In the 825 to 850 credit score tier, October originations priced from 3.99% at the 5th percentile to 13.99% at the 95th. That 10.00 point spread inside one tier is wider than the 8.21 point span of every tier average you have.
In October 2026, 120 credit unions in the CUCollaborate Data Collaborative originated 114,695 loans worth $1.91 billion. The Data Collaborative pools the AIRES files credit unions already produce, which is why these figures run all the way down to the individual loan. Sorted by credit score tier, the average rates form a tidy ladder: 5.83% at the top, 14.04% near the bottom.
Then look at the individual loans inside any one rung. We walked through these figures in the November 16 session of The Monthly Download. Some of that dispersion is product mix and belongs there. Some of it is not, and the only way to tell which is which is to look at the loans rather than the averages.
In this article, we will cover:
- What the credit score tier ladder looks like on average
- How wide rates ran inside each tier in October
- How much of that spread is explained by product mix, and how much is not
- Why the no-score bucket is the largest origination channel in the data
- What to check in your own file
The Tier Ladder Looks Clean
Here is October’s origination activity by credit score tier.
Figure 1: Originations and average rate by credit score tier, October 2026
| Credit score tier | Originations | Originations ($M) | Average rate |
|---|---|---|---|
| Exceptional II (825 to 850) | 4,241 | 118.0 | 5.83% |
| Exceptional I (800 to 824) | 7,436 | 209.2 | 6.10% |
| Very good II (770 to 799) | 10,451 | 309.5 | 6.27% |
| Very good I (740 to 769) | 10,467 | 276.5 | 6.65% |
| Good II (705 to 739) | 12,790 | 276.0 | 7.27% |
| Good I (670 to 704) | 12,432 | 218.1 | 8.12% |
| Fair II (625 to 669) | 12,784 | 147.1 | 9.99% |
| Fair I (580 to 624) | 6,936 | 47.5 | 12.92% |
| Poor II (500 to 579) | 4,604 | 21.1 | 14.04% |
| Poor I (300 to 499) | 362 | 1.4 | 13.33% |
| No score | 32,192 | 288.0 | 7.47% |
| Total | 114,695 | 1,912.5 | 7.36% |
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. 120 credit unions with an AIRES file dated October 2026. Average rates are weighted by current balance.
This is the table most credit unions already have. It behaves the way risk-based pricing should behave: each step down in credit quality costs the borrower more, with one inversion at the very bottom where the Poor I average sits below Poor II on only 362 loans.
It is also the table that hides everything useful.
Inside the Top Tier, Rates Ran From 3.99% to 13.99%
Same loans, same month. This time every individual origination is ranked by its rate, without weighting for balance.
Figure 2: Rate range on individual originations by credit score tier, October 2026
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. Percentiles are unweighted across individual loans originated in October 2026. The average is weighted by current balance. Credit score tiers are labeled by score band.
Three things fall out of this chart.
The averages sit below the medians in every tier. In the top tier the average is 5.83% and the median is 6.25%. The average is weighted by balance, so large loans pull it down. Your headline rate is being set by your mortgages and your autos, not by the typical loan you write.
A top-tier member at the 75th percentile paid 9.99%. That is the same rate as the average member in the 625 to 669 tier. One in four of your very best borrowers is priced like a Fair II borrower.
The ladder steps are smaller than the spread within each step. From Exceptional II to Good I, the average rises 2.29 points. Inside Exceptional II alone, the 5th to 95th range covers 10.00 points.
How Much of That Spread Is Product Mix
Most of it, and that part is doing its job. An 830-score member taking a credit card should pay more than the same member taking a used auto loan. Holding the tier constant in October, the credit card rate was 11.40% and the used auto rate was 5.50%. That 5.90 point gap is collateral, not mispricing.
So hold the product constant instead and read down each column.
Figure 3: Average origination rate by credit score tier and loan type, October 2026
| Credit score tier | Credit card | New auto | Used auto | First mortgage | Junior mortgage |
|---|---|---|---|---|---|
| Exceptional II (825 to 850) | 11.40% | 5.44% | 5.50% | 6.71% | 6.76% |
| Exceptional I (800 to 824) | 11.68% | 5.53% | 5.69% | 6.20% | 6.51% |
| Very good II (770 to 799) | 12.45% | 5.68% | 5.79% | 6.20% | 6.24% |
| Very good I (740 to 769) | 12.32% | 5.93% | 6.17% | 6.46% | 6.64% |
| Good II (705 to 739) | 13.95% | 6.41% | 6.79% | 6.68% | 6.67% |
| Good I (670 to 704) | 15.37% | 7.02% | 7.56% | 6.97% | 7.18% |
| Fair II (625 to 669) | 16.88% | 8.18% | 9.34% | 7.73% | 7.49% |
| Fair I (580 to 624) | 17.18% | 10.87% | 11.93% | 8.55% | 9.49% |
| No score | 15.46% | 5.95% | 7.59% | 6.84% | 6.10% |
| All tiers | 14.93% | 6.12% | 6.92% | 6.53% | 6.70% |
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. Rates are weighted by current balance. Tiers with too few originations to report in a column are omitted from this table.
The two auto columns behave exactly as designed. Used auto climbs from 5.50% to 11.93% as scores fall, a 6.43 point ladder with no reversals.
The two mortgage columns do not. On first mortgages, the 825 to 850 tier paid 6.71% while the two tiers below it both paid 6.20%. On junior mortgages, the top tier paid 6.76% while the 770 to 799 tier paid 6.24%, and members with no credit score on file got the best rate in the column at 6.10%.
The No-Score Bucket Is More Than a Quarter of Your Originations
The largest single origination bucket in October was not a credit tier. It was the members with no credit score on file: 32,192 loans, 28% of all originations across the 120 credit unions.
Its weighted average rate was 7.47%, which reads as unremarkable sitting between Good II and Good I. Then look at the distribution. The weighted average is lower than the 5th percentile, which is 7.74%. A quarter of these loans priced above 13.90%, the typical one priced at 18.00%, and the 95th percentile reached 25.99%.
That happens because the dollars and the loan count are in different places. A handful of large secured loans carry the balance: no-score first mortgages priced at 6.84% and new autos at 5.95%. The volume is in small unsecured credit. The average no-score origination was about $8,900, roughly half the $16,700 average across all tiers.
The top of that distribution is consistent with payday alternative lending. PALs across all tiers averaged 25.49% in October, which lines up with the 90th and 95th percentiles here.
If your credit union holds a low-income designation or a CDFI certification, this bucket is the population those designations exist to serve. It is 28% of what you originate, and the number you are almost certainly reporting on it is 7.47%.
What This Does and Does Not Prove
We are not going to tell you the mortgage inversions in Figure 3 are mispricing. Mortgage pricing moves with lock date, term, loan-to-value, and whether the loan is sold or held, and a single month of originations is a thin sample at the top of the credit range. What we will say is that the inversion is in your file too, or it is not, and the tier average will never tell you which.
We also cannot prove from these figures alone that the top of the no-score distribution is payday alternative lending. The rates line up, but that is the obvious place to look rather than a finding.
What the data does establish is the shape. Rates inside a single credit tier vary more than rates between tiers, and the balance-weighted average sits below the median in every tier we measured.
What Your Credit Union Should Do Now
- Report the median next to the average on every pricing exhibit. Where they diverge, the average is being set by a few large loans and is not describing the loan you typically write.
- Pull the 25th and 75th percentile for each tier. The width of that band is your actual pricing discipline. A tight band means the grid is being followed. A 10-point band means something other than the grid is setting rates.
- Read each loan type column top to bottom and look for reversals. Your auto columns will almost certainly behave. Check the mortgage columns, where ours did not.
- Break out no score as its own reporting line, not as a footnote. At 28% of originations it is larger than any credit tier you have, and its average rate describes almost none of the loans inside it.
- Tie the no-score distribution to your designation strategy. If you hold or are pursuing a low-income designation or CDFI certification, the pricing you deliver to thin-file members is the substance behind the application.
All five of these come out of the AIRES file you already submit. None of them require new systems or new data collection.
Watch the Session
Watch the full November 16 session on YouTube. It runs 30 minutes and covers the macroeconomic backdrop, origination and deposit pricing, member attrition, and delinquency. It is hosted by Luis G. Dopico, PhD, Partner and Chief Economist at CUCollaborate, and Abigail Pereira, Director of Product.
The next Monthly Download is Tuesday, December 15, 2026 at 3:00 PM ET, on which members carry your profitability and which ones leave first. Register for the December session.
How CUCollaborate Can Help
The figures in this article come from AIRES files that credit unions already produce. The Data Collaborative pools them so you can see your own rate distributions next to credit unions like yours, by tier, by loan type, and by percentile rather than by average alone. A call report cannot produce any of this, because the individual loan is not the unit of observation in a call report.
We can run the same cuts against your file. Book a 30-minute walkthrough and we will go through your rate distribution by tier and show you where your no-score book sits.



