Home › Gambling laws › Sports betting by country

Sports Betting Laws by Country 2026: Ranked by Model

Every guide to sports betting law sorts countries into legal and illegal, then quotes a tax rate. Both moves are wrong, and they fail in the same way: they describe the statute instead of the market. The axis that actually predicts anything is channelization — the share of money staked that runs through operators a state can tax and supervise. Measured that way, a state monopoly can outperform an open licensed market by more than 36 percentage points, and the lowest headline tax rate in this entire page belongs to a country whose real burden ranks tenth of thirteen.

What does every country-by-country betting guide get wrong?

Two things: it treats “legal” as a property of a country rather than of a product, and it compares tax rates that are percentages of completely different quantities. Both errors point the same direction — they make a page look authoritative while telling you nothing about where the money goes.

Take the first. Australia licenses and taxes online sports betting, and criminalises online casino under the Interactive Gambling Act 2001. Call Australia “legal” and you have described roughly two-thirds of its online market. France licenses online sports betting, horse racing and poker, and prohibits online casino; the Autorité Nationale des Jeux has put the resulting black market at as much as €1.5 billion a year. Japan has criminalised gambling since 1907 and legally takes about ¥25 trillion in stakes annually through six statutory exceptions. None of those three countries fits in a binary column.

The second error is arithmetic. A 5.3% tax and a 28% tax sound five times apart. If the first is charged on stakes and the second on the operator’s margin, the ranking can invert entirely — and on this page, it does.

How many regulatory models are there, and who uses which?

Four, and almost every country on earth sits in one of them: open licensing, state monopoly, prohibition with statutory exceptions, and unregulated grey. The table below gives the model, the tax base — which matters far more than the rate — the regulator, and the date the current regime took effect.

Two columns in that table do all the work, and they are not the ones people read. The first is tax base: three of these countries tax the stake and the rest tax the margin, which makes their headline percentages non-comparable by construction. The second is product scope: France, Australia and Poland all appear as licensed markets while criminalising or monopolising online casino, which means their “legal” label covers only part of what their residents actually play.

Which countries tax sports betting hardest, on the headline number?

On headline rates alone the order runs France 59.3%, New York 51%, the UK 40% on remote gaming, the Netherlands 37.8%, then Denmark and India tied at 28% — with Germany dead last at 5.3%. That is the ranking almost every comparison publishes, and it is close to useless.

It is useless because a percentage means nothing without its denominator. France’s 59.3%, raised from 54.9% on 1 July 2025 as part of a social-security financing package that lifted the CSG component from 10.6% to 15%, is a share of gross gaming revenue — what the operator keeps after paying winners. Germany’s 5.3% is a share of every euro staked, whether the bettor wins or loses. On a €100 bet returned at 96%, France takes 59.3% of €4. Germany takes €5.30 of the €100.

What happens when you re-rank by the real burden?

Three countries collapse down the table and everything else shuffles up by a place or two — and the three that fall are exactly the three that tax the stake rather than the margin. This is the whole argument in one table.

The conversion is simple and stated in full so you can check it: a turnover tax expressed as a share of GGR equals the turnover rate divided by the hold. We use a 10% sports-betting hold as the reference because that is close to the observed 2025 national hold in the United States, 10.15% — a real, published margin rather than an assumption we invented. Change the hold and every turnover row moves; the GGR rows do not move at all. That asymmetry is the point.

16Best analysis. Ranks are ascending, 1st = lightest. Turnover-based rows are converted at a 10% hold; GGR-based rows are unchanged by definition. Brazil’s 27% all-in figure includes corporate taxes, not just the gambling levy, and is shown for context rather than used in the ranking. New York is a US state, not a country, and is included because it is the heaviest GGR rate in the largest single market.

Germany’s number moves the same way for the same reason. Its 5.3% is charged on stakes, so on a sports bet at a 10% hold it is about 53% of GGR, and on a 96% RTP slot — a 4% hold — it reaches about 133% of GGR, more than the operator keeps. Scope matters here and we will not blur it: those are two different products under one rate, not one number quoted two ways. The full working sits in our German gambling law analysis, alongside the tiered €1 / €3 / €5 stake reform that took effect on 1 July 2026.

India is the extreme case, and the one with a live consequence. A 28% GST on the full face value of a bet equals 560% of GGR at a 5% hold, 280% at 10% and 187% at 15% — 6.7× to 20× a 28% GGR tax, depending entirely on the margin. Below a 28% hold, the tax exceeds everything the operator earns. There is no rate at which that business survives, which is the honest reading of why India’s legal market ended: the ban that took force on 1 May 2026 erased about $41.6 billion of cumulative forecast legal revenue, but the tax had already made the model unbuildable. Our India gambling law page carries the full stake-versus-revenue table.

Where does the money actually go, and who measures it?

Only a handful of regulators publish channelization, they measure it in at least three incompatible ways, and the published spread inside a single country can reach 11 percentage points before you compare two countries at all. Here is every figure we could source, with its method attached, because the method is half the number.

Read that chart from the top and the legal-versus-illegal binary falls apart. The best-channelled market in the list is a product inside an open licensing regime. The second best is a monopoly. Two open licensed markets — the Netherlands and Brazil, whose licensed regime went live on 1 January 2025 — sit at the bottom next to a country that criminalises half of what its residents play.

Why does the same country channel 95% of one product and 68% of another?

Because the thing that determines channelization is whether the licensed product resembles the unlicensed one — and Sweden shows it inside a single jurisdiction, with every other variable held constant. This is the cleanest natural control on the page, and it is hiding in plain sight in a regulator’s annual report.

Swedish sports betting and Swedish online casino operate under the same Gambling Act, the same regulator, the same 22% tax on gross gaming revenue in force since 1 July 2024, the same licence, the same advertising rules and the same national self-exclusion register. Tax cannot explain a gap between them. Legality cannot explain it. Enforcement cannot explain it. And on Spelinspektionen’s internet-traffic method the two products channel at 95% for sports betting and 68% for online casino in 2025 — while its player-survey method puts them much closer, at 98% and 93%.

Australia says the same thing from the other side. A 2025 study found 36% of Australian online gambling occurring offshore, worth about A$3.9 billion a year in losses — reported elsewhere as US$2.5 billion, implying a conversion of roughly A$1 = US$0.64 and not an inconsistency. About a quarter of that offshore activity is on products that have no legal Australian equivalent, chiefly online casino and live in-play betting. Apply that quarter to the loss figure and roughly A$1 billion a year sits structurally beyond reach of any licensing system Australia currently operates (16Best analysis; this assumes the share of losses matches the share of activity, which the study does not state). No tax rate, enforcement budget or advertising rule touches that billion, because there is nothing legal to channel it into.

The academic literature complicates the picture in a way worth stating rather than hiding. A 2025 study in the Harm Reduction Journal by Marionneau and colleagues examined European channelling and taxation from 2015 to 2021 and found no evidence of a negative correlation between tax rates and channelling rates — for online betting, some analyses even pointed the other way. That does not rescue the legal-illegal binary; it sharpens our claim. Tax is not the lever. Product is.

How much of the world lives under each model?

At least 41.6% of the world’s population lives in a country that prohibits online sports betting — and that figure comes from counting only four of them. India, China, Indonesia and Pakistan together hold about 3.42 billion people against a world total of 8.23 billion in the UN’s World Population Prospects 2024, mid-2025 estimates.

Two honest qualifications, because this is the kind of number that gets misquoted. First, it is a population share, not an adult share; prohibition-heavy countries skew younger than the world average, so the share of the world’s adults under prohibition is somewhat lower than 41.6%. Second, it is a floor by construction — adding Iran, Saudi Arabia, Qatar, Kuwait, Brunei, Afghanistan, North Korea and the rest of the blanket-ban group would push it materially higher, and we have not included them because we could not source their populations to the same standard.

The finding that matters is what the four have in common besides the ban: every one of them hosts enormous unmeasured volume. India’s offshore betting market has been put at $100–150 billion a year by the Observer Research Foundation, with offshore platforms drawing 5.8 billion visits between April 2024 and March 2025. Treat those figures the way we treat Nigeria’s circulated ₦5.6 trillion: they are handle-scale inflow estimates, not revenue, and nobody can audit them. That is precisely the problem. Prohibition converts a measurable market into an unmeasurable one.

What do the live natural experiments show?

Six countries have changed regime inside 24 months, in six different directions, and each one tests a different part of the framework. They are the closest thing this field has to controlled experiments, and they are all running right now.

Kenya is the one that most directly contradicts the intuition that heavier taxes mean bigger collections. Cutting the betting excise from 15% to 5% and shifting the charge onto the M-Pesa rail — where the money already moves and cannot easily hide — produced a record KSh 16.53 billion in FY2025/26, 115.9% of target. The rate fell by two-thirds and the take rose. The mechanism was collection infrastructure, not rate. Full working in our Kenya gambling statistics.

Two of the six are experiments in who holds the authority rather than what the rate is. Nigeria’s federal gaming regulator was dismantled by the Supreme Court on 22 November 2024, leaving 36 states to set their own terms — and the flat 11% GGR levy that eventually emerged was written not by a legislature but by an association of state regulators. The Philippines went the other way, keeping one national body that regulates, operates and takes a percentage of its licensees’ win, so its income falls every time it enforces. Both are questions about institutional design that a legal-versus-illegal table cannot even ask.

The Netherlands ran the same experiment in reverse, and it is the cleanest year-over-year series available anywhere in this field.

What does the United States prove that no single country can?

That the divergence is caused by policy rather than by culture — because 50 states ran different policies on one population under one federal constitution after PASPA fell on 14 May 2018. No cross-country comparison can control for culture, income, sporting calendar and payment infrastructure at once. The American states control for all of them by accident.

As of May 2026, 39 states and the District of Columbia permit some form of legal wagering and 32 offer it online. Tax rates within that single country span a wider range than most of Europe: New York takes 51% of gross gaming revenue, while Illinois raised its online rate to as much as 40% and then became the first US state to add a per-wager excise — 25 cents on each of a licensee’s first 20 million bets a year, then 50 cents on every bet above that. Note what that Illinois fee is: a turnover tax wearing a different hat, charged per ticket rather than per dollar, and therefore heaviest on exactly the small-stake bettor the rate schedule was meant to spare.

The one number that travels furthest from the American experiment is the margin. The national sportsbook hold rose from 9.1% in 2023 to 9.2% in 2024 to 10.15% in 2025, so revenue compounded at about 23.4% a year while handle grew about 16.9% — GGR growing 1.38× faster than the money wagered (16Best analysis). That matters for every country in the table above, because it means a GGR tax is levied on a base that is widening faster than the market itself, and a turnover tax is not. The state-by-state detail sits in our US online gambling revenue by state analysis, and the fastest-moving product category in several of these markets is covered in our esports betting statistics.

Why do these country comparisons disagree?

Because six distinct things are routinely called the same thing, and almost no source says which one it means. This section is the disambiguator. If you take one part of this page, take this one.

• Statute versus enforcement. Many countries prohibit betting and have never prosecuted a bettor. India’s PROGA reaches offshore operators and payment processing; Indonesia blocks sites by the million; both still host large unmeasured markets. A page that reports the statute has reported a rule, not a behaviour, and the two are different objects. We do not tell readers how to bet from a prohibited jurisdiction — we report that prohibition demonstrably fails to end betting, which is an argument about policy design, not an invitation.
• Turnover versus GGR versus profit. The single most consequential distinction in gambling law. A tax on turnover is charged on every unit staked; a tax on GGR is charged on what the operator keeps; a tax on profit comes after costs. Germany’s 5.3%, Poland’s 12% and India’s 28% are the first kind. France’s 59.3%, New York’s 51% and Brazil’s 12% are the second. Comparing across the two without converting produces exactly the inverted ranking most guides publish.
• The conversion depends on the hold, and the hold depends on the product. The same 5.3% turnover tax is about 53% of GGR on a sports bet at a 10% hold and about 133% on a 96% RTP slot. Any effective-rate figure that does not state its hold assumption is not checkable. Ours is 10%, anchored on the 10.15% US national hold in 2025.
• What counts as legal when the answer differs by product. France, Australia and Poland license sports betting and prohibit or monopolise online casino. Japan bans gambling and exempts six products by separate statute. Pakistan bans gambling and permits horse racing and some state lotteries. The Philippines licenses domestic online play and criminalises offshore operations. “Legal” is a property of a product in a country, never of a country.
• State monopoly is not the same as prohibition, and it is not the same as no market. Norway, Finland until mid-2027 and Singapore all run monopolies with real, taxed, supervised volume. Norway’s monopoly channels 86–88% by GGR, which is better than several open licensed markets. Filing monopolies under “restricted” and open licensing under “legal” produces a ranking that inverts the observed outcome.
• Channelization is measured at least three ways and they do not agree. Sweden publishes 84% as the average of a player-survey method giving 89% and a web-traffic method giving 78% — an 11-point spread inside one regulator’s own report. The method dependence widens at product level: the online-casino figure is 68% on the traffic method but 93% on the survey, so the headline “sports 95 vs casino 68” gap is largely a traffic-method artefact even though both methods agree on the direction. Germany’s GGL models 77% from blockchain and payment data while industry analysts argue 40–60%. The Netherlands measures by GGR, which produces lower numbers than player-count methods because heavy players leave first. Never compare a GGR-based channelization figure with a player-based one.
• Onshore volume is counted and offshore volume is estimated. The licensed number comes from tax returns. The unlicensed number comes from surveys, traffic modelling or payment inference — estimates of something built not to be counted. That is why India’s offshore market is quoted anywhere between $100 billion and $150 billion, and why we treat all such figures as orders of magnitude rather than measurements.
• Passed, in force and enforced are three different states. India’s ban took force on 1 May 2026, not 2025. The UK’s 40% remote gaming duty is in force from April 2026; its 25% remote betting rate is not — that arrives in April 2027. Finland’s licence applications open on 1 March 2026 while Veikkaus keeps its online monopoly into 2027. Brazil’s law was signed in December 2023 and went live on 1 January 2025. Check which state a “new law” is in before quoting it.

One last caution on currency, since this page ranks by percentage precisely to avoid it. Where absolute sums appear they are in local currency first: A$3.9 billion of Australian offshore losses was reported elsewhere as US$2.5 billion, implying about A$1 = US$0.64 at the time of that study, and a few cents of movement in that rate changes the dollar figure by more than most of the trends being discussed. Rank by rates. Quote sums in the currency they were collected in.

Key takeaways

• Legal versus illegal is the wrong axis. It describes the statute. Channelization — the share of stakes running through operators a state can tax — describes the market, and the two routinely disagree.
• A monopoly can out-channel an open market. Norway reports 86–88% by GGR; the Netherlands has fallen below 50%. Brazil, the newest fully regulated major market, still has 41–51% of activity outside the licensed system.
• Only the tax base moves the ranking. Re-rank thirteen markets by effective burden and exactly three fall — Germany by nine places, Poland by eight, India by three. All three tax the stake. The other ten only rise because those three moved past them (16Best analysis).
• Germany’s 5.3% is the lowest headline rate here and about the fourth-heaviest real one — roughly 53% of GGR on a sports bet at a 10% hold, and about 133% on a 96% RTP slot.
• Product design channels better than tax rates. Sweden runs one law, one regulator and one 22% tax across both products and still records 95% channelization in sports betting against 68% in online casino — a 27-point gap that tax cannot explain.
• Raising the rate can lower the take. Dutch receipts fell to 83% of the prior year under a rate 12.1% higher, implying a taxable base down about 26% — within a point of the 25% GGR decline the industry reported independently (16Best analysis).
• Lowering it can raise the take, if you fix collection. Kenya cut betting excise from 15% to 5%, moved the charge onto the mobile-money rail, and collected a record KSh 16.53 billion in FY2025/26.
• At least 41.6% of the world — about 3.42 billion people across four countries — lives under prohibition, and every one of those four hosts large unmeasurable volume. Prohibition does not end betting; it ends the data.
• The United States is the control group. One country, 39 states plus DC with legal wagering, tax rates from single digits to 51%, and a hold that rose from 9.1% to 10.15% in two years — so GGR now compounds 1.38× faster than handle.

Frequently asked questions

Which country has the highest sports betting tax?

It depends on whether you mean the headline rate or the real burden. On headline rates, France is heaviest among major regulated markets at 59.3% of gross gaming revenue for online sports betting, in force since 1 July 2025, ahead of New York at 51% and the UK at 40% for remote gaming from April 2026. On effective burden the answer changes completely, because India taxes 28% of the full face value of every bet rather than the operator margin, which equals about 280% of gross gaming revenue at a 10% hold, and Poland taxes 12% of turnover, which equals about 120% on the same basis.

Is sports betting legal in most countries?

Some form of it is legal in most countries, but the label is misleading because legality attaches to products rather than to countries. France, Australia and Poland license sports betting while prohibiting or monopolising online casino. Japan has criminalised gambling since 1907 under Penal Code Articles 185 and 186 and exempts six products by separate statute, through which roughly 25 trillion yen is staked each year. Meanwhile India, China, Indonesia and Pakistan prohibit online sports betting and together hold about 3.42 billion people, 41.6% of the world population of 8.23 billion in the UN mid-2025 estimates.

What is channelization and why does it matter more than legal status?

Channelization is the share of gambling activity that takes place with licensed operators rather than unlicensed ones. It matters more than legal status because it measures what actually happens rather than what the law says. Norway, which runs a state monopoly, reports channelization of 86 to 88% by gross gaming revenue for 2025. The Netherlands, which runs an open licensed market, has seen channelization by gross gaming revenue fall from 58% in the first half of 2024 to below 50%. The monopoly captures more of its market than the liberalised one.

Why is Germany's 5.3% tax considered high?

Because it is charged on stakes rather than on the operator margin. A 5.3% tax on every euro wagered equals about 53% of gross gaming revenue on a sports bet at a 10% hold, and about 133% on a virtual slot running a 96% return to player, where the operator keeps only 4 euros of every 100 staked. That makes Germany the lowest headline rate in our thirteen-market comparison and roughly the fourth-heaviest once converted to a common base.

Does banning sports betting stop people betting?

The available evidence says it relocates the activity rather than ending it, and removes the data along with the tax. India banned online money games with effect from 1 May 2026 and the Observer Research Foundation has put the country's illegal betting market at 100 to 150 billion US dollars a year, with offshore platforms drawing 5.8 billion visits between April 2024 and March 2025. Those are inflow-scale estimates that cannot be audited, which is precisely the problem: prohibition converts a measurable market into an unmeasurable one.

Did the UK raise gambling taxes in 2026?

Yes. Following the Autumn Budget of November 2025, remote gaming duty rose from 21% to 40% of gross gambling yield from April 2026. A separate remote betting rate of 25%, up from the general betting duty rate of 15%, is legislated for April 2027 and is not yet in force; it excludes horse racing bets, pool bets, spread betting and bets placed at self-service betting terminals. Bingo duty is being abolished and casino gaming duty bands were frozen for 2026-27.

Which regulator publishes the most reliable channelization figures?

Sweden's Spelinspektionen is the most transparent, because it publishes both underlying methods rather than only the result. Its 2025 figure of 84% for the competitive market is the average of a player-survey estimate of 89% and a web-traffic estimate of 78%, an 11 percentage point spread inside one report. It also breaks the figure down by product: on the web-traffic method sports betting is 95% against 68% for online casino, while the player-survey method puts the same two at 98% and 93%, so the size of that product gap depends heavily on which method you read. Germany's GGL publishes 77% modelled from blockchain and payment data, which industry analysts including Regulus Partners and H2 Gambling Capital dispute, arguing the true figure is closer to 40 to 60%.

Sources

• House of Commons Library — Budget 2025: Gambling taxation (CBP-10440)
• iGaming Business — UK sector hit with 40% remote gaming duty, new remote betting tax from 2027
• Deloitte Taxscape — Autumn Budget 2025: gambling duties
• iGaming Business — Netherlands faces €200m gambling tax black hole as rate rise backfires
• iGaming Business — Dutch gambling tax rise yields far less revenue than forecast
• iGaming Business — Channelisation in Sweden slips to 84% in 2025, according to Spelinspektionen
• SBC News — Sweden channelisation rate down to 84% as struggle continues
• GVZH Advocates — Sweden increases gaming tax from 18% to 22% from July 2024
• Focus Gaming News — Norwegian regulator hails drop in participation in unlicensed offshore gambling
• Marionneau, Matteucci, Vieira Lima, Nikkinen and Selin — Channelling and taxation in European online gambling markets, Harm Reduction Journal, 3 January 2025
• iGaming Business — FDJ expecting €45 million impact from French gambling tax increases in 2025
• Yogonet — France delays online casino legalisation amid resistance from physical casinos
• DLA Piper — Italian online gambling licences: 46 operators approved
• SBC News — Italy nets 46 licences for new €7m online gambling regime
• Finnish Government — Finland to open some gambling services to competition
• Bird & Bird — Finland ending its online gambling monopoly in 2027
• Inside Asian Gaming — Australia’s illegal offshore gambling market now worth US$2.5 billion annually
• Tribuna — 36% of Australia’s online gambling now offshore as annual illegal losses hit A$3.9bn
• Revenue NSW — Point of consumption tax
• Nelson Mullins — Illinois: first state to implement a per-bet excise tax on sports gambling
• Track360 — US sports betting state-by-state map 2026
• iGaming Business — Is it the right time for Poland to liberalise iGaming?
• Euro Weekly News — Poland’s iGaming paradox: a regulated monopoly fuelling a €1 billion market
• iGaming Business — Illegal operators still compete with Brazil’s regulated betting market
• United Nations — World Population Prospects 2024: Summary of Results
• Observer Research Foundation — India’s illegal online betting market estimated at US$100–150 billion
• Business Standard — India a hub for illegal online betting; sites got 5bn visits in FY25
• Pakistan Code — Prevention of Gambling Act, 1977
• Singapore Statutes Online — Gambling Control Act 2022
• Gambling Regulatory Authority of Singapore — Official site
• The National — UAE’s first regulated online gaming website begins operations
• Pinsent Masons — New regulations and regulator signal new era for legal gaming in the UAE
• iGaming Business — DOCV: German online-slot channelisation at 20–40%
• Next.io — Online casino defies decline of other verticals in 2025 Denmark report