eThesis - School of Business electronic theses

In this thesis, I study the relationship between social media discussion and the stock index performance in Finland. My work relates to previous literature that uses Google searches or social media messages to explain trading activity and future stock returns. Previous research has acknowledged the relationship between social media discussion and the future returns, but the results vary due to the source of the messages and the timing of the research period. The contribution of this thesis is to provide more understanding of the dynamics of the effects of an increase in social media discussion level to stock index turnovers and returns.

I design a vector autoregression model for social media discussion, index return and index turnover. I calculate the impulse response functions from the vector autoregression model and I use the social media discussion impulses to index turnovers and returns in order to analyze the dynamic responses of the three key variables. The dataset used in the study consists of Finnish social media messages and daily index values and turnovers of NASDAQ OMXH Helsinki from January 2008 to December 2014. The OMX Helsinki stock index is split between three subindices divided by company size.

My findings show a robust relationship between social media discussion and stock index returns and turnovers. The increase in social media discussion is followed by an increase in stock index turnover for all company sizes. Increase in social media discussion is also followed by an increase in stock index return for smallcap companies and the return effect is reversed after a two-day window of positive return. I also present a simple trading strategy using the social media data. The strategy does not provide stable abnormal returns. The importance of entrance time into the strategy and transaction costs are the main limits to arbitrage.