The Korean Economic Forum
Construction of Financial Sentiment Dictionary and Effectiveness of Financial Sentiment Index
So-yeon Lim (Ewha Womans University) and In-bae Ki (Ewha Womans University)Year 2023Vol. 16No. 1
AbstractThis paper investigates whether the investors’ psychological factor has a predictive power for the stock market movement from the perspective of the behavioral economics. We made several contributions in this literature. First, we constructed the financial sentiment dictionary for the Korean text analysis to its name for the first time. It is based on the most popular financial sentiment dictionary of Loughran and McDonald (2011), but much expanded by Word2Vec model. Second, different from the related previous studies that mostly used the monthly data for financial sentiment index (FSI), we built its daily data to explain the high frequency stock return movement. Lastly, we built two kinds of FSI, that is, ‘KOSPI FSI’ for the overall Korean stock market and ‘Samsung Electronics FSI’ for an individual stock price, and explore their predictive powers. From this exercise we found the following results. (i) When the standardized daily KOSPI FSI increases by one standard deviation, the daily KOSPI return rises by 0.60%. (ii) When the standardized daily Samsung Electronics FSI increases by one standard deviation, the daily excess return of Samsung Electronics rises by 0.75%. (iii) However, the KOSPI FSI can not explain the excess return of Samsung Electronics. Nor can the Samsung Electronics FSI the KOSPI returns. Such results enforce the reliability and validity of the FSIs constructed in this paper.