| Abstract [eng] |
Changes in stock markets can be caused by different factors: updated macroeconomic scenarios, financial results of companies, information provided by media. The main source where nowadays investors can receive information is social media. Hence, changes in social media can also affect changes in stock market. Better understanding of social media impact could help to better conceive the stock market moving direction. Also, social media comparison with other information sources gives additional knowledge about the main factors which causes market movements and what should be main focus for investors. The object of research: The impact of social media for stock market returns and trading volumes. The aim of master’s thesis: Investigate social media impact on US stock market changes before and during COVID-19 pandemic. First of all, we discuss how and if US stock market dynamics has changed since the start of pandemic. We discuss different existing positions about what was the main reasons which caused the movements in stock market, why those changes are relevant and what is scientific problem behind it. After analysis we take a look at already existing theoretical solutions by other authors and investigate how they define relationship between stock market, social media and financial ratios. In this research the impact of social media is divided in few different categories: the impact of internet seaches, the impact of social networks and COVID-19 official statistics. In addition we take a look to models which uses more classical and older media sources like television and newspapers. Overall, the main focus in this research goes to application of Google Trends and Twitter sentiments. Afterwards long-term and short-term relationship analysis is implemented between stock market parameters and mentioned variables. Analysis investigate relationships on individual stock, industry sector and whole market level. The findings in analysis shows that relationship between social media sentiments and stock market returns are very unstable and varies depending on time. Also, sentiments show bigger volatility than stock returns. Based on findings multivariate regression analysis is implemented so we could compare differences of predictability with social media sentiment and with financial ratios only. Regression analysis implemented on three diffirent periods: before COVID-19, in the beginning of of pandemic and when pandemic situation is stabilized. Results indicate that social media sentiment could help to better describe changes in stock market, however model itself is more appropriate for forecasting changes in volumes and volatility, but not in stock market returns. |
