Cyclical Patterns in the Variance of Economic Activity

Abstract

This article models the conditional mean and variance of real gross national product (GNP) and its components using asymmetric exponential generalized autoregressive conditional hetero-scedasticity, a model previously applied only to financial variables. The results imply that the variance of real GNP is higher following negative innovations than positive innovations and that this asymmetry arises in the cyclically sensitive sectors. Further evidence links this asymmetry to the phase of the business cycle: The conditional variance appears to be largest around business-cycle troughs. In addition, shocks to the conditional variance of GNP and its components typically persist for long periods. The evidence of asymmetry in conditional variance is robust to a variety of alternative specifications.