We present new evidence on the predictability of aggregate market returns by developing two new prediction models, one risk-based, and the other purely statistical. The pricing kernel model expresses the expected return as the covariance of the market return with a pricing kernel that is a linear function of portfolio returns. The discount rate model predicts the expected return directly as a function of weighted past portfolio returns. These models provide independent evidence of predictability, with R2 of 16-19% for 1-year returns. We show that innovations in the pricing kernel are associated with the cash flow component of the market return.