Protracted Depression in the Mature American Economy
THE PRESENT and the subsequent chapter examine, in the light of the available American materials, certain hypotheses concerning the causes of "stagnant trends" and of "protracted depressions" in mature economies. It is not essential here to discuss in detail the problem of the limitations of time-series analysis. Economic time series do not relate to homogeneous universes, and, consequently, statistical generalizations derived from them are, strictly speaking, always "invalid." Nevertheless, the study of factual material is helpful in the same sense as that in which experience is helpful in everyday life. What the "correct" conclusions are that should be read from given factual material frequently is as controversial a question as is that concerning the "correct" behavior on the basis of a given experience in everyday life. Yet experience in both realms is of value.
The problem here is not that of finding the determinants of some observable economic equilibrium, but rather that of forming an opinion on the likelihood of certain influences raising or lowering the range of fluctuations, and of certain processes changing the duration of periods of low activity in relation to periods of high activity.
Considering the significant trend toward rising output per manhour, the achievement of a high level of employment in the long run requires a substantially rising trend in physical output. The avoidance of severe cyclical unemployment requires that the range within which output fluctuates should be high, or, if this range reaches down to low levels, that the periods during which output remains low should be brief. It is a fact that in certain stages of