Limited attention and disagreement in finance social-media platform

Wong, Gabriel 2021. Limited attention and disagreement in finance social-media platform. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF - Accepted Post-Print Version Download (2MB) | Preview
	PDF (Cardiff University Electronic Publication Form) - Supplemental Material Restricted to Repository staff only Download (225kB)

Abstract

This thesis examines retail investors' ability to predict return with behavioural constraints such as limited attention and disagreement through a unique dataset from StockTwits, a finance social media platform. To quantify StockTwits tweets into sentiment values, I borrowed from the computer science literature using the state-of-the-art machine learning algorithm, Word2Vec, to learn and classify tweets into three categories {Bullish, Neutral, Bearish}. With the rich user and stock level information from StockTwits, it was found that StockTwits sentiment predicted positively and significantly future stock returns. More importantly, such sentiment predictability decreased as the number of stocks users follow increased. This is in line with the limited attention explanation that users with more complex portfolios are inferior in assimilating information due to time constraints. I also explored a long-standing puzzle in asset pricing literature where the high risk does not deliver a higher return as proxied by CAPM beta. Hong and Sraer (2016) suggest uncertainties in fundamentals generate investor disagreement as the reason for the observed anomaly. Individuals who feel pessimistic are constrained from short selling, leading to lower future returns for these stocks. I formally test this hypothesis by using aggregate disagreement calculated from StockTwits, and confirms such findings.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Business (Including Economics)
Subjects:	H Social Sciences > H Social Sciences (General)
Uncontrolled Keywords:	Social media, Sentiment, Behavioral finance, Limited attention, Stock returns, Disagreement, CAPM
Date of First Compliant Deposit:	15 June 2022
Last Modified:	06 Jul 2023 01:56
URI:	https://orca.cardiff.ac.uk/id/eprint/150474

Actions (repository staff only)

Edit Item