Does Twitter Activity Proxy for Idiosyncratic Information ? Evidence from the Indian Stock Market

Abstract

This study analyzed the impact of stock-specific information, proxied by Twitter activity, on the stock returns of listed Indian firms. Twitter activities at different levels of information assimilation were considered through tweet-count, retweet-count, and favorite-count as proxies for information supply, information propagation, and information affirmation/validation, respectively. Day-to-day price movement and price movements in two sub-periods: the off-market-hours and the market-hours were considered separately for the analysis. It was also tested whether Twitter activities had an asymmetric impact on stock returns on days with positive and negative sentiments. The study was carried out with data of over 2.4 million tweets about 437 Indian firms listed on the Bombay Stock Exchange for 124 trading days. Panel data analysis with random and fixed effects was employed to test whether the Twitter activity is a significant price mover. The results showed that all three measures of Twitter activity significantly impacted stock returns on a day-to-day basis, especially during the overnight period. However, during market hours, only tweet-count had a significant impact on stock price movements. Further, the results revealed that Twitter activity had the most significant impact on the price of a stock when the market sentiment about the stock was negative. This study is the first large-scale study in the Indian context and opens up the possibilities of further research on these lines.