Strong and weak cross-sectional dependence in factor models for returns in event studies

Cross-sectional dependence in abnormal returns is investigated using a recently introduced exponent of cross-sectional dependence. The abnormal return is defined as the actual return minus the normal return predicted by a factor model. The factor model is valid for computing abnormal returns provided there is no remaining factor structure (strong cross-sectional dependence), so that abnormal returns are at most weakly cross-sectionally dependent. We compare five commonly used factor models for returns. The models are the market model, three-factor model, three-factor model extended with momentum factor, five-factor model and five-factor model extended with momentum factor. The models are fitted to US stock returns from Bear Stearns’ collapse and Lehman Brothers’ bankruptcy in 2008. These data provide evidence that the cross-sectional dependence in returns due to common factors other than the market factor is strong. Contrary to recommendations in the literature to use the market model, we propose the use of multi-factor models to compute abnormal returns in event studies.