The conditional CAPM and the cross section of expected returns : evidence for the Canadian market

In this study we test a conditional version of the CAPM, proposed by Jagannathan and Wang (1996), that allows betas (Ý) to vary over time as proxied by the yield spread between three-month Prime Corporate Paper and the three-month T-Bill rate ([Special characters omitted.] ). The model also includes a measure of the sensitivity of human capital to the market Ý as proxied by the lagged return on Total Labor Income ([Special characters omitted.] ), and SIZE (log of share price times number of shares outstanding) as explanatory variables. Our objective is twofold: (a) to test this model's ability to better explain the cross-sectional variation of monthly returns on 25 SIZE - and beta -sorted portfolios of Canadian common stocks over the period from June 1965 to December 1992 (330 months); and (b) to compare the performance of this conditional CAPM with the unconditional CAPM (SLB model). For portfolio formation, we use a methodology similar to Fama and French (1992). For the estimation procedure, we use the two-step approach of Fama and MacBeth (1973) as well as the more powerful GLS time-series cross-sectional estimation approach. (Abstract shortened by UMI.)