Research the Relationship of Public Information, Stock Returns and Institutional Investors’ Trading Behaviors by Text Mining Technique: Evidence from Taiwan Listed Electronic Stock

• Main Authors: Yu-Ru Shih, 施喻繻
• Other Authors: Jon-Chi Shyu
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/65800711445047627417

碩士 === 國立臺灣科技大學 === 企業管理系 === 103 === The study used the top 15 companies of market value in Taiwan Electronic Stock Market, using Panel Data regression model and text mining technique－LIWC to examine the relationship between Firm-Specific, Industry and Macroeconomics public information, stock returns and institutional investors’ trading behaviors from 2014/1/1 to 2014/12/31. The empirical results of this study are: (1)Firm-specific and Macroeconomics information are positively related to stock returns. The relationship means that good daily firm-specific and macroeconomics news would make positive stock returns but bad daily firm-specific and macroeconomics information would make negative stock returns. (2)Firm-specific information is positively related to foreign investors’ and security dealers’ trading behavior. In addition to the good(bad) daily macroeconomics news would attract foreign investors to buy(sell) electronic stocks. Particularly, all kinds of information have no influence on investment trusts’ trading behavior. (3)Firm-specific is negative and Macroeconomics information is positive related to retail investors’ trading behaviors which is opposite to foreign investors’ and security dealers’ trading behaviors.