By Lia Samson
Eller MBA ’13
The question of whether retail stock buyers can predict stock market prices is difficult to address. By their nature, retail buyers’ actions are challenging to aggregate: these buyers are making decisions and trading stock as individuals rather than as a part of a corporation or fund.
“People have been asking questions for some time about whether and why retail investors in aggregate can predict future stock returns,” explained Assistant Professor of Finance Eric Kelley, “and there is not general agreement on the answers to these questions.”
In a paper published in the Journal of Finance, Kelley and coauthor Paul Tetlock of Columbia University nevertheless identify a variety of trends that retail buying is able to predict.
Using a proprietary data set containing over 200 million orders and accounts for $2.6 trillion in trades from 2003 to 2007, Kelley and Tetlock discovered three main trends. First, both passive and aggressive retail orders can predict returns over the subsequent month; second, aggressive retail orders can predict the occurrence of positive future news; and third, passive orders are related to prior and contemporaneous negative returns.
In this study, Kelley said, the most important predictor of a given stock’s future price by retail traders is net buying (when shares are bought more than sold) by passive and aggressive orders. The latter are trades made immediately, and the former are trades that occur if and when a stock reaches a certain price.
The authors demonstrate that net buying both types of orders of a given stock was positively related to a future increase in the stock’s price. The correlation is strongest in the first week after the time period for which the trades is assessed, but the relationship never becomes negative, even after a long period of time. This means retail buyers provide valuable market information, if only their aggregate actions could be easily assessed.
Aggressive orders from retail buyers may provide more information than the stock price by implying positive health of the overall firm. Kelley found evidence of this when he analyzed the tone of all articles related to a given firm on the Dow Jones newswires. By comparing the occurrences of net buying from aggressive retail orders of a company’s stock to the tone of articles about that company, he and Tetlock found the tone to generally be positive.
When net buying occurs with passive orders, the stock has generally provided negative returns in the recent past and at the time of the orders. Since this stock’s price is subsequently predicted to improve, this suggests the stock has been undervalued and the price is being corrected.
“In general, when predictability in stock returns is first documented, widespread knowledge of its existence could result in the pattern going away,” said Kelley. “However, since buying and selling by retail traders is not publicly observable in real time, that wouldn’t necessarily be the case for our findings.”
Learn more about the dynamic faculty and research of the Department of Finance.