Lets test CAPM! We measure the market risk premium at 4% If we examine the data and find that, across a set of beta sorted portfolios, an increase in beta of 1 yields an increase in return of 6%, we can conclude that This is evidence in favor of CAPM because the reward per unit of systematic risk is exactly as predicted by the CAPM. CAPM does not appear to hold across these portfolios because the reward per unit of risk is too high This does not supply evidence about the CAPM model. O CAPM does not appear to hold across these portfolios because the reward per unit of risk is too low