Comparing traditional news and social media with stock price movements; which comes first, the news or the price change?

Smith, Stephen and O’Hare, Anthony (2022) Comparing traditional news and social media with stock price movements; which comes first, the news or the price change? Journal of Big Data, 9. 47. ISSN 2196-1115 (https://doi.org/10.1186/s40537-022-00591-6)

[thumbnail of Smith-OHare-JBD-2022-Comparing-traditional-news-and-social-media-with-stock-price-movements]
Preview
 Text. Filename: Smith_OHare_JBD_2022_Comparing_traditional_news_and_social_media_with_stock_price_movements.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo
Download (1MB)| Preview

Abstract

Twitter has been responsible for some major stock market news in the recent past, from rogue CEOs damaging their company to very active world leaders asking for brand boycotts, but despite its impact Twitter has still not been as impactful on markets as traditional news sources. In this paper we examine whether daily news sentiment of several companies and Twitter sentiment from their CEOs have an impact on their market performance and whether traditional news sources and Twitter activity of heads of government impact the benchmark indexes of major world economies over a period spanning the outbreak of the SAR-COV-2 pandemic. Our results indicate that there is very limited correlation between Twitter sentiment and price movements and that this does not change much when returns are taken relative to the market or when the market is calm or turbulent. There is almost no correlation under any circumstances between non-financial news sources and price movements, however there is some correlation between financial news sentiment and stock price movements. We also find this correlation gets stronger when returns are taken relative to the market. There are fewer companies correlated in both turbulent and calm economic times. There is no clear pattern to the direction and strength of the correlation, with some being strongly negatively correlated and others being strongly positively correlated, but in general the size of the correlation tends to indicate that price movement is driving sentiment, except in the turbulent economic times of the SARS-COV-2 pandemic in 2020.

CORE (COnnecting REpositories)