| Summary: | 博士 === 國立高雄第一科技大學 === 管理研究所 === 96 === This study modifies the procedure of forming reference portfolios first proposed by Fama and French (1992) for the measuring of long-term price performance. Using the Taiwanese data, we evidence the size and book-to-market ratio are the appropriate firm characteristics that can be used for constructing reference portfolios. However, this study finds that the priority of using firm characteristics in sorting firms, the amount of reference portfolios, and the use of firms that list on large exchange in ranking samples affect the quality of reference portfolios. We suggest that disregard the exchanges in ranking firms, sort firms firstly by book-to-market ratio, and form twenty-five reference portfolios produce better reference portfolios in Taiwan. Using the event study framework, buy-and-hold abnormal return calculation and bootstrapped skewness adjusted t-statistic as Lyon, Barber, and Tsai (1999) suggested in measuring long-rum abnormal stock returns, the reference portfolios we constructed yield well-specified test statistics in random samples. Contrarily, the reference portfolios created following the method of Ikenberry et al. (1995) do not. This finding implies the number of reference portfolio set should be subjected to the market capability and the technique of forming reference portfolio varies across the markets.
|
|---|
