Standard tests for abnormal returns are invalid if the abnormal returns are cross sectionally correlated. We model the cross sectional correlation in the abnormal returns by a spatial autoregressive (SAR) model. The SAR model estimates the cross sectional correlation in the abnormal returns from the event day, or event period. Tests which are robust to an event-induced increase in the volatility and cross sectional correlation of the abnormal returns are proposed. Empirical applications to US stock returns around Bear Stearns' collapse and Lehman Brothers' bankruptcy in 2008 are provided as illustrations.
|Title of host publication||68th European Meeting of the Econometric Society|
|Number of pages||38|
|Publisher||EEA-ESEM - European Economic Association & Econometric Society|
|Publication status||Published - 25.08.2014|
|MoE publication type||B3 Article in conference proceedings|
|Event||68th European Meeting of the Econometric Society (ESEM) - Toulouse, Toulouse, France|
Duration: 25.08.2014 → 29.08.2014
Conference number: 68
- 511 Economics
- Abnormal return
- Cross sectional correlation
- Event study
- Spatial autoregressive model
Ahlgren, N., & Antell, J. (2014). Tests for Abnormal Returns in the Presence of an Event-Induced Increase in the Cross Sectional Correlation. In 68th European Meeting of the Econometric Society EEA-ESEM - European Economic Association & Econometric Society.