The Neo-Goodwinian model, reconsidered

This paper estimates the relationship between aggregate demand and the functional distribution of income in the U.S. economy using a series of aggregative VAR models. Like most previous aggregative studies, it finds evidence of Goodwin cycle effects - i.e. profit-led demand and a profit-squeeze effect - for the U.S. economy in baseline estimates using assumptions traditionally used in the aggregative literature. However, the results of other specifications suggest that these observed Goodwin cycle effects likely reflect a misinterpretation of procyclical variation in labor productivity - one of the main components of the wage share. When correcting for the cyclical effects of demand on productivity, the results differ dramatically; estimates are indicative of wage-led demand, and the effects of demand on distribution are mixed or insignificant. These findings suggest that evidence of Goodwin cycle effects is likely the result of biased estimates. Instead, it appears that the short-run relationship between the wage share and demand should be viewed as a combination of wage-led demand and procyclical productivity effects.