New testing approaches for mean-variance predictability

We propose parametric tests for serial correlation in levels and squares that exploit the non-normality of financial returns. Our tests are robust to distributional misspecification. Furthermore, our mean predictability tests can be robustified against time-varying volatility. Local power analyses confirm their gains over existing methods, while Monte Carlo exercises assess their finite sample reliability. We apply our tests to quarterly returns on the five Fama-French factors for international stocks, whose distributions are mostly symmetric but fat-tailed. Our results highlight noticeable differences across regions and factors and confirm the numerical sensitivity of the usual tests to influential observations. (c) 2020 Elsevier B.V. All rights reserved.

Automated summary

This paper proposes parametric tests for serial correlation in financial returns that are robust to non-normality and distributional misspecification. Empirical results show that the proposed methods have better local power and sample reliability compared to existing methods.