Does the CAPM Predict Returns?

Abstract

We evaluate the cross-sectional performance of the CAPM in different expected market regimes. We show that CAPM-betas positively predict portfolio and individual stock returns when market returns are expected to be high, which occurs about 50% of the time. Consequently, the product of beta and the expected market return (CAPM) predicts asset returns out-of-sample, and the predictive power of the CAPM outperforms that of alternative factor models. Strategies exploiting the joint predictive power of beta and the market return predictor have average returns increasing with beta, and Sharpe ratios up to twice those of the corresponding buy-and-hold strategies.