Being able to predict movements in the stock market – with any level of accuracy — has drawn a lot of attention lately. I’m personally glad to see Professor Bollan – the author of the famous Twitter-sentiment-stock-market-predictor paper – back in the fray.
Of course, I’m not objective– my Tradestreaming book was early on the scene to take a look at how research and investors are finding ways to use social media to make better — smarter — investment decisions.
This time, Bollan takes the discussion a step forward in looking for the connecting between social media and investing– by looking at many of the tools investors use to predict future stock market moves.
These indicators, like the Investors Intelligence or Daily Investor Sentiment, measure investor mood. Behavioral finance stresses that factors like emotion and mood impact investor decision making and therefore, markets.
Bollan’s new paper, Predicting Financial Markets: Comparing Survey,News, Twitter and Search Engine Data compares a set of best forecasting tools to see which are most accurate and useful for investors.
How to use social media to invest
Bollan’s findings include:
- Google search helps predict market moves: Bollan found that looking at Google search volume and changes in activity is significantly correlated to closing values in the Dow Jones Industrial Average, trading volume, and volatility. Joey Engelberg studied how to use Google search volume to invest — listen to my interview with him.
- Investor surveys aren’t predictive: Investors Intelligence is not predictive of financial indicators, nor is generalized news sentiment
- Follow Twitter to investing profits: Bollan found that Twitter Investor Sentiment (the percentage of bullish comments) and Twitter Volume of Financial Search Terms of the previous 1-2 days are very statistically significant predictors of daily market returns, while Daily Sentiment Indicator isn’t.
the predictive power of Twitter’s two sentiment indicators outperformed survey sentiment as well as news media analysis. Moreover, we found that before the highly downward movement of DJIA in the end of July and August 2011, Tweet volumes of ﬁnancial terms started to increase several weeks earlier than Google volumes did. This indicates a potential efﬁciency gain of Twitter over GIS
Why is this so interesting?
Well, because studying the predictive power of web information, data — some of which isn’t directly related to investing – is still young. As investors have moved their discussions from message boards to Twitter, there’s a ton of aggregate information that appears to be useful in predicting future stock movements. But this is just the tip of the iceberg.
Read Predicting Financial Markets: Comparing Survey,News, Twitter and Search Engine Data (Huina Mao, Indiana University-Bloomington, Scott Counts, Microsoft Research, and Johan Bollen, Indiana University-Bloomington)