长期事件研究方法论——一个综述

本文对国外长期事件研究方法论文献进行了全面的梳理,并着重介绍和分析了期望收益模型(或收益基准)选择、异常收益度量,以及检验统计量的设定与检验力等。研究发现,在各期望收益模型、两种度量异常收益方法—AAR(或CAR)与BHAR的选择,以及各检验统计量是否存在错误设定等问题上,国外学界并未达成共识,争论仍在继续。

Long-run Event Study Methodology-A Survey

We survey external methodology papers about longrun event studies,and focus on analyzing and introducing "the choices of expected return models(or retum benchmark)","the approaches to measure abnoumal performance",and "the empirical power and specification of test statistics".Our results show that as for which expected return models or approaches to measure abnormal performance—AAR(or CAR) and BHAR is more appropriate,and whether test statistics being misspecificated,there are different answers from the academic field,and the controversies still continue.