Comsefaz puts Brazilian families’ betting “losses” at R$ 62.5 billion, but the model ignores official SPA data

The number has been making the rounds in policy circles and the press: R$ 62.5 billion in “losses” for Brazilian families from sports betting in 2025. The thing is, that estimate is not a direct measurement of betting activity, and it sits 69% above the official gross revenue declared by licensed operators to the Ministry of Finance’s Secretaria de Prêmios e Apostas (SPA-MF).

1. The R$ 62.5 billion figure comes from the 3rd Update of the Boletim Fiscal dos Estados Brasileiros, prepared by Comsefaz (the National Committee of Finance Secretaries) together with the Celso Furtado International Center for Development Policy.
2. The bulletin’s own authors say the number is an order-of-magnitude estimate, not a precise causal measurement. For high-risk payment teams, that distinction matters: this is a policy signal, not a clean market size figure.
3. The official number for the same year is R$ 36,959,783,379.70 in gross revenue declared by licensed operators to SPA-MF. That gap is not a rounding error; it reflects a methodology that is measuring something broader than bets.
4. Comsefaz does not start from betting operators’ declared data. It derives the estimate from Pix transfers from individuals to companies classified by the Banco Central under “artes, cultura, esporte e recreação” in the EPAE (Estatísticas de Pagamentos por Atividade Econômica) system.
5. That category is much wider than betting. It also includes gyms, sports clubs, event producers, cinemas, streaming platforms, and other businesses with no connection to bets. In other words, the model attributes all unusual growth in that aggregate from 2025 to betting, without statistically separating the betting component from the rest.

The methodological wrinkle is straightforward: the model was trained on data only up to August 2024, before Brazil’s betting regulation took effect in January 2025, and then used an ARIMA projection to define what “normal” payment behavior would have looked like. Anything above that projection was treated as betting-related. If other parts of the category also expanded in 2025, that growth gets booked as “bets” too. For PSPs and acquirers, that is the difference between a policy narrative and a usable operating metric.