Comments and Discussion.

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.