THREE ESSAYS ON THE EMPIRICAL ESTIMATION OF WAGE-LED AND PROFIT-LED DEMAND REGIMES

This dissertation explores three questions related to empirical estimation of the relationship between aggregate demand and the functional distribution of income—i.e. the share of income going to labor (the wage share) vs. the share going to capital (the profit share). Previous studies exploring this relationship tend to find different results depending on the methodological approach that they follow. Aggregative studies, which estimate a bi-directional system of the wage share and demand, tend to find evidence of profit-led demand and a profit squeeze. Structural studies, which separately estimate the effects of the wage share on the components of aggregate demand while treating distribution as exogenous, more often find that demand is wage-led. Chapter 1 tests whether aggregative estimates are biased if they omit key variables or fail to account for the cyclical effects of demand on productivity—one of the two main components of the wage share. It finds no evidence of omitted variable bias. However, when the cyclical effects of demand on productivity are accounted for, the short-run relationship between the wage share and demand is found to be characterized by wage-led demand and cyclical productivity effects, rather than profit-led demand and a profit squeeze. Chapter 2 testswhether structural studies are biased if they do not account for endogenous effects of demand on the wage share or the systemic relationships between the different components of aggregate demand. Overall, no evidence of such bias is found. In fact, estimates of systems with endogenous distribution for the two models for which valid instruments could be found indicate that demand is wage-led. Furthermore, the results indicate that demand becomes more wage-led when the models are estimated in this way. Chapter 3 set out to test whether the effects of the wage share on demand differ in the short and long run. However, its findings suggest that the wage share is not a strong predictor of long-run output growth. It argues that more attention should be paid to the main components of the wage share—labor productivity and the real wage—which appear to be stronger determinants of growth.