Reading the Market Reaction: What the Event Study Can—and Cannot—Tell Us
This chapter examines the market valuation of U.S.-listed payment incumbents around the decisive GENIUS Act vote. The specific question is whether the observed stock-price gap supports a claim about expected future competition, rather than a claim that stablecoin payment volume had already displaced conventional providers. Understanding that distinction lets readers use the IMF study's estimates without turning a carefully bounded event study into an overconfident industry forecast.
A vote chosen for information, not drama
Legislative calendars contain many announcements, negotiations, and votes. Most are poor experiments for learning what investors believe because several uncertainties change at once, or because the event is so widely anticipated that prices have already adjusted. The IMF study chooses the House of Representatives' July 17, 2025 approval of S.1582—the bill that became the GENIUS Act—because it had a more useful combination of timing and information.
The Senate had already approved the bill, and the House took up the Senate's version without amendments. The House vote was therefore the last major legislative step before the bill went to the president. Signing was expected, but passage in the House was not certain. There had been disagreement over whether to advance S.1582 or an alternative, and last-minute negotiations made the outcome meaningful to investors. When the House passed the bill, the market learned that a federal payment-stablecoin framework was much more likely to become law than it had been immediately before the vote.
The setting also mattered. The vote took place during “Crypto Week,” when financial markets were already paying close attention to crypto legislation. That attention raises the chance that the marginal investor understood the news and traded on its implications. It does not make the event perfectly clean: attention to several related bills could create confusion. The study addresses that concern by comparing the response around the GENIUS Act with other legislative events and by testing whether the repricing appeared specifically after S.1582 rather than after crypto legislation generally.
The event was not a surprise in the simple sense of “nobody saw it coming.” Prediction-market data put the probability of stablecoin legislation passing in 2025 at about 93% before the House vote in the study's scaling calculation. But a high probability is not certainty. When the remaining uncertainty is resolved, a smaller final update can still matter for firms whose valuations depend on the policy becoming real.
The counterfactual is other financial firms
The study does not ask whether payment stocks went up or down in isolation. It compares firms classified as transaction and payment-processing businesses with other financial-sector firms over a narrow window around the vote. That comparison is the counterfactual: what might payment firms' returns have looked like if the legislative news had affected the broader financial sector but not payment competition in particular?
The sample contains 35 listed U.S. payment firms. The study uses stock prices at 15-minute intervals during regular trading hours and estimates firm and time effects. Firm effects absorb persistent differences between companies, such as one firm's business model or another's normal valuation level. Time effects absorb movements affecting the financial sector at a given interval. The relevant result is the additional movement of payment firms relative to the comparison group after the vote.
The pre-event pattern is important. Before the decisive vote, payment firms and other financial firms followed a similar path, and the estimated differences were not statistically significant. That does not prove that the groups were identical in every respect. It makes the post-vote divergence less likely to be a continuation of an obvious short-run trend already underway.
The study observes a ten-trading-hour window around the vote rather than comparing months before and after the law. That narrow design is a trade-off. It reduces the time available for unrelated developments to affect payment firms differently from other financial firms, but it also means the estimate is about an information shock, not about the slow process by which a law changes products, contracts, customer behavior, and revenue. The short window is useful precisely because it asks what investors revised when the probability of enactment changed, before subsequent execution could be mistaken for the effect of the vote itself.
The comparison group is also a disciplined approximation, not a perfect twin. Other financial firms can have their own exposure to digital assets, interest rates, or changes in regulation. If the GENIUS Act affected banks, asset managers, or insurers at the same time, comparing payment firms with them could either dampen or complicate the estimated contrast. That is why the paper tests alternative control groups and different specifications rather than relying on a single benchmark. The resulting claim remains relative: payment firms performed worse than the chosen comparison set after the relevant news, under a design intended to isolate payment-specific exposure.
The logic can be pictured as a sequence rather than as a claim that one line on a chart proves causality:
That structure matters for interpretation. If the entire market had sold off, or if financial firms broadly had moved together, a fall in payment stocks would tell us little about stablecoin competition. A relative gap is more informative because it asks whether the news mattered differently for the firms whose future cash flows are most connected to payments.
What the immediate estimate measures
After the vote, payment firms' stock prices fell relative to those of other financial firms and remained lower through the rest of the event window. Weighted by market capitalization, the average difference was about 1.3 percentage points. Multiplying that return difference by the firms' pre-event market value produced an estimated $21.5 billion decline in market capitalization relative to the comparison group.
The figure is best understood as a valuation translation. It converts a percentage-point return difference into dollars using the size of the firms before the event. It does not mean that $21.5 billion was removed from customers' accounts, that payment revenue fell by $21.5 billion, or that the firms lost that amount in economic resources. Market capitalization is the price investors are willing to pay for expected future cash flows multiplied by the number of shares. A change in it can reflect a revision to the expected level of future profits without any immediate change to current operations.
The study examines one reason the price move might have occurred. A new policy could make future cash flows less valuable, or it could make those cash flows more uncertain. The authors use 30-day and 24-month put-implied volatility as outcome measures and find no significant relative change after the vote. That result is consistent with a level effect: investors marked down what payment firms might earn, rather than simply widening the range of possible outcomes around the same expected earnings.
“Consistent with” is the right strength of language. Option-implied volatility is a market measure with its own assumptions, and an event study cannot observe every investor's reasoning. The evidence does not prove that all investors agreed about the same mechanism. It does show that the pattern fits a competition story more closely than a story in which the vote merely made payment firms look generically riskier.
Why scale at all—and why not overtrust it
The immediate return difference captures only the new information released by the vote. If investors had already incorporated most of the law's expected effect into prices, the event-window response would be smaller than the response to a hypothetical surprise in which the probability of passage moved from zero to one. The IMF authors therefore use prediction-market probabilities to approximate how much expectation changed around the event.
The calculation is conceptually simple. Let the stock response be the response to the change in the perceived probability of enactment. If the probability increased by only a small amount, scaling that response to a full-probability change produces a larger implied effect. In the study's baseline calculation, the probability moved from roughly 93% before the vote to certainty after it. Under a linear relationship between the probability update and the return, the observed response implies an overall effect of about 18%, or approximately $300 billion in incumbent payment-firm value.
The paper reports a sensitivity range of 13% to 27%, or $220 billion to $470 billion, when it changes legislative milestones and event-window widths. The range is useful because it shows the result is not tied to one exact window. It is also a warning: estimates that depend on scaling are not as directly identified as the event-window response.
The linearity assumption may fail in either direction. If investors wait for near-certainty before acting, the final increment from “very likely” to “certain” could have an unusually large effect, causing a linear scaling rule to overstate the full impact. If investors had already fully priced the law once its probability became sufficiently high, the final price response could be muted, causing the same rule to understate it. The authors therefore present the larger number as a cautious approximation, not as a measured loss of future revenue.
There is a second reason to keep the estimate in its category. The study measures what public-market investors expect about incumbent payment firms. It does not measure whether users actually adopt stablecoins, whether merchants change their settlement choices, or whether a lower-cost rail eventually expands total payment demand. Expectations can be predictive, but they can also be wrong, especially when the technology, rules, and business responses are still developing.
Robustness changes confidence, not category
Several checks strengthen the interpretation of the immediate relative repricing. The result is similar with alternative comparison groups, controls for different trends by firm size and profitability, and narrower or wider event windows. It also appears around a broader set of congressional events that advanced the GENIUS Act. Placebo tests are consistent with the effect being concentrated in windows containing stablecoin-related legislative news rather than appearing at arbitrary times.
These checks address specific alternative explanations. A different control group asks whether the result depends on comparing payment firms with an unusual set of financial companies. Trend controls ask whether larger or more profitable firms were already moving differently. Alternative windows ask whether the estimate is produced by one fifteen-minute interval or a narrow choice of cutoff. Placebos ask whether a similar “effect” appears when no relevant stablecoin news occurred.
The checks also clarify what kind of mechanism the study can identify. They make it less likely that the result is only a calendar artifact or a continuation of firm-size trends. They do not, by themselves, determine whether investors were responding to expected lower fees, new entrants, the cost of compliance, or the possibility that incumbents would fail to adapt. Those mechanisms belong to the next layer of analysis. The event study tells us that the market assigned a different value to payment exposure after the legislative information arrived; it does not observe the internal forecast used by each investor to reach that valuation.
That distinction is especially important for the subject of this series. A payment company can be exposed to stablecoin competition in one business line and benefit from lower settlement costs in another. The aggregate return can therefore be negative even when some firms, products, or customer segments are expected to gain. A study of the average treatment effect is a starting point for asking which capabilities explain the dispersion—not a substitute for that question.
But robustness does not turn an event study into a direct operating forecast. Other information can arrive near a major vote, even when the event window is short. The study's sample is listed U.S. firms, not private payment providers, foreign firms, consumers, or merchants. The cross-border subgroup is relatively small. The prediction-market scaling remains assumption-dependent. And blockchain transaction records do not, by themselves, reveal whether a transfer was a genuine payment for goods and services rather than trading, rebalancing, or activity among wallets controlled by one entity.
Those limits do not make the market reaction useless. They define its proper use. The evidence supports the claim that investors expected the GENIUS Act to increase competitive pressure on incumbent payment firms, with the cleanest estimate tied to the immediate relative return. It supports a larger full-law interpretation only conditionally, through the probability-scaling exercise. It does not support a claim that stablecoins had already captured a particular share of payments or that a fixed dollar amount of incumbent revenue is certain to disappear.
That calibrated result is enough to motivate the next question. If the market revised payment firms' future cash flows downward, what features of the stablecoin rail made the competition channel more convincing than the opportunity to use the rail as a cheaper tool? Answering that requires moving from identification to payment economics.