On the Predictability of Stock Returns in Real Time

Abstract

Researchers have documented an abundance of evidence that stock returns are predictable ex post. We address in this study whether the cross section of stock returns is predictable ex ante. We ask if a real-time investor could have used book-to-market equity, firm size, and one-year lagged returns to forecast stock returns during the 1974 to 1997 period. Using a recursive out-of-sample method, we find that the market was difficult to beat in real time. Our findings suggest that the current notion of predictability in the literature is exaggerated.