Who gets a payment extension: the longest loans, by almost 10 to 1
Prime auto loans written at 76 months or longer get a payment extension in 0.76% of eligible loans. Prime loans written at 60 months or less: 0.08%.
That is 9.7 times. Split the same book by vehicle age instead and the spread is 2.1 times. Term is the sharper sort by a long way, and it is the field I would reach for first.
I have spent most of the last two months on what happens after an extension and almost none on the simpler question of which loans get one at all. So this post is the profile. What the loans look like, what the borrowers looked like at origination, and where prime and subprime part ways. Everything here is the observed record in SEC ABS-EE loan-level filings through April 2026. It describes which loans servicers chose to extend. It does not say what the extension did.
Term length is the sharpest cut
Split the prime book by original term and the extension rate climbs at every step.
| Original term | Extension rate |
|---|---|
| 60 months or less | 0.08% |
| 61-72 months | 0.26% |
| 73-75 months | 0.47% |
| 76+ months | 0.76% |
The steepest multiple is the first step. Going from a 60-month note to a 61-to-72-month note more than triples the rate. The steepest absolute move is the last one, 0.29 points between the 73-75 band and the 76-plus band. That single step is bigger than the entire 60-month rate, three times over.
Original term is also not a behavioral field. It is set on day one and it never moves after that. Whatever a long term is standing in for, and it is standing in for several things at once, the tape hands it to you at origination.
Older cars, same direction
Vehicle age at origination sorts extensions too, just less violently.
| Vehicle age at origination | Extension rate |
|---|---|
| 0-1 years | 0.28% |
| 2-3 years | 0.30% |
| 4-5 years | 0.36% |
| 6-7 years | 0.48% |
| 8-9 years | 0.57% |
| 10+ years | 0.57% |
Bottom to top that is the 2.1 times from the top of this post. The climb is clean through the 8-9 year band and then stops dead, with the two oldest buckets landing on the same 0.57%. The first two bands are nearly identical as well. Almost all the movement is in the middle, between a four-year-old car and an eight-year-old one.
The comparison against term is the point of this table. Same direction, but it flattens out where term keeps climbing.
Before anyone tries to reconcile these against the headline number: every band in both tables sits below the pool-wide extension rate of about 2.6%. Not a contradiction. The pool rate blends prime and subprime, and both of these tables are prime only.
Loan-to-value and payment-to-income point the same way
The other two origination fields I checked sort in the expected direction, in both segments.
Extension rate rises monotonically with origination loan-to-value (LTV) in prime and in subprime. No band bucks the trend in either segment. We have published the subprime LTV detail before, in the piece on how subprime extensions are rising fastest at high LTV, and the prime book behaves the same way at a much lower level.
Extension rate also rises with payment-to-income (PTI) at origination, again in both segments. That is consistent with what PTI does elsewhere in this data, where it ranks risk about as well as any single field we carry.
Neither result surprised me. I checked them because a profile with a hole in it is not a profile. So the picture holds together. Longer terms, older collateral, higher advance rates, heavier payments against income, four fields all leaning the same way, and not one of them needing a month of servicing history to observe. You can run this sort on a tape the month it closes.
Prime and subprime extend different loans
This is where the field stops meaning one thing.
In prime, 61% of extensions go to loans that were fully current when the extension was granted. Another 16% go to loans less than 30 days late. Those two together make the 77% figure you will see quoted as current-or-under-30, and they are worth keeping apart, because a loan that has missed nothing and a loan that is a couple of weeks late are not the same borrower.
Subprime runs the other way. Extensions there concentrate in loans that are already delinquent.
Same field, same filing element, two populations that barely overlap.
That matters if you are aggregating. A pooled extension rate across both segments averages two behaviors that do not mean the same thing, and any read built on the blended number inherits the mix. I would not compare a prime shelf's extension rate to a subprime shelf's without saying which population each one is drawn from.
It matters for what you ask next, too. When a delinquent loan gets extended, the question is whether it comes back, which we track as the Extension Default Rate. When a fully current loan gets extended, and in prime that is most of them, there is no delinquency to cure. The question has to be something else.
Usage varies by issuer
Pool-wide, about 2.6% of eligible loans get an extension. Individual issuers sit a long way off that number in both directions. Carvana runs near 1.5%. Bridgecrest, its servicing arm, runs near 3.5%. I checked that pair twice.
They belong next to each other because they are related, and more than a factor of two separates them. Two shelves drawing on overlapping collateral can report very different extension rates. The field carries servicing policy as much as borrower characteristics, and a low rate is not by itself a statement about the book.
What the profile adds up to
Extensions are not spread evenly across the book. Term sorts them harder than anything else I looked at, by more than I expected going in. Vehicle age, LTV, and PTI lean the same way with much less force.
The caveat I would staple to all of it is the segment split. A prime extension and a subprime extension are the same filing element describing two different situations, and the prime one is mostly not a distress event. That is a description of who servicers extend. Nothing in it says what happens after, and I would want the description settled before anyone tries to price the thing.
Methodology
The metric here is the extension-event rate: loans granted a payment extension in the period divided by eligible loans, weighted by eligible loans. Eligible loans are active, non-terminal reported loans with a balance above $1. Delinquency status is measured in days past due (DPD). Fully current at extension means 0-10 DPD in the month the extension was granted; the band above it is 11-29 DPD.
Source is SEC ABS-EE loan-level filings through April 2026, parsed into LoanTape's loan-level cache and deduped to one record per loan per reporting month by latest filing. Prime is FICO 660 and above at origination, subprime is below 660. Original term, vehicle age, LTV, and PTI are all origination-time fields as reported in the filings.
The term and vehicle-age tables are prime only. The pool-wide 2.6% rate and the issuer figures cover both segments. Everything in this post is descriptive: it reports which loans received an extension, not what the extension changed.
This runs on the same loan-level infrastructure behind LoanTape's ABS-EE dataset, Form 10-D data, and ABS remittance data. The subscriber version, with issuer, trust, and post-extension outcome cuts, is in the Modification Effectiveness Deep Dive. For the loan-level files behind it, see LoanTape pricing.