Credit unions added new members at 10.6% a year and lost existing ones at 8.3%. Net growth came to 2.3%. Almost everything that decides whether your membership grows happens inside that gap.
Those figures come from the CUCollaborate Data Collaborative, which pools the AIRES files credit unions already produce. This cut covers 4,851,189 member records at 70 credit unions in October 2026. We walked through them in the November 16 session of The Monthly Download. The gross flows are more than four times the net result.
The instinct when growth stalls is to open more accounts. The account-level data points the other way. Opening an account is not the same as adding a member, and these files show exactly where the difference sits.
In this article, we will cover:
- Why the net membership number hides the two flows that matter
- The attrition rate on members whose only product is a small savings account
- What happens to retention as a loan gets paid down
- Three other places the average hides something
- What to do with all of this
The Net Number Hides Two Much Bigger Ones
Across the 70 credit unions in the October sample, 387,689 members left over the year and 498,116 joined. The difference is 110,427, which is the 2.3% net rate.
That framing matters because the two flows are not equally hard to move. Acquisition costs money and competes against every other financial institution in the market. Retention is a question of what you already know about members you already have.
One percentage point off the attrition rate across this sample is roughly 48,500 members a year. That is 44% of the entire net gain, bought without a single additional acquisition dollar.
Savings-Only Members Leave Faster Than You Replace Them
Split the member base by what each member actually holds and the averages come apart immediately.
Figure 1: Annual member attrition by member type, October 2026
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. 4,851,189 member records across 70 credit unions. Rates are annualized.
The second group excludes anyone already counted in the first, so the two sub-$1,000 groups do not overlap. Together they hold 1,968,065 members, about 41% of the base, and both shrink.
The cliff is at $1,000. Among members with no loans, attrition falls from 13.2% below that line to 3.7% above it. Nothing about the product changes across that threshold. What changes is whether the member has anything at stake.
A member whose only relationship is a $40 savings account is, functionally, a lapsed member who has not closed the account yet. You are counting them, and at 17.0% a year they are leaving faster than you can open replacements.
The Loan Is What Retains, Until It Is Paid Off
Now take the members who do have loans and sort them by how much of the original loan amount is still outstanding.
Figure 2: Attrition by share of the original loan amount still owed
| Members with loans at start of period | Members | Attrition | New member rate | Net |
|---|---|---|---|---|
| Current balances over 50% of original | 1,025,159 | 3.0% | 18.8% | 15.8% |
| Current balances 10% to 50% of original | 331,804 | 5.0% | 1.8% | -3.2% |
| Current balances 5% to 10% of original | 29,822 | 9.5% | 2.6% | -6.9% |
| Current balances under 5% of original | 150,084 | 5.8% | 4.0% | -1.8% |
| All members | 4,851,189 | 8.3% | 10.6% | 2.3% |
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. 70 credit unions. Rates are annualized.
Read the attrition column and ignore the rest for a moment. A member early in a loan leaves at 3.0% a year. A member who has paid that loan down to between 5% and 10% of the original leaves at 9.5%, more than three times the rate.
The practical reading: the loan is doing your retention work, and it stops doing it before the member finishes paying. By the time the balance is under half the original amount, attrition has already begun climbing.
Three More Places the Average Hides Something
The same pattern holds on three other cuts, and none of them needs a table.
Indirect auto members leave faster than direct ones. Take members whose entire relationship is one regular share account and one auto loan. An indirect borrower early in the loan leaves at 4.6% a year. Once that loan is paid down past 90%, the rate is 21.4%. For direct borrowers the same two points are 3.5% and 13.5%. Both climb as the loan runs out, but the indirect member leaves at a rate 7.9 points higher at the stage that matters most. That gap is the price of a relationship that started at a dealership rather than with you. The direct borrower chose your credit union. The indirect borrower chose a car, and your credit union came attached to it.
The retention break is at 30 days past due, not at 60. A member current on all loans leaves at 3.3% a year. A member 1 to 30 days late leaves at 4.3%, barely different. Past 30 days it nearly triples to 12.4%, and past 60 days it reaches 29.6%. Most credit unions already work that window as a collections problem. The attrition data says it is also the retention window with the most at stake, and it opens well before the loss becomes a credit question.
Every additional product pulls attrition below the average. Members holding a junior mortgage leave at 1.4%, a personal loan at 2.4%, a certificate at 3.7%, a money market at 3.8%, a credit card at 4.2%, and checking at 4.8%. Every one of those sits under the 8.3% all-member rate. Checking is the weakest retainer on that list and it still holds attrition to 4.8%, which matters because it reaches 2,148,251 members, 44% of the base. It is the only product here with enough reach to move the total on its own.
What This Does and Does Not Prove
These are relationships in historical data, not causes. A member who takes out a junior mortgage was already a more committed member before the mortgage existed, and some of that 1.4% belongs to the member rather than to the product. Selling a junior mortgage to a disengaged member will not produce a 1.4% attrition rate.
Two of the groups above are small. The paid-down indirect and direct auto groups hold 9,486 and 3,074 members, so treat those exact rates as directional rather than precise. The direction is consistent across both channels and it is steep.
One column in Figure 2 deserves a caveat of its own. New members arriving with new loans land in the top row by definition, so the 15.8% net figure there is partly an artifact of where new borrowers enter. The attrition column has no such problem: it tracks the same people across the period, and it rises steadily as the loan runs out.
What the data does establish is the ranking, and the ranking is consistent. The deeper and more recent the obligation, the more it retains.
What Your Credit Union Should Do Now
- Stop reporting net membership change on its own. It is the difference between two flows that are four times its size. Report gross attrition and gross acquisition next to it, every month.
- Size your sub-$1,000 single-product bucket. In this sample it was about 41% of all members and it was shrinking. If yours is similar, your reported membership count is overstating the relationship base you actually have.
- Build the trigger at 50% paid down, not at payoff. Attrition starts rising well before the final payment. A refinance, a second product, or a conversation at the halfway mark lands while the member is still yours.
- Report indirect and direct auto separately. They behave differently enough that blending them hides a 7.9-point gap in attrition at the stage that matters most.
- Treat the 30-day mark as a retention event. Attrition nearly triples between 1 to 30 days late and 30 to 60. Whatever outreach happens at day 45 should be happening at day 20.
None of these require new data. They require reading the AIRES file you already produce at the member level rather than the portfolio level.
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.
Luis walks the headline attrition figures in that session and notes that the member-type detail did not fit in the half hour. That detail is what this article covers.
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
Every figure in this article comes from AIRES files that credit unions already produce. The Data Collaborative pools them so you can see your own attrition curves next to credit unions like yours, broken out by member type, product held, and credit score tier. Call report data cannot produce any of these cuts, because the member is not the unit of observation in a call report.
We can run these same cuts against your AIRES file. Book a 30-minute walkthrough and we will go through your attrition by member type and show you where your sub-$1,000 bucket sits.

.jpg)

