波動率、交易次數與投資人情緒

Abstract

在過去研究股價波動率與交易變數的文獻顯示出股票波動率與交易次數呈現正相關，過去實證研究發現把股價波動率拆解成連續部份與跳躍部份後，僅有連續波動率與交易次數有正相關，而跳躍波動率與交易次數為負相關。本論文進一步的討論在考慮投資人情緒下，股票波動率與交易次數的關係是否會改變。本論文利用過去文獻的方法把投資人情緒回歸在總體經濟變數上，進一步把投資人情緒拆解成和總體經濟變數相關的部份與殘差的部分，其中把殘差部份視為純粹的投資人情緒。我們發現若考慮投資人情緒與總體經濟變數相關的部份下，會削弱連續部份的股價波動率與交易次數有正相關性，另外我們還發現在考慮純粹投資人情緒下，會強化連續部份的股票波動率與交易次數正相關性。

In previous researches about the relation between realized volatility and the number of trades (denoted by NT), we can find that there is a positive relation between realized volatility and NT. Furthermore, Giot, Laurent and Petitjean (2010) reveal that after decomposing realized volatility into continuous and jump components, only continuous component shows a positive and significant impact on NT and jump component has a negative relation with NT. I reexamine the relation between realized volatility and NT under the consideration of investors’ sentiment. Following former researches, I decompose realized volatility into continuous component and jump component. Moreover, I follow the method in Sibley, Wang, Xing and Zhang (2016) to decompose BW-sentiment index (Baker and Wurgler’s sentiment index) into the component of BW-sentiment index related to economic fundamentals and the component of BW-sentiment index unrelated to economic fundamentals viewed as “pure” sentiment. I show that the positive relation between continuous component and NT fades away as the component of BW-sentiment index related to economic fundamentals is considered. Moreover, I find that “pure” sentiment enhances the positive relation between continuous component and NT but provides little effect on the relation between jump components and NT.