Payment extension: a servicer lets a borrower move one or more monthly payments to the end of the loan. The loan tape reports a running count of payments moved; when that count goes up, the loan got an extension, and the rise is how many payments moved.
The measures in plain words
- Extension rate = loans with at least one payment moved in the last 12 months ÷ eligible loans. Eligible means active and reported the month before, not paid off or charged off. It counts loans, not extensions, and not dollars: the question is how many borrowers needed one.
- Payments moved = among those loans, the share that moved 1, 2, or 3 or more payments over the 12 months.
- Extensions by lateness = extensions in a month ÷ eligible loans, split by how many days late each loan was the month before.
- Share ever extended = loans of a vintage extended at least once so far ÷ loans of that vintage still reported, at each month on book.
Worked example: eight loans
Eight loans made in the same month, followed through months on book 1 to 12. "late" means 30 or more days past due; "+N" means the count of payments moved rose by N that month (after which the loan reports current). Every loan was current the month before month 1.
| Loan | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| A | ||||||||||||
| B | +1 | |||||||||||
| C | ||||||||||||
| D | late | late | +2 | |||||||||
| E | late | +1 | late | +2 | ||||||||
| F | late | |||||||||||
| G | ||||||||||||
| H | late | late | late |
- Extension rate: B, D and E had a payment moved: 3 ÷ 8 = 37.5%. Counting extensions instead (E had two) would give 4 ÷ 8 = 50.0%, which overstates how many borrowers needed help.
- Payments moved: B moved 1, D moved 2, E moved 1 + 2 = 3. One loan each in the 1, 2 and 3+ groups: 33.3% each.
- By lateness the month before: loans went into 7 months late (D twice, E twice, F once, H twice; H's last month has no following month in the window). 3 of those months brought an extension: 3 ÷ 7 = 42.9%. Loans went into 89 months current, and 1 brought an extension (B): 1 ÷ 89 = 1.1%.
- Share ever extended: 1 of 8 by month 3, 2 of 8 by month 5, 3 of 8 (37.5%) by month 9.
The real numbers
All issuers, loans only, through September 2026. Prime is FICO 660 and up at origination, subprime under 660.
Extension rate, 12 months to September 2026:
| Eligible loans | With a payment moved | Extension rate | Moved 1 / 2 / 3+ payments | |
|---|---|---|---|---|
| Subprime | 2,116,908 | 524,984 | 24.8% | 24.2% / 53.9% / 21.9% |
| Prime | 5,226,332 | 173,530 | 3.3% | 42.8% / 45.8% / 11.3% |
Subprime, worked through: 524,984 ÷ 2,116,908 = 24.8%. About one subprime borrower in four had at least one payment moved in the year, and most of those moved two payments or more.
Extensions in a month, by days late the month before (six-month average to September 2026):
| Days late the month before | Subprime | Prime |
|---|---|---|
| Current | 0.84% | 0.17% |
| 1-10 days | 2.9% | 1.9% |
| 11-29 days | 4.6% | 3.3% |
| 30-59 days | 8.2% | 6.1% |
| 60-89 days | 10.7% | 8.0% |
| 90+ days | 7.6% | 4.7% |
A subprime loan 60-89 days late is about 13 times as likely to be extended in a month as one that is current.
Share ever extended at month 22 on book (the latest age the 2024 loans have all reached):
| Vintage | Subprime | Prime |
|---|---|---|
| 2019 | 34.8% | 6.5% |
| 2022 | 29.2% | 3.2% |
| 2023 | 30.6% | 3.2% |
| 2024 | 31.2% | 3.5% |
Compared at the same age, 2024 subprime loans have been extended a little more often than 2022 and 2023 loans, and less than 2019 loans, whose second year ran into 2020. At their latest points the older vintages look far higher (2019 subprime reaches 42.9% by month 48) only because they have had longer.
Common pitfalls
- Counting extensions instead of loans. A borrower extended twice is still one borrower. The extension rate counts each loan once; the payments-moved mix shows how deep the help went.
- Ignoring lateness. Most extensions go to loans that are already behind, so a pool with more late loans will show more extensions with no change in servicer policy. Compare like with like, by days late the month before.
- Comparing vintages at different ages. The share ever extended only climbs, so an older vintage's last point is always higher. Compare vintages at the same month on book, and only at ages every loan in the vintage has reached.
Where to see it
- Monthly State of Auto deck, extension pages: the extension rate by month, payments moved, extensions by days late, and the share ever extended by vintage, prime and subprime.
Source: LoanTape, SEC Form ABS-EE loan-level data, through September 2026.