British economist A.W. Phillips found that, over a span of 96 years, periods of unemployment below the natural rate tended to coincide with periods of rapidly rising wages, hence prices. Although the U.S. experience of the 1960s fitted the Phillips hypothesis fairly well, by the 1970s unemployment and inflation were both rising. Has the structure of the U.S. economy changed in such a way that prices and wages now respond more sluggishly to unemployment?

In fact, the traditional Phillips hypothesis fails to explain our recent experience because it assumes that expectations about future prices do not change. At the natural rate of unemployment, and in the absence of inflation, competition in the market will tend to raise money wages in step with the increase in labor productivity — about two per cent a year in the U.S. But suppose the unemployment rate falls, creating a tight labor market that forces firms to bid up anticipated real wages faster than the two per cent increase in labor productivity.

If anticipated real wages are bid up seven per cent, money wages will also rise by seven per cent, in the absence of prior inflation, leading to a five per cent increase in unit labor costs, hence a five per cent increase in prices. Market participants will then adjust their price expectations to a five per cent price inflation. Even if unemployment subsequently returns to the natural rate, money wages will still rise five per cent (in anticipation of continued five per cent inflation) on top of the two per cent increase in real wages corresponding to the increase in labor productivity.

What remains unchanged is the tradeoff between anticipated changes in real wages and unemployment, rather than between changes in money wages and unemployment, as the traditional Phillips hypothesis supposed. In the context of this revised tradeoff, demand factors explain wage and price inflation as well in the 1970s as they did in previous decades. The recessions of 1970 and 1974 did not follow the pattern of earlier recessions in reducing inflation to acceptable levels merely because the rapidly accelerating inflation of the late 1960s and early 1970s had created an acute inflationary psychology.