Control of Beta Reliability in Studies of Abnormal Return Magnitudes: A Methodological Note
探讨在市场异常绩效研究中，当效应幅度较小时，控制贝塔可靠性（即回归系数）的方法是否有助于改进市场模型的应用。
Research involving applications of the market-based abnormal performance methodology has evolved from relatively simple analyses of the signs of abnormal returns (Ball and Brown [1968]) into more sophisticated analyses regarding the magnitudes of the variables being studied (Beaver, Clark, and Wright [1979] and Pincus [1983]). This use of the latter approach leads to issues in the measurement of the abnormal returns and the statistical techniques to be employed. Brown and Warner [1980] investigated five different abnormal return measurement methods and found that in situations where the magnitude of the effect being investigated is large (5% in month 0) all five of the techniques adequately picked up the abnormal performance. They also concluded that the simple one-factor market-model approach was adequate in such situations, and even simpler techniques may yield the same results. However, there remains an issue of whether modifications in the market-based methodology may be beneficial in cases in which the magnitude of the effect being investigated is small. In this paper I show that a technique to control for beta reliability (i.e., the regression coeffi-