Further evidence on the asymmetric behavior of unemployment rates over the business cycle

Abstract

Sichel recently reported that Neftci's claim to have found evidence in favor of asymmetry in the quarterly aggregate unemployment rate was based on an error. Using a modified version of Neftci's test, this paper is the first to report evidence of such asymmetry using a Markov procedure. The test is then applied to a group of quarterly sectoral unemployment rates. The results suggest that the primary source of asymmetry in the aggregate unemployment rate is the cyclical behavior of the unemployment rate in the manufacturing sector.