Whispers, walk-downs, jumps, and drift: The complex behavior behind corporate earnings announcements
Whispers, walk-downs, jumps, and drift: The complex behavior behind corporate earnings announcements
Trading on earnings announcements means understanding and mastering a unique and highly complex set of behavioral dynamics. Is it even worth trying?
Earnings are certainly not the only measure of a stock’s value. But for most public companies, they clearly represent the single most important measure of the internal health of the company, its relative valuation, its long-term growth trend, its sensitivity to the economy, and its overall proficiency at increasing value for shareholders.
Since everyone knows this, however, and the markets anticipate earnings information well before announcements, is there any way to gain an edge in the earnings game?
The prediction, analysis, and subsequent market activity surrounding earnings announcements might well be the most consuming activity in corporate finance.
It is also one of the most behaviorally convoluted.
Earnings-related activities represent a highly evolved dynamic that reflects the motives of a wide array of participants ranging from corporate managers, analysts, and brokers to the full gamut of traders, funds, and investors, not to mention untold numbers of high-speed computers. To believe that one can outplay the markets in this activity would require deep insights into the behavior of all participants, a mastery of game theory, and a huge dose of hubris. The significant trend into passive index investing conveniently bypasses the earnings game entirely, but for active managers and investors, earnings cannot simply be ignored.
The dynamics surrounding earnings reports are arguably more about behavior than fundamentals. To appreciate the behavioral complexity of earnings, we can learn from a much simpler game that also revolves around behavior.
Professional poker tournaments have shed new light on the intricate behavioral dance that underlies what many of us formerly viewed as just another game of chance. If that were the case, it would not be possible for players to consistently win, but some do.
They can do so because there are two factors in poker that are entirely behavioral. Players can alter their bets and they can elect to fold, check, call, or raise through each hand. A number of players have demonstrated that they can consistently win in tournament play by mastering the behavioral aspects of the game rather than relying solely on pulling the best cards. This offers valuable insight for those playing in the earnings game as well.
Winning consistently at poker involves understanding and effectively reacting to behavior. The behavior comes in the form of the bet, which is a signal to the other players about the strength of a player’s hand. Knowing this, however, players complicate matters by occasionally faking those signals with a bluff.
A more intuitive form of behavioral observation is therefore necessary to gain an edge in repeated play. That observational skill might entail reading facial expressions, sensing whether a player is perspiring, or taking note of how fast or slow they move their chips to the center of the table. These are among the subtle “tells” that a skilled player might use to win. Importantly, at the highest levels of tournaments, skilled players can also make informed judgments based on the patterns other players have exhibited in their play—trying to assess if a move is consistent with a player’s past style or not. (This, of course, opens multiple levels of gamesmanship.) The challenges of reading tells and determining which ones are real becomes a honed skill. Only a handful of poker pros have mastered it.
The earnings game makes poker seem trivial by comparison. In poker, you are only dealing with a half-dozen other players, all of whom are in plain sight. The earnings game has potentially thousands of players, many of whom you can’t easily see at all. In poker, everyone at the table has the same goal. With earnings, you have a variety of players with different, even conflicting, goals, and you often don’t even know what those are.
In poker, the odds of drawing cards are fixed and known. With earnings, the numbers are known by some in advance and can be, to a degree, manipulated to achieve desired results. In poker, the players are the only ones who place bets. With earnings, you have thousands of observers who place bets and who consequently change the size of the win pool for the players. Moreover, in poker, each round has a discrete beginning and end, while the earnings game begins well before the announcement and continues long afterward.
Corporate managers know what their earnings picture looks like in real time and, while they generally manage their businesses to longer-term strategic objectives, they are mindful of how markets can react to quarterly earnings disclosures. Accordingly, firms use a variety of means to manage expectations around earnings announcements, including issuing guidance, holding earnings calls, and speaking directly to investors through investor-relations channels. Their behavior is governed by their efforts to balance a responsibility for transparency with a desire to keep their stock price on a steadily rising, low-volatility trajectory. And, of course, each company is different in how they approach this.
Corporate behavior becomes more complicated when firms feel the need to manage not just the expectations but the earnings themselves.
It is no secret that companies have, and do exercise, the ability to control both revenues and expenses through tactics such as accrual strategies, price changes, or expense reductions to help achieve a desired earnings result—all within accepted accounting practices.
Academic research has addressed this subject in some depth. A paper by Bhojraj, Hribar, Picconi, and McInnis (2009) claims, “There is growing evidence that managers are willing to sacrifice economic value to meet short-term earnings objectives.” Another by Graham, Harvey, and Rajagopal (2005) reported, “A majority of managers would forgo a project with a positive net present value (NPV) if the project would cause them to fall short of the current quarter consensus forecast.” That challenges earnings watchers to determine not only what the number might be but whether it is “high quality” (i.e., the product of a healthy business) or “low quality” (the product of short-term manipulation).
The former study also found a strong correlation between companies that reduce discretionary expenditures in order to beat short-term forecasts and the degree of equity issuance and insider selling on their stock in the following year. This suggests that short-term discretionary tactics are employed not just to placate investors but to optimize stock price for management’s compensation incentives. Jason Voss, head of consulting firm Active Investment Management Consulting and co-author of the recent book “Return of the Active Manager,” notes that compensation-related earnings behavior may only occur now and then, but that when it does, it can be significant. One scenario Voss tunes in to carefully is when a new CEO is brought in from the outside, as such recruitments are almost always tied to a major stock or option incentive package for the CEO. The details of these incentives are generally disclosed, though buried in the footnotes of the company’s financials when they are initially created. That might be a year or two ahead of when the incentive will have its effect on earnings.
The preannouncement picture gets further complicated by another key player in the earnings game: sell-side analysts, who inject unique behavioral factors into the mix. Companies care about how they are perceived by their shareholders, and they harbor a predictably consistent bias toward maximizing their share price and minimizing volatility to the extent they can proactively influence either of these variables. This involves a delicate balance between maintaining an ever-positive long-term picture and the vicissitudes of short-term operating reality. An excessively rosy long-term picture sets up the potential for short-term disappointment and loss of trust by shareholders.
Meanwhile, analysts have their own unique role in the process, which only partially aligns with the company. On one hand, they behave as impartial third parties who dig deep into company fundamentals to produce insights on behalf of institutional and other investors. But it is also widely known that analysts are biased toward making the companies they cover look favorable as a long-term investment, particularly when the companies are investment-banking clients at the analyst’s firm.
In addition, analysts also don’t want to be so optimistic that short-term results disappoint or cast doubt upon their credibility. The resulting behavior is what Richardson, Teoh, and Wysocki (1999) found as “strong evidence of a switch from upward-biased to downward-biased forecasts of annual earnings as the announcement date approaches.” The industry knows this as the earnings estimate “walk-down.”
A walk-down is the art of transitioning rosy long-term projections into short-term expectations that are more realistic. It recalls Lewis Carroll’s quote by the Mad Hatter, when he said to Alice, “Never jam today, but a case of jam tomorrow.” Corporate managers communicate with the analyst community to nudge them toward lower quarterly estimates that eventually result in a consensus estimate the company can meet or just slightly exceed. This is what causes an unusually large number of estimates that are subsequently beaten by one penny.
From a behavioral perspective, the effect of a one-penny earnings beat is the equivalent of pricing a car at $29,999.99. Despite our denials, our brains react much better to numbers like these than the ones just one cent higher (or one cent lower in the case of earnings). Our loss-averse subconscious will view a one-cent beat as comforting news, while a one-cent miss might raise concerns. (Not to mention that “a miss is a miss” in breaking news headlines). The walk-down is widely practiced, yet many observers still fall prey to typically human reactions when it occurs, often neglecting to look any further for more details.
In addition, analysts are as human as the rest of us, and studies show that they too are plagued with subconscious impulses too hard to resist. For one, they are known to herd. In other words, their estimates cluster more with that of their peers than randomness would suggest. That makes sense when you consider the effect on job security of being wrong when your peers are right.
Amid the goings-on between companies and analysts, the market creates its own earnings anticipation. This tends to revolve around “whisper” numbers, which are perceived by many investors as unofficial leaks from companies to brokers to key clients. This may have been the case in the old days, though that may be more sales legend than truth. Either way, Sarbanes-Oxley put the kibosh on brokers whispering anything to select clients about upcoming earnings. Nonetheless, you can’t stop newsletters or social media, so whispers still tend to circulate and can thus impact stock price. Stocks then move up or down in accordance with not just the rumored consensus estimate itself but how everyone thinks everyone else will react to it.
As in the poker game, gaining an edge in earnings announcements therefore requires a skillful read of the other players in the game—the source company, the analysts, the walk-down, the whispers, and sizing up the resulting market view as either accurate or misguided. All of this behavior takes place before the earnings are even announced. Then the real action begins.
Even the announcements themselves are behaviorally influenced.
Remember the tell about how fast or slow a player moves their poker chips to the center of the table? There is an analogous tell in the earnings world. Eric So, a professor at MIT/Sloan, studied the earnings announcements of companies that either advanced or delayed their announcements from the scheduled date. Dr. So found that there was a statistically meaningful tell in that “advancers” had more positive surprises and “delayers” had more negative surprises in their announcements. Surprise, surprise. Another tell by companies involves the issuing of an earnings announcement separately from disclosing the financial statements that include all of the fine print.
The seconds and minutes following an earnings announcement are then like the big bang, even when released after regular hours trading. The earnings big bang is largely the result of short-term speculative trading influences dominated by automated trading, options unwinding, and speculative trading, which tend to play out before most people can take out their magnifiers for a deeper look. The immediate post-announcement behavior is often so chaotic, it will take a stock in the opposite direction or even on a multi-directional flip-flop before establishing a more rational reaction to its earnings.
It is here, though, after earnings are disclosed, that academics and astute investment managers have discovered at least one viable behavioral pattern to exploit. The action is described as “underreaction-overreaction,” and it revolves around a recorded phenomenon called “price drift.” The underlying behavior behind price drift is somewhat intuitive and relates to momentum. It says that when new information on a stock is made public (not limited to earnings), a stock will have an immediate reaction (a “jump” or “dive”), in keeping with efficient market theory. It has been shown, though, that many stocks will continue to drift further in the same direction for days or weeks more, often to a point beyond the level justified by the new information.
While some investment managers have successfully established an edge by exploiting this behavior, one should certainly not expect random purchases of stocks with earnings beats to yield a winning trade each time. Such behavioral anomalies are typically characterized by slim statistical margins and must be implemented consistently to bear fruit.
Another behavioral anomaly surrounding earnings announcements can be found in options. Prior to announcements, options in the nearest expiration on the underlying stock will typically exhibit a sharply higher level of implied volatility, which manifests itself in abnormally higher prices, given the current stock price, time remaining, and historic volatility of the stock. This happens regardless of which direction the market expects the stock will go after the announcement. The resulting “volatility skew” can frequently see implied volatilities of near-term options jump by 50%–100% or more, causing the price of out-of-the-money options to rise by an equal amount. As soon as earnings are announced, the stock makes its initial move and the implied volatilities collapse back to baseline. While this often results in unforeseen losses for those with long options positions who are unaware of the phenomenon, more astute options players can arbitrage the skew against longer-dated options on the same stock with more normal implied volatilities to profit from the chaos.
The behavioral crosscurrents surrounding earnings announcements are overwhelming. It would make little sense for a manager to divert their focus away from long-term strategy, risk management, and portfolio optimization in order to get embroiled in an earnings game they can’t consistently win.
There is simply no clear path to a high-probability strategy surrounding earnings trades. That said, however, there may certainly be specific situations where deep knowledge of the company, the analysts, and the market offer a one-time advantage in either accumulating or decreasing a stock position. With some additional analysis, it might also prove worthwhile to incorporate price drift into a planned trade as well, or to write options against a position to suppress earnings-related volatility and take advantage of the skew.
It also may be worth remembering the poker expression, “If you can’t tell who the sucker is at a poker table—you are the sucker.”
Richard Lehman is the founder/CEO of Alt Investing 2.0 and an adjunct finance professor at both UC Berkeley Extension and UCLA Extension. He specializes in behavioral finance and alternative investments, and has authored three books. He has more than 30 years of experience in financial services, working for major Wall Street firms, banks, and financial-data companies.