Post-Earnings Announcement Drift: Forty Years of Evidence and the Current Debate

Post-earnings announcement drift (PEAD) refers to the tendency for a stock's price to continue moving in the direction implied by an earnings surprise in the days and weeks following the public announcement of reported earnings. First documented in the 1960s and studied extensively over four decades, PEAD represents a potential challenge to efficient market hypotheses and remains at the center of debates about information processing, investor behavior, and market pricing. Despite sustained research and numerous proposed explanations, no consensus on the underlying mechanism has emerged, and the phenomenon continues to puzzle academics and practitioners alike.

The Core Concept

PEAD is rooted in a straightforward observation: when a company reports earnings that exceed or fall short of the consensus forecast, the stock price does not instantaneously adjust to a new equilibrium. Instead, the price continues to move in the same direction as the surprise for a period spanning days or weeks. This pattern is inconsistent with the strong form of the efficient market hypothesis (EMH), which holds that prices immediately incorporate all available information ^[1]. The drift is measured as an abnormal return: the return realized in excess of what a risk model (typically based on systematic beta or factor exposures) would predict.

The persistence of PEAD across decades and market conditions suggests a systematic feature of price discovery. Three broad classes of explanation have been proposed. The first is rational but gradual learning: investors may require time to decompose the current earnings signal into permanent and transitory components and to update forecasts of future cash flows. The second is behavioral underreaction: investors may systematically update their beliefs by less than the magnitude of the news, creating a phased adjustment that manifests as drift. The third is methodological: the apparent drift may partly reflect inadequacies in how abnormal returns are measured and controlled for in event studies.

How It Works Mechanically

An earnings announcement presents the market with reported earnings per share (EPS) and often forward guidance or management commentary. Analysts and investors compare the reported figure to their consensus forecast, generating an earnings surprise: the signed difference between actual and expected EPS. On announcement day, the stock typically experiences a price movement corresponding to the sign and magnitude of the surprise. Under EMH, this adjustment should be complete and immediate. The stock should then move only in response to new information unrelated to the earnings release.

The empirical observation is different. In the post-announcement period, stocks that announced positive surprises generate positive abnormal returns on average, while stocks that announced negative surprises generate negative abnormal returns. A buy-and-hold strategy implemented on the trading day after announcement (and held for several trading days) would, on average, capture returns in excess of risk-adjusted benchmarks. The daily magnitude of drift is typically small, but effects compound over the drift window.

Measurement of drift requires controlling for confounding factors. Returns on announcement day are mechanically larger than post-announcement drift, so researchers separate announcement-day returns from drift. Controls for systematic risk (beta), factor exposures (size, value, momentum), and cross-sectional differences in expected returns are standard. Studies typically sort announcements into quintiles or terciles by surprise magnitude and measure the drift separately for each group. This approach allows researchers to isolate the relationship between surprise size and subsequent drift.

A Worked Example with Historical Context

The PEAD phenomenon was first systematically documented in early studies of earnings information. Ball and Brown's 1968 analysis of annual earnings announcements showed that the market appeared to incorporate earnings information gradually rather than instantaneously ^[2]. This foundational observation motivated decades of subsequent research.

By the 1980s, more refined event studies were possible. Build, Olsen, and Shevlin's 1984 analysis examined quarterly earnings surprises and established that drift typically persisted for multiple weeks following announcement ^[3]. The effect was solid across firm sizes, sectors, and time periods examined. These findings were striking because markets were becoming more efficient, information dissemination was faster, and computer technology had improved price discovery mechanisms, yet drift remained observable and economically meaningful on the scales researchers could measure.

A key question emerged: would PEAD decline as markets modernized? The 1990s and 2000s saw dramatic improvements in information speed (corporate websites, electronic news feeds, algorithmic trading). Yet studies of announcements in those decades continued to detect drift. By the 2010s, with real-time earnings releases and social-media dissemination, one might expect PEAD to shrink further or vanish entirely. Empirical investigations of modern data have continued to document positive autocorrelation in post-announcement returns, suggesting that the phenomenon remains observable in contemporary markets ^[4].

Competing Explanations and the Current Debate

The persistence of PEAD across four decades of market evolution has generated competing frameworks for its interpretation. The rational explanation begins with the observation that earnings are not self-contained signals but raw data from which future earnings must be inferred. An earnings report reveals current profitability but requires interpretation: how much of the current earnings surprise represents a permanent shift in the firm's earning power versus transitory factors (one-time items, cyclical swings, accounting noise)? Informed investors may rationally require time to decompose the signal, update forecasts of future earnings growth, and incorporate implications for competitive position and strategic positioning. Under this view, drift is not a mispricing but the correct, gradual adjustment of price to fundamental value.

Behavioral explanations, grounded in the underreaction hypothesis, propose that investors update beliefs conservatively in response to earnings news, moving less than Bayesian updating would dictate ^[5]. The literature in behavioral finance suggests that anchoring on past forecasts, resistance to changing established views, and limited attention all contribute to incomplete initial adjustment. Under this framework, PEAD represents the eventual correction of an initial underreaction. A subsidiary debate concerns whether correction is complete (prices eventually reach fair value) or whether drift overshoots, creating opportunities for reversal trading.

A third explanation, increasingly emphasized in recent literature, attributes a portion of observed PEAD to measurement artifacts. Standard event study methods may inadequately control for cross-sectional differences in expected returns driven by firm characteristics (size, profitability, growth opportunities) and liquidity effects. Earnings announcements are accompanied by increases in trading volume, wider bid-ask spreads, and shifts in liquidity. These microstructure effects could generate apparent drift even absent any repricing of fundamental value.

The role of technological change and algorithmic trading is now central to the debate. If PEAD reflected true investor underreaction or slow information processing, one would expect modern high-frequency traders and algorithmic systems to have arbitraged the pattern away, accelerating price discovery. Yet empirical studies of recent data continue to document drift. Some researchers interpret this as evidence that PEAD is rooted in fundamentals, not mispricing. Others argue that behavioral factors remain powerful even in markets dominated by algorithms, or that algorithms themselves exhibit systematic biases in how they process earnings information.

Limitations

Several critical limitations constrain the confidence with which PEAD can be interpreted or exploited. First, while PEAD is statistically significant across large samples and long time series, the economic magnitude of drift per trading day is often modest. When transaction costs, bid-ask spreads, management fees, and taxes are considered, the net return to a PEAD-based trading strategy may be marginal or eroded entirely. Academic papers typically report results on equal-weighted or value-weighted portfolios of many firms, where trades are executed without explicit cost modeling. A practical trader implementing such a strategy faces execution costs that samples of hundreds of firms may not fully capture.

Second, PEAD exists most clearly in academic samples and controlled event studies, not consistently as a persistent, exploitable anomaly in live trading. The transition from observed pattern to profitable strategy requires forecasting, on an ex ante basis before announcement, which earnings surprises will drift strongly and which will not. Most PEAD studies are ex post: they condition on the magnitude of surprise and measure drift afterward. Forward prediction is considerably harder, as the same surprise (in absolute magnitude) may drift differently depending on broader market conditions, firm characteristics not fully captured in standard event study models, or the narrative context of the announcement.

Third, the lack of theoretical consensus about PEAD's mechanism severely limits the ability to forecast its future prevalence or magnitude. If PEAD is primarily driven by behavioral underreaction, technological and regulatory changes that enhance transparency and speed information dissemination should shrink it. If PEAD reflects rational learning about permanent earnings components, it should persist as long as current earnings contain noisy information about future fundamentals. Without clarity on the mechanism, predictions about whether PEAD will strengthen, weaken, or disappear remain speculative.

Fourth, most PEAD research focuses on large-cap stocks in developed markets with transparent financial reporting, substantial analyst coverage, and active institutional trading. In smaller-cap stocks, emerging markets, or thinly covered sectors, PEAD may operate differently or may not exist in measurable form. Evidence on these settings is limited.

Finally, the debate over explanation remains fundamentally unresolved. Distinguishing between rational, gradual learning and behavioral underreaction requires assumptions about what information investors should immediately process, which itself is contested. Methodological refinements have reduced the magnitude of attributed PEAD in some studies, but the core empirical regularity has proven solid to numerous specification tests.

Summary

Post-earnings announcement drift has been a documented empirical regularity for nearly six decades. The phenomenon is inconsistent with the strong form of market efficiency and has generated substantial research into information processing, behavioral finance, and market microstructure. Despite the maturity of the literature and sustained academic and practitioner interest, no single explanation commands consensus. The current debate centers on whether PEAD primarily reflects rational, gradual learning about future cash flows; behavioral underreaction; liquidity and microstructure effects; or some combination. The magnitude of drift has not disappeared in modern, high-speed markets, yet whether it remains economically exploitable remains contested. For researchers and practitioners, PEAD exemplifies the complexity of market pricing: even simple phenomena like the adjustment to public news can persist despite theoretical expectations, substantial research effort, and technological change.

Key definitions

Earnings Per Share (EPS): The portion of a company's profit attributable to each share of stock, typically compared against consensus analyst forecasts to compute earnings surprises.

Earnings Surprise: The difference between reported earnings and the consensus forecast; positive surprises exceed forecasts, negative surprises fall short.

Efficient Market Hypothesis (EMH): The proposition that security prices fully and instantaneously reflect all available information, making it impossible to achieve consistently abnormal returns using trading rules based on that information.

Abnormal Return: The return realized on a security in excess of the return predicted by a risk model or benchmark, isolated to measure the effect of a specific event such as an announcement.

Event Study: An empirical method that measures the effect of a specific event (such as an earnings announcement) on a security's return by comparing the return in the event window to a control period and adjusting for systematic risk.

Underreaction: In behavioral finance, the tendency for investors to adjust prices or beliefs by less than the information-theoretic response to new data, implying further price movement in the same direction will eventually occur.

References

1. Fama, Eugene F., "Efficient Capital Markets: A Review of Theory and Empirical Work," Journal of Finance, Vol. 25, No. 2 (1970).

2. Ball, Ray and Brown, Philip, "An Empirical Evaluation of Accounting Income Numbers," Journal of Accounting Research, Vol. 6, No. 2 (1968).

3. build, George, Olsen, Chris, and Shevlin, Terry, "Earnings Releases, Anomalies, and the Behavior of Security Returns," The Accounting Review, Vol. 59, No. 4 (1984).

4. Barberis, Nicholas, Shleifer, Andrei, and Vishny, Robert W., "A Model of Investor Sentiment," Journal of Financial Economics, Vol. 49, No. 3 (1998).

5. Hirshleifer, David and Teoh, Siew Hong, "The Psychology of Valuation and Market Puzzles," Journal of Finance, published as working paper via SSRN (2003 and subsequent revisions).

6. National Bureau of Economic Research, "Earnings Surprises and Anomalies," NBER Working Papers, ongoing. Https://www.nber.org/

Educational research on historical data only. Not investment advice, not a signal, and never a performance promise. Past results do not predict future performance. Every reference is link-verified before publication and every paper is re-audited weekly against the library's editorial standard.

Last reviewed by the PropLedger research pipeline: 2026-10-04. Educational research on historical data, not financial advice.