Manufacturing Earnings and Cycles: New Evidence

In the time domain, the observed cyclical behavior of the real wage hides a range of economic in uences that give rise to cycles of di ering lengths and strengths. This may serve to produce a distorted picture of wage cyclicality. Here, we employ and develop frequency domain methods that allow us to assess the relative contribution of cyclical frequency bands on real wage earnings. Earnings are decomposed into standard and overtime components. We also distinguish between consumption and production wages. Frequency domain analysis is carried out in relation to wages alone (the univariate case) and to wages in relation to a selected range of cyclical economic indicators (multivariate). We establish that all key components of real wages are strongly pro-cyclical but display signi cant co-variations with more than one frequency band. Moreover, components are by no means uniformly associated with each of the chosen proxies for the cycle. JEL Classi cation: E32, J31 We are grateful to Michael Burda, Christian Dustmann, Campbell Leith, Jacques Melitz, Albrecht Ritschl, Harald Uhlig, and seminar participants at Humboldt University and the Glasgow-Strathclyde Joint Seminar in Macroeconomics for helpful comments and suggestions on this and earlier versions of the paper. We would also like to thank John Stinson and Sharon Gibson of the Bureau of Labor Statistics, U.S. Department of Labor for providing unpublished data relating to hours of work and consumer prices respectively. Corresponding Author: Robert Hart, Department of Economics, University of Stirling, Stirling FK9 4LA, UK, email: [email protected].