Demographics and the Term Structure of Stock Market Risk

This paper examines the consequences for the term structure of stock market risk of the significance of demographics in capturing the time varying mean of the dividend-price ratio and in predicting stock market returns. A potential role for demographic variables has never been considered in the ongoing debate on the slope of the term structure of stock market risk. Intuitive reasoning, formal modeling and empirical evidence show that demographic trends are a slow-moving information variable, that determines the slow moving mean of the dividend-price ratio and has a forecasting power for stock market returns that icreases with the horizon. We show that the forward solution of the dynamic dividend growth model augmented with demographics delivers a negtive sloping term structure of stock market risk.Direct regressions of returns at different horizon on the relevant predictors are much better suited to capture this feature of the model than VAR based multi-period iterated forecasts. These results are very little affected by parameters’ uncertainty, as a parsimoniuos parameterization is very precisely estimated in the relevant empirical model. Moreover, they are robust to the existence of "imperfect predictors" as forecast of returns in the direct regression approach involve only currently observable variables and no projections of future variables.