Search

The article performs a statistical appraisal of the impact of the wage-price controls in the U.S. The condition of the U.S. economy improved in almost every respect after the initiation of the wage-price control program on August 15, I971. Real gross national product grew rapidly, unemployment finally began to decline, and the rate of inflation moderated. But the coincidence of timing does not necessarily mean that controls are an essential condition for prosperity, or that the August 1971 message was the key that unlocked the floodgates behind which real aggregate demand had been restrained. The determination of the four basic macroeconomic magnitudes--nominal (current dollar) income, real output, prices, and unemployment--is usefully separated into three subproblems: (1) the determination of nominal income, (2) the division of that nominal income between real output and prices, and (3) the relationship between real output and unemployment. Before the achievement of the control program can be evaluated, a criterion for its success should be established. By my standard, controls can be judged successful if the value to society of the reduction in inflation they achieve relative to that which would have occurred without them is greater than the direct and indirect costs imposed by the control program.

This article presents comments by the author on an article titled "Understanding Inflation in the 1980s," by Robert Gordon and published in the January 1985 issue of the journal Brookings papers on Economic Activity. According to the author, the only part of the theory called into question by recent experience is markup pricing, which seems to fail in the case of imports. U.S. prices have fallen less than predicted by a simple model in which the U.S. price of imported goods is a simple markup over the foreign price restated in dollars at the prevailing exchange rate. The U.S. pricing of imports seems to be quite sensitive to conditions in U.S. markets, less sensitive to costs of production. That makes good economic sense, but it goes against the spirit of the markup hypothesis, in which costs are the predominant determinant of price. The heart of Gordon's paper is a set of regressions of the rate of price inflation on various determinants. It is a good idea to ask what, in principle, can be learned from these regressions. In general, a regression of one endogenous variable on another tells nothing about the structural relations of the two variables.