An Empirical Study on Information Reputation and Performance among Security Analysts

• Main Authors: Jeff Ding, 丁介甫
• Other Authors: Li-Hsien Chien
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/32951199269496270191

碩士 === 國立中興大學 === 應用經濟學研究所 === 93 === This study measures the effectiveness of common stock investing recommendations, suggested by certain Investment Consulting Corporation, by testing if the performances of the following three different conditions are equal statistically: the payment patterns, the magnitudes of performance, and the different memberships. During April and May of 2004, total 773 recommendations made by 3 individual analysts in some investment consulting corporation were adopted in this article. Market Model is adopted to evaluate the abnormal returns (ARs) and the cumulative abnormal returns (CARs) from these recommended portfolios. Three major conclusions are： 1. Statistic results show that ARs and CARs of the given recommended portfolios are not significantly positive after the event-date. The hypothesis that the securities that recommended by the given analysts outperform the Taiwan stock index during analytical period do not support by statistical results in this study. 2. Performance difference exists significantly among individual analysts by using F test. This implies that even three analysts use the same informative support services given by the corporation, the gaps of performance among the recommendations from different experts can be found because of their particular stock selections and specific habitual operation patterns. 3. The relationship between membership expenses and extra returns are not significantly supported by statistical results during these two months. Investors who pay the high membership fee are not necessary get high return from the recommendation. Thus, the effectiveness of recommendation information is not consistent with the level of memberships. To ensure the higher returns from investment, investors need to choose analyst(s) first.