A1118
Title: Why do asset pricing models fail in equilibrium
Authors: Alejandro Lopez Lira - University of Florida (United States) [presenting]
Abstract: The equilibrium limitations of asset pricing models are examined with rational, risk-averse investors who have incomplete information about expected returns and covariances. In equilibrium, the Sharpe ratio of the achievable optimal portfolio is bounded, including those implied by asset pricing models. In contrast, the Sharpe ratio of the optimal portfolio under the true data-generating process grows with the difficulty of the estimation problem. Mispricing is related to their squared difference and increases as the estimation problem becomes more challenging. The model makes several novel predictions: (1) pricing errors of parsimonious factor models increase with the number of assets or state variables, even without uncertainty about expected returns, hence diminishing the cross-sectional correlation between betas and average returns; (2) anomalies are expected to proliferate in high dimensions; and (3) multiple high Sharpe ratio strategies can coexist without subsuming each other. Strong empirical support is found for these predictions.
