Consumer credit
Subprime auto arrears reach 6.4 per cent, the highest reading of the cycle
Sixty day delinquencies in securitised subprime auto pools have risen for eleven straight months. The 2024 vintage is performing worse at the same age than the 2019 one did, and refinancing is not available to fix it.
In this story
Illustrative. Bank Season is an editorial prototype. This story, its sources, the issuers named in it and every figure it quotes are invented to demonstrate the publication. Nothing here is reported fact or investment advice. Read the disclosure.

Subprime auto lending is the part of consumer credit that breaks first, and it has been breaking quietly for the better part of a year. Sixty day delinquencies across the Verity Research index of securitised subprime pools reached 6.4 per cent in August, the eleventh consecutive monthly increase and the highest reading in a series that starts in 2011. The previous peak, 6.1 per cent, arrived in the first quarter of 2024 and was followed by four months of improvement. This time there has been no improvement.
6.4%
Sixty day delinquency rate, subprime auto pools
Verity Research index, August, above the 6.1 per cent peak of early 2024
The distinction that matters is between a payment problem and a balance sheet problem. A borrower who misses a payment because of a temporary income gap cures the loan within two months and the pool absorbs it. A borrower whose fixed costs now exceed income does not cure, and the cure rate is the number that has changed. Verity puts it at 31 per cent of sixty day delinquencies returning to current within ninety days, against 44 per cent two years ago.
Sixty day delinquencies, securitised auto pools
Per cent of outstanding balance, monthly, Verity Research index.
Prime borrowers are fine. Their arrears have risen from 0.31 to 0.42 per cent over the year, which on a base that low is a rounding adjustment rather than a signal. The two lines diverging is the useful observation, because a genuine consumer downturn shows up in both. What is happening instead is a squeeze concentrated at the bottom of the income distribution, where a car payment competes directly with rent and where the wage growth that offset it for three years has slowed to 3.6 per cent and is still falling.
Cumulative net loss by origination vintage, subprime auto
| Vintage | Loss at 12 months | Loss at 20 months | Loss at 36 months | Average loan to value at origination |
|---|---|---|---|---|
| 2019 | 4.1% | 7.8% | 12.4% | 112% |
| 2021 | 2.6% | 5.1% | 8.9% | 118% |
| 2022 | 3.9% | 7.4% | 12.1% | 124% |
| 2023 | 4.8% | 8.5% | 13.6% | 121% |
| 2024 | 5.3% | 8.7% | n/a | 119% |
| 2025 | 5.9% | n/a | n/a | 117% |
Verity Research pool data. The 2025 vintage figure is measured at twelve months and is the highest first year loss in the series.
Loan to value at origination has actually improved slightly since 2022, which rules out the simplest explanation. Lenders are not underwriting worse collateral. They are lending to borrowers whose real income has been flat for two years against rent and insurance costs that have not been. The Thornbury Auto Receivables shelf, one of the larger subprime issuers, disclosed in its August servicing report that extensions granted as a share of the pool rose to 4.1 per cent from 2.7 per cent a year earlier, which delays the loss recognition rather than preventing it.
Cure rates are the number I would put in front of anyone who thinks this is noise. Borrowers used to catch up. A third of them catch up now, and the rest are running an arithmetic problem rather than a timing problem.
Lorena Vasquez, director of consumer credit research at Verity Research
31%
Share of sixty day delinquencies curing within ninety days
Against 44 per cent two years ago
Securitisation spreads have moved accordingly but not dramatically. Triple-B subprime auto paper trades at 248 basis points over Treasuries, 41 wider than in June, which is in line with the broader widening in credit rather than ahead of it. Investors are treating this as a known deterioration priced into structures that were built with enough subordination to absorb it, and on the 2019 and 2022 vintages that judgement has been correct.
The macro reading is narrower than the headline invites. Sixty per cent of households have no exposure to this market and the aggregate consumption data has not turned. What the series measures is the condition of the household that has no savings buffer, and that household has been running down its cushion for eight quarters while hiring in leisure and hospitality, the sector that employs a large share of them, has turned negative on the Verity payroll panel. The arrears number is the receipt for both facts.


