動態資料包絡分析法探討台灣共同基金效率

Abstract

本文利用資料包絡分析法(DEA)來衡量台灣境內共同基金的績效表現。Tone (2001)發表差額變數衡量法(SBM)衡量差額變數的影響以評估效率值。Tone和Tsutsui (2010)使用差額變數衡量架構延伸動態資料包絡分析法(dynamic DEA)而出的動態資料包絡分析-差額變數衡量法考量連結(carry-over)對連續期間的影響評價總效率值。縱橫斷面資料期間從2008年到2013年，合計77檔台灣開放一般型共同基金。我們選擇淨值、管理費、經理費和銷售費作為投入變項，年度報酬率為產出變項，連結是基金規模。我們不只分析總效率值與個別期間效率值的差異，也比較差額變數衡量法與動態資料包絡分析-差額變數衡量法的差別。我們發現兩模型評價的個別年效率值沒有顯著差異，但考量連結的動態資料包絡分析-差額變數衡量法能透過長期的最佳化效果評價精準的總效率值，應用考量各期間影響關係的動態資料包絡分析-差額變數衡量法判別投資工具是較佳的選擇。

This thesis applies Data Envelopment Analysis (DEA) to evaluate the performance of mutual funds in Taiwan. Tone (2001) proposed the slacks-based measure to measure the effect of slacks to evaluate the efficiency score. Tone and Tsutsui (2010) extends the dynamic DEA within the slacks-based measure framework to consider the effect of carry-over between the consecutive terms. A panel dataset contains 77 onshore mutual open-end equity funds in Taiwan from 2008 to 2013. We choose net asset value, custodian fee, management fee, and Front-End load as our input variables, and the return of fund as the output variable. The fund asset is the carry-over variable. We not only analyze the difference between overall score and individual term scores but compare the difference between SBM and DSBM. We find out that the efficiency scores in individual year which are evaluated by two models are no significant difference. However, owing to the concern of carry-over, DSBM measures the specific overall score based on the long run optimization. It will be better to apply DSBM to judge the investment vehicle performance with the consideration of the effect of consecutive terms.