Expected returns on value, growth, and HML

In this paper, I analyze the predictability of returns on value and growth portfolios and examine time variation of the expected value premium. As a primary tool, I use the filtering technique, which accounts for time variation in expected cash flows and explicitly exploits the constraints imposed by the present value relation. I demonstrate that returns on value and growth portfolios are predictable, and the predictability is stronger for growth stocks. Applying the filtering technique to the HML portfolio, I build a novel powerful forecaster for the value premium. The new forecaster appears to be only weakly related to business cycle variables.