This paper uses an equilibrium multifactor model to interpret the cross-sectional pattern of postwar U.S. stock and bond returns. Priced factors include the return on a stock index, revisions in forecasts of future stock returns (to capture intertemporal hedging effects), and revisions in forecasts of future labor income growth (proxies for the return on human capital). Aggregate stock market risk is the main factor determining excess returns; but in the presence of human capital or stock market mean reversion, the coefficient of relative risk aversion is much higher than the price of stock market risk. The capital asset pricing model (CAPM) measures the risk of an asset by the covariance of the asset's return with the return on all invested wealth, also known as the "market return." The price of market risk is determined by the risk aversion of investors. However, the CAPM has come under attack from several directions. The assumptions used to derive the model have been criticized, and the main auxiliary assumption—the assumption that the market return can be adequately proxied by a stock index return—has been challenged. In response to these critiques, many economists have estimated multifactor models in which risk is measured by covariances with several common factors. These models require only very weak theoretical assumptions and appear to give a good empirical fit to the cross section of asset returns. However, multifactor models do not give clear answers to the two questions posed at the beginning of this paper. This paper responds to the Merton and Roll critiques in a more structured fashion. It develops a simple discrete-time asset pricing model that allows for both changing investment opportunities and an important component of wealth—human capital—whose return may not be well proxied by the return on a stock index. The model links the vast time-series literature on asset returns to the equally vast cross-section literature, responding to Fama's call for "a coherent story that relates the variation through time in expected returns to models for the cross-section of expected returns." The model is an alternative to the consumption-based capital asset pricing model (CCAPM). The CCAPM handles the Merton and Roll critiques of the CAPM by using the covariance with aggregate consumption instead of the covariance with the market as a measure of risk. However, the CCAPM has at least two weaknesses. The paper uses a log-linear approximation to the budget constraint to get a closed-form solution for the consumption of a representative investor facing conditionally lognormal and homoskedastic asset returns, and maximizing the objective function proposed by Epstein and Zin (1989, 1991) and Weil (1989). From this, it is easy to derive an asset pricing formula that makes no reference to consumption, instead relating assets' returns to their covariances with the market return and news about future market returns. In response to the Roll (1977) critique, the paper extends the Campbell (19This paper uses an equilibrium multifactor model to interpret the cross-sectional pattern of postwar U.S. stock and bond returns. Priced factors include the return on a stock index, revisions in forecasts of future stock returns (to capture intertemporal hedging effects), and revisions in forecasts of future labor income growth (proxies for the return on human capital). Aggregate stock market risk is the main factor determining excess returns; but in the presence of human capital or stock market mean reversion, the coefficient of relative risk aversion is much higher than the price of stock market risk. The capital asset pricing model (CAPM) measures the risk of an asset by the covariance of the asset's return with the return on all invested wealth, also known as the "market return." The price of market risk is determined by the risk aversion of investors. However, the CAPM has come under attack from several directions. The assumptions used to derive the model have been criticized, and the main auxiliary assumption—the assumption that the market return can be adequately proxied by a stock index return—has been challenged. In response to these critiques, many economists have estimated multifactor models in which risk is measured by covariances with several common factors. These models require only very weak theoretical assumptions and appear to give a good empirical fit to the cross section of asset returns. However, multifactor models do not give clear answers to the two questions posed at the beginning of this paper. This paper responds to the Merton and Roll critiques in a more structured fashion. It develops a simple discrete-time asset pricing model that allows for both changing investment opportunities and an important component of wealth—human capital—whose return may not be well proxied by the return on a stock index. The model links the vast time-series literature on asset returns to the equally vast cross-section literature, responding to Fama's call for "a coherent story that relates the variation through time in expected returns to models for the cross-section of expected returns." The model is an alternative to the consumption-based capital asset pricing model (CCAPM). The CCAPM handles the Merton and Roll critiques of the CAPM by using the covariance with aggregate consumption instead of the covariance with the market as a measure of risk. However, the CCAPM has at least two weaknesses. The paper uses a log-linear approximation to the budget constraint to get a closed-form solution for the consumption of a representative investor facing conditionally lognormal and homoskedastic asset returns, and maximizing the objective function proposed by Epstein and Zin (1989, 1991) and Weil (1989). From this, it is easy to derive an asset pricing formula that makes no reference to consumption, instead relating assets' returns to their covariances with the market return and news about future market returns. In response to the Roll (1977) critique, the paper extends the Campbell (19