台灣地區共同基金績效之效率分析

Abstract

本研究使用四階段資料包絡分析法分析2006年到2010年六十家台灣地區共同基金的效率。除了基金本身所具有的特質外，基金經理人的特質亦會影響基金的效率。本研究之模型係採用基金年報酬率為產出變數，基金費用率和基金淨值之年化標準差兩個投入變數。在第一階段使用DEA模型估計各基金之效率。第二階段透過Tobit複回歸模型來估計各種環境變數，是否會對理論投入值與實際投入值所產生的差額數值產生顯著性的影響。第三階段使用在第二階段中所得到的估計參數來調整總投入差額。最後一個階段利用經過調整後之投入值計算基金之效率值。在本研究結果顯示，女性經理人較男性經理人的績效表現好，資歷長的基金經理人所操作的基金績效並不會比資歷短的經理人來得好。

This thesis applies a four-stage data envelopment analysis (DEA) approach proposed by Fried et al. (1999), in order to study the operational environment-adjusted efficiency of sixty mutual funds in Taiwan from 2006 to 2010. In addition to the characteristics of funds, the characteristics of managers also significantly affect performance of mutual funds. There are one output (actual annual return) and two inputs (expense ratio and standard deviation) in the variable-returns-to-scale (VRS) DEA model. In the first stage, the VRS DEA is used to compute the performance of each fund. In the second stage, the Tobit regression is applied to adjust the inputs by excluding effects caused by environmental factors. In the third stage, the parameter estimated from the second stage is used to predict the total input slacks. In the last stage, the DEA model is run with slack-adjusted data in order to find out the performance of fund. The empirical results reveal that female manager will have lower risk and better performance than male mangers. Experienced managers may not perform better than new managers.