自1993年芝加哥選擇權交易所(CBOE)所提出波動率指數成為投資人恐慌指數的代表與近年來行為財務學逐漸成為大家所重視的議題,投資人情緒成為影響股價超額報酬的重要因素。本研究以代表投資人恐慌情緒的波動率指數、籌碼面的融資、融劵餘額、外資法人買賣超及週轉率作為投資人情緒代理變數,及總體經面的貨幣供給及金融業隔夜拆款利率,並利用3種資料頻率(日、週、月)分別以Granger因果檢定與分量迴歸模型探討投資人情緒在不同市場報酬下的影響關係。其實證結果如下 一、 因果檢定結果波動率指數僅在日資料對股價超額報酬有影響、股價超額報酬僅在短期間(日、週)對融資餘額具有領先落後關係,股價超額報酬僅在短期間單向影響融劵餘額,股價超額報酬對週轉率有長期性的影響效果。 二、 分量迴歸實證模型結果:波動率指數與融資餘額在三個資料頻率模型皆為顯著,波動率指數與股價超額報酬有負向關係,波動率增加代表投資人恐慌程度增加,市場相對報酬表現較差;融資餘額與股價超額報酬有正向關係,融資餘額增加反有助於大盤表現,外資買賣超僅在短期間對市場報酬有正向影響。 三、 由分量迴歸模型中日資料大部分解釋變數皆為顯著,週資料受前期報酬影響最大,由月資料可得知總體經濟變數也具有影響力。不論在日、週、月資料,在低分量解釋能力較佳,顯示模型大盤相對不好時其有較好的解釋能力。 關鍵字:波動率指數、投資人情緒、Granger因果檢定、分量迴歸
In recent years, Behavioral Finance has become increasingly a topic of great importance. And it was proposed the investor’s Sentiment Indices are important factoris affecting stock returns. In this study, we selected seven variables to proxy investor’s sentiment indices, including the VIX index, margin trading, short sales, and volume of foreign institutional investor, stock’s turnover, the overall money supply and the interbank overnight rate. This paper examines the investor’s sentiment indices on the stock returns in Taiwan by using three kinds of data frequency (daily, weekly, monthly). We use the Granger causality test and quantile regressions to investigate if and how the stock returns in Taiwan are different at all quantiles. Our results are summarized as follows. First, the Granger causality test shows that, on average, the major trend in stock returns in Taiwan is determined by the VIX index in the short term (daily). The stock excess return in Taiwan affects margin trading and short sales in the short term (daily, weekly) and stock’s turnover in the long term (monthly). Secondly, the quantitle regression shows that the VIX index and margin trading are Statistics significant in three kinds of data frequency. The VIX index has negative influences upon the stock returns in Taiwan, but margin trading is positively related to the stock returns in Taiwan. Finally, the investor’s Sentiment Indices are important factors to affect stock returns by using the daily data. It is important to consider AR (p) terms when we use the weekly data. And the macroeconomic factors, like the overall money supply and the interbank overnight rate, are important and positively related to stock returns by using the monthly data. However, the explanatory power of the model is better in the lower quantile. Keywords: VIX ,Investor Sentiment, Granger causality test,Quantile Regression