Can traders extract an edge from sentiments expressed about stocks via public Internet message boards? In the March 2014 draft of their paper entitled “Investor Sentiment from Internet Message Postings and Predictability of Stock Returns”, Soon-Ho Kim and Dongcheol Kim investigate whether combined investor sentiment as expressed in Yahoo!Finance stock message boards predicts returns, volatilities and trading volumes of associated stocks. They measure sentiment of posters in two ways: (1) standardized sentiments that some posters explicitly include in their messages (Strong Buy, Buy, Hold, Sell or Strong Sell); and, (2) sentiment from all posts inferred by a learning algorithm trained on the subset with explicit standardized sentiments. They group standardized sentiments as buy (Strong Buy or Buy) and sell (Strong Sell or Sell). They ignore the Hold sentiment. They combine posted sentiments in two ways using either the normalized number of buys minus number of sells or the logarithm of an adjusted ratio of buys to sells. When an author posts multiple messages for the same stock within a measurement interval, they use only the most recent message. They analyze predictive power of combined sentiment at daily, weekly and monthly horizons. Using sentiments from more than 32 million messages for 91 stocks posted on Yahoo!Finance message boards by over a half million authors, and contemporaneous daily prices and trading volumes for these stocks, during January 2005 through December 2010, they find that:
- About 26% of messages include the standardized sentiment, with strong proponents dominating. Of this subset, Strong Buy and Buy constitute about 60% and 12%, respectively. Strong Sell and Sell constitute about 19% and 2%, respectively. Hold constitutes about 7%.
- There is no evidence that combined sentiment predicts stock returns at the individual stock level or in aggregate at daily, weekly or monthly horizons. Combined sentiment does exhibit positive correlation with past returns. More specifically:
- There is no evidence that combined sentiment predicts earnings surprises and stock returns around quarterly earnings announcements.
- Although message board activity increases before extreme stock price changes, combined sentiment does not predict the direction of the change.
- There is no difference in (lack of) ability to predict the direction of stock price changes by poster age, poster gender or length of message.
- After controlling for serial correlation and past return, there is no evidence that combined sentiment predicts stock return volatility or trading volume at daily, weekly or monthly horizons.
In summary, evidence does not support belief that traders can gain any edge by examining sentiments expressed about stocks in public message boards.
There may be private forums with informed investors that exhibit predictive sentiments. (See “Professional Investor Groups Sharing Value (or Moving Markets)”).
