What prediction markets are, how they differ from betting, and why prediction market taxes are still contested by academics, the IRS and Indian rules.
A contract priced at 63 cents, and a tax form with no obvious box
Picture someone with two tabs open. In one, a sportsbook has a cricket match at 1.85. In the other, a prediction market is quoting a contract on the same kind of question at 63 cents, with a live order book, bid-ask spread and other traders on the opposite side. The second tab feels like a brokerage. The first feels like a bet. Come tax season, the honest problem is that nobody can tell them which line of the return the 63-cent contract belongs on.
That is the whole story of prediction market taxes right now. The money is real and growing fast, the mechanics look like finance, the outcome depends on an event rather than a company’s earnings, and the tax authorities have not committed to a position. My view is simple: the uncertainty is not a harmless technicality. Classification decides the rate you pay, whether losses count for anything, and whether one industry gets a quiet tax advantage over its competitors.
What prediction markets are, in plain terms
A prediction market is a platform where people buy and sell contracts tied to the outcome of a future event. Each contract settles at a fixed value if the event happens and at zero if it doesn’t. Prices move with demand, so the price itself reads as a rough probability.
How an event contract actually works
Take a contract that pays $1 if a named candidate wins an election. If it trades at $0.63, the market is pricing roughly a 63% chance. Buy it at $0.63 and you make $0.37 per contract if you’re right, and you lose the $0.63 you paid if you’re wrong. Sell it at $0.63 and the mirror applies: you keep the $0.63 you received if the event doesn’t happen, and you lose $0.37 per contract if it does, because you collected $0.63 upfront but have to pay out $1.
Two details separate this from a fixed-odds bet. First, you don’t have to wait for the event. If the price moves to $0.80, you can sell and take the difference, the way you’d close a trade. Second, you’re trading against other users, not against a house that sets the price and carries the risk. The platform typically earns fees and settlement, not a built-in margin on every position.
Where people trade them
The two names that come up most often are Kalshi, which operates as a regulated exchange in the United States under the Commodity Futures Trading Commission, and Polymarket, which grew up in crypto. Volume has not been small. The professors discussed below cite global monthly trading volume across leading platforms rising from under $5 billion in September 2025 to around $24 billion by April 2026.
For readers in India, a flag worth planting early: there is no licensed domestic prediction market framework, and real money online gaming sits under central and state rules that have been tightening rather than loosening. Nothing here is an invitation to sign up anywhere. It’s an explanation of how the tax argument works.
Why prediction market taxes are contested
Because the same transaction can be described three defensible ways, and each one carries a different tax consequence.
- Gambling. You staked money on an uncertain event you don’t control. That’s the textbook description of a wager.
- Investing or trading. You bought a transferable contract on an exchange, held a position, and could sell it before settlement. That’s the textbook description of a trade.
- Derivatives. An event contract is a financial instrument whose value derives from an external outcome, close cousin to a binary option, and in the US it sits under a derivatives regulator.
All three readings fit. None is obviously wrong. And tax codes, including India’s, were drafted long before anyone could buy a 63-cent contract on a weather outcome from a phone.
This is where the academic pressure comes in. Jay A. Soled, Distinguished Professor of Taxation at Rutgers Business School, and Mirit Eyal-Cohen, Joseph D. Peeler Professor of Law at the University of Alabama School of Law, argue in their George Washington Law Review article, titled “Betting on Tomorrow”, that most people using prediction markets are essentially gambling and should be taxed accordingly: gains as ordinary income, losses subject to the same restrictions that apply to traditional gambling losses. Their warning is that leaving it vague risks handing prediction market platforms an advantage over licensed sportsbooks offering economically similar products. As they put it, the issue is “too significant to ignore”, and they want the IRS to take a formal position rather than leave users guessing.
Betting vs prediction markets: what actually differs
Market structure
In a sportsbook or casino game, the operator is your counterparty and the price carries a built-in margin, the overround in betting and the house edge in casino games. A 2.7% house edge on European roulette is a mathematical certainty over enough spins. On a prediction market, the price is set by competing orders, and the platform’s revenue comes mainly from fees and spread rather than from being on the other side of your position.
| Feature | Prediction market | Sportsbook or casino game |
|---|---|---|
| Who you trade against | Other users via an order book | The operator |
| How the price is set | Supply and demand | Operator, with a margin or house edge built in |
| Exit before the outcome | Yes, you can sell the position | Rarely, beyond partial cash-out features |
| Operator revenue | Fees and spread | Margin on every bet, over time |
| Typical subject matter | Elections, economics, policy, sport | Sport, races, casino games |
Economic function
The defenders of prediction markets lean on price discovery. A market price aggregates what many informed people believe, which is genuinely useful information for journalists, businesses and researchers, in a way that a roulette wheel never is. That argument has real content.
It also has a limit, and this is where I part company with the more enthusiastic version of it. Serving a useful information function does not change what the individual participant is doing with their money. You can stake money on an outcome you can’t control, with a chance of losing all of it, inside a structure that aids price discovery. Commodity futures markets do both at once. “This market is economically useful” and “this user is gambling” can both be true, and the tax question is about the user.
The possible outcomes, and what each would cost
Below are the classifications under debate in the US, with the practical consequence of each. Treat this as a map of the arguments, not a ruling, because no authority has settled it.
| If treated as… | Gains | Losses |
|---|---|---|
| Gambling winnings | Ordinary income, taxed at normal rates | Deductible only against winnings, with tight limits |
| Investment property | Capital gains, with short and long term distinctions | Offset against capital gains, with carry-forward rules |
| Regulated futures contracts | Blended capital gains treatment under the futures rules | Generally offset within the same regime |
| Business trading activity | Ordinary business income | Expenses and losses treated under business rules |
Soled and Eyal-Cohen are pushing for the first row. Many platform users would prefer the second or third. The gap between them, on a profitable year, is not trivial.
Why the Indian answer is blunter, and what that means
India’s income tax framework treats winnings from betting, gambling, card games and online money games far less forgivingly than capital gains. Such winnings are taxed at a flat 30% rate on the winning amount, with no basic exemption slab applied to them and no deduction for expenses, and tax is deducted at source by the payer where the rules apply. Losses from these activities cannot be set off against other income or carried forward. Capital gains, by contrast, sit under an entirely different set of rates, holding period rules and set-off provisions.
So the same classification fight matters here too, only the stakes are sharper. If an Indian resident’s prediction market activity is characterised as winnings from betting or an online money game, the flat 30% applies and losing positions give you nothing back. If it were somehow characterised as a capital asset trade, the arithmetic changes completely. Add the question of whether gains arose on an offshore platform, how foreign income and asset disclosure obligations apply, and whether the platform is even permitted to serve Indian users, and you have a reporting problem that no app dashboard will solve for you.
The authoritative texts are the ones to read, not summaries: the Income Tax Department for Indian rules and the IRS for the US position, if and when it issues one.
What careful participants actually do while this is unsettled
Four habits, none exotic, all worth more than an opinion from a forum thread.
- Keep your own records. Date, platform, contract, position size, price in, price out or settlement, fees, and the currency involved. Export statements regularly rather than assuming they’ll be downloadable in two years.
- Track gross, not net. Many tax regimes care about total winnings rather than your season’s profit. A spreadsheet that only shows the final balance is useless under a gross-income rule.
- Record the currency and conversion. If funds moved through crypto or foreign currency, the conversion steps are part of the trail, and may create their own reporting events.
- Get professional advice before filing, not after a notice. A chartered accountant who has handled online gaming or foreign asset reporting is the right person to decide how your specific activity should be characterised.
And accept that the answer may change. A single formal position from a tax authority, or one piece of legislation, could redefine how every past and future position is reported. Writing your records now so they survive a reclassification later is the cheapest insurance available.
FAQ
What are prediction markets?
Platforms where users trade contracts on the outcome of future events. Each contract settles at a fixed value if the event occurs and at zero if it doesn’t, and the trading price works as an implied probability.
How are prediction markets taxed?
There is no settled answer. Depending on classification, gains could be treated as gambling winnings, ordinary income, capital gains or derivatives income. In India, winnings from betting and online money games face a flat 30% rate with no loss set-off, which is why characterisation matters so much. Confirm your own position with a tax professional.
Are prediction markets gambling?
Economically, a participant is staking money on an uncertain outcome, which is the core of gambling. Structurally, the exchange model resembles financial trading. Tax professors Jay A. Soled and Mirit Eyal-Cohen argue most users are effectively gambling and should be taxed on that basis.
What is the difference between betting and prediction markets?
In betting, the operator is your counterparty and sets prices with a margin built in. In a prediction market, you trade against other users at order-book prices and can usually sell your position before the event resolves.
One last point, because it belongs in any article like this: event contracts can lose their entire value, and no amount of research changes the fact that you are risking money on an outcome you don’t control. Stake only what you can afford to lose, use deposit and loss limits where platforms offer them, and if the activity stops feeling like a decision and starts feeling like a compulsion, seek help from a qualified support service. This article is educational and is not tax, legal or financial advice.
