The Nuances of “Buy the Rumor, Sell the News”

by Erin Peterson (7/23/16)
Erin Peterson


Traders are often urged to follow the age-old advice to “buy the rumor, sell the news.” In other words, before a big announcement on a stock, trade based on what the likely news is. Once information has officially been released, reverse the trade, reaping short-term profits.

Recently, a trio of researchers, including Pamela Moulton, associate professor of finance in the School of Hotel Administration, and Roni Michaely, Rudd Family Professor of Management and professor of finance at Johnson, investigated the subtleties of this adage in one very specific case. The pair, along with Ohad Kadan,a researcher at Washington University at Saint Louis, studied the patterns of buying and selling by different categories of traders in the days just before, and just after, sell-side analysts publicly issued buy or sell recommendations for specific stocks. “[The public announcement of these analyst recommendations is] a time when important information comes out, and some — but not all — traders may know something about it in advance,” Moulton says. “Who’s benefiting? And who is potentially being harmed?”

It turns out that traders with advance information typically do “buy the rumor” — but only a portion of them “sell the news.” And those most affected by these short-term trades were more likely to be institutional traders — not individual investors.

The research team was able to study these questions thanks to a unique slice of proprietary data from the New York Stock Exchange (NYSE) that was available to Moulton, a former academic economist for the exchange.

The NYSE data included 11 years of daily buy and sell volumes, along with information about whether buyers and sellers were classified as retail investors, market-makers, or institutions. (Typically, public datasets do not reveal the type of trader involved in transactions.)

Roni Michaely

Roni Michaely

A few subcategories were included within these larger groups. For example, institutional trades were further subdivided into three categories. The first was program trades (often used by index funds): orders for multiple stocks, unmotivated by news of any specific stock. The other categories were proprietary trades, in which NYSE member firms trade for their own account, and agency trades, in which these same firms enter orders for their customers. The final two categories of trades represent news-driven traders who are most likely to have brief informational advantages over other investors.

As the team dug into the data, they found compelling trends. Four days before an analyst announced an upgrade on an individual stock, proprietary and agency institutional purchases spiked. On the day of the upgrade announcement, only proprietary traders sold a significant portion of their earlier purchases to snare a quick profit. Agency traders appear to hang on to the stocks beyond the day of the announcement, benefiting from the early information, but not immediately selling to net a profit.

Moulton, Michaely, and Kadan mined multiple insights from this discovery.

First, the gains for proprietary traders were significant: The average return from “buying the rumor” a few days before an analyst upgrade and then “selling the news” on upgrade day (or doing the reverse set of transactions for a downgrade) was 1.71 percent. Annualized, the numbers jump to a remarkable 106.9 percent.

Second, the timing of the preannouncement trades was consistent with the idea that proprietary and agency traders were being “tipped” about analyst recommendations several days before they were publicly announced.

Moulton is careful to outline that this “information leakage” in the form of tipping is just one of multiple possibilities. “Certainly, many institutions employ their own analysts. It’s possible that at the same time analysts on Wall Street are changing their minds from neutral to buy, for example, a portfolio manager’s own analyst, who is looking at the same public sources of information, might also be making the same judgment,” she says. “But it’s also possible that someone at a Wall Street investment bank might tell their best institutional clients when they’re planning to upgrade a recommendation on a specific stock. Both things are possible, and, at least during our sample period, neither was illegal.”

Either way, the fact remains that both proprietary and agency traders are benefiting from this early information. But who’s on the losing side?

It would be easy to imagine that retail investors — individuals who don’t have early access to the information — are typically losing out in these situations. However, the team’s research suggests that this is not the case. “We find that individuals are not the major seller to [informed traders] before the recommendations are announced,” says Michaely. “Instead, we find that the sellers are those traders who just trade on indices and don't pay attention to particular stocks.” In fact, these program investors are buying and selling in a nearly mirror image to their informed counterparts.

Pamela Moulton

Pamela Moulton

Moulton suggests that the patterns she and her co-authors discovered aren’t necessarily worrisome. Instead, they may be indicative of profoundly different investing strategies. Proprietary and agency traders are sensitive to short-term moves and trade accordingly, but program traders who seek to make profits over longer time frames are less likely to be interested in taking advantage of the market’s daily hiccups. “For those who are interested in the long-term [time horizon], knowing that a stock’s price might go up 50 cents tomorrow isn’t relevant,” explains Moulton. “They’re going to hold positions for months or years, and they don’t care about upgrade recommendations before they’re publicly released. They understand that they’re not always going to be on the ‘right’ side of a specific trade, but their long-term goals are important enough that it still makes sense for them to [behave this way]. To suggest they should change would be like asking a leopard to change its spots.”

The research team also conducted several placebo tests to verify that their results were most likely the result of early information acquisition. For example, the team studied buying and selling patterns prior to earnings announcements. Because of insider trading laws and regulations, it is very unlikely that anyone would have access to earnings information in advance. As expected, the researchers did not find the same “buy the rumor, sell the news” behavior as they did around analyst recommendations.

The team’s findings may be relevant to other scenarios in which investors have short-term informational advantages, such as high-speed traders who pay for news feeds that give them information fractions of a second faster than other traders, for example. “The effects we document around analyst recommendation changes occur over a period of days,” says Moulton. “But one could imagine the same thing happening over shorter time periods of hours, minutes, or even seconds.”

To learn more, access the working paper here: Trading in the Presence of Short-Lived Private Information: Evidence from Analyst Recommendation Changes.