Auto-loan originations hit a record in the second quarter of 2026, roughly $211 billion, the highest quarterly figure in the New York Fed's data, as lenders relaxed their underwriting standards and approved more borrowers with lower credit scores. At the same time, more borrowers fell behind on their payments, with subprime auto delinquencies at levels one analyst described as rivaling or exceeding those of the Great Recession.

On its face, that combination is a puzzle and a warning at once. Why would lenders extend more credit, and to riskier borrowers, at precisely the moment more people are failing to pay? The easy answer is that lenders are being reckless, chasing volume as the ground softens beneath them, and there is a thread of truth in it. But the deeper driver is an affordability problem the credit system is absorbing rather than solving, and following how it absorbs that problem reveals where the risk is actually going.

The record isn't strength; it's a symptom

A record in loan volume sounds like a sign of health, but here it is closer to the opposite. The origination figure is not adjusted for inflation, and more importantly, it is being driven not by booming prosperity but by cars that have grown very expensive. The average new-vehicle monthly payment now sits around $750, on top of elevated interest rates and sharply rising insurance and fuel costs.

People are not borrowing more because they are flush. They are borrowing more because cars cost more, and for most Americans a car is a necessity for getting to work and running a life, not a luxury they can simply skip. So the record reflects strain at least as much as demand. When the volume of borrowing sets a record because the thing being borrowed for has become barely affordable, the record is measuring a problem, not celebrating a boom.

What "weakening standards" actually means

"Lenders relaxed their underwriting standards" sounds like straightforward recklessness, but it is worth unpacking what the phrase actually consists of, because the specifics reveal the mechanism at work. Alongside approving somewhat lower credit scores, the central move is stretching the loan: longer terms, with 84-month loans, once unusual, now common; smaller down payments; higher loan-to-value ratios.

And the purpose of the stretch is specific. It is to bring the monthly payment down to a level the borrower can actually handle. When a car costs more than someone can afford, the lender cannot lower the price, so it lengthens the loan instead, spreading the same debt across more months so that each month's bite is smaller. That, more than lending to people who obviously cannot pay, is what "weakening standards" largely amounts to here. It is a restructuring of the loan so that the monthly number looks manageable, whatever the total underneath it.

The affordability problem doesn't vanish; it changes form

Here is the crux of it. Stretching the loan makes the monthly payment affordable, but it does not make the car any cheaper or the borrower any richer. The underlying problem, that the vehicle costs too much relative to the income buying it, does not go away. It is transformed into the structure of the loan.

A longer term means the borrower stays underwater, owing more than the car is worth, for far longer, and remains exposed to any disruption for that whole extended stretch: a lost job, a medical bill, a jump in insurance. So the risk is not eliminated by the stretch. It is deferred, pushed further out into the future, and relocated, moved out of the monthly payment, where everyone can see it, and into the loan's length and structure, where it is easy to overlook. The payment looks fine. The risk is hiding in the term. This is the quiet trick of a stretched loan: it converts a problem of price, which is visible, into a problem of duration and exposure, which is not.

The risk lands on the people least able to carry it

And it does not land evenly. The distress is sharply concentrated. Prime borrowers, those with strong credit, are largely doing fine, their delinquency rates stable and healthy, while subprime borrowers are in their worst shape in more than three decades. A New York Fed researcher described this as a K-shaped economy, one with a great many households living paycheck to paycheck and squeezed hardest by inflation.

The stretching dynamic falls most heavily on exactly those borrowers, because they are the ones for whom the payment had to be stretched furthest to work at all. Someone with ample income can absorb an expensive car without an 84-month loan; someone without it cannot, so the longest, thinnest, most exposed loans cluster among the people with the least cushion to withstand them. And the loosening of standards means more of those borrowers are being approved, at the very moment their cohort is already breaking. The system is extending more credit to the group in record distress, on precisely the terms that defer their risk into a future they may be least equipped to survive.

The costs the loan doesn't count

There is a further wrinkle that makes the risk even easier to understate. A lender underwrites the loan payment: can this borrower make the monthly payment on this debt? But the true cost of owning a car is far more than the loan alone. Insurance premiums have climbed steeply, driven by higher repair costs tied to materials, tariffs, and labor, and fuel remains expensive, with both rising faster than headline inflation.

So a borrower who can genuinely afford the loan payment the lender modeled can still be pushed into delinquency by the other costs of ownership the underwriting never captured. The lender solves for one number, the monthly payment on the debt, but the borrower lives with the entire bill, and the widening gap between the two is one more place the real risk hides. A loan can be perfectly affordable on the lender's spreadsheet and quietly unaffordable in the borrower's actual life.

Why this isn't 2008, and what it is

Perspective is owed here in both directions. This is not the makings of a 2008-style systemic crisis, and it would be wrong to imply otherwise. Auto lending is far smaller than mortgage lending, the losses, though real, are more contained, prime borrowers remain healthy, and unusually strong used-vehicle values are currently propping up the collateral behind these loans, so lenders recover more when they repossess. Regulators say they are watching, and some subprime specialists have already pulled back. This is a serious consumer-distress story more than a threat to the financial system.

Nor are the lenders simply villains. They are responding to genuine market pressure, and a stretched loan does let some people obtain a car they truly need to work and live. But the honest reading is still cautionary. Loosening credit into rising delinquency is the classic late-cycle mistake, the kind that reliably produces higher losses down the road, and here those losses are being seeded disproportionately among the households with the least room to absorb them. Restraint is not the same as alarm, and both the reassurance and the warning are true at once.

In the end, "volumes rise as standards weaken" reads like a story about reckless lenders, and there is a strand of that in it. But it is better understood as what a credit system does when the thing it finances becomes unaffordable. It does not lower the price; it stretches the loan, quietly converting an affordability problem it cannot fix into a risk problem it can defer. The monthly payment comes down, and everyone can point to that as progress. The risk itself does not disappear. It moves, into a longer term, a longer stretch spent underwater, and onto the balance sheets of the borrowers least able to carry it. That is less a crisis in the making than a slow accumulation of stress in the corner of the economy that can least absorb it, tucked inside loans engineered to look affordable one month at a time. The figure worth watching was never the record origination total. It is who is being lent to, on what terms, and whether the steady income and stable costs those stretched loans quietly assume will actually hold. And for a borrower, the plainest caution is also the oldest: a loan long enough to make an unaffordable car affordable is long enough to keep you underwater for years, and the monthly payment is never the same thing as the price.

Primary sources

  1. American Banker for the report that new auto-loan volumes hit a record roughly $211 billion in the second quarter of 2026, the highest in the New York Fed's data though not inflation-adjusted, as lenders relaxed underwriting standards and approved more lower-credit-score borrowers while transitions into serious delinquency rose, the New York Fed researcher's "K-shaped economy" characterization, and the observation that many households live paycheck to paycheck.
  2. Bankrate's Ted Rossman on subprime auto delinquencies rivaling or exceeding Great Recession levels and on the broader affordability squeeze from elevated gas prices and sharply rising insurance costs, driven by repair costs tied to materials, tariffs, and labor, outpacing headline inflation, and the New York Fed's statement that it will continue to monitor the trend.
  3. Experian commentary, via industry coverage, for banks becoming the largest auto lender and being in growth mode.
  4. Fitch and Black Book analyses for average monthly payments near $750 to $760, the rising prevalence of 84-month terms, record subprime delinquency alongside healthy prime performance, and strong used-vehicle values supporting collateral.
  5. Bridgeforce for the characterization of loosening credit amid rising delinquency as a late-cycle mistake and the decline in median new-auto credit scores.
  6. Snell & Wilmer's summary of Dealertrack data on subprime origination share, 72-month-plus term prevalence, and declining down-payment requirements.
  7. CarEdge for the multi-decade subprime delinquency record and the caution that ultra-long loan terms leave borrowers underwater for years.