I am a PhD Economics student at Cardiff University. My research focuses in the area of Behavioural finance, Empirical asset pricing, and Corporate finance.
My job market paper implements textual analysis such as Natural Language Processing to gauge tweet sentiment and predict future stock returns.The paper is about a long-standing debate in investment community for generalist vs specialist, defined by large (small) number of past stocks coverage, would provide better forecast. Using finance social media platform “StockTwits”, we identified when user’s past stock coverage is high, together with sentiment they generate predicts lower future return and earnings. Consistent with limited attention hypothesis, high stocks coverage users suffer from informational disadvantage due to behavioral time constraint. Such effect is especially pronounced when firms are opaque, hard to analyze, and complex.