A Study in the Non-linear Relationship between Media Content and Stock Return

• Main Authors: Cheng, Wei-Ying, 鄭瑋瑩
• Other Authors: Chiu, Huei-Yu
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/30823435714500868519

碩士 === 輔仁大學 === 金融與國際企業學系金融碩士班 === 103 === The study in the relationship or causal impact between financial news and stock return usually uses linear models to describe how and to what degree they affect each other. Literature review had proved that pessimism in financial news medias would influence stock return. Therefore, this study applied non-linear threshold vector autoregressive model to investigate if there is non-linear effect among the two most important variables: media content and stock return. Dow Jones Industrial Average return and newspaper articles from the Business Day pages of the New York Times ranging from Jan. 1 to Dec. 31, 2013, total 252 trading days, are taken in this study. To measure how pessimistic the media content is, this study quantified financial news content to determine the difference in positive and negative word counts and its proportion to total words defined as pessimism factor. The empirical results indicated that there are two thresholds in the former type of media proxy variable, while there’s no threshold effect in the later one. The paper found that how the media content and stock return interacted partly corresponded to literature review but some are different from existing literature or general perception.