Introduction

The past decade has witnessed a substantial change in the field of macroeconomics. Fixed price and wage models of the simple Keynesian variety have been displaced by flexible price and wage models. The Phillips curve, which until the end of the 1960s was viewed as representing a stable trade-off between inflation and unemployment, has come to be regarded as a temporary phenomenon that lasts only as long as inflationary expectations differ from actual inflation.

In a world with a stable trade-off, the government could pick, at least within some range, any desired combination of inflation and employment. Recognition that the trade-off is temporary altered this view drastically: The consensus became that in the long run there is no policy-exploitable trade-off between inflation and unemployment. Monetary policy may temporarily boost employment, but after a while employment will return to its original level and the rate of inflation will be higher. Moreover, if policy makers try to use monetary policy to maintain employment systematically at a level higher than it would otherwise be, the Phillips trade-off will become more biased toward inflation and ultimately vanish even in the short run.

Because many modern explanations of the Phillips curve rely on the relationship between actual and expected inflation, the modeling and empirical measurement of inflationary expectation have become important tasks for macroeconomics. Initially expectations were considered adaptive: Yesterday's expectation is altered partially in response to the current forecast error.