The Structuralist Theory of Employment

The postwar era showed us that we knew much less about employment determination than we thought. In this country, I remember, economists estimated in the mid-1950's that unemployment could fall to 4.5 percent or less without bringing an inflation problem. Later, it took an unemployment rate around 5 percent to keep inflation stable, as 1964 and 1973 illustrated. By the mid-1980's, it apparently took an unemployment rate of more than 6 percent: witness the turnaround of the inflation rate early in 1987. In Western European economies with few exceptions and in Canada, the rise of unemployment has been much greater, reaching a higher level from a generally lower starting point. Among the OECD countries, the secular increase in joblessness typically exceeds the increase found in the average recession, and it is far more destructive, since its social ill-effects are somewhat cumulative. I began to try to understand this sea change in unemployment ten years ago. My first effort, on the slump in Western Europe, with Jean-Paul Fitoussi (Fitoussi and Phelps, 1988), invoked wage stickiness nominal or real, to show how the overseas shock to real interest rates that began in 1981 could drive Europe's unemployment rate above the natural rate. We did not explore how such an external shock might alter the natural unemployment rate itself. I soon sensed, though, that there was a permanent component to the rise in Europe's unemployment. If I was to rescue the concept of the natural rate from the growing discontent with it, I needed to understand how events had driven up the natural rate in Europe and elsewhere in the West. To model the natural rate I decided, naturally enough, to return to the road I started