基金經理人情緒與基金績效預測－拔靴追蹤因果關係檢定

Abstract

基金績效是否可以預測，過去相關文獻主要是從基金的過去報酬、基金流量、經理人的選股能力與外在特質等進行研究，本研究則從基金經理人的內在心理角度來分析。本文首先確認了32 檔中小型基金經理人情緒與基金績效之間，存在著橫斷面相依與異質性，其次，採用Kónya (2006) 的拔靴追蹤因果關係模型，檢視中小型基金經理人情緒與基金績效的四種因果關係：樂觀、保守、回饋與中立，並建立四種基金投資組合。實證結果發現，不論持有期是一年或二年，中立因果關係型的投資組合，即基金經理人情緒與基金績效互相沒有影響的報酬最佳；回饋因果關係型的投資組合，即基金經理人情緒與基金績效互相影響的報酬最低。在進一步考慮風險變數下，發現基金經理人投資策略不受情緒與系統風險影響的基金，有更優異的績效表現。

If fund performance could be predicted, previous related reference documents had studied the past compensation of fund, fund flows and security selection ability of manager and external characteristics, but this study is analyzed from internal psychological point of view. This paper firstly confirms that there is cross section dependence and heterogeneity between 32 middle and small size fund managers' sentiment and fund performance. Next, five years as a ection moves year by year and there are totally 9 sections, it adopted the Kónya (2006) Bootstrapping Causality Model and inspects the four causalities of managers' sentiment and fund performance within these sections: optimistic, conservative, feedback and neutrality, and established four fund portfolio. It proves that no matter the holding period is one or two year, the investment portfolio of neutrality causality that is the fund managers' sentiment and fund performance are not affected by each other, the compensation is the best; the investment portfolio of feedback causality that is the fund managers' sentiment and fund performance are affected by each other, the compensation is the worst. In the further analysis, the investment strategy of the fund manager is not influenced by sentiment and system risk, and has the better performance of compensation.