Post-Earnings Announcement Drift: The Edge Academics Documented but Traders Miss

In 1968, two accounting researchers named Ray Ball and Philip Brown published a study that, without announcing itself as such, broke the efficient market hypothesis. They showed that stock prices responded to earnings announcements — but the response was not instantaneous. Prices kept moving in the direction of the surprise for weeks afterward. Markets were absorbing the information slowly, inefficiently, and in a direction you could measure in advance.

Two decades later, Victor Bernard and Jacob Thomas formalized the phenomenon into what is now the canonical literature on post-earnings announcement drift, or PEAD. Their papers in the Journal of Accounting Research (1989) and Journal of Accounting & Economics (1990) did something rare in financial economics: they documented an anomaly so robust, so replicable, and so persistent across decades that it survived every rational explanation academics threw at it. It still survives.

For the momentum trader, PEAD is not an intellectual curiosity. It is one of the most well-documented statistical tailwinds in equity markets — a pre-identified pool of stocks with a measurable probabilistic edge over the next sixty to ninety days. The overlap between PEAD candidates and the stocks most likely to form high-tight bases and break out of sound volatility contractions is not coincidental. It is structural. The same institutional slow-reaction that drives the academic drift drives the chart pattern.

What the Research Actually Shows

The core finding is elegantly simple. Sort all earnings announcements by the magnitude of their surprise relative to consensus expectations. Put the top decile — the biggest positive surprises — in one bucket and the bottom decile — the biggest misses — in another. Hold each bucket for the sixty trading days after the announcement. The top decile outperforms the bottom decile by a margin that academics have consistently measured in the range of six to ten percentage points over that roughly three-month window.

This is not a one-decade phenomenon that vanished once arbitrageurs discovered it. Bernard and Thomas found it persisted from the 1970s through the late 1980s. Subsequent researchers — including Chan, Jegadeesh, and Lakonishok in 1996, and dozens of studies since — have found it persists through the 2000s and 2010s, across market regimes, across country borders, and across different specifications of what counts as an earnings surprise. It has been called, with only minor hyperbole, "the granddaddy of all market anomalies."

Three empirical features of PEAD matter most for a structural trader. First, the drift is monotonic across surprise deciles. The top decile drifts up the most, the ninth decile drifts up less, the fifth decile barely drifts, and the drift gradually inverts into the bottom deciles. This is a clean dose-response curve, not a lottery effect driven by a few outliers. Second, the drift is roughly symmetric. Large negative surprises generate drift downward of comparable magnitude to the upward drift from large positive surprises. This matters for short-side discipline — a stock that gaps down on a catastrophic miss tends to remain weak for months. Third, the majority of the drift occurs in the first thirty to forty-five days, with a tapering effect extending into the ninety-day window. This is the structural reason most institutional allocators care about the first quarter post-announcement.

Why PEAD Persists — The Behavioral Explanation

If PEAD were a pure data artifact, it would have disappeared once Bernard and Thomas published their findings. It did not disappear, which means something structural is preventing arbitrage. The leading explanations are behavioral and institutional, and they are worth internalizing because they map directly onto what the VCP methodology is already exploiting on the chart.

Anchoring and conservatism. Analysts do not revise their forecasts in a single day after an earnings beat. They revise them gradually, in steps, over the following weeks and months, as they absorb the surprise, reassess the business trajectory, and become willing to stake their reputations on a more optimistic number. Each upward revision prompts another wave of institutional buying. The drift is partly a cascade of analyst revisions chasing a reality that already happened.

Confirmation lag and information diffusion. Not every institutional portfolio manager reads every earnings call transcript the night of the release. News spreads. Conference presentations, sell-side notes, buy-side meetings — each touchpoint introduces a new set of institutional buyers who were not ready to act on announcement night but are ready two weeks later. The marginal buyer keeps arriving.

The "surprise in the surprise" effect. Bernard and Thomas themselves identified a striking secondary pattern: positive earnings surprises tend to be followed by another positive surprise at the next earnings announcement. The market systematically underreacts to the information content of current earnings because it does not fully appreciate the implication for future earnings. A company crushing estimates this quarter is more likely than average to crush estimates next quarter. Analysts remain calibrated to the old trajectory.

Implementation frictions. PEAD is concentrated in smaller and mid-cap stocks where liquidity is thinner and institutional positioning takes longer. The largest institutional players cannot deploy capital into these names quickly without moving price. The very frictions that prevent arbitrage are what preserve the anomaly for traders operating at a scale where liquidity is not the binding constraint.

Quantifying the Surprise — The SUE Score

The academic literature does not measure earnings surprise as a simple "beat versus miss" binary, and neither should a serious trader. The standard quantification is Standardized Unexpected Earnings, or SUE. The idea is to measure the surprise not in absolute dollar terms, not even as a percentage of expected earnings, but relative to how surprising surprises have historically been for that specific company.

A SUE of +3.0 means the surprise was three standard deviations larger than the typical surprise for that company. That is statistically remarkable. A SUE of +0.4 means the surprise was within the normal noise of what analysts miss by every quarter — not worth trading on. The SUE framework does for earnings what z-scores do for any other distribution: it separates signal from noise by normalizing against the company's own history.

In Bernard and Thomas's original work, the top SUE decile — roughly corresponding to SUE values above +2 — produced the cleanest drift. Subsequent research has refined this further. Chan, Jegadeesh, and Lakonishok showed that combining SUE with post-announcement price reaction strength creates a more selective filter: stocks that surprise big AND whose price confirms the surprise with strong immediate action drift further than stocks that surprise big but trade flat. This is where the chart finally enters the story.

The PEAD Drift Curve

Before examining the synthesis with structural analysis, it is worth visualizing what the academic data actually looks like. The following diagram shows the stylized cumulative abnormal return profile for each SUE decile over the sixty trading days following an earnings announcement. The shapes of these curves are the empirical reality every practitioner must internalize.

Where PEAD and VCP Converge

The reason PEAD matters to a structural trader — and not merely to a quant running a decile-ranked factor portfolio — is that the mechanism driving the drift is the same mechanism driving the most powerful chart patterns. Institutional capital absorbs new information slowly. That slow absorption, visualized on a chart, looks like a base. It looks like volatility contracting as the transition from shock to consensus plays out. It looks like a breakout when absorption is complete and the revaluation becomes the new trend.

There is a spectrum of quality within earnings-related moves. A stock that gaps up on earnings into thin air, with no base, no relative strength, no prior accumulation footprint — this is raw PEAD exposure, statistically valid but structurally fragile. A stock that gaps up on a massive SUE surprise, clears a well-defined volatility-contraction base, and exhibits institutional volume characteristics in the post-gap consolidation — this is PEAD and structure, and the probabilities compound.

The synthesis is not that one framework validates the other. The synthesis is that they are measuring the same phenomenon through different lenses. The academic literature sees institutional slow-absorption through the window of price drift over sixty days. The VCP practitioner sees it through the window of volume profile, range contraction, and breakout confirmation. A high-SUE stock building a tight base in the weeks after its announcement is showing both footprints at once.

A Composite Framework for Earnings-Catalyst Plays

The framework below illustrates how the variables combine. It is not a proprietary scoring system from any specific program — it is an illustrative synthesis of what the academic literature and the structural literature collectively suggest matters. Treat the weightings as a starting point for your own testing, not as received truth.

Factor	Signal Strength	Weight Category	Source Lens
SUE score > +2.0	Strong	Primary	Academic (Bernard-Thomas)
Revenue surprise aligned with EPS surprise	Strong	Primary	Academic (quality of beat)
Guidance raise accompanying the beat	Confirming	Primary	Academic / Fundamental
Gap-up open on 2×+ average volume	Strong	Secondary	Structural (confirmation)
Close in upper third of day-one range	Confirming	Secondary	Structural (absorption)
Analyst revisions upward within 5 days	Confirming	Secondary	Academic (revision cascade)
Prior base within Trend Template criteria	Strong	Primary	Structural (VCP)
Post-gap volatility contraction forming	Strong	Primary	Structural (VCP)
Earnings gap retraces > 50% in first 10 days	Warning	Disqualifier	Structural (failed absorption)

The value of the composite view is that it filters ruthlessly. Raw PEAD exposure on any random beat is a statistical edge but a noisy one. Combining SUE with structural confirmation selects a narrower set of candidates where multiple independent frameworks agree. The hit rate on the narrower set is materially higher, which is what allows position sizing to scale confidently.

Implementation Realities and Common Mistakes

The trader who reads the PEAD literature and concludes that buying every earnings beat will generate alpha will be disappointed. The literature's decile-sorted returns are gross of costs, frictions, and the hundreds of small execution decisions that separate academic abstraction from trading reality. Several mistakes are predictable enough to deserve naming.

Mistake two: Treating every beat equally. A company that beats by a penny on a single-penny surprise is not a PEAD candidate. A company whose SUE is +3.5, whose revenue surprise aligns with EPS, and whose guidance was raised, is. The decile structure matters. Without SUE discipline, a trader is capturing the average drift across all beats, which is considerably weaker than the top-decile drift that makes PEAD interesting.

Mistake three: Ignoring what the market is telling you about the beat. A stock that beats substantially and sells off on announcement day is sending a signal the academic literature recognizes: the market already priced in more than the analysts did, and the "surprise" is actually a disappointment. The SUE score may say top decile, but the price action says the setup is broken. Bernard and Thomas's follow-up work showed that pairing the SUE with post-announcement price reaction filters out exactly this failure mode.

Mistake four: Holding through the next announcement without reassessment. The PEAD window is sixty to ninety days. The next quarterly announcement is roughly that far out. Traders who hold mechanically into the subsequent earnings event are re-introducing gap risk into a position that was supposed to be riding a statistical tailwind. The structural approach is to recalibrate well before the next announcement — either by trimming size, tightening stops, or exiting if the chart deteriorates.

Mistake five: Confusing magnitude of gap with magnitude of drift. A stock that gaps up 18% on earnings has front-loaded a large portion of its post-announcement move into a single day. A stock that gaps up 4% on a comparable SUE has more drift left to deliver. The academic drift is measured in cumulative abnormal return — if most of the reaction happened on day one, there is less left for days two through sixty. The best PEAD setups, structurally, tend to be moderate-gap-with-strong-base candidates, not spectacular-gap-with-nothing-underneath.

Calibrating Expectations: What PEAD Is Not

Academic anomalies have a way of being oversold in trader education. It is worth stating plainly what PEAD is and is not, so the framework is applied with appropriate humility.

PEAD is a statistical tendency, not a deterministic outcome. Any individual stock in the top SUE decile can decline over the next sixty days. The edge is visible at the portfolio level, across many positions, over many cycles. A single trade that goes against you is not evidence the anomaly has failed. This is the same epistemological discipline that applies to any edge — treat the sample of one as noise and the sample of many as signal.

PEAD is not a timing tool. It does not tell you when to enter, where to place stops, or how much to risk. It tells you which pool of candidates has a tailwind. Structural methodology fills in the timing, the risk controls, and the sizing. Using PEAD as a standalone entry signal is like reading a weather forecast that says "rain likely this week" and then standing outside for seven days waiting. The forecast is information; it is not an operating plan.

PEAD is not regime-invariant. In strong uptrends, top-decile SUE stocks drift beautifully. In distribution-heavy markets, even earnings winners can struggle as the tide works against them. The same market regime filters that govern any momentum strategy — distribution day counts, breadth, the condition of the major indices — apply to PEAD-driven trades as well. Academic literature measures the drift across many regimes and reports the average; practitioners live in the current regime, which can be materially better or worse than the average.

The Broader Principle

The deeper lesson of PEAD is not about earnings. It is about the way information is absorbed into prices. The efficient market hypothesis, in its strongest forms, insists that new information is incorporated instantaneously and completely. Fifty years of research have shown this is not how the world works. Information diffuses. Analysts revise in stages. Institutional capital rotates on a lag. Consensus shifts on a timeline measured in weeks, not minutes. The gap between the arrival of information and its full absorption is where structural traders live.

VCP, CANSLIM, stage analysis, and every momentum methodology worth studying are, at their core, frameworks for detecting this absorption process on the chart. They recognize that institutions buy in waves, that supply and demand imbalances persist, that sentiment shifts gradually, and that the patterns these forces leave behind are readable with discipline. PEAD is the academic formalization of exactly this idea applied to one specific catalyst — earnings — using one specific lens — relative return drift. It is not a separate edge from pattern-based trading. It is the same edge, documented differently.

The sophisticated trader does not choose between academic finance and chart-based methodology. They synthesize. They use the academic literature to identify where the probabilistic winds are blowing and the structural literature to navigate those winds with precision. When the two converge on the same stock at the same time — a top-decile earnings surprise building a tight base in a healthy market regime — the resulting setup is greater than the sum of its parts.