The relationship of social media messages with stock returns and volatility

The aim of this study is to investigate the effects of messages conveyed on social media on stock returns and volatility. Accordingly, companies in the BIST30 and Sub-Market segments are included in the scope. Official Twitter accounts of licensed brokerage firms operating in Turkey and Twitter accounts belonging to private individuals with the highest number of followers and to whom the most messages about companies were sent on social media during the research period were preferred as message sources. Official accounts are taken as corporate (institutional) news sources and other accounts are taken as non-institutional news sources. Sentiment analysis of the obtained messages was performed, and the effect of the news on stock returns and volatility according to sentiment polarization was analyzed using the GARCH model. Data analysis was carried out according to the emotional states of the messages conveyed by institutional and non-institutional news sources. Within the framework of the GARCH model, it has been determined that returns and volatility are affected as a result of the news about the relevant companies. In line with the assumption that institutional investors are interested in the news transmitted from corporate / institutional accounts, and non-institutional investors are interested in the news transmitted from corporate and mostly non-institutional accounts; clues have been obtained regarding the separation of investor behavior in terms of two different types of investors and the explanation of this separation within the scope of behavioral finance. This study contributes to the literature on whether the effects of news transmitted from corporate and non-corporate sources are different when examining the impact of social media news on stocks. In this regard, an efficient method is proposed for acquiring social media data and separating news sources as corporate or non corporate. It contributes to the literature in terms of examining the news impact by distinguishing between institutional and non-institutional sources of social media messages regarding intraday returns, especially for Borsa Istanbul. It is thought that this research study is an empirical research study that can benefit the formation and development of social sentiment funds.