Regulus: ‘Verhoging kansspelbelasting kan staatskas € 55 miljoen kosten’

Netherlands: GGR tax increase proposals – whack-a-whale?

One of the first policy announcements of the new right-wing Dutch government coalition set to form a government is to increase gambling taxes, including online GGR tax from 30.5% to 37.8%. The government hopes to raise €202m from both landbased and online gambling tax increases on a projected 2024 gambling tax base of €1bn, implying the government expects no material shift in customer behaviour caused by the tax. The Dutch online gambling trade body, NOGA, by comparison has warned of a significant potential shift to the black market. So, who is right?

We have long argued that black market risks are more nuanced than many gambling industry lobbyists present, but that does not mean they are not real (see our January 2021 blog ‘What creates a Black Market’). The simple key to black markets are policies which create distortions that prevent customers from doing what they want to do legally when a workable offer is more-or-less readily available in an offshore environment. Crucially, the higher the value of the customer, the more likely both they and black market industry participants are to jump regulatory or UX hurdles in order to continue gambling the way they want to. On this basis, it can be seen that separate Dutch parliamentary proposals to ban slots and other ‘high risk games of chance’ are a very clear case for significantly growing the black market. However, the proposed tax hike is more complex. 

The prima facie case against high taxes being distortive and causing black markets is Pennsylvania. When PA set its online slots tax at 56% of ‘GGR’  in 2019, practically every stakeholder and commentator howled that this would be a total disaster for the state’s domestic market. That it has barely even impacted the market can be seen on two levels. First, PA annualized run-rate online gaming GGR per capita for Q124 is US$160. The equivalent reported figures for NJ and MI are US$245 and US$234 respectively – optically much higher, but the comp state GGR figures contain bonuses. If we apply a 30% incentive discount to NJ and MI online gaming GGR to make them comparable with PA figures, we get to an equivalent US$170 and US$164. In other words, PA is trending exactly in line with other US online gaming markets in terms of overall online gaming revenue per capita despite very high revenue taxes on a key product. Second, while only PA of the major gaming states breaks down revenue for slots vs. table games, a slots mix of 71% (incuding poker) is exactly in line with international averages and what we would expect to see based upon industry experience. Consequently, we can state with some certainty that there is no product distortion and no revenue distortion from PA’s 56% online slots tax; just more money for the state and less for the licensee on that major online gaming product line. On the face of it, the PA example appears to end the debate in the favour of the Dutch government in waiting, but it doesn’t.

PA is generous in two critical aspects of its online gaming tax policy: online taxes are levied on revenue net of bonuses; and table games (including poker) are taxed at a much more lenient 16% of revenue. The PA blended rate of tax on GGR is therefore c. 30%, or pretty much where the Netherlands is now. By comparison, the effective online gaming tax rate on revenue is 23% for NJ and 27% for MI (which allows a 10% bonus deduction): lower than PA, but nowhere near as different as the headline 15% and 20% GGR rates suggest.

However, the issue is not just a question of where the blended tax rate comes out but also what the tax does to the market environment for the consumer. A GGR tax that includes bonuses drives up the cash cost of offering ‘free money’ as incentives: it is no longer free to the licensee. To an empirical level of c. 20% GGR taxes it would seem that most operators find it safer to take the hit than risk their competitive position with lower incentives. We already know that the Netherlands is a competitive gaming environment: the re-entry of Kindred into the market has caused at least Toto, Holland Casino and Entain revenue in the market to fall in 2023, with other less visible losers also likely. Efficiently or not, incentives are a key way to attract customers in a competitive market: the tax cost of a €100 sign on bonus will increase from €31 to €38 under the proposed tax change – reducing the ability of licensed operators to offer incentives.

From a customer standpoint, we know that many regular online gamers like to hunt the internet for attractive bonuses: it is a core engine of the affiliate market. We also know that VIPs like to be treated with bonuses as a retention method. The scale of this activity is hard to quantify in the Netherlands but given that the top 1% of customers usually represent c. 30-40% of revenue in a gaming-led product environment, we can guess it is significant even in a maturing market, with bonuses likely representing c. 25% of GGR in 2023, we believe (the US bonusing environment tends to be more aggressive than in Europe). The effective tax rate on 37.8% of GGR if Dutch operators to not cut incentives is therefore 50% of net revenue (37.8% of 133). We doubt whether the government-in-waiting recognises it is proposing an obviously distortive business tax of 50%, but that is the effect nevertheless. For Dutch online licensees, the only way to mitigate a tax this high is to cut incentives – there is a dangerously straight line of mitigation down to no bonuses at a still high but arguably workable 37.8% revenue tax (workable, that is, if not offering bonuses was a real consumer choice with no revenue consequences).

It is hard to imagine Dutch regular gamblers and VIPs simply accepting lower bonuses in the licensed environment and continuing to play. While the KSA has had many successes in stopping operators in Malta especially from targeting Dutch players, it is far harder to stop Dutch players from using English-language sites licensed outside the EU – in the self-governing Dutch Caribbean island of Curacao for example. The KSA believes its channeling rate is currently c. 90% – implying an illegal market of c. €150m in GGR terms and c. €80m in revenue terms. This is already a sizeable base of working illegal supply to expand, albeit a relative channeling success. The whack-a-mole of finding and finding operators breaching Dutch law has been partially working. However, if Dutch licensed operators are forced to cut incentives to manage a growing tax burden, we anticipate that many high value customers and regular customers will be drawn to the 33%+ incentive levels found in the black market. The channeling rate could therefore fall well below 80%. Given continued mass market adoption, the optics of this will be an ex-growth market rather than a declining one, but it will be a failure nevertheless. There is also a dangerous risk to measuring channelling by the number of players: the players that are most likely to seek the black market are by default higher value. In extremis a 90% channelling rate by number of players could suggest an 80% black market in revenue terms. Following the money matters at least as much as following the traffic, especially for judging fiscal policies.

Even putting to one side the safer gambling problems caused by encouraging to use the black market, the tax increase could be a material failure in purely fiscal terms also. The incremental value of 37.8% tax in 2025 on an unchanged online gambling market according to our forecasts would be €133m (on 25% bonusing), supporting the government’s overall yield calculations. However, if domestic licensees dropped incentives from 25% to 10% to mitigate the cost increase on non-cash revenue, the incremental yield would be just €20m even if there were no change to channelling. Far worse, a 10ppt shift in channelling caused by the improved relative attractiveness of black market bonuses would cost the existing tax base c. €55m per annum. Ie, the proposed Dutch tax increase could very well make current online gambling tax yields go down, not up, based upon some simple and conservative player behaviour assumptions.

In short therefore, we believe that NOGA is right. An increase in GGR tax to 37.8% does make the Dutch market uncompetitive and is likely to grow the channeling gap to the extent that the tax increase becomes counter-productive on its own fiscal terms as well as significantly reducing player safety in the Dutch market. 

As a policy recommendation, we strongly believe that GGR taxes are distortive and dangerous at the level at which they make incentives too expensive for the domestic market to offer freely (empirically c. 20%). Online gambling taxes should be levied on revenue net of bonuses, not GGR, allowing policy-makers to set tax rates without causing counter-productive market distortions. Incentives need to be regulated, but they should not be taxed or policy-makers risk undermining the very markets they are trying to improve.