Comment on Proposed Rulemaking -- Prediction Markets; Public Interest Determinations

RIN 3038-AF65 | Docket No. CFTC-2026-0024

Submitted by Micah Warren
Associate Professor of Mathematics, University of Oregon | micah541@proton.me

July 24, 2026

Introduction

I submit this comment to address a specific and underexamined category of event contract that I believe the Commission's proposed framework should address explicitly: pre-primary election contracts.

I argue that pre-primary election contracts are contrary to the public interest under the Commission's proposed factors, and that this conclusion follows from basic principles of market microstructure, risk management, and information theory---not from a general opposition to prediction markets. I further argue that the empirical record of prediction market accuracy in elections does not support the price discovery justification that underlies the Commission's permissive approach to election contracts generally.

Recent research by the Bitcoin Policy Institute^[1] documenting foreign influence campaigns in American civic life illustrates how readily foreign-aligned actors can amplify genuine domestic sentiment to manufacture the appearance of political will stronger than independently exists. Pre-primary prediction markets hand these actors a purpose-built instrument: a legal, lightly surveilled venue in which a single large bet generates media coverage indistinguishable from organic candidate momentum, with no contribution limits, no disclosure requirements, and no enforcement mechanism within the Commission's reach. This is not a speculative concern --- it is the predictable application of a documented playbook to a newly available tool.^[2]

I urge the Commission to adopt a bright-line prohibition: no event contracts settling on the outcome of any election, or on any measure of candidate viability, may be listed until after primary voting in the relevant jurisdiction has concluded. The arguments that follow explain why this rule is warranted and how it responds to the Commission's proposed public interest factors.

Background: Relevant Proposed Provisions

This comment addresses the following specific provisions of the proposed rulemaking:

• Proposed §§ 40.11(a)(5) and 40.11(a)(6) --- Public Interest Factors. The Commission proposes a multi-factor test for determining whether event contracts involving Enumerated Activities are contrary to the public interest, including factors relating to price discovery utility, manipulation susceptibility, settlement integrity, information leakage, and compliance infrastructure. The Commission specifically invites comment on whether additional factors should be included. This comment argues that pre-primary election contracts fail several of these factors, and that the Commission's analysis of the price discovery factor relies on an empirical premise---that prediction markets accurately aggregate information about political outcomes---that is not supported by the available evidence.

• Commission's Request for Comment on Additional Public Interest Factors. The Commission explicitly invites comment on whether there are additional general factors that should be considered in public interest determinations. This comment identifies two such factors that are specific to pre-primary election contracts and that the proposed framework does not currently address: (1) the capacity of regulated domestic markets to launder manipulation originating in offshore, unregulated markets; and (2) the media amplification mechanism by which prediction market prices are converted into political outcomes independent of their informational accuracy.

I. Pre-Primary Election Markets Are Not Performing Price Discovery --- They Are Performing Attention Allocation

The Commission's primary justification for permitting election-related event contracts is that they serve a price discovery and information aggregation function. This justification does not hold at the pre-primary stage.

At the pre-primary stage, the actors with genuine insight into a candidate's viability --- internal polling, donor commitments, opposition research, family and staff decisions --- are a small tight group with incentive to conceal or strategically distort that information. The remaining market participants are effectively guessing. What results is not a Hayekian aggregation of dispersed knowledge but something closer to a poker table.

This matters because the Commission's proposed price discovery factor asks whether event contracts can "meaningfully facilitate risk transfer and price discovery." Pre-primary markets cannot. They can, however, do something else entirely: allocate media attention. Because mainstream news organizations now routinely report prediction market odds as a proxy for candidate viability, a sufficiently large bet on an obscure candidate generates news coverage, donor interest, and retail betting activity --- regardless of whether the bet reflects any genuine information. The market is not discovering truth; it is manufacturing salience.

This dynamic is compounded by the degree to which prediction market odds have been integrated into mainstream information infrastructure. As Packin and Rabinovitz note in a 2026 Science policy forum, "major financial platforms such as Google Finance and mainstream media outlets including Reuters and CNN now display real-time [prediction market] odds, covering reported probabilities of elections, geopolitical escalation, or armed conflict, alongside standard economic indicators, granting speculative probabilities the appearance of objective forecasts rather than potentially manipulable signals."^[3] The authors further observe that "reflexivity theory posits that traders' beliefs shape market probabilities, which in turn reshape beliefs; thin liquidity amplifies this reflexivity: Once reported, even modest price shifts influence subsequent behavior and institutional responses" --- precisely the self-reinforcing cycle that makes pre-primary markets uniquely dangerous as manipulation vehicles.

A wealthy backer can place $5 million on a Senate candidate nine months before a primary, watch that candidate appear in national coverage as a "surging contender." The political outcome --- elevated name recognition, media legitimacy, donor attention --- is achieved whether or not the trade is profitable. This is manipulation whose cost is bounded by the size of the bet and whose benefit is unbounded in political terms. Again, this is not Hayekian price discovery, it is trading the map against the territory.^[4]

The attention-purchasing mechanism is further amplified by well-documented patterns in donor behavior. Research by Meisels, Clinton, and Huber finds that campaign donors systematically penalize moderate candidates and reward ideologically extreme ones, with the most ideologically activated donors showing the strongest responsiveness to perceived candidate momentum.^[5] This creates a compounding dynamic: a wealthy actor who pumps a pre-primary candidate's market odds does not merely generate generic media coverage --- they generate a viability signal that disproportionately activates the most ideologically extreme segment of the donor pool. The result is not just attention for one candidate but a systematic pull on the entire primary field. Pre-primary prediction markets therefore function not only as manipulation vehicles for individual campaigns but as structural mechanisms for distorting primary field composition in ways that advantage extreme candidates over moderate ones --- a harm that falls entirely outside any individual contract's public interest analysis but is a predictable consequence of permitting this contract category to exist.

This dynamic extends beyond domestic actors. Recent research by the Bitcoin Policy Institute documents a structurally analogous mechanism in a different domain: foreign-aligned organizations embedding themselves within authentic American political movements to amplify and redirect domestic sentiment toward foreign strategic objectives.^[6] Their analysis concludes that "legitimate local grievances can be identified, amplified, and redirected by organized actors whose objectives are not local, environmental, or even democratic" and that "the presence of authentic civic concern does not preclude the presence of foreign-aligned influence." The mechanism they document --- an organized, foreign-aligned actor amplifying genuine domestic sentiment at the moment of maximum downstream effect --- maps directly onto the pre-primary prediction market context. A foreign-aligned actor need not manufacture political support for a candidate from scratch. They need only place a large bet at the right moment: before the primary field has stabilized, before disclosure obligations attach, and before the Meisels et al. donor-activation dynamic has run its course. Genuine voter sentiment exists; the foreign actor's contribution is to distort its apparent scale and intensity through a prediction market that media organizations now treat as an objective signal. The Commission should treat this not only as a market integrity concern but as an election security concern. Pre-primary prediction markets create a legal, lightly surveilled, and highly leveraged instrument through which foreign-aligned actors can purchase influence over American primary elections --- with no contribution limits, no disclosure requirements, and no enforcement mechanism available to the Commission.

To be excruciatingly clear: Foreign adversaries interested in the implosion of the United States have an avenue to insert both extreme right wing and extreme left wing candidates into major races at the primary stage. This amplifies division exponentially by allowing folks on the left to claim that "the Right supports white supremacists" while the right can claim "The Left supports violent revolutionary marxists" even when neither claim is true (beyond at the extreme margins) -- the desired fruits for the adversary would be an unraveling of democracy as both sides act as if they face an existential crisis of American values.

The Commission's proposed factor framework does not currently account for this dynamic. The Commission should treat this attention-purchasing function as an independent public interest concern when evaluating pre-primary election contracts.

II. The Offshore Laundering Problem

The Commission cannot regulate offshore prediction markets, and those markets will exist regardless of what the CFTC does. This is sometimes offered as a reason to permit domestic markets --- better to have regulated price discovery than unregulated. I argue the opposite is true at the pre-primary stage.

A prohibited domestic market means that offshore manipulation has no legitimate domestic vehicle -- Mainstream media is unlikely to partner with, or even report results from, offshore unregulated markets. A permitted domestic market creates exactly such a vehicle. The mechanism is as follows:

An actor manipulates an offshore market --- say, moving a candidate's odds from 4% to 30%. Domestic market participants observe this discrepancy. Many will treat the offshore signal as genuine information, either because they believe offshore markets are less subject to political bias or simply because price discrepancies look like arbitrage opportunities. They buy the domestic contract up toward the offshore price. The regulated, CFTC-sanctioned market now reflects the manipulated offshore signal. News organizations report the regulated market number as fact. The manipulation is complete, laundered through the legitimacy of a domestic regulated exchange.

Note that this cannot be addressed by disclosure requirements or surveillance, because the source of the distortion is outside the Commission's jurisdiction.

Packin and Rabinovitz identify the precise asymmetry at stake: "Foreign nationals are prohibited from contributing to US elections, yet may trade political event contracts referencing electoral outcomes... Residents directly affected by election results may thus be excluded while cross-border actors face fewer constraints. This asymmetry is compounded by thin liquidity: Low trading volume and few participants mean even small trades can substantially shift prices and manufacture the appearance of consensus, influencing political expectations rather than simply reflecting them, turning electoral 'forecasts' into instruments of influence rather than collective judgment." A domestic regulated market does not solve this problem; it provides the instrument of influence with a CFTC imprimatur.

One apparent tension in this analysis deserves explicit resolution, because conflating two distinct mechanisms obscures the argument. Cross-market arbitrage --- buying a contract on one exchange where it is priced lower and simultaneously selling the same contract on another where it is priced higher --- is a genuine carry trade: the outcome of the underlying event is irrelevant because the two positions offset. The profit is locked in at execution. This mechanism is efficient, rapid, and attracts well-capitalized traders precisely because it carries negligible risk. A manipulated offshore price therefore creates a cross-market spread that riskless carry capital will close by buying the domestic contract upward. Correction-based arbitrage --- betting against a price you believe misrepresents the true probability --- is categorically different. A trader who believes a candidate has a 20% chance of winning when the market prices them at 40% is not executing a carry trade. They are taking a directional, outcome-contingent position whose payoff depends entirely on being right. If the candidate wins --- through luck, genuine viability the trader underestimated, or continued manipulation --- the trader loses. This position is subject to risk constraints detailed in Section III, and sophisticated capital has rational reasons to avoid it. There is therefore no contradiction in asserting simultaneously that offshore manipulation will propagate rapidly into domestic markets (carry trade arbitrage works) and that manipulated domestic prices will not be corrected by informed capital (opinion-based arbitrage does not work automatically). The two claims invoke different mechanisms. Markets are efficient at the former and structurally impaired at the latter.

This dynamic represents an additional public interest factor the Commission should consider: whether permitting a domestic regulated market creates an amplification mechanism for offshore manipulation that would not otherwise exist. The Commission has authority to deny manipulative actors a domestic vehicle even when it cannot regulate the offshore source.

III. The Arbitrage Correction Argument Fails Here

The standard defense of prediction markets against manipulation claims is that mispriced contracts attract corrective capital: if a contract is priced wrong, sophisticated actors will bet against it until the price converges to the true probability, hence "arbitraging" the difference. This argument is often treated as decisive. But this argument fails to consider fundamental differences in arbitraging spot markets and futures markets.

Sophisticated capital --- hedge funds, quantitative traders --- does not maximize expected value per bet. It maximizes long-term portfolio growth, which requires sizing positions using something like the Kelly Criterion^[7] and accounting for covariance with other portfolio positions. A fund that believes a candidate has a 20% chance of winning a primary when the market prices them at 40% will not necessarily bet against that contract. The expected-value argument for the trade may be strong, but the Kelly-optimal position size depends on how the trade covaries with the rest of the portfolio, what the variance of the outcome is, and how much of the fund's risk budget is already committed to correlated positions.

The result is that the actors most capable of correcting manipulation are also the actors most likely to pass on the trade. What remains in the market are retail participants without the tools to identify manipulation and sophisticated actors doing the manipulating. The self-correcting property that makes prediction markets valuable in liquid, information-rich environments is structurally absent in prediction markets.

This is not a theoretical concern. The Vanderbilt study discussed in Section VII found that arbitrage opportunities across platforms actually peaked in the final two weeks before the 2024 election --- the opposite of what efficient market theory predicts. Correction did not occur even when the opportunities were obvious and the stakes were highest.

IV. The Claimed Predictive Accuracy of These Markets Is Empirically Unsupported and Actively Misleading

A persistent claim in favor of election prediction markets is that they "outperform polls." This has become something close to received wisdom in certain media and policy circles. A careful look at the evidence --- including a recent study specifically designed to test this claim --- suggests it is not supported, and that the framing itself reflects a fundamental confusion about what polls are and what prediction markets do.

The comparison to individual polls is the wrong benchmark.

A single poll is a sample of voter intentions at a moment in time. It does not give a probability of election outcome. Comparing a prediction market's probability estimate to a single poll's margin is not a meaningful test of forecasting accuracy --- it is a comparison between a probabilistic forecast and a raw data point. The relevant comparison is to aggregated, model-adjusted forecasts such as those produced by major election modelers (for example 538 or Princeton Election Consortium). On that comparison, the case for prediction market superiority largely disappears.

The Vanderbilt study.

In January 2026, researchers Joshua Clinton and TzuFeng Huang at Vanderbilt University published "Prediction Markets? The Accuracy and Efficiency of $2.4 Billion in the 2024 Presidential Election," analyzing over 2,500 individual prediction market contracts across four major platforms during the final five weeks of the 2024 campaign.^[8] Their findings directly contradict the industry narrative:

• Polymarket --- the largest platform by volume, with over $2 billion in 2024 election trading --- achieved only 67% accuracy.
• Kalshi achieved 78% accuracy.
• PredictIt, which operates under a per-contract bet cap, achieved 93% accuracy.

The inverse relationship between capital concentration and accuracy is not a minor footnote. It is the central finding. The platform where wealthy actors could place unlimited bets performed worst. The platform with structural limits on position size performed best. The authors conclude that "high trading volume does not guarantee superior forecasting accuracy" and that concentrated capital can act as a pollutant to price discovery rather than a signal. This finding is directly consistent with the manipulation dynamics described in Sections I through IV.

Arbitrage failed when it mattered most.

The Vanderbilt study found that identical contracts showed divergent prices across exchanges simultaneously --- a basic arbitrage opportunity that should not persist in an efficient market --- and that arbitrage opportunities peaked in the final two weeks before Election Day, precisely when they should have been lowest. Markets were growing less efficient over time, not more.^[9]

The 2024 presidential race.

Polymarket had Trump at approximately 57% on election eve; major poll aggregators had the race at approximately 50-50. Trump won. The claim that Polymarket therefore "beat" the aggregators is being made as if a 7-percentage-point difference on a single binary outcome constitutes evidence of superior calibration. It does not. With one data point, you cannot statistically distinguish a 57% model from a 50% model. The sample sizes required to make confident comparative claims about probabilistic forecasters --- across elections that happen every four years, in conditions that change each cycle --- simply do not exist.

But the story is much worse: Polymarket gave Donald Trump a 27% chance of winning the popular vote, which he won. This is a significant miss. Prediction markets leaned Trump, correctly predicting a toss-up electoral college, but missed badly on popular vote. This is not a claim to keen insight.

Where prediction markets did shine was after the votes were cast and the results were being tabulated. Many people observed that as the results trickled in from precincts across the country, prediction markets converged quicker than say, the New York Times needle. This points to decentralized quants working in aggregate can perform faster than a single desk such as the New York Times, but does not suggest any sort of prescience whatsoever or predictive power about the main question: "how will voters vote?"

The broader historical record.

Prediction markets had Hillary Clinton at approximately 80% on election eve in 2016. They had Remain winning Brexit. For a mechanism the industry claims efficiently aggregates all available information, these failures are not minor calibration errors. They are fundamental misses at the moments of highest stakes --- and they are not being honestly accounted for in the industry's claims of superior accuracy.

What this means for the Commission's analysis.

The Commission's proposed framework asks whether event contracts provide meaningful price discovery and information aggregation. The empirical evidence on election markets specifically says no: accuracy is inversely correlated with the size of capital deployed, arbitrage fails to correct inefficiencies even over weeks at the highest-stakes moments. The prediction market industry has been extraordinarily effective at marketing its products as truth-telling machines. As Packin and Rabinovitz observe, "scientific papers demonstrating that diverse, independent judgments improve forecasting accuracy are repurposed as marketing collateral for systems designed to maximize engagement, not epistemic rigor." The Commission should not mistake effective marketing for empirical validation.

V. A Proposed Bright Line

The Commission's proposed framework applies a multi-factor balancing test. I support this approach for general election contracts, where the information environment is richer, the participant pool is larger, and the manipulation dynamics are less acute.

For pre-primary contracts, however, I urge the Commission to adopt a bright-line prohibition: no event contracts settling on the outcome of any election, or on any measure of candidate viability, may be listed until after primary voting in the relevant jurisdiction has concluded.

The conclusion of the primary is an unambiguous, publicly verifiable date that applies uniformly across all candidates and jurisdictions. It marks the point at which the field is set, the information environment becomes substantially richer, and the legitimate price discovery and hedging functions prediction markets can serve become more plausible.

This bright line would not prevent general election markets. It would specifically eliminate the period of maximum manipulation risk --- the pre-primary window where information is most concentrated, markets are most illiquid, and the political return on manipulation is highest.

VI. Response to Anticipated Counter-Arguments

"These markets help voters make better decisions."

There is no evidence that prediction market odds at the pre-primary stage reflect genuine candidate viability rather than the financial preferences of the betting public, which skews heavily toward wealthy, financially sophisticated, and demographically unrepresentative participants. A candidate leading pre-primary prediction markets is not meaningfully more likely to win the general election than a candidate leading in equivalent-stage polling.

"Prohibition will just push activity offshore."

Yes --- and that is preferable to a domestic market that launders offshore manipulation into mainstream media coverage. Offshore markets without a domestic regulated counterpart have no legitimate vehicle to influence domestic discourse at scale.

"The Commission can address manipulation through surveillance."

The manipulation described here often involves no violation of domestic law --- a wealthy individual placing large legal bets is not subject to surveillance under existing frameworks. And the offshore laundering mechanism is by definition outside the Commission's surveillance reach.

"Markets always converge to truth through arbitrage."

As explained in Section III and confirmed by the empirical findings in Section IV, this does not hold in practice for election prediction markets. The Vanderbilt study directly observed arbitrage opportunities increasing rather than decreasing over time, at the moment of highest stakes.

Conclusion

Pre-primary election contracts fail the public interest factor tests the Commission proposes. They do not provide genuine price discovery. They are uniquely susceptible to manipulation by wealthy actors whose political return on investment exceeds any financial loss on the trade. They create an offshore laundering mechanism that the Commission cannot surveil or regulate. They enable head fakes, pump-and-pump schemes, and attention-purchasing strategies with no analogue in legitimate financial markets. And the corrective mechanism the industry relies upon --- arbitrage by sophisticated capital --- is structurally absent because the actors most capable of correcting manipulation have every rational reason to stay away.

I urge the Commission to adopt the following specific measures in the final rule:

1. Adopt a bright-line prohibition on pre-primary election contracts, defined as any contract settling on election outcomes or candidate viability before primary voting in the relevant jurisdiction has concluded. Primary voting day is an unambiguous, uniformly applicable, publicly verifiable trigger that eliminates the manipulation risks specific to the pre-primary period without restricting general election markets.

2. Add offshore market amplification as an explicit public interest factor under §§ 40.11(a)(5) and (a)(6), requiring the Commission to consider whether permitting a domestic contract creates a laundering mechanism for manipulation originating in unregulated offshore markets.

3. Add media amplification as an explicit public interest factor, requiring the Commission to consider whether prediction market prices in a given contract category are routinely reported by news media as signals of real-world probabilities, and whether that reporting creates manipulation incentives disproportionate to the contract's informational value.

4. Require empirical substantiation of price discovery claims for election contracts. Where an applicant claims that a proposed election contract serves a price discovery function, the Commission should require evidence sufficient to support that claim, including calibration data, and should not treat the general reputation of prediction markets as a substitute for contract-specific evidence.

This rule is narrow, administrable, and directly responsive to the public interest concerns the proposed framework identifies. It would not prohibit general election markets. It would deny manipulative actors the domestic regulated platform they need to launder political influence into mainstream discourse.

Key Sources Referenced

• Hayek, F.A. (1945). The use of knowledge in society. American Economic Review, 35(4), 519--530.

• Thorp, E.O. (2008). The Kelly criterion in blackjack, sports betting, and the stock market. In S.A. Zenios & W.T. Ziemba (Eds.), Handbook of Asset and Liability Management (pp. 385--428). North-Holland.

• Packin, N.G. & Rabinovitz, S. (2026). Prediction markets as a public health threat: Scientific framing, gambling-like design, and regulatory gaps create risks of a new behavioral addiction and democratic manipulation. Science, 392(6795), 257--260. https://doi.org/10.1126/science.aee3932

• Meisels, M., Clinton, J.D., & Huber, G.A. (2024). Giving to the extreme? Experimental evidence on donor response to candidate and district characteristics. British Journal of Political Science, 54(3), 851--873. https://doi.org/10.1017/S0007123423000650

• Clinton, J.D. & Huang, T. (2026). Prediction Markets? The Accuracy and Efficiency of $2.4 Billion in the 2024 Presidential Election. Vanderbilt University. https://ideas.repec.org/p/osf/socarx/d5yx2_v1.html

• Lyman, S. et al. (2026). China-aligned ground game: Stalled or blocked $23.6 billion in American AI infrastructure. Bitcoin Policy Institute. https://www.btcpolicy.org/articles/china-aligned-ground-game-stalled-or-blocked-23-6-billion-in-american-ai-infrastructure

• Bitcoin Policy Institute (2026). Foreign influence campaign against American AI, Part II: Singham ground game. https://www.btcpolicy.org/articles/foreign-influence-campaign-against-american-ai-part-ii-singham-ground-game