Earnings Surprise: Why Stocks Keep Moving After Earnings

What Is an Earnings Surprise?

An earnings surprise occurs when a company's reported earnings per share (EPS) differs from the consensus estimate compiled from sell-side analyst forecasts. If actual EPS exceeds the consensus, it is a positive surprise, commonly called a "beat." If actual EPS falls below consensus, it is a negative surprise, or a "miss."

This sounds simple, but the mechanism through which consensus estimates form, the ways surprises are measured, and the market's reaction to those surprises are anything but straightforward. Understanding earnings surprises is essential for any serious equity trader because they drive some of the largest and most persistent stock price movements in the market.

The central puzzle is this: if markets are efficient and all available information is already priced in, why do stocks continue to move in the direction of the surprise for weeks or even months after the announcement? This phenomenon, known as post-earnings announcement drift (PEAD), has been documented and replicated in dozens of academic studies since the late 1960s. It remains one of the most robust anomalies in financial economics.

How Consensus Estimates Are Built

Before we can talk about surprises, we need to understand what stocks are being surprised against. The consensus estimate is the average (or median) EPS forecast from sell-side analysts who cover a given stock. These estimates are aggregated by several major data providers:

• I/B/E/S (Institutional Brokers' Estimate System) — originally founded by Lynch, Jones & Ryan in 1976, now owned by LSEG (formerly Refinitiv). I/B/E/S is the most widely used source in academic research on earnings surprises and PEAD.
• FactSet — aggregates estimates from hundreds of brokerages and provides its own consensus calculations. Widely used by institutional investors and hedge funds.
• Bloomberg — the Bloomberg consensus is compiled from analyst submissions through the Bloomberg terminal and is the default reference for many trading desks.

Each provider surveys sell-side analysts at major brokerages — firms like Goldman Sachs, Morgan Stanley, JPMorgan, and others — who publish earnings models for the companies they cover. The number of analysts covering a stock varies enormously: large-cap stocks like Apple or Microsoft might have 30-40 analysts submitting estimates, while a small-cap company might have only 2-3, and many micro-caps have none at all.

This coverage gap matters significantly for surprise dynamics. Stocks with thin analyst coverage tend to produce larger surprises and stronger post-announcement drift, because there is less pre-announcement information impounded in the price.

The Whisper Number

Beyond the official consensus, there exists an informal "whisper number" — the buy-side's real expectation, which may differ from the published sell-side consensus. If the whisper number is higher than the consensus (because buy-side analysts think the company will beat by more than the street expects), then merely meeting the consensus can actually disappoint the market. This is why you sometimes see stocks sell off on what looks like a beat — the actual number beat the published consensus but missed what informed investors were truly expecting.

The Founding Research: Ball and Brown (1968)

The academic study of earnings surprises began with Ray Ball and Philip Brown, who published "An Empirical Evaluation of Accounting Income Numbers" in the Journal of Accounting Research (Vol. 6, No. 2, pp. 159-178) in 1968. This paper is widely regarded as the founding work of PEAD research and one of the most influential papers in accounting history.

Ball and Brown studied 261 NYSE-listed firms over the period 1946-1966. They classified each earnings announcement as either "good news" (actual earnings above expected) or "bad news" (actual earnings below expected), using the prior year's earnings as the expected value. Their key findings were:

• Stock prices began moving in the direction of the earnings news well before the announcement, suggesting that the market partially anticipated the information.
• However, a meaningful portion of the price adjustment occurred after the announcement date. Prices continued drifting in the direction of the surprise for at least two months.
• This post-announcement drift implied that the market was slow to fully incorporate the information contained in earnings reports.

This finding was remarkable because it directly challenged the emerging efficient market hypothesis. If earnings are public information after the announcement, prices should adjust instantly. The fact that they did not — that there was money left on the table for weeks afterward — suggested a systematic inefficiency.

Beat, Miss, and the Magnitude Effect

Not all earnings surprises are created equal. The magnitude of the surprise matters enormously for the subsequent stock price reaction.

Academic research has consistently shown that the relationship between surprise size and drift is approximately proportional up to a point. Bernard and Thomas (1989), in their landmark study "Post-Earnings-Announcement Drift: Delayed Price Response or Risk Premium?" published in the Journal of Accounting Research (Vol. 27, Supplement, pp. 1-36), sorted stocks into deciles by the size of their earnings surprise. They found that the hedge portfolio (long the top surprise decile, short the bottom decile) generated approximately 4.2% over the 60 trading days following the announcement, or about 25% annualized.

The practical implication is clear: traders should focus on the magnitude of the surprise, not just whether the company beat or missed. A 1-cent beat on a $2.00 consensus is a very different signal from a 20-cent beat on the same consensus.

Measuring the Surprise: SUE

The standard academic measure of earnings surprise is Standardized Unexpected Earnings (SUE):

SUE normalizes the raw surprise by the historical volatility of that company's forecast errors. This is important because a 5-cent surprise means something very different for a stock that routinely surprises by 10 cents versus one that has historically been within 1 cent of estimates. By standardizing, SUE makes surprises comparable across companies with different earnings volatility profiles.

An alternative measure is the simple analyst forecast error:

This percentage-based measure is simpler and more widely used by practitioners, though it has the disadvantage of blowing up when the consensus estimate is near zero.

Revenue Surprise vs. EPS Surprise

Historically, the financial media and most traders focused almost exclusively on EPS surprises. But over the past two decades, revenue surprises have become increasingly important — and for good reason.

EPS can be managed through a variety of mechanisms that do not reflect genuine business performance: share buybacks reduce the denominator, one-time gains inflate the numerator, and aggressive accounting choices (revenue recognition timing, depreciation schedules, restructuring charges) can shift earnings between quarters. Revenue, by contrast, is much harder to manipulate. You either sold the goods and services or you did not.

Research has shown that revenue surprises carry incremental information beyond EPS surprises. Jegadeesh and Livnat (2006), in "Revenue Surprises and Stock Returns" published in the Journal of Accounting and Economics (Vol. 41, No. 1-2, pp. 147-171), found that revenue surprises predict future stock returns even after controlling for earnings surprises. Stocks that beat on both revenue and EPS show stronger drift than those that beat on EPS alone.

This hierarchy matters for implementation. A trader who only screens for EPS beats will capture many signals that are low-quality — companies that financially engineered their way to a beat without genuine top-line growth. Incorporating revenue surprise as a filter significantly improves signal quality.

The Power of Guidance

Perhaps the most underappreciated component of an earnings announcement is forward guidance — the company's own projection for the next quarter or full year. Academic and practitioner research both confirm that guidance has a powerful effect on post-announcement drift.

Companies that beat current-quarter estimates and raise their forward guidance see substantially stronger drift than those that merely beat. The logic is intuitive: a beat tells you the past quarter was better than expected, but raised guidance tells you the company's management — who have the best information about the business — believes the improvement is sustainable.

Conversely, a company that beats the current quarter but lowers guidance often sees its stock sell off despite the "beat." The market is forward-looking, and guidance revision is one of the strongest signals of future earnings trajectory.

Management guidance vs. consensus creates its own layer of surprise. When a company guides Q2 EPS to $1.50-$1.60 and the consensus settles at $1.55, the actual "bar" the company needs to clear is not just the consensus but also the upper end of its own guidance range. Beating both the consensus and the company's own guidance is the strongest possible signal.

The Surprise Cascade: Cross-Firm Predictability

One of the most fascinating aspects of earnings surprises is that they are not independent across firms. An earnings surprise at one company can predict drift at related companies that have not yet reported. This phenomenon was first documented by Foster, Olsen, and Shevlin (1984) in "Earnings Releases, Anomalies, and the Behavior of Security Returns" published in The Accounting Review (Vol. 59, No. 4, pp. 574-603).

The mechanism is straightforward: companies in the same industry face similar macroeconomic conditions, supply chain dynamics, and demand patterns. If a semiconductor company reports a blowout quarter driven by strong data center demand, it is informative about what other semiconductor companies will report when their earnings come in weeks later.

This creates a surprise cascade where early reporters in an earnings season effectively preview what late reporters will announce. Traders who recognize this pattern can position in late-reporting stocks based on the surprises of early reporters in the same sector.

The cross-firm predictability of earnings surprises is strongest within industries but also extends to customer-supplier relationships, geographic peers, and even factor exposures. A positive surprise from a major retailer is informative about consumer spending, which in turn predicts surprises at consumer goods companies, logistics providers, and payment processors.

Post-Earnings Announcement Drift: The Continuation Effect

The single most important phenomenon associated with earnings surprises is post-earnings announcement drift (PEAD). First documented by Ball and Brown (1968) and rigorously quantified by Bernard and Thomas (1989), PEAD is the tendency for stocks to continue drifting in the direction of the earnings surprise for 60-90 days after the announcement.

PEAD has been replicated in dozens of studies across multiple countries and time periods. It is present in U.S., European, and Asian equity markets. It persists in both large-cap and small-cap stocks, though it is stronger in smaller, less-followed names. It survives controls for risk factors including beta, size, value, and momentum.

Why does PEAD exist? The leading explanation is that investors underreact to the information in earnings announcements. Bernard and Thomas (1990), in a follow-up study "Evidence that Stock Prices Do Not Fully Reflect the Implications of Current Earnings for Future Earnings" published in the Journal of Accounting and Economics (Vol. 13, No. 4, pp. 305-340), showed that the market behaves as though investors do not fully adjust their expectations to account for the autocorrelation in quarterly earnings. When a company beats estimates in Q1, it is more likely than chance to beat again in Q2, but the market prices this serial correlation too slowly.

Other contributing factors include:

• Institutional constraints: Many institutional investors cannot trade immediately after earnings due to compliance review periods, position sizing rules, or fund mandate restrictions.
• Analyst inertia: Sell-side analysts are slow to fully revise their models after earnings surprises. Studies show that post-earnings estimate revisions predict further drift, suggesting that even professional forecasters underreact.
• Behavioral biases: Anchoring (investors anchor to pre-announcement expectations), conservatism (updating beliefs too slowly), and limited attention (investors cannot process all information simultaneously during a busy earnings season).
• Liquidity and transaction costs: The stocks with the strongest PEAD signals are often small-caps with wide spreads, making it expensive to fully exploit the anomaly and preventing arbitrage from eliminating it.

Measuring Surprise: A Practitioner's Toolkit

For traders looking to implement earnings surprise strategies, several measurement approaches are available, each with different strengths:

In practice, the most robust approach combines multiple measures. A stock that has a high SUE, gaps up on heavy volume, and sees the company raise guidance is a much stronger signal than one that merely beats consensus by a small margin.

The Earnings Gap as a Proxy

For stocks without analyst coverage (and therefore no consensus estimate to surprise against), the earnings-day gap serves as a useful proxy. If a stock opens 5% higher on 3x normal volume on the morning after an earnings report, the market is telling you the news was substantially better than expected, even if there is no formal consensus estimate to compare against.

Alpha Suite uses both formal SUE computation (when analyst estimates are available via market data feeds) and gap-volume proxies (for broader coverage across the full universe of SEC filers). This dual approach ensures that surprise signals are captured even for the many small-cap and micro-cap stocks that lack meaningful analyst coverage — precisely the universe where PEAD effects tend to be strongest.

Management Guidance vs. Consensus: A Separate Layer

An increasingly important dimension of earnings surprise analysis is the gap between management guidance and analyst consensus. When these two diverge, it creates a distinct source of information.

Consider a company that guided Q2 EPS to $1.50. Analysts, perhaps skeptical, set their consensus at $1.45. When the company reports $1.52, the "surprise" relative to consensus is +$0.07 (a 4.8% beat), but relative to the company's own guidance, it is only +$0.02. The market's reaction will depend on which benchmark investors were truly anchored to.

Companies that consistently beat both their own guidance and the street's consensus develop a reputation for "sandbagging" — deliberately setting low guidance to ensure they beat it. The market learns this pattern, and over time, the drift following beats from habitual sandbaggers diminishes. The strongest drift comes from companies with no history of consistent beats that suddenly report a large surprise, suggesting genuine unexpected improvement rather than expectations management.

Practical Considerations for Surprise-Based Trading

Implementing an earnings surprise strategy requires attention to several practical realities:

• Timing: The first 15-30 minutes after the market opens following an earnings report are typically the most volatile and the most expensive to trade. The PEAD literature suggests that most of the anomalous drift occurs over days and weeks, not minutes. Patient entry (waiting for the initial volatility to settle) generally improves execution quality.
• Position sizing: Earnings can move stocks 5-20% in either direction. Any individual earnings event is uncertain, but the statistical edge comes from trading many events over time. Position sizing should reflect this — no single earnings trade should risk more than 1-2% of the portfolio.
• Sector neutrality: During earnings season, surprises tend to cluster by sector. If technology companies are broadly beating, a strategy that goes long every tech beat will accumulate unintended sector concentration. Neutralizing sector exposure (or at least being aware of it) reduces the risk of correlated drawdowns.
• Stale estimates: Consensus estimates can become stale if analysts have not updated their models recently. Before trading on a "surprise," verify that the consensus reflects recent analyst activity, not stale estimates from months ago.

The Insider Connection

One of the most powerful amplifiers of earnings surprise signals is insider trading activity around the announcement. Lakonishok and Lee (2001), in "Are Insider Trades Informative?" published in the Review of Financial Studies (Vol. 14, No. 1, pp. 79-111), documented that insider purchases predict future returns even after controlling for earnings surprises.

The convergence is intuitive: if a stock reports a positive earnings surprise and the company's own officers and directors were buying shares in the weeks preceding the announcement (as reported on SEC Form 4 filings), the two signals are reinforcing. The earnings data tells you the company performed better than expected; the insider buying tells you that people with privileged information about the company's trajectory were willing to put their own money at risk.

Alpha Suite's signal engine evaluates every earnings surprise in the context of surrounding insider activity. When a stock with recent cluster buying by multiple insiders also reports a positive earnings surprise, the combined conviction score is substantially higher than either signal alone. This cross-referencing of public earnings data with insider transaction patterns is a core part of the platform's quantitative framework.

Earnings Surprises in the Current Market

The nature of earnings surprises has evolved significantly over the past two decades. Several structural changes have reshaped the landscape:

• The beat rate has risen: In recent years, approximately 75-80% of S&P 500 companies beat consensus EPS estimates in any given quarter. This is partly because analysts have become more conservative (they would rather be surprised positively than negatively) and partly because companies actively manage expectations through guidance. The result is that merely beating consensus is no longer particularly informative — the market has learned to expect beats.
• The bar has shifted: Because beats are expected, the market now penalizes companies that beat by less than expected. A company that beats consensus by 2% when the market was expecting a 5% beat can actually see its stock decline. This "expectations of expectations" dynamic makes earnings trading more complex than simple beat/miss binary analysis.
• Alternative data: Credit card data, satellite imagery, web traffic, and app download metrics now allow sophisticated investors to estimate earnings before the announcement. This pre-announcement information reduces the surprise element for widely followed stocks, which may be one reason why PEAD has diminished (but not disappeared) in large caps.
• Speed of information processing: Algorithmic trading systems now parse earnings releases within milliseconds and execute trades before human traders can read the headline. The initial price reaction to earnings is faster and more efficient than it was when PEAD was first documented. But the subsequent drift — the slow incorporation of the earnings information into analyst models, institutional portfolios, and market expectations — still takes weeks.

The Bottom Line

Earnings surprises are one of the most important drivers of stock price movements. The fundamental insight, first documented by Ball and Brown in 1968 in the Journal of Accounting Research, remains valid today: the market is slow to fully process earnings information, and prices continue to drift in the direction of the surprise for weeks after the announcement.

The key to exploiting this phenomenon is understanding that not all surprises are equal. The magnitude of the surprise matters more than the direction alone. Revenue surprises carry incremental information beyond EPS surprises. Forward guidance amplifies or dampens the drift. Surprises cascade across related companies within an industry. And the combination of earnings surprises with insider buying patterns produces the strongest signals of all.

For traders willing to do the work of measuring surprises carefully, combining multiple information sources, and managing risk appropriately, earnings surprises remain one of the most reliable edges in equity markets.