Is the persistence of shocks to output asymmetric?

This paper provides evidence that the persistence of positive shocks to output is not significantly different from the persistence of negative shocks. This contrasts the results of Beaudry and Koop (1993) Journal of Monetary Economics 31, 149-163 showing asymmetry in the effects of innovations to GNP. An unobserved components model modified to incorporate a threshold process is used to examine GNP as well as industrial production data. Its ability to discriminate between positive and negative output shocks and separately estimate their enduring effects provides for a more cogent test of symmetry in the persistence of shocks. (C) 1998 Elsevier Science B.V. All rights reserved.