Local substitution and habit persistence: Matching the moments of the equity premium and the risk-free rate

This paper studies the empirical properties of introducing consumption complementarity and/or substitutability over time in a Lucas-style asset pricing model. Specifically, I investigate whether the model can replicate a selected set of observed US asset return moments over the 1890-1999 period. Firstly, I find that local substitution substantially improves the habit persistent model's ability to fit the asset return moments. Secondly, combined effects of local substitution and long-run complementarity over consumption nearly explain the equity premium and the risk-free rate means and volatilities. I conclude that both habit persistent and local substitution are required to solve the standard financial empirical puzzles. However, these results imply slightly high values of relative risk aversion in consumption and in wealth.