存續者偏誤、基金分類與績效持續性之研究

A Study of Survivorship Bias, Fund Categories and Fund Performance Persistence

基金績效持續性與否是非常典型的重要開放性議題。本文同時考量存續者偏誤、績效衡量指標不同與績效等級分類的差異，對台灣境內基金績效持續性進行討論。研究結果顯示(1)股票型基金績效之年持續性，將受到存續者偏誤的影響。(2)考量風險調整與否之基金績效差異，對整體基金績效年持續性之檢視結論將相反；對樣本不含清算合併之股票型基金及其他類型基金而言，衡量績效不同對持續性檢測有差異；債券型基金則非常穩定。(3)債券型基金之績效最佳組與績效最差組皆相對於中等績效基金，更具有兩年績效持續性之強度。(4)本文支持台灣境內共同基金具有某種程度的績效持續性，且強弱程度會因投資標的分類、績效評估指標不同呈現不一致的情況。

Mutual fund performance persistence is an important issue. However, most of the empirical studies do not provide consistent results. This study focuses on factors that measuring the mutual fund performance persistence, such as survivorship bias, definition of performance, and method of return categories. The empirical results of study support the following notions. First, survivorship bias and different performance affect performance persistence of stock funds and balance funds. Second, there exists a certain degree of persistence in mutual fund performace for different performace and categories. Finally, funds within the best and the worst performance groups have more persistent perfomace in mutual funds classification algorism.