workers seek to maintain their real incomes by adding higher inflation to their wage bargains and prices, causing joblessness to rise again.

Thus, every economy presumably has a natural level of sustainable unemployment below which inflation tends to accelerate. What Akerlof, Dickens, and Perry find, however, is that this level declines when inflation is low and stable. At such times, companies and people tend to ignore past inflation and thus don't fully offset it in wages and prices. As a result, companies hire more people and sell more goods, which means less unemployment and more output.

As evidence, the researchers cite survey data from 1954 to 1999, which show that employees and employers are far more likely to incorporate inflationary expectations into wage and price hikes in periods of high inflation (over 4%) than in low-inflation periods (under 3%). They then use these data to estimate the trade-offs between unemployment and inflation over the postwar period.

The results suggest that the natural rate of unemployment is about 5 1/2% when core inflation is running over 6%. But when inflation is in a stable, moderate range of between 2% and 4%, unemployment can be safely kept as low as 4%. At that point, reducing inflation still further would raise unemployment, while pushing unemployment below 4% would boost inflation.

That, so it seems, is the lesson Greenspan's pragmatic Fed learned. When unemployment fell below 5 1/2% in the mid-1990s, core inflation was only 3% or so, and the Fed didn't panic. Its calm restraint allowed the New Economy to flower and millions of Americans to join the ranks of the employed.

Before it's here, it's on the Bloomberg Terminal.