NextFin News - Brazil's delinquency rate on credit operations climbed to a record 4.7% in May, the highest level since the central bank's data series began in 2011, and 82% of families now carry debt - a credit squeeze that is becoming the clearest economic headwind facing President Luiz Inacio Lula da Silva as he seeks re-election on October 4.
The combination is what makes the numbers politically dangerous: a record share of Brazilians carrying debt, a record share falling behind on it, and a central bank that is cutting rates only grudgingly because inflation has already breached the top of its target. Lula's re-election campaign was built on protecting the poor and expanding credit to the working class. Instead, the cost of credit is doing damage that politics cannot easily fix.
The Numbers: A Credit Market Under Strain
The deterioration is broad-based and confirmed by the central bank's own data. The 90-days-past-due rate for the whole financial system reached 4.7%, up 1.0 percentage point over 12 months. In the household segment it stood at 5.6%, up 1.2 points in a year; in the corporate segment, 3.2%, up 0.5 points. In non-earmarked credit - the free-market lending that carries the highest rates - loans more than 90 days overdue reached 6.2% of the portfolio, up 0.9 points in 12 months.
The weakness is concentrated exactly where it hurts the poorest borrowers most: revolving credit-card balances, overdrafts, and non-payroll-deductible personal loans. Revolving credit-card late payments jumped 2.4 percentage points in a single month, to 63%, while the average annual percentage rate on those balances reached 439.9% in May. Non-payroll-deductible personal loans saw delinquency rise to 10.5% from 10.1%, and overdraft contracts to 16.1% from 15.6%. The only bright spot was installment credit-card debt, where the rate fell to 12.5% from 12.7%.
The debt burden behind those numbers is equally severe. A separate survey showed 82% of Brazilian families carried debt in July 2026, a record, while delinquency on household loans hit 5.6% in June - the worst reading on record. The central bank's own figures show the household debt-to-income ratio at 49.8% in May, up 0.9 points in a year, with the debt-service ratio at 28.5%, up 1.3 points over the same period. Roughly half of the income of indebted families is already committed before the next bill arrives.
"The increase in delinquency reflects interest rates that are still very restrictive, strained income, costly debt rollover, and deterioration in higher-risk free-market credit lines," said Jeferson Bittencourt, chief macroeconomist at ASA and a former Treasury secretary.
Two caveats matter, and both cut in opposite directions. First, a new central bank rule - Resolution 4,966, which took effect at the start of the year - changed how overdue loans are classified and provisioned, mechanically pushing the reported delinquency rate higher. Luiz Fernando Castelli of the Brazilian Federation of Banks estimates the headline figure would be almost one percentage point lower without it. Second, the government's Desenrola debt-renegotiation program has removed more than 15 billion reais of old debt from the books, but much of that stock was already excluded from the central bank's delinquency statistics because it was overdue by more than a year.
In other words, the record is partly mechanical - but the underlying trend was worsening before the rule change, and the renegotiation program is not creating new current borrowers.
Market Reaction: Rates, Currency, and the Pricing of Risk
Brazilian assets have priced the tension between political risk and monetary easing. The benchmark Ibovespa index closed at 174,913 points on August 27, up 0.19% on the day, while the real traded at 5.16 per dollar - a level that reflects both the country's high real interest rate and persistent fiscal uncertainty. The currency's weakness matters directly for the delinquency story: a softer real feeds imported inflation, which keeps the central bank's hands tied and borrowing costs high.
The policy rate tells the same story as the credit data. The central bank cut the Selic benchmark to 14.00% on August 5, a fourth consecutive 25-basis-point reduction from the near-two-decade high of 15% reached earlier this year. But even after those cuts, the real interest rate - the policy rate minus inflation - sits near 8% to 9%, among the highest in any economy with GDP above $1 trillion. That is the number that actually matters to a borrower: a 14% policy rate translates into a 33.4% economy-wide average lending rate, and 440% on a revolving card.
Longer-dated interest-rate futures, which price the path of the Selic, have been volatile through the election cycle. A poll of 38 respondents found the median expectation that the bank would hold at 14.00% until the start of 2027 - a slower easing path than the government would prefer, and a signal that markets do not expect rapid relief for indebted households.
Why This Is Different From a Normal Cyclical Dip
The first question any credit cycle raises is whether the deterioration is cyclical - a temporary squeeze that reverses when rates fall - or structural, a regime shift that persists. Brazil's case points to structural, for three reasons.
First, the cost of credit is not a passing shock. Brazil's real interest rate is a deliberate feature of the country's inflation-targeting framework, earned after decades of high inflation and bought back at the price of slow growth. The central bank cut rates only grudgingly because inflation, measured by the 12-month IPCA, reached 4.64% in early May - above the 4.5% ceiling of the target range for the first time since October 2025. The bank has raised its own 2026 inflation projection to 5.2%, and economists surveyed weekly expect 5.04% at year-end. With inflation above target, the central bank cannot cut aggressively without risking its credibility, and credibility is the one asset it cannot afford to lose.
Second, the debt-service burden is embedded in household balance sheets. The debt-service ratio of 28.5% is a system-wide average; for the marginal borrower in revolving credit it is far higher. A cumulative rate cut of 100 basis points from the peak does little to relieve a household paying 440% on a card balance. The relief that would matter - a fall in spreads - requires competition and regulation, not just a lower Selic.
Third, the deterioration is self-reinforcing. Borrowers who fall behind on high-cost credit are pushed into even higher-cost credit - the revolving-card trap - which raises the probability of the next default. That is a balance-sheet mechanism, not a sentiment cycle. It does not mean-revert on its own; it requires either a sustained fall in the cost of credit or a rise in real income, and neither is arriving quickly.
"The program helps clean up credit records and renegotiate debts with discounts, but it does not solve the underlying problem: families still have almost 50% of their income committed to debt, are paying average interest rates of 33.4% a year - and revolving credit cards are back at 440%," said Rafael Rondinelli, an economist at MAG Investimentos.
The cyclical leg is real but subordinate: the rate shock and softer income growth of 2025-26 accelerated the deterioration. The structural leg is the level of rates and spreads that Brazil's concentrated banking system charges, and that level is not about to collapse. This is a regime, not a dip.
The Political Transmission: From Balance Sheets to Ballots
The market for Brazilian credit is also the market for Brazilian votes, and the transmission runs through the same households. Lula's core constituency is the working class and the poor - precisely the borrowers concentrated in revolving credit, overdrafts, and non-payroll personal loans. When 63% of revolving balances are delinquent, the pain is not abstract; it arrives monthly, in statements that cannot be paid.
Recent polling still shows Lula ahead, but by a margin that leaves room for erosion. A survey published on August 20 put him at 40.4% against 33.4% for Flávio Bolsonaro, the Liberal Party senator endorsed by his jailed father, former president Jair Bolsonaro. A PollingData survey on August 16 showed 40.4% to 34.0%, and a Nexus poll on August 14-16 showed 41% to 36%. The lead is real, but it is not insurmountable in a two-round system where the economy is the dominant issue and the first round is on October 4.
The fiscal backdrop limits Lula's room to respond. Stabilizing Brazil's debt by 2031 would require a fiscal effort of at least 2.5 percentage points of GDP - roughly 350 billion reais, or $68 billion - according to Roberto Secemski, chief Brazil economist at Barclays. Markets are skeptical that either candidate can credibly alter the debt trajectory, which keeps long-term rates high and the central bank's hands partially tied. That is the trap: the fiscal constraint that keeps rates high is the same constraint that keeps delinquency rising.
The government's answer so far is Desenrola, which allows reductions of up to 90% in outstanding balances, backed by the Operations Guarantee Fund. More than 15 billion reais has been renegotiated. But the central bank's own statistics chief, Fernando Rocha, acknowledged the program's limits: negotiations on debts between 361 and 720 days overdue are genuine, he said, but they "did not affect our statistics. Potentially, Desenrola will produce numbers larger than the reduction in delinquency."
That is the political problem in one line: the government can claim credit for renegotiations that the official delinquency data does not fully capture, while voters experience the delinquency that the data does capture.
The Counter-Thesis: Why the Squeeze May Not Decide the Election
The strongest argument against reading the delinquency surge as decisive is that the record is, in part, an accounting artifact. Resolution 4,966 may account for nearly a full percentage point of the 4.7% headline, according to the banks' federation. Strip that out, and the adjusted rate is still elevated but no longer a series high. On that view, the political story is being oversold.
There is also monetary relief in the pipeline. The central bank cut rates for a fourth consecutive meeting in early August, and the median market expectation is for the policy rate to hold at 14.00% into 2027 rather than rise. Rate cuts work with a lag; by election day, debt-service costs on new borrowing could be modestly lower. And the Desenrola program, for all its statistical quirks, has put cash back in the hands of indebted households - a tangible benefit that polls may register more clearly than credit data.
Finally, the polls themselves argue against panic. Lula's lead of 4 to 14 percentage points across major surveys is substantial, and Flávio Bolsonaro carries the baggage of his father's name in a polarized electorate. Credit conditions are one input into the vote, not the only one; foreign policy disputes, public security, and the personal appeal of the candidates all weigh heavily. Jair Bolsonaro's ineligibility after being indicted and imprisoned for plotting a coup also fragments the right-wing vote among multiple candidates, including governors Romeu Zema and Ronaldo Caiado, which improves Lula's first-round position.
These points are valid, but they do not fully answer the structural case. The accounting adjustment explains the level, not the direction: delinquency had been worsening before the rule change, and it is worsening across every breakdown of the data. Rate cuts of 25 basis points at a time cannot quickly unwind a 440% revolving rate. And Desenrola renegotiates old debt; it does not prevent new defaults among borrowers whose income has not caught up with their bills. The counter-thesis explains why the election may not be lost; it does not explain why the credit data would improve.
What to Watch: The Signal That Would Prove This Wrong
The falsifying signal is specific. If the central bank's headline 90-days-past-due rate, after accounting for the roughly one-percentage-point regulatory effect of Resolution 4,966, rises for two consecutive monthly prints following the August rate cut - that is, an unadjusted reading at or above about 5.7% in the June or July data - then the "mechanical artifact" defense fails, and the deterioration must be read as structural and ongoing.
A secondary signal sits in the quarterly household debt-service ratio published by the central bank. If that ratio prints above its historical range while the Selic rate remains above 13%, the squeeze is confirmed as a balance-sheet problem that rate cuts alone will not fix. A third signal is the real: a sustained break above 5.40 per dollar would feed inflation and force the central bank to pause easing, closing the only relief valve available to indebted borrowers.
For investors, the exposure runs through Brazilian consumer lenders, retailers with captive finance arms, and the banks' loan-loss provisions - the non-earmarked portfolio, where 90-day delinquency is already at 6.2%, is the first place provisions will rise. For policymakers, the exposure runs through the October ballot. The base case is that delinquency stabilizes at an elevated level rather than accelerating into a crisis - enough to erode Lula's support among the working poor, not enough to overturn his polling lead outright. The upside case for Lula is that inflation falls faster than expected, allowing faster cuts and a year-end credit relief that voters feel before they vote. The downside case is that inflation stays above target, the central bank holds rates high, and defaults keep climbing into the second round.
Brazil's credit market is not flashing a recession warning; it is flashing something more politically specific. The voters most likely to punish the government are the ones already behind on their bills, and no amount of renegotiation can put them back in good standing while the cost of borrowing stays at 440%. Lula may yet win re-election - but the debtors are no longer voting from strength.
Explore more exclusive insights at nextfin.ai.
