Mardi | 2015-01-13
Sully 5, 16h-17h20
Patrick GAGLIARDINI – Elisa OSSOLA – Olivier SCAILLET
We build a simple diagnostic criterion for approximate factor structure in large cross-sectional equity datasets. Given a model for asset returns with observable factors, the criterion checks whether the error terms are weakly cross-sectionally correlated or share at least one unobservable common factor. It only requires computing the largest eigenvalue of the empirical cross-sectional covariance matrix of the residuals of a large unbalanced panel. The panel data model accomodates both time-invariant and time-varying factor structures. We develop the theory for large cross-section and time-series dimensions. No restriction is imposed on the relation between both dimensions. The empirical analysis runs on returns for about ten thousands US stocks from July 1964 to December 2012. Among several multi-factor models proposed in the literature, we cannot select a model with zero factors in the errors.