Extension Default Rate: the share of extended loans that are 30 or more days past due, or charged off, two months after the moved payments end.
The formula in plain words
- Check month = the month of the extension + the number of payments moved + two months.
- Extension Default Rate = extended loans that are 30+ days late in their check month, or charged off by then ÷ extended loans whose check month has been reported.
- Loans whose check month has not arrived yet are left out of both the top and the bottom.
- The published rate counts loans. Weighting each loan by its balance at the extension gives a second reading of the same loans.
What an extension does
A payment extension lets a borrower who has fallen behind skip a payment. The missed payment moves to the end of the loan and the loan is reported as current again. Being current right after an extension says little about how the loan will do, which is why the rate waits until two months after the moved payments.
Worked example: eight loans
Eight loans got an extension in March. Each moved one payment, so each is checked in June (March + 1 + 2).
| Loan | Balance at extension | Lateness when extended | Status in June |
|---|---|---|---|
| A | $18,400 | Current | Current |
| B | $12,600 | Current | Current |
| C | $9,200 | Current | Current |
| D | $22,800 | 18 days late | Current |
| E | $15,300 | 24 days late | 30+ days late |
| F | $8,700 | 47 days late | Charged off |
| G | $13,900 | 38 days late | 30+ days late |
| H | $11,100 | 33 days late | Current |
| Total | $112,000 | 3 of 8 slipped |
- By count: 3 ÷ 8 = 37.5%.
- By balance: ($15,300 + $13,900 + $8,700) ÷ $112,000 = $37,900 ÷ $112,000 = 33.8%. The larger loans did a little better here, so the balance reading is lower.
Two more loans, I ($16,500, current) and J ($7,400, 12 days late), were extended in May. Their check month is August, and the data ends in June. They are left out. Counting them as good would give 3 ÷ 10 = 30.0%, which is too low.
Split by lateness when extended, the same eight loans give:
| Lateness when extended | Slipped | Rate |
|---|---|---|
| Current | 0 of 3 | 0% |
| 1-29 days late | 1 of 2 | 50.0% |
| 30-59 days late | 2 of 3 | 66.7% |
The real numbers
Every extension in our loan tapes granted from November 2024 to October 2025, all issuers. That is the latest 12 grant months in which at least 99% of extensions, at every issuer, had reached their check month. Prime is FICO 660 and up at origination, subprime under 660.
| Extended loans with a result | 30+ days late or charged off | Rate by count | Rate by balance | |
|---|---|---|---|---|
| Subprime | 701,934 | 279,803 | 39.9% | 40.0% |
| Prime | 222,621 | 40,548 | 18.2% | 18.4% |
Subprime, worked through: 279,803 ÷ 701,934 = 39.9%. By balance at the extension: $5,955,379,123 ÷ $14,882,899,971 = 40.0%. On real pools the two readings are close; in the eight-loan example they differ because the larger loans happened to stay current.
By lateness when extended, rate by count:
| Lateness when extended | Subprime | Prime |
|---|---|---|
| Current | 10.1% | 4.5% |
| 1-29 days late | 25.4% | 15.2% |
| 30-59 days late | 51.6% | 41.2% |
| 60+ days late | 71.6% | 61.0% |
By loan-to-value at origination, rate by count:
| LTV at origination | Subprime | Prime |
|---|---|---|
| Under 100% | 36.9% | 13.6% |
| 100-125% | 39.5% | 18.5% |
| Over 125% | 42.9% | 20.9% |
Lateness when extended moves the rate far more than loan-to-value does.
Common pitfalls
- Comparing pools with a different mix of lateness. A loan that was already 30-59 days late when extended is several times as likely to slip again as one that was current. A pool's rate moves when the mix of who gets extended moves, even if nothing else changes. Compare like with like: the same lateness group, or the same issuer against its own history.
- Counting recent extensions as good. An extension from last month cannot have a two-month result yet. Leaving it in the bottom of the fraction as if it were fine makes the rate look lower than it is.
- Reading it as a loss rate. A loan that is 30+ days late two months on has not necessarily been lost. Some catch up later. The rate is an early reading on how extended loans are doing, not a forecast of loss.
Where to see it
- Monthly State of Auto deck, Extension Default Rate pages: prime and subprime over time, and by issuer.
- Dashboard, M+2 Bad Outcome page: the same two-month outcome for extended loans, with issuer, FICO, LTV and term filters.
Source: LoanTape, SEC Form ABS-EE loan-level data, through September 2026.