基金投資整合決策支援模式

針對共同基金績效研究所採用的年化標準差、β條款及Sharpe等單一指標，其不足之處在於無法涵蓋所有會影響基金績效表現因子。雖然Murthi等人提出以資料包絡分析法（Data Envelopment Analysis, DEA）來改善其缺點，但傳統DEA無法針對不在效率前緣上的基金加以排序，所以不能進一步以此協助基金管理者處理預算配置等投資決策。故本研究提出一資金投資整合決策支援模式，改以核糊多目標DEA來改善不能直接排序的特性，並以模糊多目標資料包絡法視窗分析來讓基金管理者了解基金歷年來績效走勢的表現，更進一步以多目標投資支援模式，來整合基金排名的結果與基金管理者自我的偏好，依據不同的投資目標，給予不同基金管理者所需的最佳投資組合建議。為驗證本模式的可用性，本研究以民國91至93年間31家台灣科技型股票基金為實例，分別以兩種不同的案例來給予基金管理者不同的支援與建議。分析的結果顯示本研究所提出的模式，確實能以一整合性的流程給予基金管理者完整的協助，並從中發現國內科技型股票基金尚不能完全排除由基金經理人等所造成的非系統風險影響的現象。

關鍵字

共同基金； DEA ；模糊多目標規劃

並列摘要

An integrated decision support method is proposed to provide not only the ranking of mutual funds but also a suitable portfolio. The conventional data envelopment analysis (DEA) has been applied to evaluate mutual funds. Although DEA copes with the weakness in using annualized standard deviation of return, β coefficient and Sharpe index etc., the application is limited to evaluation only. No ranking or budget allocation decision is suggested further. The proposed method adopts a fuzzy multiple objective programming into DEA to provide ranking among mutual funds. A window analysis is also addressed to show annual trend for each fund. After incorporating the ranking and preference of the decision maker, a multiple objective model is developed. The best portfolio could be derived according to the goals set by the decision maker. The empirical data of 31 equity funds of technology in Taiwan during 2002 and 2004 with two different conditions is employed to demonstrate the proposed model. The informative results such as ranking, trend and budget allocation are provided. Furthermore, it also finds that the non-system risk is an important factor in rating the mutual funds.