A Practitioner’s Defense of Return Predictability

Abstract

Revisiting the issue of return predictability, the authors show that there is substantial predictive power in combining forecasting variables. Applying correlation screening to combine 20 variables that have been proposed in the return predictability literature, the authors demonstrate forecasting power at a six-month horizon. They illustrate the economic significance of return predictability through a walk-forward simulation, which takes positions in the SPDR S&P 500 ETF Trust (SPY) proportional to the model forecasted equity risk premium. The simulated strategy yields annual returns more than twice that of the buy-and-hold strategy, with a Sharpe ratio four times as large. To eliminate look-ahead bias, the authors perform additional simulations, while including variables only as they are discovered in the literature. Results show similar annual returns and Sharpe ratios. Although a market-timing strategy outperforms the market, the authors maintain that it is difficult to implement.