Post-Settlement Price Persistence in Event Contract Markets

Prices reflect three distinct structural mechanisms, namely capital lock-up discounting, settlement manipulation, and oracle latency frictions. Each produces a different pattern around the point when an event becomes known and a contract becomes payable. Understanding those differences is essential to reading settlement-window prices as financial claims.

Why prediction market prices do not simply track event probabilities

The textbook interpretation of a binary contract treats its price as the market’s probability that the event will occur. That reading is useful but incomplete. A contract’s value depends not only on whether it pays, but also on when payment becomes available and how the payable outcome is determined. Traders can therefore be sure about the event while still unsure about payment, leaving the winning claim’s price below face value even after most doubt about the event is gone.

That separation changes what a price movement means. A persistent discount need not indicate lingering doubt about the outcome, and a late price change need not reflect a better-informed forecast. Both can arise from the terms under which the claim trades and settles. The key distinction is structural rather than informational, because prices can move without new information about the event itself.

How the scale and architecture of event contract markets make structural frictions consequential

Kalshi’s volume in crypto event contracts passed a billion dollars a month for the first time in early 2026. Nasdaq filed with the SEC in March 2026, and Cboe filed its main binary index options expansion in April 2026, to list binary fixed-payout options on equity indices. At that scale, settlement-window distortions affect economically significant flows.

Meanwhile, contract horizons have compressed from a day to an hour to minutes, with Polymarket offering very short windows. Short horizons amplify settlement-window effects per unit time. Yet the platforms do not share a settlement architecture. Polymarket uses an oracle process, while Kalshi settles centrally, typically within a few hours. Those differences set how long capital stays locked up and which institution decides that a claim is payable.

The CFTC’s June 10, 2026 Notice of Proposed Rulemaking names market integrity as a factor in judging the public interest. Its proposed inquiry includes whether a contract “presents particular risks of manipulation, settlement integrity deficits, or information leakage”. The economic question of what a price means now intersects with the regulatory question of whether the contract’s structure supports an orderly market.

How locked collateral creates a maturity-dependent discount below true probability

In a collateralized market, a near-certain dollar is still a delayed dollar if collateral remains locked until oracle settlement. The delay has a real opportunity cost. A winning claim can thus trade below face value even when traders have little doubt about the outcome. All else equal, a longer expected wait for settlement produces a larger discount, creating a maturity-dependent wedge between the quoted price and the event probability.

Gebele and Matthes (2026) use the times when persistent near-certain contracts actually settled to work out an implied settlement-discount term structure. They summarize it as an annualized settlement wedge, or ASW. The recovered wedges are positive, maturity-dependent, and time-varying. Adjusting prices by these curves reduces the near-certainty horizon gradient by roughly 48–88%, indicating that much of the raw maturity pattern reflects priced settlement frictions.

The Polymarket contract “Will Russia capture Kostyantynivka by August 31, 2025?” illustrates that distinction. Its shaded carry window began at the final transition to a midpoint of 0.9995 and ended at oracle settlement. Before that window, trades still reflected two-sided price discovery. Inside it, outcome uncertainty was effectively gone, with bids near 0.999 and asks at par. The discount that remained reflected the wait for redemption after doubt about the event had faded.

Market architecture changes the cost of that delay. negRisk conversion compresses discounts by recycling part of a position into synthetic collateral, while yield-bearing collateral flattens the term structure by reducing the opportunity cost of lock-up. The same event can therefore generate different prices under different settlement and collateral arrangements. This mechanism is passive, since no participant needs to distort the outcome for the discount to arise. Settlement manipulation, by contrast, needs someone who profits by changing the reference that sets the payment.

How asset-price contracts create a structural incentive to manipulate the settlement price

An asset-price event contract creates that incentive when its holders can trade the underlying asset used for settlement. A holder of a near-the-money contract benefits from pushing the spot price just past the strike in the final seconds, then allowing it to revert. The settlement reference becomes a target for profitable distortion. This is the cash-settlement manipulation of Kumar and Seppi (1992), operating inside a retail prediction market at a five-minute horizon.

Dai, Jia, and Yu (2026) build a formal sequential game-theoretic model of the two markets. Their model shows the manipulator’s order pooling with prediction-market liquidity traders, reducing the information content of late trades. The harm extends into the underlying market. Manipulative orders lower the adverse selection faced by the spot market maker, inducing more liquidity supply even as spot prices become less informative.

Longer horizons make the per-contract manipulation profit too small relative to its cost, so manipulation does not arise in equilibrium. Reducing collateral lock-up would not, by itself, remove a holder’s reason to push a tradable settlement reference across the strike.

U.S. law already places a constraint on that design. Commodity Exchange Act Designated Contract Market Core Principle 3 requires every listed contract to be “not readily susceptible to manipulation”. CFTC Staff Advisory 26-08 (2026) reaffirms that requirement for event contracts. Traders may expect a distorted price, but the contract must also be judged on whether it gives them a predictable chance to create one.

How oracle latency and adjudication uncertainty keep prices from converging at settlement

Oracle-settled contracts pay on a certified report of an event. The distinction remains after the outcome becomes clear. If certification is uncertain or the resolution institution can be influenced, prices reflect the chance that the oracle will certify the correct result. A contract can consequently remain discounted throughout a dispute window even when further information about the underlying event would add little.

Latency and adjudication introduce different risks within that window. Latency prolongs the period during which capital cannot be redeemed, while adjudication uncertainty leaves the payable outcome in doubt. Disputed markets on platforms with token-vote oracles can remain unresolved for days. Where truthful resolution constrains the rents available for information, quotes can remain below certainty, with the last-cent discount pricing the threat of capture.

A 2026 study of disputed Polymarket markets finds that settlement departs from the audited label in 12.4% of 233 markets with an objectively clear audited winner under the primary coding model. The scope of that finding is limited to the platform’s disputed markets. It nevertheless establishes that event truth and certified payment can diverge. Where oracle voters also have financial stakes in the markets they judge, settlement risk overlaps with manipulation risk because the decision-maker can benefit from the decision.

Kalshi’s Khamenei market paid the last price before his death rather than par, illustrating a problem distinct from waiting for an oracle decision. A shorter timeline would not by itself change the contractual payment rule. The relevant uncertainty is therefore not always how long settlement takes, but what the resolution process treats as the amount owed.

Why a rational-discounting reading of these frictions does not eliminate the concern

The strongest objection treats these deviations as potentially correct pricing. Collateralized contracts are delayed contingent claims. A discount can rationally compensate for capital immobilization and the possibility of erroneous settlement. Gebele and Matthes make that argument for the settlement wedge, and their ASW adjustment provides empirical support for it. A price below par is not sufficient evidence that the market is wrong.

That objection holds most clearly for capital lock-up. An analyst who ignores the ASW may mistake compensation for waiting for a systematic underestimate of the event probability. Even with rational discounting, the quoted price still mixes beliefs about the event with the cost of getting paid, so it cannot be read as a simple probability.

Settlement manipulation requires a different assessment altogether. A price push followed by a post-settlement reversal is not compensation for waiting to receive money. Dai and colleagues show that manipulators gain mostly at retail traders’ expense, transferring wealth from liquidity traders without adding information. Calling this price pattern a risk premium would hide what causes it. The distortion exists because an agent can profit by moving the settlement reference.

The rational-discounting account is only partial for oracle risk. Participants can price a probability of erroneous settlement, but an unobservable conflict of interest is harder to incorporate accurately. When they cannot observe who is voting or the voters’ financial stakes, recognition that capture is possible does not resolve the information disadvantage. A discount may be rational from the participant’s position while the adjudication process remains compromised.

What each mechanism implies for how market participants should read and trade settlement-window prices

Near-certain contracts on collateralized platforms should be read against the platform’s ASW. Post-event carry, the observed practice of buying near-certain shares at a discount and redeeming at par, earns its return from that delay. It also immobilizes capital and depends on clean resolution. Yield-bearing collateral changes that comparison because it has a lower opportunity cost and thus a flatter discount curve.

Dai and colleagues find a settlement-window spike that reverses afterward, mostly in the roughly 6% of cycles where the contract price still pointed to a near-even outcome just before settlement. That reversal distinguishes the pattern from an informative reassessment. Manipulation concentrates in the final moments and in near-the-money cycles, while longer-horizon contracts do not show the same pattern. Spot traders exposed to those final seconds also face degraded price discovery, even if the underlying market appears more liquid.

For oracle and rule-based frictions, the central distinction is between knowing the event and knowing the payment. Large gaps between prices on platforms can last after an event because the platforms have different settlement paths, even when their forecasts agree. To read those gaps, traders need to look at both the resolution rule and the oracle timeline. The Khamenei case illustrates the former, while days-long disputes on token-vote platforms illustrate the latter.

The CFTC’s proposed public-interest test brings checks on settlement integrity closer to being part of contract approval. Treating settlement-window deviations as random will systematically misread collateral discounts, leave short-horizon traders exposed to manipulative price pushes, and understate resolution risk in disputed markets. These are structural forces with distinct signatures. Identifying which one is operating is the first line of defense.