A study of the changing variability of economic activity in the United States

Apparent improvement in the stability of real aggregate economic activity in the United States over the past century has attracted the recent interest of macroeconomists. Attention has focused on the cyclical properties of two episodes of United States history: the pre- and post-war periods. While there is at least tentative agreement that output has been more stable in the post-war period, a debate has emerged over the role played by declining wage and price flexibility in producing the increased stability. In this study, I attempt to resolve this critical issue by exploring theoretically and empirically a class of models that can explain both the post-war decline in wage and price flexibility and the increased stability of the post-war economy.

The theoretical portion of the study develops an overlapping contracts model in which the degree of wage and price rigidity, and therefore the duration of business cycles, is endogenously determined. I demonstrate that this model is consistent with the main stylized facts arising in many pre-war/post-war comparisons of macroeconomic stability: (i) output is more stable in the post-war period, (ii) wages and prices are less flexible in the post-war period, (iii) output movements are more persistent in the post-war period, and (iv) the positive correlation between price and output innovations is lower in the post-war period than in the pre-war period.

The empirical portion of the study formulates and tests empirical models of real activity that are consistent with the theoretical model described above. The empirical models assign a critical role to errors in predicting the underlying determinants of aggregate demand and aggregate supply. The results suggest that the data are consistent with the models. This is true for both the post-war period and the pre-war period. Furthermore, my proxies for aggregate demand disturbances show a dramatically lower variance in the post-war period. The theoretical model predicts that this reduction should be accompanied by reduced flexibility in wages and prices and a greater response of real activity to aggregate demand shocks. Indeed, comparisons of parameter estimates obtained for the two periods confirm the fact that prices respond less and output more to aggregate demand disturbances in the post-war period than in the pre-war period. Overall, of course, output has become more stable in the post-war period since the reduction in the variability of aggregate demand shocks has dominated the increase in the size of response to a particular shock, as predicted by the theory.