The research isn’t new. Making it practical is.
A note on what this page is not: it is not a claim that academic research proves any trading strategy is profitable. What the literature establishes is narrower and more useful — that earnings are a distinct, measurable market event, and that several recurring price and volatility phenomena around them have been documented, each in specific samples and periods. EarningsWatcher’s role sits after that: giving traders the data and tools to research, quantify and test those phenomena themselves.
1. Why earnings are a special market event
Most market-moving news is a surprise. Earnings announcements are different: the date is known well in advance, only the content is uncertain. That combination — a scheduled release of unscheduled information — is what makes earnings unusually amenable to research. The same event repeats for thousands of companies, four times a year, with a known clock.
The research consequence is a separation that runs through this whole literature: a stock’s day-to-day “background” volatility on one side, and the concentrated, event-specific uncertainty of the announcement itself on the other. Dubinsky, Johannes, Kaeck and Seeger (2019), in The Review of Financial Studies, model exactly this decomposition: they separate earnings-announcement price uncertainty from normal diffusive volatility and find that announcement uncertainty is quantitatively important for how options on the stock are priced. In their framework, an option expiring after an earnings date embeds a measurable premium for the event that an otherwise identical option expiring before it does not.
This is the foundation under a concept traders use daily: the implied move — the size of the reaction the options market is pricing for a specific report — is, in research terms, an estimate of that event-specific component.
2. Implied volatility before and after earnings
The earliest empirical work here goes back further than most traders expect. Patell and Wolfson (1979), in the Journal of Accounting and Economics, documented that anticipated earnings announcements are reflected in call option prices: implied variance rises as a known announcement date approaches. The options market, in other words, was pricing the calendar decades before retail platforms displayed it.
Donders and Vorst (1996), studying the European Options Exchange in Statistica Neerlandica, report the full cycle: option-implied volatility increases in the days before scheduled earnings announcements and drops sharply once the news is out. Truong, Corrado and Chen (2012), in the Journal of International Financial Markets, Institutions and Money, provide more recent evidence in the same direction — a general rise in implied volatility ahead of announcements followed by a significant decline immediately afterward.
These papers are the academic counterpart of two things traders name informally: the pre-earnings IV build-up (“IV rush”) and the post-earnings collapse (“IV crush”). The literature documents the pattern’s existence and shape; it does not say that any particular way of trading it is profitable after costs. The rise before the event is, at least in substantial part, rational: uncertainty genuinely is higher, and the drop afterward reflects uncertainty being resolved rather than a free premium evaporating.
3. What options are pricing vs. what actually happens
Implied volatility is a forward-looking quantity: it is the market’s price for future uncertainty, extracted from option premiums. Realized volatility — and, around earnings, the realized post-announcement move — is what actually happens. Comparing the two is one of the oldest frameworks in quantitative options research.
Goyal and Saretto (2009), in the Journal of Financial Economics, is a useful anchor here, with an important caveat: it is not an earnings study. They examine the cross-section of option returns sorted on the difference between historical realized volatility and implied volatility, and find that this gap contained information about subsequent option returns in their 1996–2006 sample. The reason to cite it on this page is narrower than its headline result: it establishes implied-versus-realized comparison as a legitimate, peer-reviewed research framework — the same framework a trader applies when they line up a stock’s implied move against its historical earnings moves.
The caveat matters as much as the citation. A stock whose options price a ±8% earnings move when it has averaged ±5% over ten years is not automatically “overpriced” — the market may know something the average doesn’t: a pending product decision, a guidance reset, litigation. Historical averages describe the past distribution; they do not by themselves establish mispricing. The research framework says compare; it does not say the comparison alone is an edge.
Option returns around earnings specifically
A smaller literature looks directly at option returns around announcements. Chung and Louis (2017), in the Journal of Empirical Finance, examine option returns around earnings announcements and document return patterns for specific portfolio constructions in their sample. As with any portfolio-level result, it does not generalize to “buying options before earnings makes money” — transaction costs, spreads and the particular construction rules carry much of the outcome, and results of this kind often attenuate outside the studied window.
4. Post-earnings announcement drift
Not everything about earnings resolves at the print. One of the most persistent findings in accounting research is that prices have, in many samples, continued to adjust in the direction of the earnings surprise for weeks after the announcement — post-earnings announcement drift, or PEAD.
The classic reference is Bernard and Thomas (1989) in the Journal of Accounting Research, who framed the central question in their title: is the drift a delayed price response, or compensation for risk? Their evidence favored a delayed response to the information in earnings — prices under-reacting at the announcement and completing the adjustment afterward. PEAD has since been re-examined for decades; later work debates its magnitude after trading costs, its concentration in smaller and less liquid names, and whether it has weakened as it became widely known. It remains one of the most documented anomalies in the literature precisely because it has survived that scrutiny as a phenomenon, even while its tradability is argued.
Note the boundary: Bernard and Thomas is a stock-price study, not an options study. It belongs to the post-earnings section of the map — the part traders explore through post-earnings drift and reaction records — not to the option-pricing evidence above.
5. Information transfer and sympathy moves
An earnings report is rarely about one company. When a major semiconductor firm reports demand falling, the news carries information about its suppliers, customers and competitors — none of whom said anything that day.
Foster (1981), in the Journal of Accounting and Economics, is the foundational study of this “intra-industry information transfer.” Examining earnings releases and the stock returns of non-announcing firms in the same industry, Foster documents significant information transfer — announcement-period returns of industry peers move in ways related to the announcing firm’s news — for subsets of firms and industries. Later research refined the picture: transfer is stronger where firms share economic exposures, and it is not uniform.
This is the academic ancestor of what traders call sympathy plays — studying how a reporter’s release moves its peers. The research supports the premise that peer reactions are a real, measurable channel; it does not support the idea that every peer reaction is predictable. The honest framing is statistical: some reporter–peer pairs show stable historical relationships, many do not, and the work is in measuring which is which.
6. From research to a repeatable process
Step back from any single paper and a common structure appears, and it is the structure this entire site is organized around:
- Repeated scheduled events create comparable observations. Every listed company reports on a known calendar, quarter after quarter. Unlike one-off news, each earnings event is an observation in a long, growing panel.
- Comparable observations allow distributions, not anecdotes. With enough observations you can ask distributional questions: how large is this stock’s typical reaction, how often does the realized move exceed the implied one, how did implied volatility behave in the days before past reports.
- Distributions generate hypotheses, and hypotheses can be tested. A documented tendency is not an edge until it survives your own testing — across names, periods, and with realistic assumptions about costs. That is what backtesting and stress-testing are for.
That is the philosophy in one paragraph: the literature above establishes that the phenomena are real enough to study; the practical work — measuring them on today’s names, on current option chains, with your own assumptions — is the trader’s. The methodology page describes how EarningsWatcher computes the quantities involved, and the earnings calendar is where the panel of upcoming observations lives at any moment.
7. Professional options literature
Peer-reviewed papers are one layer; a second layer is professional literature written by practitioners, which tends to focus less on documenting phenomena and more on the working machinery: how to measure volatility well, how to evaluate whether an apparent edge is real, and how to size positions when it is.
Euan Sinclair’s Volatility Trading (2nd edition, Wiley) is a standard example of the quantitative practitioner approach: measuring and forecasting volatility, defining and estimating an options edge statistically, evaluating trades after the fact, and managing risk and position size. His Positional Option Trading (Wiley) extends the treatment to option structures, return distributions and statistical methods for position-level decisions. Neither book is academic proof of any specific strategy — that is precisely their value. They are about the discipline of testing and sizing, which applies to earnings volatility the same way it applies to any other volatility trade.
8. References and further reading
Peer-reviewed papers first, professional books after. DOIs link to the publisher’s record.
