Social Media and Financial Market

Abstract

Researchers and practitioners use tweets and comments on social media for emotion artificial intelligence because social media users tweet or post voluntarily. These researchers and practitioners assume tweets tweeted at different time are homogenous. This study constructed sentiment proxies from StockTwits tweets posted during NYSE non-trading hours and during NYSE trading hours. This study examined whether they predict daily S&P500 Futures, Emini Dow Futures, and Emini NASDAQ100 Futures returns using an ordinary least squares regression with Newey-West standard errors. This study found: (1) Collective sentiment at different hours of tweeting exerts a different level of influence on futures returns. (2) Positive sentiment and negative sentiment have a different influence on futures returns. (3) Returns predictability differ across the three futures contracts. Investors, businesses, and companies providing social media sentiment proxies should consider the time the tweets were tweeted and positive/negative sentiment in using social media tweets for insights.