Prediction markets as a public good

In game theory, a situation whereby one party knows something relevant to a situation that another party doesn’t is called “having asymmetrical information”. This knowledge comes with advantages, such as the possibility to update priors before others (previous assumptions), act more quickly on new information, and greater possibility for maximum rent extraction. Take financial markets, perhaps the thing most often associated with the common understanding of asymmetrical information. Typically, markets are opposed to information imbalances — they diminish trust in the market, which in turn can lead to liquidity crises. They must often tolerate these imbalances, however, due to their ubiquitous nature. But what happens when asymmetrical information is a core component of a market’s model?

Enter the crypto-based, hyper-gamified prediction markets of today. They’re relatively easy to access, typically requiring only an account, the completion of a short know-your-customer (KYC) verification based on location, and a deposit of funds in cryptocurrency or fiat. As advertised, they provide an outlet for free individuals to gamble on current events like sports, celebrity relationships, and even political events. However, their popularisation has been marred by high-profile cases involving the use of insider information. Insider information, a type of asymmetric information advantage, is distinct in three ways: it involves information that is non-public, reasonably price sensitive, and precise.

Take a government official. An unannounced policy decision is, by definition, non-public. The same can go for regulatory rulings, troop movements, intelligence assessments, etc. In a typical market environment the effect of such a decision on price is at least variably indirect. Those who wish to leverage said insider information must still translate the proposed effect of the policy on a financial instrument through which they can make money. I think of this as a kind of “friction”, or how difficult it is for an insider information-leveraging agent to convert information into profit. Modern prediction markets collapse this chain by offering binary contracts on the choices of policymakers, reducing a substantial amount of volatility associated with operationalising insider information.

This article argues that the gamification of prediction markets, in their current form, presents a category of national security risk that existing financial regulation is poorly equipped to address. Most salient is not the familiar problem of officials profiting from market-moving information, but rather the specific structural shift these markets introduce: by turning policy decisions into binary, readily executable contracts, they offer officials stronger incentives and a more streamlined process to benefit monetarily from decisions that should otherwise be made with the goal of maximising social welfare. The argument develops in three parts. The first sketches the relevant regulatory and legal environment that precedes the modern prediction market environment. The second places focus on the market environment itself and explores relevant counterarguments posed by some who support the injection of insider information into prediction markets. The third considers policy angles: the existing national security vulnerabilities, where existing securities, commodities, and anti-corruption frameworks fall short, and what a more coherent policy response may look like.

The lead-up

To understand the modern regulatory landscape that prediction markets exist in, you must start in 1992, the year that the Professional and Amateur Sports Protection Act (PASPA) was implemented (102d Congress, 1992). This piece of legislation prohibited state authorisation of sports gambling, with few exceptions for states like Nevada. However, there is much speculation about whether this piece of legislation was effective or did more harm than good by driving the market for sports wagering underground. Further, the time between the beginning of PASPA and the mid-2010s saw a significant shift in the attitudes towards the legalisation of sports betting (Minton, 2018). This is why in 2018, following the case of Murphy v. NCAA, the federal ban on state-implemented betting regulation was lifted. What drove this was concern over conflict with the Tenth Amendment to the US Constitution’s mandate that “powers not delegated to the [federal government] … are reserved to the states [or] the people” (Harvard Law Review, 2018).

Following the repeal of PASPA, the sports gambling landscape changed considerably. Many states moved to legalise various forms of sports betting, including those related to professional and even collegiate sports. Today, 11 states (incl. DC) prohibit the practice outright, while 40 states allow a mixture of direct, online, and prop betting practices (Sutton & Goldman, 2025). However, the landscape did not remain static for long after 2018. Alongside the introduction of new state-level legislation came the typical experimentation one might expect from opportunistic actors in a tumultuous market. For example, the term “sweepstakes casino” was popularised, referring to sites that offer traditional house games but which do not use “real money”, opting instead for a dual-currency system where virtual cash can be bought or exchanged for fiat. By dodging the new regulation, these types of casinos bypassed typical licensing and tax structures, saving millions (American Gaming Association, 2024, p. 1).

Exploring the modern prediction market calculus

Kalshi and Polymarket, founded in 2018 and 2020 respectively, are the two biggest names in the modern prediction market landscape. While the distinction between the two isn’t terribly important for the purposes of this article, it is worth mentioning their structural dissimilarities: Kalshi operates on a fiat-based, more tightly regulated model, while Polymarket (legally based in Panama) primarily uses cryptocurrency and allows for a wider range of topics to wager on (Metzger, 2026). Kalshi and Polymarket are both regulated by the Commodity Futures Trading Commission (CFTC). However, there is a caveat in the latter’s case whereby only its US arm (QCX LLC/Polymarket US) is subject to direct oversight.

These platforms have demonstrated that what they offer is ripe for insider trading, and reports have revealed “dozens” of instances where such trading was flagged (Thompson & Yaffe-Bellany, 2026). So far, it hasn’t been the biggest issue for them — Polymarket’s CEO, Shayne Coplan, even alludes to the necessity of insider trading in prediction markets. In an interview with 60 Minutes late last year, Coplan discusses the derived accuracy of the various available markets’ predictions, using the performance vis-à-vis the 2024 US election as a backdrop. Polymarket generally outperformed many established polls, such as those conducted by Siena College and Quinnipiac University. Coplan’s argument is essentially that services like his provide a public good, where insiders are compensated for providing their advanced information by way of market signals (60 Minutes, 2025).

Kalshi is more hostile towards the idea of hosting those with insider information, regularly posting ads which announce its aversion to it and compliance with the CFTC. However, this may not matter in the long run. Robin Hanson, the economist responsible for conceptualising and popularising the modern prediction market, tends to echo Coplan’s sentiments. He writes in his original 1999 column: “Such markets pool the information that is known to diverse individuals into a common resource, and have many advantages over standard institutions for information aggregation, such as news media, peer review, trials, and opinion polls” (Hanson, 1999, p. 16). As such, modern prediction markets provide value in three ways: monetary value to those that are able to consistently beat the market, entertainment value to all users, and, as proponents would argue, societal value in the ability to price in insider information more quickly than any news agency or pollster. I argue that the latter of the three is most crucial to the service’s long-term viability and survival.

While this argument isn’t without its legs, there are three reasons why it ultimately falls flat. First, these services only realistically operate as information aggregates when a significant number of individuals hold domain-specific/tacit knowledge. And while a single individual with the same knowledge can produce the same odds with their wager as a group of dispersed individuals (given a proportionally sized wager), this is a much weaker justification given that the market is simply front-running an announcement rather than aggregating many small signals that would’ve otherwise gone unnoticed. Second, the fact that the operationalisation of insider information has been empirically shown to be harmful to equity markets raises the bar for the proponents of prediction markets as a benefit to social welfare. While prediction markets don’t have to deal with a house against whom players bet or the price of an underlying asset, they are still subject to issues of liquidity. A popular loss of trust by retail investors [bettors] would still widen the bid-ask spread substantially and result in market-making becoming intractable.

Third, and perhaps most interesting to me, is the exploration of the counterfactuals. Two are worth distinguishing; defenders of prediction-market insider trading seem to slide between them without acknowledging that they answer different questions. The first asks whether a prediction market with insider trading is more efficient than one without. Nobody would seriously argue that more relevant and reliable information ingested by the market isn’t likely to produce a more accurate price (though information varies in quality). But this comparison is not hypothetical, and its answer is partially supplied by the industry itself. Kalshi’s entire regulatory strategy is built around excluding insider participation, and its existence demonstrates that the choice between a market with insiders and one without is an explicit commercial decision. The fact that this major US contract market has chosen the latter suggests that the platform doesn’t believe that the within-market efficiency gain survives contact with regulators, the legislature, public legitimacy, etc. The second counterfactual is more demanding: a prediction market with insider information versus no prediction market at all in which the public relies solely on media, agencies, and institutions to source their data. Existing forecasting institutions are gatekept — intelligence assessments are classified, central bank communications are ambiguous, regulators disclose stuff on their own timetable, and news agencies operate under editorial constraints that leave entire categories of signals unreported. A continuously updated, numerically legible probability accessible to anyone with an internet connection is, on its face, a democratising improvement on that landscape. The price, however, can reflect a range of things besides the underlying information, including liquidity, market structure, speculation, and attempts at manipulation. That being said, the accessibility argument is tangible, and an honest assessment of the public-good case must concede it in my opinion. What it cannot do, however, is carry the weight that defenders ask of it. The public benefit may justify prediction markets, but it does not by itself justify allowing insiders to trade in those markets. Therefore, one can sever the accessibility defence from the insider-trading defence. Once this is done, the case for permitting officials to trade on their own decisions must stand on its own terms.

What the theoretical approach to these situations reveals is that, at the very least, the burden for proponents of modern prediction markets as social welfare contribution machines is perhaps higher than it seems on the surface. What remains is a series of open questions: what must prediction market services comply with? Due to the flexibility and range of options when it comes to creating futures contracts, must the punishment for those found to be leveraging insider information be greater than the historical standard? And, at the end of the day, should decision-makers be allowed to quietly leverage their insider information if it means the public has the chance to pick up on those signals prior to any official announcement?

Policy implications and recommendations

Before diving into the policy implications and suggesting some recommendations, first consider two high-profile cases associated with the use of insider information on modern prediction markets:

1.    In late 2025, an individual with editing rights to the Institute for the Study of War’s (ISW) live Russo-Ukrainian War map made an edit indicating that Russian forces had taken the city of Myrnohrad. A few hours later, they had won hundreds of thousands of dollars from a bet placed earlier on Polymarket; the service relies on the ISW for the settling of its conflict-related wagers (Cleveland-Stout, 2025).

2.    A US soldier, who made a similar amount of money betting on the US’ brief invasion of Venezuela in January 2026, was recently charged with the use of classified information for profit (Office of Public Affairs, 2026).

These are the pioneers. Relatively unconnected but able to leverage insider information with adequate speed, they take advantage of the regulatory gray zone that is modern prediction markets and profit before anything can be done. So, the question stands: what happens when a connected individual submits the same wager? Some would argue this issue already runs deep, all the way to the level of the current administration (see: Michaels et al., 2026). Further, the cases reveal various vectors of security concern: signal leakage, modern prediction market services as a target for foreign intrusion, and the distortion of public decision-making to benefit market outcomes rather than social welfare.

To understand what could be done, I’ll first canvas what already exists. The CFTC (through the Commodity Exchange Act) already holds blanket jurisdiction over event (futures) contracts, though runs into issues of enforcement (Better Markets, 2026). The SEC also holds some responsibility, though it is limited with respect to event swaps. Since 2012, the STOCK Act has prohibited members of Congress and their families from leveraging insider information, though calls for a blanket ban on stock ownership become louder by the day (119th Congress, 2025). Taken together, they provide suitable mitigation but fall short in the way of preventing the harms associated with insider trading at large.

The main theoretical issue worth combatting is the lowered friction offered by modern prediction markets. Screening for decision-exposed individuals and those connected may be worth the effort and relatively easy to implement for some models, such as that already employed by Kalshi. Position limits may also be tractable, as they reintroduce a degree of friction for larger investors. Finally, greater platform awareness of the contract design is necessary. While contracts would remain binary, those whose outcomes are concentrated on a single or small set of decision-makers should be subject to heightened scrutiny or prohibited altogether. This could also include delays in payouts in the case of a concentrated decision.

The next recommendations concern the counterfactuals. Following the discussion in the latter half of the second section, this article is actually bound to accept the theoretical argument that prediction markets are a valuable tool for tacit information aggregation; effective regulation then attempts to preserve this aspect while stripping malicious insiders from the pipeline. First and foremost, the CFTC requires updated jurisdiction clarification over event contracts. This will reduce ambiguity substantially. This, paired with some of the transparency requirements outlined in the previous paragraph, creates a suitable environment to begin rooting out and disincentivising connected individuals. Another thing to consider is the competition space — platforms will find themselves relatively overregulated and outcompeted by offshore competitors, so protections for the regulated market are required to mitigate this.

While these proposals are, in my opinion, promising, there exist limitations that are integral to the space. When it comes to platforms like these — many of which are based outside of the US, with crypto, and with perhaps reduced moral reservations — regulation can sometimes be impossible and requires a careful approach. Two options are relevant. The first is international coordination, though many smaller nations whose primary draw is their tax-haven status are unlikely to approach the table. The second, and the one that is likely more tractable in the short term, is the tightening of barriers to entry (like KYC) to reward those domestically compliant and create a difficult user process for those non-compliant.

As seen with PASPA, gambling and markets generally are a hard area to regulate. New technologies facilitate underground markets, there remains an aspect of addiction to the gamification, and their access is connected to ideas of libertarianism and free trade. This article approaches the issues of modern prediction markets with this in mind: this form of gambling is likely here to stay, and critics of their abolishment raise valid points about their effect on social welfare. As such, the mitigations address those who seek out these services for concentrated personal gain while hardening the sector against external actors who face little worry about the composition of their books’ trading volumes. In doing so, it presents an interesting suite of considerations for policymakers as they continue to deliberate over how to approach this issue.

References

102d Congress. (1992). Professional and Amateur Sports Protection Act (S.474). Professional and Amateur Sports Protection Act. https://www.congress.gov/bill/102nd-congress/senate-bill/474/text

119th Congress. (2025). End Congressional Stock Trading Act (H.R.1908). https://www.congress.gov/bill/119th-congress/house-bill/1908/amendments

60 minutes. (2025, December 1). Polymarket CEO says his prediction market is “the most accurate thing we have as mankind right now.” [Video]. YouTube. https://www.youtube.com/watch?v=ZA5bY4K9XPs

American Gaming Association. (2024). Regulatory Vigilance Critical to Ensure “Sweepstakes” Don’t Threaten Consumers and Undermine Gaming Regulation. https://www.americangaming.org/wp-content/uploads/2025/02/Sweepstakes-Memo-Final.pdf

Better Markets. (2026, May 13). “Prediction Markets,” Gambling, the CFTC & Regulation: Facts, Fiction & The Law. https://bettermarkets.org/analysis/prediction-markets-gambling-the-cftc-regulation-facts-fiction-the-law/

Cleveland-Stout, N. (2025, December 16). Think tanker altered Ukraine war map before big Polymarket payout. Responsible Statecraft. https://responsiblestatecraft.org/isw-polymarket-ukraine-war-map/

Hanson, R. D. (1999, May/June). Prediction Markets. IEEE Intelligent Systems, 16. https://mason.gmu.edu/~rhanson/decisionmarkets.pdf

Harvard Law Review. (2018, November). Murphy v. National Collegiate Athletic Association. https://harvardlawreview.org/print/vol-132/murphy-v-national-collegiate-athletic-association/

Metzger, B. (2026, April 4). The 4 biggest differences between Kalshi and Polymarket. Business Insider. https://www.businessinsider.com/polymarket-kalshi-prediction-market-key-differences-regulation-trading-crypto-2026-3

Michaels, D., Scechtman, J., & Schwartz, B. (2026, May 16). Suspicious Betting in Washington Is on the Rise—and Authorities Are Playing Catch-Up. Wall Street Journal. https://www.wsj.com/finance/regulation/government-prediction-markets-betting-2126dcb3

Minton, M. (2018). Legalizing Sports Betting in the United States (243). Competitive Enterprise Institute. https://www.cei.org/wp-content/uploads/2018/03/Michelle-Minton-Legalizing-Sports-Betting-in-the-United-States-1.pdf

Office of Public Affairs. (2026, April 23). U.S. soldier charged with using classified information to profit from prediction market bets. Department of Justice | United States Department of Justice. https://www.justice.gov/opa/pr/us-soldier-charged-using-classified-information-profit-prediction-market-bets

Sutton, B., & Goldman, C. (2025, November 16). U.S. sports betting: Where all 50 states stand on legalizing online sports betting sites, proposed legislation. CBS Sports. https://www.cbssports.com/betting/news/u-s-sports-betting-where-all-50-states-stand-on-legalizing-online-sports-betting-sites-proposed-legislation/

Thompson, S., & Yaffe-Bellany, D. (2026, May 13). Dozens of Polymarket Bets Show Signs of Insider Trading, The Times Finds. The New York Times - Breaking News, US News, World News and Videos. https://www.nytimes.com/2026/05/13/technology/polymarket-insider-trading.html