Firms’ financial leverage can largely explain the value effect, however, we document that the expected return-beta relationship of equity in unconditional tests of the CAPM is more generally distorted by leverage. We show that unlevered betas explain the cross-sectional variation in average unlevered returns for various portfolio sorts. The model outperforms the traditional CAPM and the Fama and French (1993) three-factor model. We address the implications for levered returns with a two-beta model that accounts for the leverage-risk premium interaction. The robustness of the results and the theoretical underpinnings advocate for the use of unlevered returns in asset pricing tests. 

VIEW PAPER