Strategy #0000000014 · Cross-sectional momentum and events
Earnings momentum
An announcement is the signal’s start; execution time determines whether it was tradable.
Post-earnings-announcement drift (PEAD) asks whether prices keep moving in the direction of an earnings surprise after the report. A common approach ranks standardised surprises, buys positive-surprise firms and shorts negative-surprise firms after the information is public.
This maps event timing and possible return continuation; prices do not always move in one direction after a release.
Three ways in
Understand the story. Then inspect the risk.
Start with a familiar situation, check the payoff, then test how the numbers change.
The idea · in two minutes
Jack hears good news and wants to buy; Rose asks when the market heard it
A company reports EPS of 1.20. The same quarter last year was 1.00, and the standard deviation of past seasonal changes is 0.10. Under this simplified seasonal random-walk benchmark, the surprise is two standard deviations. Rose ranks it as positive news rather than calling any positive EPS a beat.
Another company’s EPS is below its year-ago quarter and its standardised surprise is −1.5. Jack wants to trade before both releases. Rose points out that this is future information. If a report comes after the close, a backtest entering at that day’s close has seen news the market did not yet have; at minimum, entry must wait for the next tradable time.
If the high-surprise group later earns positive abnormal returns while the low-surprise group lags, the spread is consistent with drift. It may also reflect risk, liquidity or measurement choices. A hypothetical table cannot show that the effect still works after today’s trading costs.
- 正向 surprise 假設Ranked in the positive-surprise group; the follow-on return is assumed, not observed.
- 接近預期假設Actual EPS matches the benchmark forecast, so there is no surprise-ranking edge.
- 負向 surprise 假設Ranked in the negative-surprise group; the figure only illustrates direction.
Illustrative research examples only. They are not live signals or a recommendation.
The mechanics · payoff and exposure
A standardised surprise must use earnings information and a scaling estimate that were available at the announcement time.
This page uses a simplified seasonal-random-walk proxy: EPSᵢ,q−EPSᵢ,q−4 divided by the ex-ante standard deviation of seasonal changes. Research also uses analyst forecast errors; the definitions are not interchangeable and must use point-in-time inputs.
Greeks: the directional intuition
- Event clock
- Measures sensitivity to implied volatility.
- Forecast scale
- Measures sensitivity to implied volatility.
- Drift window
- Measures sensitivity to implied volatility.
- Trading costs
- Measures sensitivity to implied volatility.
A standardised surprise must use earnings information and a scaling estimate that were available at the announcement time.
Practical context · published observations
High-minus-low SUE scenario: read the record with its caveats.
The chart uses the published figures or explicitly labelled study inputs below. A historical index or model replay is not the same as an investor’s realised account return.
未使用真實公司公告、分析師預估或報價；不含公告跳空、延遲成交、借券、交易成本和持有期間中的新資訊。 Source ↗ · Methodology ↗
Change assumptions and inspect the scenario
Inputs illustrate the strategy mechanics; they are not live quotes, forecasts or investment advice.
Variants · same family, different risk levers
Sixteen earnings-momentum studies change the surprise or information environment.
Variants examine surprise streaks, attention, revenue confirmation and call transcripts. Preserve original release times and data vintages for each.
| Related strategy | Relationship | What changes | New cost or risk |
|---|---|---|---|
| Announcement return | 公告後報酬 | Ranks short-horizon abnormal returns after the earnings release, rather than pre-event price trends. | Event dates and benchmark choice can mistake gaps or market moves for drift. |
| Earnings consistency | 獲利穩定性 | Uses the direction or consistency of earnings across several periods. | Past earnings stability does not ensure continuation; accounting differs by industry. |
| Earnings surprise streak | 驚喜連續性 | Counts consecutive earnings beats or misses to capture a surprise sequence. | Forecast revisions and duplicate event records can distort the streak. |
| Standardised earnings surprise | 標準化意外盈餘 | Scales actual earnings minus a seasonal benchmark by historical forecast-error volatility. | Fix the benchmark and scale window; tiny denominators amplify noise. |
| Earnings streak length | 連續期長度 | Uses the number of consecutive earnings-growth quarters, not surprise magnitude. | Long streaks can coincide with stretched valuations or cycle peaks. |
| Gross-margin-to-sales growth | 營業基本面組合 | Compares gross-margin growth with sales growth to identify margin improvement. | Mix, one-off items and accounting classifications can mimic improvement. |
| Friday earnings announcements | 公告日注意力 | Tests whether lower attention to Friday releases delays price adjustment. | Trading calendars, holidays and firms’ timing choices confound the test. |
| Driven to Distraction | 同日事件注意力 | Studies whether competing events weaken the immediate response to earnings. | Needs point-in-time event data; news intensity also relates to firm risk. |
| Earnings seasonality | 季節性驚喜 | Compares the quarter with the same quarter a year earlier to address seasonality. | When seasonal patterns shift, last year is a poor forecast benchmark. |
| Information in scheduled timing | 發布時點訊號 | Uses scheduled announcement timing as a signal about information or uncertainty. | Timing is chosen by firms; selection is not causal evidence. |
| High-frequency post-event underreaction | 盤中反應速度 | Uses high-frequency prices to examine adjustment after news. | Asynchronous quotes, spreads and microstructure noise can dominate. |
| International earnings momentum | 跨國盈餘動量 | Forms earnings-surprise portfolios across countries using local reports. | Accounting, time zones, shorting rules and FX complicate comparisons. |
| Earnings momentum plus short-term reversal | 反轉控制 | Adds a short-term reversal signal to separate drift from price bounce. | Weights and event timing can overfit microstructure effects. |
| Low analyst-dispersion earnings momentum | 預測分歧篩選 | Tests earnings momentum among stocks with lower analyst-forecast dispersion. | Low dispersion does not guarantee forecast accuracy; sparse coverage weakens the measure. |
| Revenue-surprise confirmation | 營收交叉確認 | Adds revenue surprises to test whether sales support the earnings signal. | Recognition, deferrals and acquisitions reduce revenue comparability. |
| Earnings-call investor responses | 電話會議文字訊號 | Uses call responses or text features after the release to predict drift. | Transcript delays, model drift and training leakage require control. |
Do not trade at a price before the report was public; use the next tradable time for after-hours releases and audit revised fundamentals.
Read beyond the summary
References
Primary sources and research used to describe the strategy and its limits.
- 01Bernard & Thomas (1989), Post-Earnings-Announcement Drift: Delayed Price Response or Risk Premium?research
- 02Chan, Jegadeesh & Lakonishok (1996), Momentum Strategies.research
- 03DellaVigna & Pollet (2009), Investor Inattention and Friday Earnings Announcements.research
- 04Hirshleifer, Lim & Teoh (2009), Driven to Distraction: Extraneous Events and Underreaction to Earnings News.research
- 05SEC, companyfacts API: XBRL company concepts and reported facts.official