Post-Earnings Announcement Drift (PEAD)

Post-earnings announcement drift (PEAD) is the tendency of a company's share price to keep moving in the direction of an earnings surprise for weeks or months after the results are announced. Shares of companies that report better earnings than expected tend to continue outperforming, and those that disappoint tend to continue underperforming, instead of adjusting fully on the day of the news.

The pattern was first documented by Ray Ball and Philip Brown in 1968 and is one of the oldest and most studied anomalies in finance. It is a challenge to the efficient market hypothesis, under which public information such as an earnings release should be reflected in the price almost at once. It forms the basis of a class of event-driven and news-based trading strategies.

The drift is studied by sorting companies on the size of their earnings surprise. The standard measure is standardised unexpected earnings:

• $\text{EPS}_{\text{actual}}$ is the reported earnings per share for the quarter.
• $\text{EPS}_{\text{expected}}$ is the expected figure, taken from analysts' forecasts or from a statistical model such as earnings in the same quarter a year earlier.
• $\sigma_{\text{surprise}}$ is the standard deviation of the company's past surprises, which puts volatile and stable companies on a common scale.

Suppose a company reports earnings of 1.30 per share when 1.10 was expected, and its past surprises have a standard deviation of 0.08. The measure is $(1.30 - 1.10) / 0.08 = 2.5$, a large positive surprise. An alternative approach uses the market's own verdict: the share's abnormal return over the two or three days around the announcement.

The fullest early description came from Victor Bernard and Jacob Thomas in studies published in 1989 and 1990. Ranking US companies each quarter into ten groups by surprise, they found that a portfolio long the highest group and short the lowest earned a substantial abnormal return over the following 60 trading days. The drift was larger for small companies than for large ones. A notable share of it arrived in the few days around the next three quarterly announcements, as though investors were repeatedly surprised by news that the previous quarter's results should have led them to expect.

The effect has since been found in many countries and periods. Eugene Fama, a leading defender of market efficiency, acknowledged in a 1998 review that it was among the anomalies that had best survived scrutiny. More recent studies report that the drift has weakened considerably in large, liquid US shares since the early 2000s, as trading costs fell and more capital was devoted to exploiting it, while it remains visible in smaller and less liquid shares and in a number of other markets.

• Underreaction. The leading explanation is that investors do not fully appreciate what current earnings imply for future earnings. Quarterly earnings changes tend to persist, but prices behave as though investors expect them not to. This account belongs to behavioral finance and is related to limited attention: drift is stronger for announcements made on days crowded with other news.
• Limits to arbitrage. Exploiting the drift requires trading many small shares. A wide bid-ask spread, price impact and the cost and difficulty of short selling can absorb the profit, so the mispricing persists where trading is expensive.
• Risk. Some researchers argue that companies with extreme surprises are riskier in ways that standard models do not capture, so the returns are compensation, not a market inefficiency. Attempts to identify such a risk have had limited success.
• Measurement. Part of the measured drift may reflect flaws in how expected returns and surprises are calculated, though the effect has proved robust to many alternative methods.

PEAD overlaps with price momentum. Companies with positive earnings surprises have usually had rising share prices, so the two signals often select the same shares. Studies that separate them find that each retains some predictive power, with earnings-based signals accounting for a considerable part of price momentum. Both are consistent with gradual incorporation of information into prices. Analyst forecast revisions, which tend to follow surprises and to arrive in sequence, are a third related signal, and the three are frequently combined in factor investing.

A strategy based on PEAD buys shares shortly after a strongly positive surprise and sells or shorts after a strongly negative one, holding for a period of weeks to a quarter. Quantitative funds typically apply it across hundreds of shares at once, so that company-specific accidents average out, and hedge the portfolio's overall market exposure. Modern versions refine the signal with the tone of the earnings call, the quality of the earnings and the reaction of the share on the announcement day.

The practical obstacles are those of any anomaly. The largest paper profits are in shares that are costly to trade, and a realistic allowance for every transaction cost removes much of the return. Announcements cluster in a few weeks each quarter, so capital is used unevenly. A backtest must use the date and time at which results actually became public and the forecasts that existed beforehand. Using later, revised data introduces look-ahead bias. As with other published anomalies, returns since discovery have been lower than those in the original samples.