Hedge funds are known to produce significantly positive OLS-regression alphas. I test whether the positive OLS alpha could be consistent with a rational outcome from a peso problem. To test this, I construct a two-state regime-switching model which can capture a potential future crash state with a small probability. By applying the time-series version of testing a peso problem, I provide evidence that an ex-ante rational expectation of zero alpha can frequently produce an expost positive OLS alpha. My results suggest that the positive OLS alpha in hedge funds can be a rational outcome.
목차
Abstract 1. Introduction 2. Data and Summary Statistics 3. OLS Regressions 4. Regime-Switching Model and Test of a Peso Problem 4.1. Bayesian Estimation using Gibbs Sampling and Test of Zero Alpha in the CAPM 4.2. Regime-Switching Model 4.3. Regime-Switching Model and Test of a Peso Problem 4.4. Results 4.5. When Both Alphas are Zero across Two States 5. Conclusion Appendix References