Title: Point-optimal sign-based tests for stock return predictability
Authors: Abderrahim Taamouti - Durham University Business School (United Kingdom)
Kaveh Salehzadeh Nobari - Durham University (United Kingdom) [presenting]
Jean-Marie Dufour - McGill University (Canada)
Abstract: Simple point-optimal sign-based tests are proposed for inference on linear and nonlinear regression models in the presence of stochastic regressors. The motivation is to build sign-based tests for linear and nonlinear predictability of asset returns. The most popular predictors of stock returns (e.g. dividend-price ratio, earning-price ratio, etc.) are known to be stochastic. The proposed sign-based tests are exact, distribution-free, and robust to heteroskedasticity of unknown form. They may be inverted to build confidence regions for the parameters of the regression function. Since the point-optimal sign tests depend on the alternative hypothesis, an adaptive approach based on split-sample techniques is suggested in order to choose the appropriate alternative. We present a Monte Carlo study to assess the performance of the proposed quasi-point-optimal sign test by comparing its size and power to those of some common tests which are supposed to be robust against heteroskedasticity. The results show that our procedures are superior. Finally, an empirical application using real data is considered to illustrates the proposed quasi-point-optimal sign tests.