Consumption Persistence and the Equity Premium Puzzle: A Resolution or Not? (Job Market Paper)

As first pointed out by Mehra and Prescott (1985), the excess return of equities over the risk-free rate, roughly 6%, is too high to be readily reconciled with a standard intertemporal model. Recently, Bansal and Yaron (2000, 2004) have demonstrated a resolution of the equity premium puzzle when high persistence in the consumption growth process is combined with the Generalized Expected Utility (GEU) specification of Epstein and Zin (1989, 1991). However, Nelson and Startz (2006) and Ma, Nelson, and Startz (2006) have shown that standard estimates of persistence are generally spurious in time series models that are weakly identified. This motivates the re-examination of the evidence for resolution of the equity premium puzzle in this paper. Using the valid Anderson-Rubin type test proposed by Ma and Nelson (2006) I show that weak identification may account for the apparent resolution and valid confidence regions and tests reject high persistence in consumption growth. Also, the possibility of integrated expectations is examined using the Median Unbiased Estimator of Stock and Watson (1998) and little supporting evidence is found. Evidently, the equity premium puzzle remains just that.