Predictability of Equity Returns over Different Time Horizons: A Nonparametric Approach

Abstract

This paper aims to test whether equity returns are predictable over various horizons. We propose a reliable and powerful nonparametric test to examine the predictability of equity returns, which can be interpreted as a signal-to-noise ratio test. Our comprehensive in-sample and out-of-sample analysis shows that the commonly used predictive variables such as short rate, dividend yields and earnings yields have good predictability power at both short and long horizons, different from both the conventional wisdom and Ang and Bekaert (2007). Contrary to Goyal and Welch (2007), an out-of-sample nonparametric forecast outperforms the historical mean model and linear predictive models.

Keywords: Asset Return Predictability, bootstrap, Hypothesis Testing, Kernel, Nonlinearity, Signal-to-Noise

JEL Classification: G17 C14 C53