Brazil’s Ministry of Social Development (MDS) acknowledged flaws in its methodology for projecting expenditures on the Continuous Cash Benefit (BPC), one of the largest mandatory spending items in the federal budget. The admission, contained in a technical note obtained by Valor through a freedom of information request, came after the Federal Audit Court (TCU) concluded that the ministry had underestimated spending by R$6.4 billion in 2024 without technical justification.
The BPC provides a monthly minimum-wage payment to low-income seniors aged 65 and older, as well as to low-income people with disabilities. Because it is mandatory spending, failing to revise projections upward allows the government to temporarily avoid cuts in discretionary expenses. However, at year-end, the government must recognize the true costs and make offsetting cuts in mandatory expenditures.
The TCU also found that BPC outlays in 2024 exceeded the amount set in the annual budget law (LOA) by R$7.6 billion. Critics have long warned that the government has a pattern of underestimating mandatory spending to mask budgetary pressures.
One example, according to the audit court, occurred in May 2024. On May 14, the ministry’s National Secretariat of Social Assistance issued a technical note calling for a R$6.4 billion upward revision in BPC spending, citing the government’s program to eliminate the backlog of claims, which was expected to accelerate approvals, while the impact of ongoing benefit reviews remained uncertain. But the very next day, another technical note reversed the increase “without presenting equivalent technical analysis,” the court said. The revision was delayed until July, when the ministry formally recognized the impact of the backlog-clearing program and concluded that reviews would not yield significant savings in 2024.
In response, the TCU recommended that the ministry strengthen its methodology for BPC projections to ensure “consistency, clarity, and transparency” regarding both rising baseline spending and expected savings. The recommendation, relayed by TCU Justice Jonathan de Jesus, was unanimously approved by the audit court in June.
Following the court’s findings, the MDS conceded that its current methodology “is not aligned with the TCU’s recommendations with respect to consistency, transparency, and reliability.” The ministry cited several shortcomings: the inability to incorporate continuous effects of benefit terminations into future-year estimates, the failure to account for reactivations of benefits after termination, and reliance on averages that produce wide variability in projections.
“To meet the TCU’s requirements and improve forecast quality, the methodology must be revised to incorporate ongoing analysis of terminations and reactivations, while ensuring tighter control over forecast uncertainty,” the note stated.
The ministry said it is revising its methodology along three axes — registry updates, income verification, and biopsychosocial assessments — which began rolling out in 2024. Registry updates are already routine; income verification is still in the planning stage and requires new regulations; and biopsychosocial assessments began in July 2025 through medical and social evaluations. “The start of these assessments made forecasting more complex, since each stage directly interferes in estimates,” the ministry said.
“To improve accuracy, the MDS, in partnership with Ipea, has developed a hybrid projection model using multiple linear regression for the short term and the Holt-Winters time series method for the medium term. This combination ensures methodological rigor, legal certainty, and greater reliability, guaranteeing that the BPC is executed with budgetary predictability,” the ministry said.