As the conflict between prediction markets and the gambling sector continues, two documents issued from opposing sides this month may hint at how the regulatory approach around these controversial platforms could develop.
The discussion was initiated by Elie Mishory, who previously held the role of Kalshi’s chief regulatory officer before taking a position as senior advisor to Paul Atkins, head of the U.S. Securities and Exchange Commission, earlier this year. Before Kalshi, Mishory also worked as a lawyer for both the Commodity Futures Trading Commission and the Internal Revenue Service.
Mishory released a research paper this month titled “Information Asymmetry and Event Integrity in Prediction Markets: A Framework for the CFTC”, where he tackled and organized the concerns around insider trading and event manipulation on prediction platforms. These issues have made national headlines recently, including the detention this week of a U.S. soldier linked to insider transactions concerning the removal of Venezuelan President Nicolas Maduro.
In Mishory’s opinion, “material non-public information” is a wide concept that ought to be split into various categories from a regulatory standpoint. Moreover, actual event manipulation, where a participant influences the event being traded on, is a completely distinct matter and must be handled separately, he contended.
“The entire purpose of an event contract is to convert scattered, unevenly held knowledge about a future outcome into a marketable price,” Mishory wrote. “So the query is not whether asymmetry exists. The query is which kind of asymmetry the law should consider harmless, which kind it should deem a disclosable market risk, and which kind it should label as abusive.”
Various categories of informational edge
When news of scandals and insider trading on prediction markets emerges, Mishory said, the immediate impulse is to pinpoint individuals tied to the underlying event and probe them. But Mishory claims this perspective is wrong, and instead regulators ought to set up a ranking of informational advantages. The main priority, he wrote, should be to “determine which informational advantages are inappropriate and ought to be prohibited” and which should not.
“The correct method is to concentrate on the information itself,” Mishory stated. “What type of information is it? Was it taken? Was it selectively disclosed because of a connection? Is it merely the trader’s own non-public awareness of facts arising from the trader’s direct involvement in the underlying event or activity? Could it have been acquired through normal investigation, synthesis, observation, or expertise? Those are the questions that count.”
Through this perspective, Mishory outlined four categories of informational advantages that prediction market participants might utilize, all of which he argued ought to be treated distinctly from a regulatory viewpoint:
- Use of non-public information that is stolen or used unlawfully
- Use of non-public personal information, or information obtained directly by participants in the underlying event
- Use of non-public third-party information, or information obtained indirectly or through ties to the underlying event
- Use of skill-based information that is gathered through the trader’s own analysis or research
Positive, negative, and in between
The unlawful use or theft of non-public information ought to be straightforward and explicitly banned, Mishory argued. In these instances, individuals who are “monetizing information that the trader had no legitimate right to use for personal trading” are operating outside the limits of acceptable market knowledge. Any exchange that permits such trading is rewarding crime and penalizing lawful traders, the paper argues.
“Once trading on stolen information is accepted, the market starts rewarding disloyalty, breach, and informational theft,” Mishory wrote. “That is a clear market-integrity issue.”
On the other hand, the use of skill-based information ought to be explicitly permitted, Mishory argued, since it was obtained legally through the trader’s own efforts.
The middle two categories – non-public personal information and non-public third-party information – are where matters become challenging. As Mishory argues, there is an inherent level of information asymmetry that cannot and should not be eliminated from trading. Financial markets and exchanges cannot demand that all participants trade with identical information, as it would hinder price discovery and liquidity.
Regarding non-public third-party information, Mishory argued that it ought to be allowed, but only with appropriate disclosures. This is because the potential edge for traders could be hard to detect but substantial enough to merit an unfair advantage. Essentially, the paper argues that traders should enter these markets at their own risk, knowing that other participants might have third-party information.
“The proper disclosure is market-level disclosure,” Mishory wrote. “The exchange’s rules, contract terms, and customer-facing risk disclosures can clarify that certain markets may include participants trading on relationship-based, non-public informational access. That allows market participants to evaluate the nature of the venue without imposing a universal reveal-your-edge rule.”
Direct knowledge allowed for trading?
Finally, Mishory asserts that non-public personal information, gathered from direct event participants, should “generally be permitted without legal prohibition and without any special market-specific disclosure obligation beyond the ordinary disclosures that markets already make”. Here too he argues that direct participants trading with their own information has always been a feature of financial markets, similar to executives trading their company’s stock, for instance.
“The presence of traders with their own non-public information, including their own MNPI, is not an unusual distortion of a CFTC market,” Mishory wrote. “It is part of the structure.”
All four of these informational categories are separate from actual event manipulation, which, in this debate, applies mainly to sports. Event manipulation damages public market trust and should lead to individual bans, but it should not be viewed as an informational issue, Mishory said.
He argued that “person-based prohibitions can be highly effective against outcome manipulation because the ability to distort the event often does attach to a category of participants. Players, coaches, referees, campaign officials, corporate decision-makers, and similar actors may pose special event-integrity risks because of what they can do, not merely because of what they know.”
Using Mishory’s framework to analyze the Maduro scandal, the U.S. special forces soldier Gannon Ken Van Dyke executed the trades using non-public personal information, which in itself would be legal. But given his ability to influence the event outcome by participating in the actual raid, he should have been barred from trading, based on the proposed breakdown.
Russell: taxonomy offers no help for enforcement
After the release of Mishory’s paper, a response was issued shortly afterward by Jon Russell, a longtime sports betting executive who previously worked for William Hill, Ladbrokes, and Betway. Russell’s response, titled “From Taxonomy to Detection”, was “not a contradiction” of Mishory’s proposal.
Rather, Russell aimed to examine Mishory’s framework from an integrity and enforcement perspective and apply it to real-world situations. Overall, he concurred with Mishory’s information taxonomy and the categories it establishes. But that framework hardly addresses the regulator’s task of identifying and stopping bad actors.
“The taxonomy tells you what you are searching for. It does not tell you how to locate it,” Russell stated.
Russell used examples to demonstrate how “information seeps” into betting and trading markets from multiple sources. For instance, an employee on a game show might have non-public information about which contestant will be eliminated, but so might an employee from a third-party phone company that compiles the voting data, and neither are easily connectable to each other.
“The market sees only an evolving price and an unusual order flow pattern,” Russell wrote. “Non-public status is not something the surveillance system can observe directly. It has to be inferred, imperfectly, from what the market should have known at the moment the bet was placed. Making that inference problem explicit is what any workable detection architecture has to start from. And it is the step that sits between Mishory’s framework and its enforcement.”
Prediction markets revisiting resolved issues?
Much of Russell’s rebuttal centers around the idea that prediction markets are reinventing the wheel of regulated sports betting. The CFTC is embarking on a prediction market rulemaking process under the direction of Chairman Michael Selig. Selig largely deferred questions about prediction markets from House Ag Committee members last week, pointing to said rulemaking.
But might that process just be an exploration of ideas and concepts that regulated sports betting has already addressed and tested, especially with regard to event integrity? Selig is the CFTC’s lone sitting commissioner, and the agency is racing to staff up after a wave of departures from the previous administration.
There are questions as to whether this typically niche agency, which might gain additional duties if a crypto framework bill is passed, has the ability or capacity to properly surveil prediction markets on everything from finance and politics to sports and pop culture.
“Mishory has built the legal architecture. The detection infrastructure that makes it enforceable in practice – data feeds, alert logic, information-sharing agreements with event owners, investigative protocols – is the piece that is almost entirely absent from the current prediction market policy debate,” Russell wrote. “That infrastructure exists in regulated sports betting. It has been constructed over twenty years through exactly the kind of trial and error that prediction markets are now beginning.”