Aggregate Supply and Demand Shocks and Asset Prices

Abstract

We explore one potential reason for the insignificant observed premium on CAPM risk: the observed covariance matrix is not drawn from the return space the model addresses. The CAPM is a pure exchange model that asks how investors on the demand side of the asset market would price an exogenously-determined return space. Returns in the model are variable because asset supply produces uncertain outcomes, but the returns researchers observe reflect variability in both supply outcomes and investor demand. We use separate instruments for aggregate demand and supply shocks, and obtain the following results. First, aggregate supply shocks and aggregate demand shocks both carry an economically and statistically significant positive risk premium. Second, assets' sensitivities to these aggregate shocks are correlated with commonly studied firm characteristics. High book-to-market firms are more sensitive to aggregate supply shocks, while return-momentum and high earnings-growth firms are more sensitive to aggregate demand shocks. This provides a clue as to why those factors are priced. Third, firms' sensitivities to supply and demand shocks are negatively correlated, with partially offsetting effects on single-beta risk estimates, thereby providing one potential explanation of the small risk premium reported in the literature. These results suggest that the supply-side risk addressed by the CAPM is priced, but single-beta estimation suffers from an identification problem.