Inflation and Unemployment in a Macroeconometric Model and Discussion

The main question of this conference is why there has recently been both high inflation and high unemployment in the U.S. economy. The purpose of this paper is to consider this question within the context of a macroeconometric model. Much of the literature on inflation and unemployment since Phillips wrote his classic paper [10] has centered around the question of whether the relationship between inflation and unemployment is stable over time. The fact that this relationship does not appear to be stable (i.e., appears to "shift" over time) has caused much puzzlement. From the perspective of a macroecono-metric model builder, however, this lack of stability is not necessarily surprising. Inflation and unemployment are two endogenous variables out of many in a model, and there is in general no reason to expect that the combined influences on any two endogenous variables in a model are such as to lead to a stable relationship between them. This holds true not only for the relationship between inflation and unemployment, but also for the relationship between such variables as unemployment and output ("Okun's law") and inflation and output. A model builder must approach the task of explaining inflation and unemployment with considerable caution. A major problem in this area is the difficulty of testing alternative hypotheses. It is relatively easy with aggregate time series data to fit the data well within the sample period, but a good within-sample fit is by no means a guarantee that the particular equation or model is a good representation of the actual process generating the data. It is also difficult to make comparisons of predictive accuracy across models because of differences in the number and types of variables that are taken to be exo-genous in models. These difficulties and the fact that inflation has not been particularly well explained in the past obviously (and justifiably) make people skeptical of any new attempt at an explanation. This paper is primarily a review of that part of my recent work ([2], [3], [4], [5], [6]) that relates to the inflation-unemployment question, and so the value added of this paper to someone who is already familiar with this work is small. Sections II-IV contain a review of the determination of inflation and unemployment in my theoretical and empirical macro models, and Section V contains a discussion of some of the important properties of the empirical model