This paper examines the relationship between aggregate demand and the wage share in the US using a vector autoregression methodology. It finds evidence of Goodwin-cycle effects - that is, profit-led demand and a profit-squeeze effect - in baseline estimates using assumptions traditionally used in the aggregative literature. However, estimates that examine the relationship between demand and the two components of the wage share (the real wage rate and labor productivity) indicate that these results are highly sensitive to ordering restrictions relating demand and labor productivity and that different types of shocks to the wage share may have differing effects on demand. The results suggest two possible interpretations of the initial Goodwin-cycle finding, depending on the assumptions used to identify the demand-productivity relationship. One suggests that the initial estimates reflect a causal relationship in which the effects are driven by a positive effect of productivity on demand and a negative effect of demand on productivity. The other suggests that the initial Goodwin-cycle finding may be spurious, as it interprets contemporaneous pro-cyclical variation in labor productivity as a profit-led demand effect, thereby obscuring an underlying wage-led relationship.
