Long-horizon Event Studies: The Taiwan Evidence

• Main Authors: Hsin-Fu Lai, 賴信甫
• Other Authors: Kuang-Ping Ku
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/20765682814370754667

碩士 === 中國文化大學 === 會計研究所 === 92 === It is intended to detect the average long-run abnormal returns after the one to three years of the event day, and also to discover the suitable long-run abnormal return computational method, the test and the benchmark. This research period was from January 1985 to December 1999. The monthly data are used to simulate the different sample group, including random sample, non-random sample, cross-sectional dependence of sample observations. Four kinds of computa-tional methods of long-run stock abnormal returns were discussed including accumulation abnormal returns, buy-and-hold abnormal returns, mean monthly calendar-time abnormal returns, the Fama-French three factor model and calendar-time portfolios. Two test, conventional t-statistic and Wilcoxon signed-rank test and two benchmark, reference portfolios, and control firm are used to study. It is found that mean monthly calendar-time abnormal returns and the Fama-French three factor model and calendar-time portfolios will be well specified. The use of the Wilcoxon signed-rank test was found to yield more empirical rejection levels exceeding theoretical rejection levels. It is suggested that (1)mean monthly cal-endar-time abnormal returns matched with control firm, and (2)the Fama-French three factor model and calendar-time portfolios based on excess return on a value-weighted market index will reduce most of the misspecification in test statistic.