Governments have spent the past decade raising the duty they charge licensed gambling operators, and several are now discovering that the activity they taxed has moved somewhere their revenue authority cannot reach. Brazil’s Congress is trying to strip casino games out of a betting market it legalised two years ago. Australian states diverge on what they permit and collect. The common thread is not a disagreement about gambling policy. It is a question about where a state’s taxing power stops once the counterparty is incorporated elsewhere.
Malaysia shows the problem in its clearest form. Its licensed sector files audited accounts and its unlicensed one files nothing at all. Ministry of Finance figures put federal gaming tax collection at roughly RM80mil a week. At the same time, several state governments have moved to remove licensed operators from their territory altogether: Kedah stopped renewing the business licences of gaming outlets through its local councils, and Perlis restricted alcohol sales at coffee shops and restaurants at the start of 2022. Both measures were argued on social grounds, and neither was presented as fiscal policy. When the prospect of a more conservative federal administration arose after the 2022 general election, RM3.38bil came off the combined market capitalisation of the six listed gaming and brewery counters within days.
What the federal government actually collects, and from whom
The taxed side of that arrangement is small enough to count. Three number forecast operators run the 4D draws: Sports Toto, Magnum and Da Ma Cai, the last of them operated by the unlisted Pan Malaysia Pools. One casino licence covers one resort. Everything else Malaysians stake sits outside the tax base completely.
Those collections are not marginal to the federal budget. Affin Hwang Investment Bank put the number forecast industry’s annual contribution at close to RM3bil in a 2022 note, and estimated Genting Malaysia’s own gaming tax bill at more than RM1bil a year.
That gap is not an enforcement failure in the usual sense. It is a structural feature of a country that built a tax apparatus for gambling in one form and never built a licensing regime for the form most of the activity later took.
The detail that shapes everything else is the base the tax is calculated on. Number forecast operators pay gaming tax on gross sales, meaning the total amount wagered, before any prize money leaves the business. A gaming analyst quoted by The Edge described three separate levels of tax reaching the NFOs: one netted off before net gaming revenue is calculated, one inside cost of sales, and then ordinary corporate income tax at 24 per cent. Pool betting duty and the sales and services tax sit on top.
Casino operations are taxed on gross gaming income rather than turnover, but at a rate that has moved in one direction. The 2018 budget raised casino duty to 35 per cent from 25 per cent and gaming machine duty to 30 per cent from 20 per cent. The annual casino licence fee went up by RM30mil to RM150mil, and the machine dealer’s licence went from RM10,000 a year to RM50,000. It was the first increase in casino duty in roughly two decades, the previous one having taken the rate from 22 per cent to 25 per cent during the 1997 and 1998 financial crisis.
The untaxed side is not an accident of geography
The standard way to describe the other half of the market is as offshore spillover, as though Malaysian money drifts by default toward whichever operator happens to accept it. The evidence does not support that reading. The operators taking the volume are localised, and deliberately so.
They quote in ringgit rather than dollars, they run Malay-language interfaces, and they publish welcome terms denominated to the sen. A Malay-language review of the casino sites serving Malaysia lists operators such as Longfu88 and DK88 that do not appear in English-language market summaries at all, next to ringgit offers quoted as precisely as RM2,988 and RM1,588. That specificity is the point. A business that sets a bonus at RM1,588 has done the work of pricing for one market, which is a different proposition from an international site that happens to accept a Malaysian card.
For the tax question, this matters more than the moral framing usually applied to it. Volume moving out of the licensed sector is not dispersing into an undifferentiated global pool where it becomes untraceable. It is consolidating with operators who have made a considered commercial decision to serve Malaysia and who carry no Malaysian tax line at all, while Sports Toto and Magnum pay duty on every ringgit staked with them before a single prize is paid out.
The measurement problem is real, and it cuts both ways
Figures for the size of the unlicensed market circulate widely and should be treated with care. Estimates in the region of RM18bil a year in stakes, with something around RM2bil in forgone tax, appear regularly in trade coverage, but they rest on methodology that is rarely published and almost never audited. The honest position is that nobody outside the operators knows the number.
What can be observed is the direction of the licensed figures, which are audited and filed. When a taxed operator reports declining sales while nothing suggests Malaysians have stopped gambling, the difference has gone somewhere. Whether that somewhere is large enough to justify the numbers quoted is a separate question from whether the movement is happening.
The same pattern has been visible in markets that did write online rules. Our coverage of Brazil’s move to strip casino games out of the betting framework it built two years ago describes a government discovering that a licensed channel is easier to tax than an unlicensed one, and harder to reopen once closed. New Jersey’s record internet gaming month makes the opposite case with the opposite policy.
What a licensing decision would actually be about
Discussion of online licensing in Malaysia tends to be conducted as a debate about permission, and that framing is why it goes nowhere. The activity is already occurring at scale. The decision in front of any government is narrower: whether the state wants a revenue relationship and a supervisory relationship with that activity, or neither.
Malaysia has chosen neither, and has done so consistently across administrations of different complexions. There is a defensible argument for that position on social policy grounds. What is harder to defend is the current allocation, in which the operators who submit to audit, pay duty on turnover, and lose outlets to state licensing decisions are competing for the same customer as operators who do none of those things.
Africa’s regulatory consolidation and Australia’s state-level divergence suggest that the question eventually forces itself, usually when a finance ministry rather than a social policy ministry starts asking it. Malaysia’s gaming tax base is concentrated in a small number of filers whose retail channel is narrowing. That is the kind of arrangement that tends to be revisited by whoever has to balance the next budget.
