Jan 4, 2018
Traders Are Surprisingly Slow to Respond to Off-hours Earnings Announcements
It can take days for investors to react, creating a potentially lucrative strategy for some.
On Wall Street, “business hours” aren’t quite what they seem.
The opening bell of the New York Stock Exchange rings at 9:30 a.m. Eastern, and the closing bell sounds at 4 in the afternoon—yet 95 percent of publicly traded companies announce their quarterly earnings outside of those official hours.
So how do traders respond to off-hour announcements?
“I have three kids,” says Matt Lyle, an assistant professor of accounting and information management at the Kellogg School. “And there are times when I’m with my kids and, if I get an email that’s important for work, I can’t really concentrate on it until I’m not distracted by other things.”
Lyle wondered whether the same was true for traders trying to process earnings information. In a new study, Lyle and coauthors indeed find that the timing of earnings announcements has a strong impact on how investors incorporate them into their investment decisions.
Specifically, the further from business hours earnings are announced, the slower the investors’ reaction, particularly for announcements made before the opening bell. This delay is not a mere matter of hours. To the authors’ surprise, it takes a full four days for investors to fully integrate the new information into their portfolios when it is announced early in the morning.
This puzzling result could prove lucrative for investors: the authors find that options traders stand to turn a steady profit if they incorporate these results into their trading strategies.
The Timing of Earnings Announcements Matters
Lyle and his colleagues—Kellogg doctoral student Christopher Rigsby, Andrew Stephan of the University of Colorado, and Teri Lombardi Yohn of Indiana University—started by exploring how investors responded to early morning versus evening announcements.
Other research has shown that when investors are distracted—on Fridays, on days with lots of other earnings announcements, or during March Madness, for instance—they are slow to react to important financial information.
With that in mind, Lyle and his coauthors hypothesized that traders would respond more quickly to post-close announcements (those released between 4 p.m. and midnight Eastern) than to pre-open announcements (those released between midnight and 9:30 a.m.). Their thinking: since many traders (especially those in further-west time zones) are still at work when the market closes, they would more likely have time to read and think about the announcement before after-work activities and family obligations took over. For the same reason, the authors suspected that the closer to business hours a company announced earnings, the quicker investors would react.
The researchers procured a dataset from an investment-research firm containing a large sample of quarterly earnings announcements from 2006 to 2014, each time-stamped down to the second. They sorted them into two categories: “pre-open” and “post-close.”
“Part of research is to document things that do not conform with the way we think the world works.”
To measure investor response, the researchers focused on volatility. Whenever a company releases its quarterly earnings, there is a certain amount of extra fluctuation in its stock’s performance, as traders buy or sell based on the new information. By measuring how long this volatility lingered after an announcement, the researchers could estimate how quickly investors reacted.
Their research confirmed their hunch—stock prices stabilized faster after a post-close announcement than after a pre-open announcement. Additionally, the further from NYSE hours an announcement came, morning or evening, the longer its stock remained in flux.
To Lyle, this suggested that investors needed extra time to consider and act on intel they received outside of work hours.
But he could not yet be sure—there was another potential explanation the researchers still needed to account for.
Maybe firms making early morning announcements were also more likely to have something to hide, the authors theorized. If so, this difference in volatility could be a reflection of what was in the announcements, rather than the time of day. (And in fact, conventional wisdom says that firms with bad news will make their announcements late at night, or on a Friday, to lessen the impact on their stock.)
But contrary to this wisdom, the team found that firms were highly consistent in when they made their announcements.
“For instance,” Lyle says, “Apple announces at 4:30, within thirty minutes of the closing bell, at the same time basically every time. GE announces at 6:30 in the morning, basically all the time. So while there’s no doubt that the firms do play a little bit of strategic games, it’s very hard to attribute our results to that.”
The authors ran statistical tests, which confirmed that nothing in the content of the earnings announcements could account for differences in stock volatility. Nor could characteristics of the firms, like its size, profitability, or time zone. The best explanation was the timing.
The Lag Lingers
Still, in an environment where computers can trade stocks in nanoseconds, how long could this added volatility linger? The researchers predicted that, even when announcements were made off-hours, investors would likely take at most a few hours to respond to them.
“That is not what we find,” Lyle says.
When they compared how stocks were faring a full day after earnings announcements, they found that companies that announced pre-open were still fluctuating more than post-close announcers. Many investors, it seemed, were just getting around to the inconveniently timed earnings report a day later.
The researchers looked further and further out. Two days after the initial announcement, the gap in volatility persisted. Three days after, the same story. It took a full four days before pre-open stock prices had stabilized as much as post-close stock prices.
This confirmed the team’s hypothesis that reactions to early morning announcements would be delayed. But why were they delayed for a full four days? Lyle is still not sure.
“I mean, it’s shocking!” he says. “It’s hard to generate a theory that will lead to this type of prediction. Which is also one of the reasons why we think the study’s important—part of research is to document things that do not conform with the way we think the world works.”
A Lucrative Finding
Although Lyle was shocked by this finding, he suspected there was one group of folks who might not be: options traders.
“We said, ‘Given that options traders care so much about volatility, if anyone’s gonna know about this result, it would probably be the options market,’” he says.
Options are financial instruments that traders can use to speculate on the performance of stocks. If an investor buys an option, she is essentially betting that a company’s stock will hit a certain price within a given time period. The more a stock’s price fluctuates, the higher the odds that it will hit that given stock price in any time period, and so the higher price that option will be.
If options traders indeed already knew that companies that announced earnings in the morning tended to experience more prolonged volatility, then options prices should adjust accordingly.
“And if that was the case, then they couldn’t make money off of it,” Lyle explains.
The authors devised two trading strategies that would profit when pre-open companies’ stock prices fluctuated more than post-close companies’ stock prices. They used a database of options prices and stock performance to model how these strategies would have performed had they been applied to real trades in the past.
Surprisingly, the strategies delivered large and significant returns—which indicated that the options market was not, in fact, aware of this quirk.
Having succeeded on paper, these strategies are now being tested in the real marketplace. Lyle shared his results with a pair of Northwestern alumni he knows who work as traders. So far, they have independently confirmed Lyle’s results and are currently devising formal trading strategies to earn traders a profit.
The study’s unexpected results have left Lyle wanting to look more closely at a related issue: why firms choose to announce earnings when they do.
Once upon a time, many announcements happened during the business day. But by the late 1990s, after-hours announcements had become commonplace. Analysts have offered competing theories about what brought about this change. Some suspect it had to do with new financial-disclosure legislation, while others tie it to the advent of electronic trading.
“Nobody’s been able to give us a clear answer,” Lyle says. “I think it’s quite curious. And maybe understanding that will also help us better understand the results that we’ve documented now.”
About the Writer
Jake J. Smith is a writer and radio producer in Chicago.
About the Research
Lyle, Matthew R., Christopher Rigsby, Andrew Stephen, and Teri Lombardi Yohn. 2017. “The Timing of Earnings Announcements and Volatility.” Working paper.