Home › Gambling in Sweden

Gambling in Sweden Statistics 2026: The 11-Point Question

Sweden is described across Europe as the model licensing regime that fell six points short of its 90% channelisation target. Open the regulator’s own report and that framing collapses. Spelinspektionen’s June 2026 report gives two channelisation figures for the same market in the same year — 89% and 78% — and the published 84% is simply their midpoint. The internal disagreement is 11 points. The shortfall everyone argues about is six. Sweden has spent seven years legislating against a number whose own error bar is nearly twice the size of the miss.

Did Sweden’s gambling reform miss its 90% channelisation target?

On the published headline, yes — by six points. But the regulator produces that headline by averaging two of its own measurements that sit eleven points apart, and it publishes no confidence interval around either. The 90% figure is real and traceable: it was set before re-regulation, in SOU 2017:30 and the government bill Prop. 2017/18:220, as the share of Swedish gambling that should take place inside the licensed system. Every assessment since has been scored against it.

What almost nobody scoring it does is read how the number is built. Here is every credible reading of Swedish channelisation for 2023–25, with the method attached, because the method is most of the answer.

Compiled by 16Best from Spelinspektionen’s Kanaliseringsgrad på den svenska spelmarknaden 2025 (15 June 2026, Table 6), its 2024 and September 2024 predecessors, and the Nordic Legal report for BOS of 13 October 2025, which reproduces the H2 Gambling Capital, ATG and SKOP figures. All rates are shares of the competitive market — betting plus online casino — not of all Swedish gambling.

What does Spelinspektionen’s own channelisation report actually say?

That its two main indicators are biased in opposite directions, that it knows it, and that it averages them anyway. This is the part of the Swedish story that has never been written up, and it is sitting in plain text in a 25-page government report.

The survey indicator asks a Verian panel of 6,744 Swedish adults what they staked at their most recent gambling session and whether the site held a Swedish licence. Of those, 4,175 had gambled at all in the past year, 903 had played a competitive-market product, and just 220 had played online casino. The regulator flags the consequence itself: the number of respondents who played on unlicensed sites is small, and “when the sample is small, a single answer can have a large impact on the result.” Sweden’s 93% online-casino channelisation figure — the high end of the national band — is computed off that base.

The traffic indicator has the opposite problem. It estimates visits to gambling sites from Sweden and converts them to turnover. But web-traffic estimation tools do not see app sessions, and Spelinspektionen’s own operator survey found that 38% of licensed turnover arrives through apps — while unlicensed sites, which cannot list in the App Store or Google Play, are almost entirely browser-based. The regulator says so directly: an indicator based only on internet traffic underestimates the licensed share.

Then there is what the measurement leaves out. Skin betting — wagering with cosmetic items from video games — took 42% of all Swedish visits to unlicensed gambling sites in 2025, the single largest destination. Spelinspektionen excludes skin-betting sites from its channelisation indicators entirely, on the reasonable ground that those sites also run trading, news and crypto services and the gambling share is not separable. The effect on the raw traffic picture is large: counting all sites, 61% of Swedish gambling visits went to licensed sites; excluding skin betting, 73%. Twelve points of the visit picture turn on a category the headline number does not contain.

What did the Swedish gambling market look like before 2019?

A state monopoly with an offshore market layered on top of it — Spelinspektionen puts pre-reform channelisation of the competitive market at “considerably lower than 50 percent.” Svenska Spel, owned by the state, ran the lotteries, the Vegas value machines and the four Casino Cosmopol venues. AB Trav och Galopp held the monopoly on horse-race betting. Everything else Swedes played online — and they played a great deal of it — ran on Malta and Gibraltar licences that Sweden could tax at nothing and supervise not at all.

That is the situation the Gambling Act (2018:1138) was written to end. The licensing system opened on 1 January 2019. Operators had to meet financial and integrity criteria, connect to a national self-exclusion register, market with måttfullhet — moderation — and pay 18% tax on gross gaming revenue under the Gambling Tax Act (2018:1139). Bonuses were restricted to a single offer at a player’s first participation in each form of gambling, a rule that catches reload and retention offers and is the one Swedish requirement foreign operators most often get wrong.

The design brief was coherent: make the licensed product legitimate enough that players stay inside it, and accept slightly less commercial freedom in exchange for supervision. Whether it worked depends entirely on which row of the table above you believe.

What happened in the first year of the licensed market?

Sweden reported 88% channelisation in 2019 — two points off target in year one, and it has never been reported higher since. The figures cited in the Finance Ministry’s April 2022 review of the reform run 50% in 2018, 88% in 2019, 85% in 2020 and 87% in 2021. Statskontoret, the Swedish Agency for Public Management, independently put 2021 at approximately 87% and 2020 at 85%.

Read that sequence carefully, because the standard telling of the Swedish story gets the shape wrong. This is not a reform that started low and failed to climb to 90%. It is a reform that arrived within touching distance of its target in year one and has drifted away from it since, through a period in which every significant policy intervention was aimed at the licensed side of the market.

One caution before anyone draws a line through those numbers: you cannot. The 2019–21 figures come from a ministry review and Statskontoret; the 2023 figure came from a four-indicator average that included H2 Gambling Capital’s estimate; the 2024 and 2025 figures come from a two-indicator average after the regulator dropped H2 from the calculation. At least three distinct methodologies produced that apparent trend, which is why every channelisation chart on this page uses bars.

Where has the money actually moved since 2019?

Into online, and out of everything else — online commercial gambling and betting grew 31.3% between 2019 and 2025 while the rest of the licensed market shrank 9.1%. This is the part of the Swedish market that can be measured exactly, because Spelinspektionen and Skatteverket publish net turnover by licence type every quarter, and net turnover here means precisely what players lost: stakes minus winnings paid back.

Spelinspektionen and Skatteverket, net turnover by licence type, quarterly workbook; 2025 marked preliminary. Year-on-year changes and the online-share column are 16Best calculations on the published figures.

Same source. Rows may not sum exactly to the total because of rounding at one decimal place.

Two things fall out of those tables that Swedish coverage rarely puts together. The first is the compound arithmetic: licensed online grew at a 4.64% compound annual rate across six years while the non-online remainder contracted at 1.58% a year. The licensed market did not grow. It rotated.

The second is the inflation adjustment nobody applies. Between 2019 and 2024, licensed net turnover rose 12.3% in kronor. Swedish consumer prices rose 24.2% over the same five years, driven by 8.4% inflation in 2022 and 8.6% in 2023. In constant kronor, Sweden’s licensed gambling market was about 9.6% smaller in 2024 than in its opening year. Every headline describing the market as growing since re-regulation has been quoting a nominal figure taken during the largest Swedish price surge in three decades.

The most physical illustration sits in the fourth row of the mix table. Casino Cosmopol turned over SEK 974.5 million in 2019 across four venues. Svenska Spel closed Sundsvall in 2020 and Gothenburg and Malmö in February 2024; on 25 April 2025, the day after the government changed the company’s ownership instruction, it announced the immediate closure of the Stockholm room, affecting around 240 staff. Land-based casino games were prohibited nationwide from 1 January 2026. The regulator’s workbook records SEK 0 for Q1 2026. Sweden did not lose its state casino monopoly to competition; it legislated the product out of existence, and the line item simply ends.

Why did Swedish channelisation fall after 2021?

Because the leak is almost entirely one product — online casino — and that product is also where the measurement is thinnest. Every estimator agrees on the direction even where they disagree violently on the level. In Spelinspektionen’s 2025 report, betting moves only three points across methods (98% survey, 95% traffic) while online casino moves twenty-five (93% and 68%). On ATG’s Q1 2024 study, betting is 77% and online casino 57%. On H2’s revised model, betting is 83% and online commercial gambling 62%.

Part of Sweden’s strong betting number is structural rather than a policy success. A large share of Swedish betting is harness racing pooled inside ATG’s tote product, which has no meaningful offshore substitute — you cannot replicate a pari-mutuel pool on a Curaçao licence. Online casino has no such moat: the games are the same titles from the same studios on either side of the licence. We work through why product similarity, rather than tax, drives that gap in our sports betting laws by country hub; the point here is narrower and more uncomfortable. The entire Swedish channelisation debate is an argument about the habits of the 3.3% of Swedish adults who played online casino in the past year — 220 people in the regulator’s own 6,744-person sample (16Best analysis of Spelinspektionen’s survey base).

The supply side has grown regardless. Spelinspektionen had identified 2,186 gambling websites without a Swedish licence as of 30 April 2026, of which 334 belong to companies already under prohibition orders. Nearly all of them are casino: 976 offer online casino only and 800 offer casino plus betting, against just six offering betting alone.

Then there is the H2 Gambling Capital revision, which deserves its own paragraph because of how it reads. In 2025 H2 rebuilt its model for the Swedish unlicensed market and cut its channelisation estimate from 91% to 72% — a 19-point downgrade with no change whatsoever in what Swedish players did. It cut Denmark from 87% to 72% at the same time, citing the growth of skin betting and crypto-settled offshore casinos since 2022. This is the same firm whose global figures anchor our gambling loss statistics, and we use its data elsewhere precisely because it is careful. A careful estimator moving a national figure 19 points in one revision is the strongest single piece of evidence that channelisation is not a measurement. It is a model output.

Sweden is not unusual in having this gap, which is worth saying plainly. The pattern repeats across Europe: Germany’s GGL reports 77% while industry estimates run 40–60%; the Netherlands has fallen below 50% on GGR; Norway’s state monopoly channels 86–88% by GGR. On the regulator’s figure, Sweden’s open licensed market (84%) sits just below Norway’s monopoly. On ATG’s figure (68%) it sits well below it. Which of those two sentences you print determines whether Sweden reads as a vindication of licensing or an argument against it.

What did raising the gambling tax to 22% actually do?

It cost licensees about SEK 736 million in 2025 — and left no visible break in the licensed revenue series at all. The rate went from 18% to 22% of gross gaming revenue on 1 July 2024, under the Gambling Tax Act (2018:1139). The government’s stated aim was roughly SEK 540 million of additional annual revenue, SEK 270 million in the part-year 2024. Operators warned it would push players offshore. ATG’s chief executive called for a differentiated rate. In October 2024 a motion was filed in the Riksdag to reconsider it.

The regulator’s quarterly data lets us test the claim directly, half-year by half-year, on the segment that actually pays it.

16Best analysis of Spelinspektionen and Skatteverket quarterly net turnover, licence category “commercial online gambling, betting and gambling on ships”. Half-year figures are the sum of the two published quarters; growth is against the same half of the prior year.

The licensed online market was contracting in both halves of 2023, at 18% tax. It has grown in every half-year since the rise to 22%. Nothing in that pattern isolates the tax — half-year growth swings between +0.8% and +5.8% and plainly responds to other things — but it does dispose of the strong version of the industry claim. If the 2024 tax rise drove Swedish players offshore, the licensed revenue series does not know about it.

The cost to operators is real, though, and bigger than the headline four points suggest. Before the change a licensee kept SEK 82 of every SEK 100 of player losses; after it, SEK 78 — a 4.9% cut in retained revenue before a single cost is paid. Run that across the actual numbers: licensed online net turnover grew SEK 1,419 million between 2023 and 2025, but operators retained SEK 14,361 million in 2025 against SEK 13,934 million in 2023 — growth of just SEK 427 million, while the state’s take on the same segment rose SEK 992 million. Of every extra krona of gross revenue the licensed online market produced over those two years, 30 öre reached the operator and 70 went to the Treasury (16Best analysis). Had the rate stayed at 18%, licensees would have kept SEK 736 million more in 2025 alone — about US$75 million at the 2025 average of SEK 9.8096 per dollar. Note the perimeters: that SEK 736 million is what one licence category paid in extra tax, and it is 36% more than the SEK 540 million a year the government forecast the change would raise across the entire market (16Best analysis). Either the forecast was conservative or the base grew faster than the Budget Bill assumed. Both readings point the same way: the rise raised more than it was sold on.

Set that against the comparison that matters. Sweden takes 22% of GGR. Germany’s 5.3% turnover tax on online slots works out to roughly 133% of GGR on a 96%-RTP game, as we set out in gambling in Germany statistics — six times the Swedish burden. Yet Germany’s regulator reports 77% channelisation and Sweden’s reports 84%: a seven-point gap on a six-fold difference in tax, comparing regulator figure with regulator figure. Whatever is driving Swedish leakage, the headline tax rate is not carrying much of the explanation.

How many Swedes have banned themselves from gambling?

134,484 people were registered on Spelpaus at the end of 2025 — roughly one in every 63 Swedish adults — and new registrations have fallen for two straight years. Spelpaus.se launched with the licensed market on 1 January 2019. One registration blocks a person from every Swedish-licensed operator, online and land-based, and from receiving their direct marketing. There is no equivalent register in most of Europe, which is why the Swedish dataset repays close reading.

Spelinspektionen annual reporting for the end-2025 stock, compiled with a July 2026 analysis of the regulator’s figures for the registration series. Adult population of 8,493,129 (mid-2025, UN World Population Prospects). The register is a stock of currently excluded people, not a cumulative count — exclusions can lapse, and the Q1 2026 reading of close to 138,000 was higher than the figure recorded at the end of June.

Now put two of the regulator’s own indicators next to each other, because they have been moving in the same direction and almost nobody reports them together. Between 2023 and 2025, new Spelpaus registrations fell 26.8%. Across the same two years, Spelinspektionen’s channelisation headline fell from 86% to 84%. Fewer people are stepping inside the protection perimeter, and a growing share of play is happening outside it.

There is a mechanism connecting those two lines, and it comes from the regulator’s own survey. Asked why they had last gambled with an unlicensed operator, respondents in the 2025 report cited: no particular reason (23%), having self-excluded via Spelpaus (13%), better perceived chances of winning (11%), games unavailable on the licensed market (11%), and more attractive bonus offers (9%). Spelinspektionen’s own conclusion names it explicitly: the most common reason given is having self-excluded through Spelpaus.se. The previous year’s report put the same reason at 21%. Both figures rest on tiny bases — the 2025 chart carries 53 responses in total — so treat the ranking as informative and the percentages as indicative.

You could multiply that share by an unlicensed-market estimate and produce a headline number in the hundreds of millions of kronor. Do not. The share is of players citing a reason, not of spend, self-excluded players almost certainly do not spend at the average rate, and the base is 53 people. The honest statement is directional and it is enough: Sweden’s flagship consumer-protection instrument is, by the regulator’s own account, the most commonly cited reason Swedes gamble outside Swedish supervision.

How common is problem gambling in Sweden?

Swelogs 2021 puts 0.5% of Swedes at PGSI 8 or above — roughly 40,000 people — with a further 0.8% at elevated risk and 3.0% at some risk. The instrument is the Problem Gambling Severity Index; the survey is the Swedish Longitudinal Gambling Study run by Folkhälsomyndigheten, the Public Health Agency of Sweden, with more than 7,400 participants in the 2021 population wave. Naming the instrument and the wave matters, because Swedish prevalence figures between 0.5% and 4% all circulate and they measure different thresholds.

Folkhälsomyndigheten, Swelogs 2021, PGSI instrument. Bands are as published by the agency; the “any degree of risk” row is the sum of the three bands above it.

The longer trend is genuinely encouraging and should be reported that way. An analysis by economist Ola Nevander published on 31 March 2026 tracks Swelogs across roughly 25 years and finds PGSI 3+ prevalence down from 2.2% of adults in 2008–09 to 1.3% in 2021, with the at-risk PGSI 1+ group down by around 200,000 people. Problem gambling among people who gamble online specifically fell from 12% in 2008–09 to roughly 4% across 2018–21. Severe problem gambling has stayed inside a 0.3–0.6% band throughout. Note the commissioner: the analysis was produced for BOS, the online gambling trade body, which has an obvious interest in a declining-harm narrative. The underlying Swelogs data is the state’s, and it says what the analysis says it says.

One comparison is worth making carefully. There were 134,484 people on Spelpaus at end-2025, against roughly 105,000 Swedes estimated at PGSI 3 or above in Swelogs 2021. The self-exclusion register now holds about 28% more people than the survey-estimated at-risk population (16Best analysis). Four years and two different instruments separate those figures, so this is an order-of-magnitude observation, not a measurement. What it suggests is that Spelpaus is not functioning solely as a clinical tool for people already in difficulty. A meaningful number of Swedes are using it precautionarily, which is arguably the best thing anyone can say about the reform.

How much does a Swedish adult lose to gambling each year?

SEK 3,325 to licensed operators in 2024 — about US$314, and 0.9% of disposable income — on the regulator’s own published per-capita statistic. That is the fully sourced floor, and it is a floor because it counts only what licensed operators booked. Whatever Swedes lost offshore sits on top of it.

Spelinspektionen publishes the SEK 3,325 figure directly on its own 2024 adult denominator. Rows two to five are ours, all on a single denominator of 8,493,129 adults (mid-2025, UN WPP) and annual average rates of SEK 10.575 per US dollar (2024) and SEK 9.8096 (2025), which is why our 2024 equivalent is SEK 3,278 rather than 3,325. The channelisation-adjusted rows gross up the competitive market only and add the monopoly and public-benefit segments unadjusted, because channelisation is measured on the competitive market. The global comparison is on the same basis — net player losses — but is a 2024 figure.

Two readings from that table. First, the currency trap: on our constant denominator, per-adult licensed losses rose 1.3% in kronor between 2024 and 2025 but 9.2% in dollars. Nearly eight of those nine points are the krona strengthening from 10.575 to 9.8096. Any Swedish gambling figure quoted in dollars without a stated rate and year is close to meaningless.

Second, the scale. On licensed play alone a Swedish adult loses about 2.4× the global average of roughly US$132 per adult on Earth — both figures 2024, both net player losses. Include the implied unlicensed market and the 2025 Swedish range runs 2.8× to 3.4× that 2024 global benchmark. In physical terms: a Swedish adult’s licensed gambling losses in 2025 come to about SEK 64 a week, every week — and that is a population average spread across everyone, including the 38% of Swedish adults who did not gamble at all in the past year (Verian for Spelinspektionen, 2026). For a sense of the other end of the global distribution, our gambling in Kenya statistics page puts estimated Kenyan betting losses at around US$10.78 per adult a year, though on a bookmaking-only perimeter against Sweden’s all-products one, so read that as orders of magnitude rather than a clean ratio. The mechanism behind the Swedish figure is the same one we trace in how much money casinos make: a small hold on a very large volume of repeat play.

What changes in Sweden in 2026?

Three tightenings landed inside eight months, and all three act on the licensed side of the market. That is the through-line of Swedish policy since 2022 and it is worth stating explicitly, because the stated problem is offshore play.

• 1 January 2026 — land-based casino games prohibited nationwide. The last venue had already closed in April 2025. The licence category now reads zero.
• 1 May 2026 — ban on credit-funded gambling. Licensees and their agents may not allow or facilitate gambling financed by credit cards, loans, overdrafts or buy-now-pay-later, and must take active measures to prevent it. Note the date: the government proposed 1 April 2026 in Prop. 2025/26:11, and the Riksdag, deciding on 17 February 2026, adopted the bill with the commencement moved to 1 May. Spelinspektionen may grant exceptions for public-benefit lotteries.
• 1 August 2026 — new Spelpaus connection rules. Under SIFS 2026:3, decided 23 April and published 29 April 2026, licensees must query the national register through purpose-specific APIs using an Actor ID and API key issued to them individually, at direct marketing, at player registration and at player login. A check counts as performed only once it has returned a result showing whether the person is excluded. Sweden also added a 10-day self-exclusion option alongside the existing durations ahead of the switch-over.
• Proposed, not law — the Gambling Act review. Marcus Isgren, chair of the National Board for Consumer Disputes, was appointed in February 2025 and reported on 17 September 2025. His central proposal replaces the Act’s “directional” criterion — which bites only on operators actively targeting Sweden — with one that catches any site playable from Sweden. If adopted it would take effect on 1 January 2027. BOS wants the one-bonus rule relaxed on the argument that restriction drives leakage; others have proposed banning online casino bonuses outright. Both sides cite channelisation. Neither can cite an agreed number for it.

Q1 2026 gives the first read on all this: total licensed net turnover of SEK 6,679.7 million, up 0.78% year on year, with the online category up 3.36%. Flat, in other words — which after seven years is the most consistent finding in Swedish gambling data.

Why do Swedish gambling figures disagree?

Because “channelisation” is at least six different measurements wearing one word, and Sweden has never settled on which one counts. Every conflict we hit building this page reduces to one of these seven.

1. Share of what? Channelisation can be measured as a share of money, of players, or of website visits, and the three diverge sharply. For online casino in 2025 alone, Spelinspektionen’s own report contains 93% (share of turnover, player survey), 81% (headline mean of turnover indicators), 68% (turnover estimated from traffic) and 54% (raw share of visits). The first three are money; the fourth is not, and putting it in the same sentence without saying so is the single commonest error in coverage of this market. A regime can look near-perfect on visits or player counts and mediocre on money, because the people who play offshore play more.

2. Every method has a known bias, and none is quantified. The survey indicator rests on 220 online-casino players out of 6,744 respondents, and the regulator warns that a single answer can swing a small sub-sample. The traffic indicator misses apps, through which 38% of licensed turnover arrives, so it systematically undercounts the licensed share. Skin-betting sites are excluded from both indicators despite taking 42% of unlicensed visits, because their gambling share cannot be separated from their trading and news activity. Three known distortions, no published error bars.

3. Who is estimating, and what do they want? The regulator’s figure is the one the government is graded on. The trade body’s figures support relaxing rules its members find costly. ATG is a licensed operator competing directly with offshore casino sites. H2 sells subscriptions to people who need the number to be right. None of that makes any of them dishonest; it does mean the 68-to-89 spread is not random noise around a true value, which is why we publish all of them side by side rather than picking one.

4. The regulator’s own series is not a series. Spelinspektionen used a single method for 2022, averaged four indicators for 2023 (player survey, internet traffic, traffic-to-turnover and H2’s number), then dropped to two indicators for 2024 and 2025. Dropping H2 mattered, because H2’s Sweden estimate had just been cut to 72%. Nordic Legal notes that had the 2023 four-indicator method been retained, the 2024 headline would have come out lower than 85%. The apparent 86→85→84 decline is partly a methodology artefact and should never be plotted as a trend line.

5. Which market is in the denominator? Sweden’s channelisation figures cover the competitive market — betting and online casino. Svenska Spel’s lottery and value machines, the public-benefit national lotteries, hall bingo and land-based commercial gambling all sit outside it. Those excluded segments were SEK 9.79 billion of the SEK 28.20 billion licensed total in 2025, or 34.7%. Quote a channelisation rate as though it applies to all Swedish gambling and you have silently reweighted a third of the market.

6. Net turnover, turnover, and what “GGR” means here. Spelinspektionen’s statistics use nettoomsättning — stakes minus winnings paid out — which is the same quantity as gross gaming revenue and is what the 22% tax is levied on. It is not amount staked. Sweden’s SEK 28.2 billion is therefore not comparable with handle-based figures from markets that report turnover, an error we take apart in our gambling in Brazil statistics page.

7. Kronor, dollars, and which year’s rate. The krona ran at SEK 10.575 to the dollar in 2024 and SEK 9.8096 in 2025. The same SEK 28.2 billion is US$2.67bn at the 2024 rate and US$2.87bn at the 2025 rate — a 7.8% swing created entirely by the currency. Add Swedish inflation of 24.2% between 2019 and 2024 and a market that grew 12.3% in nominal kronor shrank about 9.6% in real terms. Nominal krona figures, dollar figures and real figures tell three different stories about the same market, and Swedish coverage routinely mixes them.

Key takeaways

• The measurement’s own spread is bigger than the shortfall. Spelinspektionen’s 2025 report gives 89% and 78% for the same market in the same year and publishes the midpoint, 84%. The internal band is 11 points; the gap to the 90% target is six. Sweden has spent seven years legislating against a number it cannot pin down to better than a fifth of itself.
• Both ends of that band are known to be wrong, in opposite directions. The high end rests on 220 online-casino players in a 6,744-person survey the regulator itself calls unstable at small sample sizes; the low end comes from a traffic method the regulator says undercounts licensed play, because 38% of licensed turnover arrives through apps it cannot see.
• Priced in kronor, the regulator’s internal disagreement is SEK 2.92 billion a year of player losses — about 78% of everything Sweden’s public-benefit national lotteries take in. Extend the band to every published estimator and it reaches SEK 6.39 billion.
• Channelisation peaked in year one. Sweden reported 88% in 2019 and has reported lower ever since. The reform did not fail to climb; it arrived near target and drifted, while every subsequent intervention tightened the licensed side.
• The licensed market rotated rather than grew. Online is up 31.3% since 2019, everything else down 9.1%, and the total was about 9.6% smaller in real terms in 2024 than in its opening year.
• The 2024 tax rise is invisible in the revenue data and expensive on the P&L. Licensed online GGR was shrinking at 18% and has grown in every half-year at 22% — but 70 öre of every extra krona of growth since 2023 went to the state, and the four points cost licensees about SEK 736 million in 2025 alone.
• Tax rates do not explain leakage. Germany taxes online slots at roughly 133% of GGR and its regulator reports 77% channelisation; Sweden taxes at 22% and its regulator reports 84%. Six times the burden, seven points of difference — and German industry estimates run 40–60% anyway.
• Spelpaus is the reform’s clearest success and part of its leak. 134,484 registered — 1 in 63 adults, 28% more people than Swelogs estimates at PGSI 3+ — yet it is the reason offshore players cite most often, and new registrations are down 26.8% in two years.
• Harm is falling on the state’s own instrument. PGSI 3+ prevalence went from 2.2% of adults in 2008–09 to 1.3% in 2021, with severe problem gambling stable at 0.3–0.6%. That trend predates re-regulation and should not be claimed by either side of the channelisation argument.
• What Sweden should publish next is not a better single channelisation number but a stated confidence interval around it, by vertical, on a frozen method. Until then, “Sweden missed 90%” is a sentence with eleven points of slack inside a single government report.

Frequently asked questions

How big is the Swedish gambling market in 2026?

Operators holding a Swedish licence booked SEK 28.20 billion of net turnover in 2025, up 1.3 percent on 2024, according to Spelinspektionen and Skatteverket. Net turnover means players' stakes minus winnings paid out, the same quantity as gross gaming revenue, and it is what the 22 percent gambling tax is levied on. Commercial online gambling, betting and shipboard gaming accounted for SEK 18.41 billion, or 65.3 percent of the total. In the first quarter of 2026 the licensed market turned over SEK 6.68 billion, up 0.78 percent year on year. All of these figures cover licensed operators only; play with unlicensed sites is not included and is estimated rather than measured.

What is Sweden's channelisation rate?

Spelinspektionen's published figure for 2025 is 84 percent of the competitive market, down from 85 percent in 2024 and 86 percent in 2023. That 84 percent is the mean of two indicators in the same report: a player survey giving 89 percent and an internet-traffic-to-turnover model giving 78 percent. The regulator states its own conclusion as a range of 78 to 89 percent. Outside estimators go lower: H2 Gambling Capital revised its Swedish figure to 72 percent in 2025, down from 91 percent, after rebuilding its model, and ATG's web-traffic study put it at 68 percent overall with online casino at 57 percent. The government target set in Prop. 2017/18:220 is 90 percent, so the shortfall on the published figure is six points while the regulator's own two measurements sit eleven points apart.

Why do Swedish channelisation estimates disagree so much?

Because each method has a known bias and none of them is quantified. Spelinspektionen's survey indicator rests on a small sub-sample: of 6,744 adults surveyed by Verian in February and March 2026, only 220 had played online casino in the past year, and the regulator warns that a single answer can move a result at that sample size. Its traffic indicator estimates visits and converts them to turnover, but web-traffic tools do not see app sessions, and the regulator's own operator survey found 38 percent of licensed turnover arrives through apps, so the traffic method undercounts licensed play. Skin-betting sites are excluded from both indicators even though they took 42 percent of Swedish visits to unlicensed gambling sites in 2025. On top of that, the estimators have different interests: the regulator's figure is the one the government is graded on, ATG is a licensed operator competing with offshore casinos, and BOS is the online gambling trade body.

Why is Sweden's online casino channelisation so much worse than betting?

Because online casino has no product moat and Swedish betting does. In Spelinspektionen's 2025 report, betting moves only three points across methods, at 98 percent on the player survey and 95 percent on the traffic indicator, while online casino moves twenty-five, at 93 percent and 68 percent. ATG's study found betting at 77 percent and online casino at 57 percent. Online casino games are the same titles from the same suppliers whether the site holds a Swedish licence or not, so switching costs nothing, while a large share of Swedish betting is harness racing pooled inside ATG's tote product, which cannot be replicated offshore. Spelinspektionen had identified 2,186 gambling websites without a Swedish licence as of 30 April 2026, of which 976 offer online casino only and 800 offer casino plus betting, against six offering betting alone.

What is the gambling tax in Sweden?

Twenty-two percent of gross gaming revenue, calculated monthly under the Gambling Tax Act 2018:1139. The rate rose from 18 percent on 1 July 2024, with the government forecasting around SEK 540 million a year in additional revenue. Gambling for public benefit purposes, such as charity lotteries, is outside the charge. Licensed online operators kept SEK 78 of every SEK 100 of player losses after the rise rather than SEK 82. On 16Best's analysis of the regulator's data, the four-point rise cost licensed online operators roughly SEK 736 million of retained revenue in 2025, equal to 52 percent of all the extra gross revenue that segment added between 2023 and 2025.

Did Sweden's gambling tax increase push players offshore?

The licensed revenue data does not show it. Licensed online net turnover was contracting year on year in both halves of 2023, when the rate was 18 percent, and has grown in every half-year since the rise to 22 percent on 1 July 2024: up 4.05 percent in the second half of 2024, 0.84 percent in the first half of 2025 and 5.69 percent in the second half. That does not prove the tax was harmless, because half-year growth is volatile and responds to many things, but it does contradict the strong version of the claim that the rise triggered an immediate exodus. For comparison, Germany taxes online slots at a rate equivalent to roughly 133 percent of gross gaming revenue on a 96 percent RTP game and its regulator reports 77 percent channelisation, against Sweden's 84 percent at a 22 percent rate.

How many people have registered on Spelpaus?

134,484 people were registered as self-excluded at the end of 2025, and 136,955 on 29 June 2026. That is about 1.58 percent of Sweden's roughly 8.49 million adults, or one in 63. Three in four registrants are men. New registrations have fallen for two consecutive years, from 18,953 in 2023 to 16,306 in 2024 and 13,877 in 2025, a decline of 26.8 percent. The register is a stock of currently excluded people rather than a cumulative total, since exclusions can lapse. Spelpaus launched with the licensed market on 1 January 2019 and blocks a registrant from every Swedish-licensed operator at once. From 1 August 2026, under regulation SIFS 2026:3, licensees must check the register through dedicated interfaces at direct marketing, registration and login, and a 10-day exclusion option has been added alongside the existing durations.

Sources

• Spelinspektionen — Kanaliseringsgrad på den svenska spelmarknaden 2025 (report dated 15 June 2026, ref. 25Si2603. Table 6: competitive market 89% player survey, 78% traffic-to-turnover, 84% headline; betting 98% / 95% / 96%; online casino 93% / 68% / 81%. Table 2: 6,744 respondents aged 18+, 4,175 gamblers, 903 competitive-market players, 220 online casino. Table 4: licensed share of visits 61% including skin betting, 73% excluding, 54% online casino, 92% betting. Table 1: 2,186 unlicensed sites to 30 April 2026, 334 under prohibition orders. Also the 38% app-turnover footnote, the 42% skin-betting visit share, the reasons-for-offshore-play chart with 53 responses, and the stated 78–89% range)
• Spelinspektionen and Skatteverket — Omsättning på spelmarknaden per kvartal (quarterly net turnover by licence type, 2019 to Q1 2026; the primary series behind every SEK figure and growth rate on this page; 2025 marked preliminary)
• Spelinspektionen — Statistics (2024 net turnover SEK 27,842m; SEK 3,325 per inhabitant aged 18+; 0.9% of disposable income)
• Spelinspektionen — Kanaliseringsgrad på den svenska spelmarknaden 2024 (1 September 2025; 85% headline; betting 92% survey and 96% traffic, online casino 82% survey and 72% traffic; the 2024 reasons-for-offshore-play figures of 35%, 23%, 21%, 15% and 13%)
• Spelinspektionen — Den offentliga kontrollen över spelmarknaden — kanaliseringsgrad (September 2024; four-indicator method; 2023 rate of 86%)
• Spelinspektionen — SIFS 2026:3, föreskrifter om det nationella självavstängningsregistret (decided 23 April 2026, published 29 April 2026, in force 1 August 2026; checks required at direct marketing, player registration and login, via purpose-specific APIs and licensee-specific Actor ID and API key)
• Spelinspektionen — Utökat kreditförbud träder i kraft den 1 maj and Sveriges riksdag — Prop. 2025/26:11, Ett nytt förbud mot spel på kredit (government proposed 1 April 2026; Riksdag decision of 17 February 2026 set commencement at 1 May 2026)
• Nordic Legal for BOS — The Channelisation Challenge: Lessons from Two Nordic Markets (13 October 2025; the 90% target sourced to SOU 2017:30 and Prop. 2017/18:220; H2 Gambling Capital revision from 91% to 72% with betting 83% and online commercial 62%, and Denmark from 87% to 72%; ATG Q1 2024 study at 68% overall, 57% online casino, 77% betting, Denmark 90–95%; ATG Q4 2024 at 69–82% with online casino 59–74% and an unlicensed market of SEK 3.9–7.7bn a year; SKOP 2023; Statskontoret 2020 and 2021 benchmarks)
• iGaming Business — Channelisation in Sweden Slips to 84% in 2025 and InterGame — Sweden’s iGaming Channelisation Rate Declines Again (the two trade reports of the June 2026 release, one quoting the traffic indicator and one the headline; ATG Q4 2025 at 73–84% against Q4 2024 at 69–82%)
• iGaming Business — Swedish Treasury Orders Affiliate Crackdown as Channelisation Misses Target (1 April 2022; channelisation of 50% in 2018, 88% in 2019, 85% in 2020, 87% in 2021)
• iGaming Business — Sweden’s Channelisation Rate at “Critically Low” 77% (SKOP for BOS, fieldwork February–March 2023, published June 2023; 9,850 interviewed; sports betting 84%, online casino 72%)
• CMS — The Swedish Government Proposes Increased Excise Tax on Gambling from 18 to 22 Percent and SBC News — Swedish Government Proposes Gambling Tax Increase in 2024 (rate effective 1 July 2024; forecast SEK 270m in 2024 and SEK 540m a year thereafter)
• ICLG — Gambling Laws and Regulations Report 2026 — Sweden (22% of GGR under the Gambling Tax Act 2018:1139; the single first-participation bonus rule; the måttfullhet advertising standard; land-based casinos prohibited from 1 January 2026)
• iGaming Business — Svenska Spel Closes Remaining Casino Cosmopol Venue and Svenska Spel — Svenska Spel stänger Casino Cosmopol i Stockholm (ownership instruction 24 April 2025, closure announced 25 April 2025, around 240 staff affected; Sundsvall closed 2020, Gothenburg and Malmö February 2024)
• Spelinspektionen — Årsredovisning 2025 (134,484 people registered on Spelpaus at the end of 2025, an increase of almost 14,000 on 2024) and Bonusetu Press via TribuSoft — Två av Spelinspektionens mått faller i takt (23 July 2026; new registrations of 18,953 in 2023, 16,306 in 2024 and 13,877 in 2025; 75.6% male; 136,955 at 29 June 2026)
• SBC News — Spelpaus Reaches 80,000 Swedish Player Self-Exclusions and CasinoBeats — Sweden’s Spelpaus Registers 100,000 Self-Exclusions (September 2022 and October 2023 milestones)
• Folkhälsomyndigheten — Statistik över spelproblem i Sverige (Swelogs 2021, PGSI: 0.5% at PGSI 8+ or ~40,000 people; 0.8% at PGSI 3–7 or ~65,000; 3.0% at PGSI 1–2 or ~247,000; 130,000 sharing a household, 40,000 of them children; close to 2% of men against 0.7% of women)
• iGaming Business — Sweden’s Problem Gambling Rate Falls, Still “Scope for Improvement” (analysis by Ola Nevander for BOS, published 31 March 2026: PGSI 3+ down from 2.2% in 2008–09 to 1.3% in 2021; PGSI 8+ stable at 0.3–0.6%; online gamblers with problems down from 12% to about 4%)
• European Gaming — Swedish Government Appointed Marcus Isgren as Investigator of Swedish Gambling Act Review and iGaming Business — Sweden Targets Unlicensed Operations with Gambling Act Amendment (appointed February 2025, reported 17 September 2025; proposed participation criterion in place of the directional criterion, effective 1 January 2027 if adopted)
• Statistics Sweden (SCB) — Population statistics and UN World Population Prospects (8,493,129 residents aged 18+ at mid-2025, the denominator for every per-adult figure here)
• World Bank — Inflation, Consumer Prices (Annual %), Sweden (0.50% in 2020, 2.16% in 2021, 8.37% in 2022, 8.55% in 2023, 2.84% in 2024 — the basis for the 24.2% cumulative figure and the real-terms calculation)
• Exchange-Rates.org — USD to SEK Exchange Rate History (2024 annual average SEK 10.575 per US dollar; 2025 annual average SEK 9.8096 — the two rates used for every dollar conversion on this page)
• H2 Gambling Capital — global player losses of about US$573bn in 2024, roughly US$132 per adult on Earth, the benchmark used across our gambling loss statistics and sports betting laws by country