Dissecting anomalies in conditional asset pricing

This paper introduces a novel methodology for analyzing anomalies in conditional asset pricing models with time-varying risk exposures and premia. Our approach extends the conventional two- pass methodology to include both ordinary and weighted least-squares estimation in a conditional setting. We establish closed-form standard errors to statistically dissect anomalies, including a version robust to global misspecification. We introduce a novel R-squared criterion to quantify the joint contribution of large anomaly sets in explaining cross-sectional stock return variations. Our analysis highlights the significant impact of anomalies during economic and financial crises, linking them closely with market conditions.