Consumer credit has a hierarchy, and lenders have relied on it for decades. When a household runs short, the card goes unpaid before the car, and the car goes unpaid before the house, because the card is an inconvenience and the car is how a person gets to the job that pays for everything else.
That ordering has been stable enough to underwrite against. It is now bending, and the bend is showing up first in the collateral people used to defend hardest.
Longer terms hid the problem for a while
The mechanism is not complicated. Vehicle prices rose sharply and did not come back down; incomes rose less; and the industry closed the gap the way it always does, by extending the term. Seventy-two months became standard and eighty-four stopped being unusual, which held the monthly payment inside what a household would agree to while quietly changing what the loan is.
A long loan against a depreciating asset spends most of its life underwater. A borrower four years into an eighty-four-month note who needs a different vehicle rolls negative equity into the next loan, and the next loan starts further underwater than the last one. Each individual decision is rational and the sequence has no exit, which is why the average amount of negative equity carried into a new purchase has kept climbing through periods when nothing else in the credit picture was deteriorating.
Insurance is the accelerant nobody underwrote for. Premiums have risen enough to constitute a second payment on the same vehicle, and unlike the loan it reprices every renewal, so a household that budgeted carefully at signing faces a materially different total cost two years in. That is the household-level face of an insurance repricing that is behaving like a second inflation, and it lands hardest on exactly the borrowers with the least headroom.
The distributional picture matters more than the aggregate. In the upper income tiers the balance sheet is genuinely in good shape, and the broad repair of household finances is real. Subprime and near-prime auto delinquency has meanwhile reached levels not seen outside a recession, in an economy that is not in one. The average conceals two populations moving in opposite directions.
Lenders have tightened at the margin without solving the underlying problem, because the underlying problem is the price of the asset relative to the income of the buyer, and no underwriting standard fixes that. What tightening does is push marginal borrowers toward buy-here-pay-here arrangements and toward the newer point-of-sale credit that is only now becoming visible to the bureaus, which means the next round of stress will be measured with better instruments than the last.
Repossession volumes tell the same story from the servicer's side, and they come with a lag that flatters the current numbers. A lender facing a delinquent note on a vehicle worth less than the balance has every incentive to extend, defer or modify rather than recover the collateral and book the loss, so forbearance absorbs some months of deterioration before it appears anywhere. The reported delinquency rate is therefore a floor rather than a reading, and the people who work these portfolios treat it that way.
For anyone reading consumer credit as a signal, the useful point is what the ordering meant. Auto delinquency rising while mortgage performance holds does not indicate a housing problem. It indicates households burning through the buffer they keep for the thing they cannot do without, which is the last flexibility they have.





