Cash Flow Risk, Discounting Risk, and the Equity Premium Puzzle

This article investigates the impact of cash flow risk and discounting risk on the aggregate equity premium. Our approach is based on the idea that consumption is hard to measure empirically, so if we substitute out an empirically difficult-to-estimate marginal utility by a pricing kernel of observables, we can evaluate the empirical performance of an equilibrium asset pricing model in a different way. Once the pricing-kernel process is specified, we can endogenously solve for the equity premium, the price of the market portfolio, and the term structure of interest rates within the same underlying equilibrium. Embedded in the closed-form solution are compensations for cash flow risk and discounting risk. With the solution for the risk premium explicitly given, we then calibrate the model to evaluate its empirical performance. This approach allows us to avoid the impact of the unobservable consumption or market portfolio on inferences regarding the model's performance. Our illustrative model is based on the assumption that aggregate dividend equals a fixed fraction of aggregate earnings plus noise, and the expected aggregate earnings growth follows a mean-reverting stochastic process. Moreover, the economy-wide pricing kernel is chosen to be consistent with (1) a constant market price of aggregate risk and (2) a mean-reverting interest rate process with constant volatility. Estimation results show that the framework can mimic the observed market equity premium