Post-Earnings-Announcement Drift

When a company reports earnings that beat expectations, its stock jumps on the news — that part is no surprise. The surprise is what happens next: the stock keeps drifting up for weeks, even months, afterward. Firms that miss keep drifting down. The market reacts to the news, but not all at once — it underreacts, and then slowly finishes the job. This lag is called post-earnings-announcement drift (PEAD), first documented by Ball and Brown in 1968 and still going strong decades later, which makes it one of the most stubborn cracks in Market Efficiency (The EMH).

If markets were perfectly efficient, all the information in an earnings report would be in the price by the closing bell on announcement day. PEAD says the price keeps moving predictably for a month or two — so a patient trader can position after the news is public and still get paid.

Measuring the surprise

You need a clean measure of how big the earnings surprise was, standardised so it is comparable across companies. The standard tool is standardised unexpected earnings (SUE):

$$\text{SUE}_i = \frac{\text{actual EPS}_i - \text{expected EPS}_i}{\sigma_i} .$$

The numerator is the surprise itself — reported earnings per share minus what analysts (or a simple model) expected. Dividing by $\sigma_i$, the typical size of that firm's past forecast errors, turns raw dollars into standard-deviation units, so a big surprise at a volatile firm and a small one at a steady firm are on the same scale. The strategy sorts stocks by SUE and goes long the biggest positive surprises, short the biggest negative ones — a close cousin of Cross-Sectional Momentum, but triggered by a discrete news event rather than a price trend.

Worked example: a clean beat

A company was expected to earn $1.00 per share and actually reports $1.15 — a $15$-cent beat. Over the past several years the typical error in forecasting this firm's earnings has been about $5$ cents, so $\sigma = 0.05$ in dollars. Its surprise in standardised units is

$$\text{SUE} = \frac{1.15 - 1.00}{0.05} = 3.0 .$$

That is a big surprise — three standard deviations. Historically, the top SUE decile has drifted an additional few percent (net of the market) over the following 60 trading days, after the announcement-day pop. So a trader who buys the day after the report — once the news is fully public — still captures the drift. Do the same in reverse for the deepest misses, and the long-minus-short book earns the spread between the two drifts.

Why does it persist?

The leading explanation is behavioural: investors anchor on stale expectations and are slow to update, so the price crawls to fair value instead of jumping there. Limits to arbitrage keep it alive — the drift is strongest in small, illiquid, thinly-covered stocks where trading costs and short-sale frictions deter the arbitrageurs who would iron it out. A competing view is that part of the "drift" is really a risk premium, but the sheer predictability and its concentration in hard-to-trade names point mostly to underreaction plus frictions.

Where it stumbles

• It hides in the illiquid corners. The drift is largest in small, cheap-to-ignore stocks — exactly where costs and shorting frictions eat the edge. In big liquid names it is faint.
• It has decayed. As PEAD became famous, faster reactions and quant crowding shrank it; part of the effect that survives is in the costliest-to-trade segment.
• Expectations are hard to pin down. SUE depends on a clean "expected EPS." Analyst estimates are biased and stale; a bad expectation model manufactures fake surprises.
• Clustered risk. Earnings arrive in bunched seasons, so a PEAD book takes on lots of correlated event exposure at once, and drifts can reverse around the next quarter's report.