鑒於投資人情緒對市場報酬與波動存在的潛在影響，本文首度嘗試利用Engle提出的動態條件相關多變量DCC-GJR-GARCH模型來探討投資人情緒對於風險值(Value at risk，VaR)評量績效的影響。四種投資人情緒變數分別為避險、投機、大額交易及前八大機構投資人情緒，檢視不同投資人情緒變數對於估計風險值之影響。實證結果顯示，避險者情緒為市場價格變動的反向指標；投機投資人情緒與市場報酬呈現高度正向連動關係；大額交易投資人情緒部分，原油市場價格變動受到投機投資人情緒影響大於避險投資人情緒；黃金及美元指數市場而言，則是避險投資人情緒影響大於投機投資人情緒。最後，前八大機構投資人情緒對於報酬呈現負向影響，顯示前八大機構投資人主要以避險投資人為主。最後，前八大機構投資人情緒最能正確預估風險值，有利於投資者或市場參與者更有效的控管風險性資產。 In view of investor sentiment on market returns and volatility of the potential impact, this study first attempts to use Engle's Dynamic Conditional Correlation multivariate DCC-GJR-GARCH model to assess the impact of investor sentiment on Value at risk (VaR) performance. Four kinds of investor sentiment variables are hedger, speculator, extreme large trader and the top eight large institutional investors’ sentiment, this study measures different investors sentiment impact of the estimated risk value. The empirical results show that the sentiment of hedger is the contract indicator of the changes of market price. Speculators’ sentiment is highly positive related with market returns. As for the sentiment of extreme large trader, changes in market prices of crude oil are affected more by hedgers than speculators. Speaking of gold and the dollar index market, the hedgers were affected more than speculators. Besides, the sentiments of the top eight institutional investors have negative impact to market return. It shows that top eight institutional investors are mainly hedgers. Finally, sentiments of the top eight institutional investors can correctly estimate the value of risk most, in this way, investors or market participants can control risky assets more effectively.
