Many tests of asset‐pricing models address only the pricing predictions, but these pricing predictions rest on portfolio choice predictions that seem obviously wrong. This paper suggests a new approach to asset pricing and portfolio choices based on unobserved heterogeneity. This approach yields the standard pricing conclusions of classical models but is consistent with very different portfolio choices. Novel econometric tests link the price and portfolio predictions and take into account the general equilibrium effects of sample‐size bias. This paper works through the approach in detail for the case of the classical capital asset pricing model (CAPM), producing a model called CAPM+ε. When these econometric tests are applied to data generated by large‐scale laboratory asset markets that reveal both prices and portfolio choices, CAPM+εis not rejected.