A1269
Title: Robust Cauchy-based methods for predictive regressions
Authors: Jihyun Kim - Sung Kyun Kwan University (Korea, South)
Rustam Ibragimov - Imperial College Business School and New Economic School (United Kingdom) [presenting]
Anton Skrobotov - Russian Presidential Academy of National Economy and Public Administration and SPBU (Russia)
Abstract: The purpose is to develop robust inference methods for predictive regressions, addressing challenges posed by endogenously persistent or heavy-tailed regressors, as well as persistent volatility in the errors. Building upon the Cauchy estimation framework, two novel tests are proposed: One that relies on t-statistic-based group inference and another that employs a hybrid approach combining Cauchy and OLS estimation. These methods effectively tackle key issues in standard inference procedures, including size distortions arising from endogenously persistent or heavy-tailed regressors and persistent volatility dynamics. The proposed methods are straightforward to implement and broadly applicable to both continuous and discrete time models. Extensive simulation studies highlight the finite-sample advantages of the proposed methods under realistic settings. An empirical application is provided to test the predictability of excess returns for two major stock indices using the dividend-price and earnings-price ratios as predictors. The results indicate that the dividend-price ratio has predictive power, while the earnings-price ratio does not significantly predict returns.
