A reexamination of the equity-premium puzzle: A robust non-parametric approach

Recent tests of stochastic dominance of several orders, proposed by Linton, Maasoumi and Whang [Linton, O., Maasoumi, E., & Whang, Y. (2005). Consistent testing for stochastic dominance under general sampling schemes. Review of Economic Studies, 72(3), 735–765], are applied to reexamine the equity-premium puzzle. An advantage of this non-parametric approach is that it provides a framework to assess whether the existence of a premium is due to particular cardinal choices of either the utility function or the underlying returns distribution, or both. The approach is applied to the original Mehra–Prescott data and more recent data that include daily yields on Treasury bonds and daily returns on the S&P500 and the NASDAQ indexes. The empirical results show little evidence of stochastic dominance among the assets investigated. This suggests that the observed equity premium represents compensation for bearing higher risk, taking into account higher-order moments such as skewness and kurtosis. There is some evidence of a reverse puzzle, whereby Treasury bonds stochastically dominate equities at the third order, a result which potentially reflects insufficient compensation to investors for bearing the negative skewness associated with the S&P500 index.