Full Reporting in Empirical Asset Pricing: A Statistical Decision-Theoretic Approach

This study develops a full-reporting framework for empirical asset pricing grounded in statistical decision theory. In the abductive stage, simulated pseudo-factor model spaces are used to evaluate whether candidate test scores can distinguish valid from invalid model states through likelihood behavior, receiver operating characteristic analysis, and classification metrics. In the inductive stage, realized sample scores are mapped to likelihood ratios, error rates, and decision criteria, translating observed evidence into model-state predictions. Reporting this complete inferential chain enables test performance and realized evidence to be integrated into transparent assessments of factor-model validity.