Brazil's Finance Ministry Told a Records Request It Had No Impact Study the Day Before the Betting Ban

A freedom-of-information reply described by BNLData says the ministry held no assessment of what banning betting would do, though the memorandum its minister signed the same day priced the tax loss. The R$38.8 billion social-cost figure the government cited was built from English estimates, and its authors say it shows association, not causation.

• The Finance Ministry replied on 24 September to a freedom-of-information request that it held no study, technical note or estimate of the economic, fiscal, regulatory or social impact of banning fixed-odds betting, according to the trade outlet BNLData, which says it obtained the request and the reply
• President Luiz Inácio Lula da Silva signed Provisional Measure 1.394 the next afternoon; its explanatory memorandum, dated 24 September and signed by the finance and justice ministers, estimates lost tax revenue of about R$1.54 billion this year and R$5.15 billion in 2027
• The R$38.8 billion annual "social cost" cited at the launch comes from a December 2025 dossier by the Institute for Health Policy Studies (IEPS), which applied England's per-gambler cost estimates to Brazilian survey data
• IEPS says the harms it costs are "associated" with gambling because the evidence does not allow it to say they were caused by it; BNLData, an industry-focused outlet, argues the figure has been presented as a measured cost
• The findings arrive as operators ask the Supreme Federal Tribunal to suspend the ban, arguing it has no demonstrated urgency

A Reply Registered the Day Before the Measure Was Signed

The request was filed under Brazil's Access to Information Law (LAI) on 20 September, with protocol number 18800.145984/2026-44, and addressed to the Finance Ministry's Secretariat of Prizes and Betting (SPA) and its Secretariat of Economic Policy (SPE), according to BNLData. It asked whether any document existed, was being prepared or had been commissioned on the impact of three scenarios: a total ban on fixed-odds betting, a ban or restriction on online games including casino, and a ban or substantial restriction on advertising and sponsorship. The reply was registered at 17:33 on 24 September and classified as "Informação Inexistente", or information non-existent. BNLData does not name the person who filed the request. iGaming Times could not find the reply in the government's public LAI search tool, so the account below rests on BNLData's description of the document.

According to BNLData, the ministry's Citizen Information Service said it had consulted the technical units and found no documents assessing the economic, fiscal, regulatory or social impact of a total or partial ban. It listed the estimates it did not have, including the effect on tax revenue, funding for sport, club sponsorship, the migration of bettors to unauthorised operators and public health, and said producing them would require multidisciplinary studies that do not yet exist. The reply also said no unit anywhere in the federal administration had been formally charged with assessing the ban's impact, and that the SPA's remit is to regulate, authorise, monitor, supervise and sanction under the existing laws. BNLData notes the reply spoke only for the SPA, not the SPE, and that the request was not forwarded elsewhere.

BNLData sets the reply against the timeline: on 23 September Lula met Finance Minister Dario Durigan and other ministers to settle the measure, the reply came the next day, and the measure was signed in São Paulo on the afternoon of 25 September and published that night.

The Memorandum Did Put a Number on the Tax Loss

The explanatory memorandum sent to the president with the measure, EXM No. 2114/2026, is dated 24 September and signed by Durigan and Justice Minister Wellington César Lima e Silva. It says "studies estimate" that household transfers to betting platforms in 2025 "may have been tens of billions of reais", without naming the studies, and that the available evidence shows the regulatory instruments have not protected health, consumers, children and family finances adequately. To satisfy the constitutional rule requiring legislation that cuts revenue to be costed, it estimates a reduction in tax revenue of about R$1.54 billion (approximately $297 million) in 2026, R$5.15 billion in 2027 and R$5.33 billion in 2028, plus a loss of supervision fees of about R$268.8 million over the three years, and says there is no loss from licence fees because they have been paid in full. It does not show how those figures were reached. They are well below the R$9.91 billion in federal tax the sector paid in the first eight months of this year.

Where the R$38.8 Billion Came From

At the launch, according to the Ministry of Justice's release, Health Minister Alexandre Padilha "cited a survey" estimating the annual social cost of online betting at R$38.8 billion (approximately $7.5 billion), "considering loss of productivity, time off work and the impact on health services". The release does not name the source. The figure matches the IEPS dossier "A Saúde dos Brasileiros em Jogo", published on 2 December 2025 with the Joint Parliamentary Front for the Promotion of Mental Health (FPSM) and the Umane foundation.

IEPS describes its method plainly. It took the shares of the population aged 14 and over that the third National Survey on Drug Use (LENAD III), conducted by the Federal University of São Paulo in 2022 and 2023, classed as low or moderate-risk gamblers (6.5%) and problem gamblers (0.8%), which it puts at about 12.8 million people. It applied the costs per type of gambler estimated for England in 2023 by the UK government's Office for Health Improvement and Disparities (OHID), converted at a purchasing-power-parity rate of R$3.40 to the pound, to the 2025 population. The result is R$38.8 billion a year: R$17 billion for additional deaths by suicide, R$10.4 billion for lost quality of life from depression, R$3 billion for depression treatment, R$1.3 billion for homelessness, R$2.1 billion for unemployment benefits and R$4.7 billion for imprisonment, with R$30.6 billion, or 78.8%, related to health. IEPS calls the estimates conservative because they include only harms that could be costed. A footnote states that "the results are evidence of cases associated with problem gambling, it not being possible to establish causality", and the chart on IEPS's public release says the harms are "associated with gambling because, given the limited evidence, it is not possible to state that such harms were caused directly by problem gambling".

In a fact-check published on 29 September, BNLData argues that the figure is a "three-stage extrapolation": British cost data, converted at an exchange rate, applied to a Brazilian proportion taken from a drugs survey not designed to measure the causal effect of betting on health. It says the monetary value placed on lost quality of life rests on assumptions the dossier's publicity does not detail, and that comparing the estimate with the R$6.8 billion in tax collected in the first nine months of 2025 sets a projection against an audited number. It accepts that problem gambling exists and needs public policy. BNLData covers the lottery and betting industry and carries operator advertising.

A Records Reply Shows What Was Filed, Not Everything That Was Considered

A finding of "information non-existent" means the units consulted held no qualifying document, and the memorandum shows that at least a revenue estimate existed on the same day. Other departments, such as the Health Ministry, which presented health data at the launch, may hold analysis the Finance Ministry's units do not. But the substance of the reply, as BNLData reports it, is harder to explain away: nobody in government was asked to estimate how many bettors would move to illegal sites, what would happen to sport's funding or how much enforcement would cost. The memorandum rests its case on "the available evidence" and unnamed "studies" without identifying them. That gap matters now, because the operators' petition to the Supreme Federal Tribunal argues the measure fails the constitutional test of urgency, and the government's answer will have to show more than the memorandum does.

The R$38.8 Billion Transplants England's Harm Profile Rather Than Measuring Brazil's

The IEPS figure is an openly described extrapolation, not an invention, and OHID itself regards its English figures as underestimates. But its limits are real. OHID put England's total annual cost at between £1.05 billion and £1.77 billion and presented the suicide cost as a range, on its expert panel's advice, to reflect the uncertainty; the IEPS dossier gives a single figure of R$17 billion. The LENAD survey measured gambling of all kinds, in which lotteries are the most popular product, and was carried out before Brazil's licensed online market opened in January 2025. The R$38.8 billion is therefore not a cost of the licensed market the ban closes, and some of it, the lottery share and whatever moves offshore, will not disappear with it. The government's description of the figure as covering lost productivity and time off work also does not match the categories in the dossier.

Both Sides Are Arguing Across the Same Evidence Gap

The industry's own forecast, that the ban will push 30 million bettors into the illegal market, has not been independently tested either. The two questions that matter most, how much harm the licensed market caused and how much of it a ban will prevent rather than displace, were not studied by the government before it acted, and Congress now has up to 120 days to decide on the measure with no impact assessment in front of it.

Brazil ended its licensed betting market on a judgment about harm, not on an assessment of what a ban would achieve. Governments are entitled to make that judgment, but it leaves the courts and Congress weighing a measure whose costs nobody in government was formally asked to count.