Using Expectations to Test Asset Pricing Models

Abstract

Asset pricing models generate predictions relating assets’ expected rates of return and their risk attributes. Most tests of these models have employed realized rates of return as a proxy for expected return. We use analysts’ expected rates of return to examine the relation between these expectations and firm attributes. By assuming that analysts’ expectations are unbiased estimates of market-wide expected rates of return, we can circumvent the use of realized rates of return and provide evidence on the predictions emanating from traditional asset pricing models. We find a positive, robust relation between expected return and market beta and a negative relation between expected return and firm size, consistent with the notion that these are risk factors. We do not find that high book-to-market firms are expected to earn higher returns than low book-to-market firms, inconsistent with the notion that book-to-market is a risk factor. Asset pricing models seek to establish the determinants of financial assets’ expected rates of return. Classic asset pricing models, such as Sharpe (1964), Lintner (1965), and Black (1972), predict that an asset’s expected return should be positively related to its systematic market risk. Based on Merton’s (1973) ICAPM, Fama and French (1992, 1993) argue that their findings of higher returns for high book-to-market stocks and low capitalization stocks reflect compensation for risk. That is, those stocks are expected to earn higher rates of return because they are riskier. Although asset pricing models aim at explaining cross-sectional variation in expected return, researchers have been forced to use realized return as a proxy for expected return in tests of these models. The use of realized returns implies that such tests are conditioned on the joint hypothesis of rational expectations, in the sense that the average realization is a good proxy for expectation and that the null asset pricing model describes the relation between expectations and firm attributes. However, realized return may not be a perfect proxy for expected return. First, noise in realized returns is likely to be large (Blume and Friend, 1973 and Sharpe, 1978). Second, realized returns may be poor estimates of expected returns if information surprises do not cancel out over the period of study (Froot and Frankel, 1989 and Elton, 1999). Third, realized returns may also be noisy and biased estimates of expected returns due to complex learning effects. 1 1