This study out of the University of Utah attempts to discover and evaluate the predictive power of stock micro blog sentiment on future stock price directional movements. Professors Chong Oh and Olivia R. Liu Sheng construct a set of robust models based on sentiment analysis and data mining algorithms. Using 72,221 micro blog postings for 1909 stocks and 3874 distinct authors, their study reveals "not only that stock micro blog sentiments do have predictive power for simple and market adjusted returns" respectively, but also that this predictive accuracy is consistent with the underreaction hypothesis observed in behavioral finance. They establish that "stock micro blog with its succinctness, high volume and real-time features do have predictive power over future stock price movements. Furthermore, this study provides support for the model of irrational investor sentiment, recommends a complimentary investing approach using user-generated content and validates an instrument that may contribute to the monetization schemes for Virtual Investing Communities."
The Utah study follow two previous papers. The first is by Professor Eugene Soltes of the University of Chicago Booth School of Business. Published in 2009, Soltes’ paper News Dissemination and the Impact of the Business Press states,
“How information is distributed, even when public, is important. Ultimately, I find that the press has a significant effect on bid-ask spread, share turnover, and idiosyncratic volatility. Specifically, greater dissemination of firm news is found to lower bid-ask spreads, increase trading volume, and lower idiosyncratic volatility.”
Professor Terrance Odean of the University of California, Berkeley and Professor Brad Barber of University of California, Davis stated in their paper All that Glitters: The Effect of Attention and News on the Buying Behavior of Individual Investors,
“Consistent with our predictions, we find that individual investors display attention-driven buying behavior. They are net buyers on high volume days, following both extremely negative and extremely positive one-day returns, and when stocks are in the news. Attention-driven buying is similar for large capitalization stocks and for small stocks.”
"Having participated in and closely observed the StockTwits stream since the beginning, I am not surprised by the researchers’ findings. The crowdsourced nature of the StockTwits stream zeros in on and amplifies new information and analysis so quickly that often we see a stock ticker begin to trend before the majority of a corresponding move in the stock occurs.
Oh and Sheng’s work supports the underreaction hypothesis, a much studied and popular proposal among behavioral economists, which posits that asset prices tend to underreact to new information and only slowly drift over time towards new values which account for the news.
Its my suspicion that as research moves from academia to quantitative analysis performed by money managers, they will discover a variety of anomalies, only some of which might be explained by underreaction.
Tim Richards at The Psy-Fi Blog penned a thoughtful summary of the research and wrote,
I don’t know if StockTwits sentiment analysis can really predict market movements over any time period, but I will make a prediction. Someone, somewhere, is building a system to trade on the possibility. It’s an arms race, and in arms races its usually the side with the deepest pockets and the smartest people that wins. Only, in a world of globally networked social media it’s not so obvious who that’s going to be.