The Taxman Cometh
What Every Dollar Wagered on a Prediction Market Costs the States That Will Never See It
“The art of taxation consists in so plucking the goose as to obtain the largest possible amount of feathers with the smallest possible amount of hissing.”
— Jean-Baptiste Colbert, Finance Minister to Louis XIV
The Architecture of Gaming Taxation
The American system of gaming taxation is a monument to improvisation. There is no unified federal gambling tax. Instead, there is a 0.25 percent federal excise tax on authorized wagers and a 2 percent tax on unauthorized ones — a distinction that, in the prediction market context, carries implications no one in Congress anticipated when the rates were set.
On top of this federal floor, each state has built its own revenue extraction apparatus, and the variation is staggering. Nevada taxes sports betting operators at 6.75 percent of gross gaming revenue. Iowa matches it. Pennsylvania charges 36 percent. New York demands 51 percent. The gap between Nevada and New York is not a policy difference — it is a philosophical chasm between a state that views gambling as a legitimate industry and one that views it as a sin to be taxed into near-unprofitability.
For the individual bettor, the federal government takes 24 percent withholding on winnings exceeding $5,000 that also exceed 300 times the original wager.
Beginning in 2026, W-2G reporting thresholds consolidated to $2,000 for many gambling types. And in a change that received almost no public attention, starting in 2026, gamblers can deduct only 90 percent of their losses against winnings — down from 100 percent.
This is not a trivial adjustment. For a prediction market trader who wins $50,000 on one contract and loses $50,000 on another, the math no longer zeroes out. The IRS treats the winner and the loser as different people, even when they are the same person.
What Prediction Markets Don’t Pay
Kalshi is registered with the CFTC as a Designated Contract Market. Polymarket received its Amended Order of Designation in late 2025 and relaunched for US users in December. Crypto.com operates its prediction trading through a CFTC-regulated subsidiary. All three platforms argue — and the current CFTC chair, Mike Selig, energetically agrees — that event contracts are derivatives, not wagers, and that CFTC jurisdiction preempts state gaming law.
The fiscal consequence of this classification is enormous.
Prediction market platforms pay no state gambling taxes. They hold no state gaming licenses. They pay no state application fees, no investigation fees, no ongoing regulatory assessments. They fund no state gambling commissions. They contribute nothing to state problem gambling programs. They do not participate in state self-exclusion databases. They do not comply with state advertising restrictions.
They operate, in effect, as tax-exempt gambling platforms in states where licensed operators pay up to 51 percent of their gross gaming revenue to the state treasury.
Consider the scale. Kalshi and Polymarket together generated approximately $38 to $39 billion in notional volume in 2025. Eighty-five percent of Kalshi’s volume was in sports contracts — contracts that, if offered by a licensed sportsbook in New York, would generate tax revenue at 51 percent of operator margins. Robinhood, which entered prediction markets in October 2024 and traded nine billion contracts in 2025, projects this business as a $300 million annual revenue stream. Interactive Brokers’ ForecastTrader platform traded 286 million pairs in a single quarter. None of this revenue flows through state tax regimes.
The Revenue That States Actually Collect
To understand what prediction markets cost the states, you must first understand what regulated gaming delivers.
In fiscal year 2025, state-legalized sports betting generated $3.2 billion in tax revenue across thirty-nine states and the District of Columbia. Pennsylvania’s iCasino market alone produced $1.2 billion in state taxes. Michigan’s iCasino generated $791.8 million. New Jersey’s produced $542.3 million. The commercial casino industry generated $71.92 billion in total revenue in 2024, funding gaming commissions, problem gambling treatment, infrastructure, and education.
These revenues are not abstractions.
They fund teachers. They pay for addiction treatment. They support regulatory agencies that investigate operators, enforce consumer protections, and exclude bad actors from the market. When a prediction market platform offers what is functionally a sports wager under CFTC registration, every dollar wagered there is a dollar that does not flow through the state’s gaming tax apparatus.
The University of Washington research finding that 57 percent of sportsbook revenue cannibalizes existing lottery sales suggests the migration effect could be even more severe: prediction markets may not merely fail to generate state revenue — they may actively erode existing gaming revenue streams as players shift to untaxed platforms.
The Licensing Paradox
A casino operator applying for a license in New Jersey will spend $400,000 or more in fees, endure a background investigation that scrutinizes every principal’s financial history and criminal record, wait ten to sixteen months for approval, and then pay ongoing regulatory assessments that fund the state’s Division of Gaming Enforcement. This process is expensive, intrusive, and slow. It is also the reason New Jersey’s casino industry has operated for nearly fifty years without a major organized crime infiltration.
Prediction market platforms face none of this. CFTC registration as a DCM involves capital requirements, reporting obligations, and surveillance procedures designed for derivatives markets — markets where the typical participant is an energy company hedging natural gas exposure, not a twenty-two-year-old placing $50 on whether the Kansas City Chiefs will cover the spread.
The CFTC has no institutional expertise in problem gambling. It has no self-exclusion databases. It has no advertising restrictions tailored to gambling products. Its enforcement division investigates market manipulation and fraud, not whether an operator’s mobile app is designed to exploit variable-ratio reinforcement schedules. The regulatory arbitrage is not subtle. It is the business model.
Tribal Revenue and the Compact Crisis
The prediction market expansion poses what tribal gaming organizations have called an “existential threat” to tribal gaming. Under IGRA, tribal gaming operates through compacts negotiated with states — compacts that typically include revenue-sharing provisions, licensing requirements, age restrictions, and consumer protections. These compacts were bargained-for exchanges: tribes received gaming exclusivity; states received regulatory input and revenue.
If prediction markets are derivatives and CFTC jurisdiction preempts state law, the exclusivity that gives tribal compacts their value evaporates. Prediction market wagers circumvent IGRA entirely — they disregard state and tribal licensing requirements, minimum age requirements, and the consumer protections negotiated in carefully crafted compact agreements.
In November 2025, a federal court denied tribes’ motion for preliminary injunction against Kalshi, finding they had not shown likelihood of success on IGRA and Lanham Act claims. But the underlying economics are clear: every dollar wagered on a prediction market sports contract is a dollar that does not flow through a tribal compact’s revenue-sharing provisions.
The State Attorneys General’s Brief
Thirty-four state attorneys general signed a brief arguing that Congress did not intend to preempt state gambling authority in the Commodity Exchange Act. Their argument draws on the Supreme Court’s presumption against preemption of traditional state police powers and the observation that “when Congress removes the States’ historic police powers, it does not whisper in the dark of night.” The CEA contains no express preemption of state gambling laws. Section 12(e)’s savings clause explicitly preserves state authority over unregistered off-exchange activity.
The CFTC’s response, articulated by Chairman Selig in a February 2026 statement, is characteristically combative: “To those who seek to challenge our authority in this space, let me see you in court.” But the legal question is more nuanced than either side’s rhetoric suggests. If prediction markets are derivatives, and if the CEA grants the CFTC exclusive jurisdiction over derivatives, then state gambling laws may be preempted under field preemption principles. But if the CEA’s savings clause preserves state authority over activities that states have historically regulated as gambling, then CFTC registration does not immunize a platform from state gaming law. The circuit courts are split. The question is heading to the Supreme Court.
• • •
There is a version of this story in which prediction markets are regulated, taxed, and integrated into the existing gaming framework — a version in which states collect revenue, tribes maintain their compact protections, consumers receive gambling-specific safeguards, and the CFTC oversees market integrity while state gaming commissions oversee consumer protection. That version requires legislation. The current version — in which a federal derivatives regulator has unilaterally authorized nationwide gambling with no state taxes, no state licensing, and no state consumer protections — is not that version.