共同基金近年來的重要性逐漸提升,現已成為家庭財富的主要部分之一。股票型基金分散投資於市場上的各種股票,於是股票型共同基金的淨值與加權股價指數有高度相關性。為了瞭解基金經理人對於指數走勢的掌握度與專業度,本文使用VECM-GARCH(1,1)和VAR-GARCH(1,1)模型來分析基金淨值與股價指數間的共整合關係及因果關係,並在瞭解其因果關係後,利用此關係進行基金淨值或指數的預測。資料除涵蓋包含完整景氣循環之十年資料外,並從其中擷取多頭時期和空頭時期以了解在不同時期是否會有不同的行為反應。 本文的主要結論如下: (1) 數支基金的淨值與股價指數間具有共整合關係,而此種關係多出現於空頭時期。 (2) 在全樣本期間和空頭期間,基金的淨值與股價指數間多具有因果關係,但在多頭時期則多 不具關聯。 (3) 由Theil U指標可知,以VECM-GARCH(1,1)和VAR-GARCH(1,1)模型所進行之樣本外預測具有良好的預測績效。
Mutual funds instruments have become increasingly important, and shares in the funds now represent a major part of the household wealth. Stock mutual funds often diversify their investments in a wide range of stocks of specific markets. Thus, their net asset values are usually highly correlated with weighted stock price indices. To inspect mutual-fund managers’ expertise, this thesis employs VECM-GARCH(1,1) and VAR-GARCH(1,1) models to examine cointegration patterns and causality relationships between Taiwan’s stock funds and stock price indices, and use the uncovered relationship to forecast net asset values of mutual funds and/or the stock price indices. Moreover, our data spans include a 10-year whole period, and an upward and a downward interval within the whole period. Three major empirical findings of this thesis are as follows: (1) Several funds’ net asset values are cointegrated with stock price indices, and most are in the downward period. (2) In the whole and downward periods, many mutual funds’ net asset values have causal relationships with stock price indices. But in the upward period, the funds’ net asset values are seldom correlated with stock price indices. (3) Judging from the Theil U criterion, the out-of-sample forecast performances are superior using the estimated VECM-GARCH(1,1) and VAR-GARCH(1,1) models.