Testing for Negative Expected Market Return Premia
This paper adds to the literature on testing the hypothesis that the expected return premium on the market portfolio is always non-negative. The lower bound restriction is an important element in framing the case against a broad class of risk-based equilibrium models of market returns. Our goal is to merge powerful tests of the restriction, which include multiple information variables, with conservative bootstrapping that measures distance relative to the closest point in parameter space consistent with the null and a jack-knifing procedure that generates out-of-sample forecasts. We sharpen the inference achieved by incorporating multiple information variables into a single minimum expected value estimate, as this procedure can be appreciably more powerful than testing joint moment restrictions simultaneously as in Boudoukh, Richardson and Smith (1993). We find statistically reliable evidence against the non-negativity hypothesis for the excess return on the value weighted market index. The most negative, out-of-sample prediction was -2.21% in September 1973.
Expected market return premium, predictability, bootstrapping
SMU Cox School of Business Research Paper Series