Mispricing in linear asset pricing models

In the framework of a reduced-form asset pricing model featuring linear-in-instrument betas and time-varying risk premiums and allowing for missing factors, I propose a measure of mispricing that is largely free from the bias due to missing factors or missing instruments. Applying the model to U.S. equity data, I find evidence of mispricing in stock returns. A mispricing-based zero-dollar investment strategy intersecting momentum and contrarian horizons is highly profitable when applied to both firm- and portfolio-level returns, even after controlling for Fama-French factors, momentum and liquidity effects. Focusing on momentum, I find that the phenomenon is partially caused by the mispricing that does not vary with macro variables. Time-varying betas reduce the mispricing by 40% or better.