This paper uses band spectrum regression to isolate the cyclical relationship between real wages and employment in U.S. postwar monthly data. The author finds a small, but statistically significant, countercyclical pattern to real wages when real wages are measured relative to producer prices, but no cyclical pattern when real wages are measured relative to consumer prices. The spectral decomposition shows that the "cyclical" relationships reported in many time domain studies are attributable to correlations at frequencies either too high (periods less than two years) or too low (periods greater than ten years) to be associated
Your browser does not appear to support JavaScript, or you have turned JavaScript off. You may use unl.edu without enabling JavaScript, but certain functions may not be available.