🤖 AI Summary
This paper addresses the estimation challenge of semiparametric conditional latent factor models in asset pricing, aiming to disentangle the distinct roles of stock characteristics in modeling factor loadings (betas) versus pricing errors (alphas), and to explain the existence of high-Sharpe-ratio orthogonal arbitrage portfolios.
Method: We propose the first semiparametric framework that explicitly decouples the functional roles of characteristics in beta modeling and alpha identification. Conditional factors are extracted via principal component analysis on Fama–MacBeth sorted portfolios, and a two-stage regression ensures robust estimation.
Results: Although the model passes conventional specification tests, it exhibits significant, time-varying nonzero alphas. Orthogonal arbitrage portfolios constructed from these alphas deliver out-of-sample Sharpe ratios exceeding 4, revealing systematic mispricing unexplained by standard factor models.
📝 Abstract
We introduce a simple and tractable methodology for estimating semiparametric conditional latent factor models. Our approach disentangles the roles of characteristics in capturing factor betas of asset returns from ``alpha.'' We construct factors by extracting principal components from Fama-MacBeth managed portfolios. Applying this methodology to the cross-section of U.S. individual stock returns, we find compelling evidence of substantial nonzero pricing errors, even though our factors demonstrate superior performance in standard asset pricing tests. Unexplained ``arbitrage'' portfolios earn high Sharpe ratios, which decline over time. Combining factors with these orthogonal portfolios produces out-of-sample Sharpe ratios exceeding 4.
