Beating the Average: Equity Premium Variations, Uncertainty, and Liquidity

Abstract

This paper contributes to the equity premium prediction literature by studying the performance of rarely or not researched predictors. To do so, we analyze the ability of state-of-the-art liquidity and uncertainty predictors to beat the historical average when forecasting the monthly US equity premium. For this purpose, we apply an out-of-sample predictive regression approach to analyze statistical accuracy as well as economic gains of equity premium forecasts. Our findings show that the treasury-eurodollar (TED) spread, as well as the macroeconomic uncertainty measure, is able to beat the historical average and provide robust predictions in various business cycles. Moreover, these two economic predictors also beat forecasts of a classical time series model.