Habit persistence: Explaining cross sectional variation in returns and time-varying expected returns

This paper finds empirical support for the habit persistence model of Camp- bell and Cochrane (1999) along both cross sectional and time-series dimensions of the US stock market. GMM estimations show that the model is able to explain a substantial part of the cross sectional variation of returns on the 25 Fama and French value and size portfolios over the period 1932-2003, although it has difficul- ties in fully explaining the value premium, and some of the implied risk free rates are strongly negative. In addition, the model accounts for time-varying expected returns on stocks. Forecasting regressions show that the estimated surplus con- sumption ratio has strong forecasting power for future real stock returns and holds additional explanatory power relative to traditional financial forecasting variables such as the dividend yield. We also document that the Campbell-Cochrane model is particularly successful up to 1991. Including data from the 1990s reduces some- what the fit of the model.