傳統上基金投資的優點在於透過投資組合可分散風險以及有專業的研究團隊與操作經理人執行操盤,Kacperczyk, Sialm, and Zheng(2005)提出基金集中投資可為投資人帶來的價值比分散投資多,並以產業集中度為量測指標,爾後陸續有學者提出以產品、集團、個股、持有家數等不同層面探討分散投資與基金績效之間的關聯性,故本文也採取上述五種集中度的量測指標來作為研究之變數。而集中投資必定帶來高波動的報酬程度,所以本文之貢獻在於評估績效上除了報酬亦納入報酬之變異數及偏態係數做線性規劃模式,以評估出較佳的基金相對效率值,而因相對效率值之最大值為1,此為受限制資料,再使用受限制的迴歸模型來做探究,並且本研究認為基金集中投資與否,並非為單一方向的關係,一味地分散投資組合或集中投資,故使用受限制分位數迴歸來分析在不同的績效區間,其集中度量測指標與基金績效間的關係。本研究以144檔台灣境內股票型基金2005至2009年為期間,結果顯示以平均狀況而言,集中度指標對基金績效皆無顯著的影響;而在績效最差的第10百分位區間,五種集中度指標一致顯示分散投資對績效有正向的幫助,表示績效較差的基金經理人應以分散投資組合為其操盤策略;在績效最佳的第90百分位區間,個股集中度與持股家數顯示集中投資對績效有正向的幫助,代表績效好的基金能藉由精選個股為投資人創造價值。
Traditionally, the advantages of mutual funds are the benefit of diversification and professional trading. However, it seems to have a conflict between the two advantages. A growing literature provides empirical evidences that fund managers have superior information of stocks, then they would like to hold focused portfolios to improve their performances. The mutual fund performance measurements used in this paper is the shortage function in the mean-variance-skewness space (MVS).To solve the problem of the censored value of scores. The purpose of this study is to use the censored quantile regression model to evaluate the relationship between the concentration index and mutual fund performance under different efficiency quantile levels. The result shows that the fund group with the highest 10% efficiency scores have the positive relationship between scores and concentration by investing in specific stocks. In the group of the worst 10% efficiency scores, the five indicators to measure the degree of concnetration in portfolio prove that diversification is the best strategy for fund managers to execute their asset allocating.