The number that is actually in dispute
Modelling by EY, cited by the Betting and Gaming Council, puts the cost of raising Machine Games Duty to 40 per cent at up to 16,000 jobs, almost 1,500 betting shops and as many as 34 casinos, while leaving the Treasury up to £124 million worse off. The figure being campaigned on ahead of the Budget, though, is 36,000. The gap between those two numbers is where the whole gambling tax policy impact argument lives, and almost nobody separates them clearly.
Here is the distinction. The BGC says Britain’s betting shops support more than 36,000 jobs in total. That is a headcount, not a forecast. The 16,000 figure is the modelled loss under one specific scenario, a Machine Games Duty rate of 40 per cent, which sits well above the current MGD bands. Read quickly, the campaign reads as though 36,000 jobs are at risk. Read properly, it says a large workforce exists and a specific tax change could remove a sizeable slice of it.
That matters for anyone outside the UK reading this as a precedent. The mechanism being described is real and well understood. The precision of the numbers attached to it is not, because they are projections commissioned in the middle of a live lobbying fight, and they depend on behavioural assumptions nobody can verify in advance.
Why the shop feels it before the app does
High-street betting shops carry fixed costs that online operators do not: commercial rent, business rates, and rostered staff who must be present whether twelve customers come in or two. Those costs do not shrink when turnover does, so a duty increase lands on an already compressed margin rather than on a variable cost base.
An online operator facing the same percentage rise can cut marketing spend, slow product development, or shift acquisition budget between markets. A shop manager can cut opening hours and headcount, and that is roughly the end of the list. This is why betting shop jobs are the first visible casualty in almost every gambling tax dispute, in any jurisdiction with a retail estate.
The workforce profile sharpens the political edge. According to the BGC, more than 80 per cent of betting shop jobs sit outside London, in market towns, former industrial areas and seaside communities. Nineteen per cent of employees are under 25 against 11 per cent across the wider economy, and 51 per cent are under 35 against 35 per cent nationally. Women make up half the workforce. Those are the characteristics of employment that is hard to replace locally, which is precisely why the trade body leads with them.
Where the cost bases actually differ
| Cost or exposure | High street shop | Online operator |
|---|---|---|
| Premises | Rent plus business rates, fixed per site | Negligible per unit of turnover |
| Staffing | Rostered, must cover opening hours | Centralised, scales with volume |
| Main gaming duty exposure | Machine Games Duty on terminals | Remote gaming duty on gross gambling yield |
| Response to a duty rise | Cut hours, cut staff, close sites | Cut marketing, reprice, reallocate between markets |
| Ability to relocate | None, the site is the business | Licensing permitting, can shift jurisdiction |
Where the gambling tax policy impact actually lands
Gambling industry taxation does not just raise money, it decides which operators survive. Duty rates, the base they are charged on, and how they differ between channels together determine who can absorb a cost increase and who has to exit. That is the part of betting tax regulation that gets least attention and does the most structural work.
Consolidation before closure, then closure
A large multi-channel group can offset retail pressure against an online book, renegotiate leases across hundreds of sites, and carry loss-making shops while it restructures. An independent with nine shops cannot. So the first-order effect of a duty rise is rarely a clean reduction in supply. It is a transfer of market share to whoever has the balance sheet to wait, followed by closures concentrated among the smallest operators and the weakest sites.
The policy argument usually treats closures as the outcome. In practice, market consolidation is the earlier and more durable one. A market can end up with the same volume of betting, fewer employers, and less competitive pressure on terms offered to customers. Whether that is a good or bad result depends on what the tax was for, which brings us to the harder question.
The migration question nobody can size properly
Higher duty on one channel pushes volume toward cheaper channels. Domestically that means online, which is taxed on gross gambling yield rather than per terminal. Beyond the licensed perimeter it means unlicensed offshore sites, which pay no UK duty at all and apply none of the UK’s player protections. The point-of-consumption approach the UK adopted in 2014, taxing operators on the basis of where the customer is rather than where the licence sits, exists specifically to limit that leakage.
It limits it. It does not eliminate it. Offshore gambling migration is the claim most likely to be stated with more confidence than the evidence supports, by both sides. Industry submissions treat it as a predictable elasticity; critics treat it as scaremongering. In reality, nobody has a reliable measure of unlicensed market share before a tax change, which makes any precise forecast of the share after one an estimate dressed up as arithmetic. If you are assessing a market’s exposure to this, the licensing regime matters more than the duty rate, because enforcement capability decides whether offshore substitution is a genuine option for ordinary customers. The practical test is whether a market can identify and block unlicensed sites at all, which is the same question that sits behind how to spot illegal casinos and licensing red flags”>licensing red flags that expose illegal casinos.
What the 2019 reset already demonstrated
The UK has run this experiment before. In April 2019 the maximum stake on fixed odds betting terminals was cut from £100 to £2, a regulatory rather than a tax measure, but one that hit the single most profitable product in the retail estate. Remote gaming duty was raised from 15 per cent to 21 per cent in the same period, partly to cover the expected revenue shortfall. Shop closure programmes followed within months, including William Hill’s announcement that year that it would close around 700 shops, explicitly citing the stake cut.
Two things follow from that episode, and only two. First, the direction of effect is established: squeeze retail gaming revenue and a meaningful number of shops close. Second, the precise magnitude was not well predicted in advance by anyone, including the operators. Readers citing 2019 as precedent should check closure counts and duty rates against the Gambling Commission’s industry statistics and HMRC’s published duty receipts rather than against trade-body summaries, including this one.
Two arguments that do not meet in the middle
The industry case is about employment concentration. It argues that retail gambling supports a workforce that is young, half female, and overwhelmingly located outside the capital, and that a Machine Games Duty rise to 40 per cent would be self-defeating, costing the Treasury up to £124 million while removing up to 16,000 jobs, on the EY modelling the BGC cites. The economic logic is that above some rate, lost duty from closed shops and migrated volume exceeds the gain from the higher rate.
The Treasury case is about revenue and, in the wider policy debate, about harm. Gambling duties are a direct source of receipts, the sector’s gross gambling yield is large, and a government writing a Budget weighs that against sectoral employment claims it did not commission and cannot independently verify. Where the argument includes a gambling levy or harm-reduction rationale, a reduction in high-intensity machine play is not treated as a cost at all.
Both positions can be internally consistent and still irreconcilable, because they disagree about the elasticity, not the principle. The industry assumes customers and operators respond sharply; the Treasury assumes they respond less. Until a rate is actually set, neither claim is testable. That is the honest state of the debate as the Budget approaches.
The trade-off every market eventually has to price
Any jurisdiction raising gambling duty is choosing between near-certain revenue and uncertain structural damage, and the uncertainty is asymmetric. Receipts from a higher rate can be estimated reasonably well. The employment and market-structure response cannot, because it depends on operator decisions that are themselves partly strategic, made in public, by companies with an interest in the forecast being believed.
Three design questions travel to any market, whatever the local numbers. Does the duty apply evenly across retail and online, or does it tilt volume toward whichever channel is taxed more lightly? Is the tax base turnover or gross gambling yield, which determines how hard it bites on thin-margin products? And can the regulator police the offshore boundary well enough that a rate increase does not simply export the activity?
What the UK case does not yet settle is whether a duty rise large enough to shrink the retail estate also shrinks total gambling activity, or merely moves it. Those are very different outcomes for public finances and for harm, and the 2019 precedent points more toward displacement than reduction without proving it. Anyone using the 36,000 figure as shorthand for jobs at risk is reaching past what the modelling says. Anyone dismissing the retail exposure entirely is ignoring what happened the last time the UK pulled this lever.