68.9% of credit union share accounts hold under $1,000 and contain 1.1% of the money. 1.5% hold $100,000 or more and contain 45.6% of it, earning 2.21%. A three-month Treasury bill pays 4.14%.
Those figures come from the CUCollaborate Data Collaborative, which pools the AIRES files credit unions already produce. This cut covers 16,438,247 share accounts at 116 credit unions in October 2026. We covered them in the November 16 session of The Monthly Download.
The uncomfortable version of the finding is short: nearly half your funding sits in a few thousand accounts, those accounts are earning less than the member could get from the government, and the member holding them can move the money with a phone.
In this article, we will cover:
- Where your accounts are and where your deposits are
- Why some six-figure balances are earning nothing
- What you pay on your largest balances against what Treasury pays
- Why the certificate curve is nearly flat and what that means for funding
- What to check in your own file
The Accounts and the Money Are in Different Places
Here is every share account at 116 credit unions in October, sorted by balance.
Figure 1: Share accounts and balances by account size, October 2026
| Account size | Accounts | % of accounts | Balance ($M) | % of balances | Dividend rate |
|---|---|---|---|---|---|
| Under $1,000 | 11,319,849 | 68.9% | 1,165.0 | 1.1% | 0.35% |
| $1,000 to $9,999 | 3,300,320 | 20.1% | 11,701.6 | 10.7% | 0.56% |
| $10,000 to $99,999 | 1,579,678 | 9.6% | 46,772.7 | 42.7% | 1.38% |
| $100,000 or more | 238,400 | 1.5% | 49,899.3 | 45.6% | 2.21% |
| Total | 16,438,247 | 100.0% | 109,538.6 | 100.0% | 1.65% |
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. Covers 116 credit unions with AIRES share data dated October 2026. Dividend rates are weighted by current balance.
The top two rows are 11.1% of accounts and 88.3% of deposits. The bottom row is 68.9% of accounts and 1.1% of deposits.
This is why a blended cost of funds is close to useless as a management number. The 1.65% total describes no account anyone actually holds. It is the weighted midpoint between 11.3 million accounts earning almost nothing and 238,400 accounts that represent nearly half your balance sheet.
It also explains why member counts and funding stability are only loosely related. Close a thousand accounts from the bottom row and you lose roughly $103,000 in deposits. Lose a hundred accounts from the top row and you lose around $20.9 million. These are accounts rather than members, and one member may hold several, so the account count is not a membership count.
Inside Every Size Band, Some Accounts Earn Nothing
Rank the individual accounts inside each size band by the rate they are paid, without weighting for balance.
Figure 2: Dividend rate percentiles across individual accounts by account size, October 2026
| Account size | 10th | 25th | Average | Median | 75th | 95th |
|---|---|---|---|---|---|---|
| Under $1,000 | 0.00% | 0.00% | 0.35% | 0.05% | 0.10% | 4.00% |
| $1,000 to $9,999 | 0.00% | 0.00% | 0.56% | 0.05% | 0.20% | 3.69% |
| $10,000 to $99,999 | 0.00% | 0.02% | 1.38% | 0.20% | 3.21% | 3.93% |
| $100,000 or more | 0.00% | 0.15% | 2.21% | 2.90% | 3.75% | 4.22% |
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. Percentiles are unweighted across individual accounts. The average column is weighted by current balance.
The bottom row is the one to sit with. At least 10% of accounts holding $100,000 or more were paid 0.00% in October. On 238,400 accounts in the band, that is roughly 24,000 six-figure balances earning nothing.
These are not neglected accounts in an operational sense. Someone opened them, someone services them, and they are among the largest relationships at the institution. They are simply sitting in a product with no rate on it, most likely a checking account, which paid 0.15% on average across all balances in October.
The mirror image is in the top row. The 95th percentile of accounts under $1,000 earned 4.00%, better than the median six-figure balance. Somewhere in the product set, very small accounts are getting a promotional or high-yield rate that the largest balances are not.
Neither of these is visible in a blended cost of funds. Both are visible in about ten minutes of work on an account-level file.
What You Pay Against What Treasury Pays
October dividend rates by share type, set against market rates as of November 12, 2026.
Figure 3: Credit union deposit rates against market rates
| Instrument | Rate |
|---|---|
| Draft (checking) | 0.15% |
| Regular savings | 0.68% |
| Money market | 1.63% |
| Accounts of $100,000 or more, all products | 2.21% |
| IRAs | 2.89% |
| Certificates | 3.64% |
| Three-month Treasury bill | 4.14% |
| Ten-year Treasury | 5.10% |
Sources: CUCollaborate Data Collaborative, October 2026 AIRES files; Federal Reserve, November 12, 2026. Deposit figures are October and Treasury figures are November 12.
The deposit figures are October and the Treasury figures are November 12, so this is a close comparison rather than an exact one. It is close enough for the point.
Every deposit product on that list pays less than a three-month Treasury bill. The member holding $100,000 or more at an average of 2.21% is giving up 1.93 points against an instrument that carries no credit risk and settles in days. That is roughly $1,930 a year on a $100,000 balance.
For most of the last fifteen years this gap did not matter, because the alternative paid close to nothing. The alternative is not paying nothing now, and buying a Treasury bill is something a member can do from a phone in a few minutes. The gap is large enough and the friction is low enough that the comparison gets made.
The question is not whether your rates are competitive against the credit union down the road. It is whether the member sitting on your most valuable 1.5% of accounts has run that comparison yet.
The Certificate Curve Is Nearly Flat, and the Money Is Going Short
Certificates are the one place credit unions do compete on rate. Here are the certificates opened during October, by maturity.
Figure 4: Certificates opened during October 2026, by maturity
| Maturity | New accounts | New dollars ($M) | Rate on new accounts |
|---|---|---|---|
| Three months and under | 3,462 | 142.9 | 3.67% |
| Four to eleven months | 9,811 | 416.8 | 3.71% |
| One year | 11,794 | 476.2 | 3.91% |
| Over one year, under five | 4,428 | 223.4 | 4.18% |
| Five years | 240 | 9.0 | 4.27% |
| Total | 29,735 | 1,268.3 | 3.87% |
Source: CUCollaborate Data Collaborative, October 2026 AIRES files. 116 credit unions. Rates are weighted by current balance.
From three months to five years, the rate rises 0.60 points. Over a comparable stretch of the Treasury curve, from the three-month bill at 4.14% to the ten-year at 5.10%, the market moved 0.96 points.
So credit unions are paying a thinner premium for duration than the market is, and members are responding accordingly. Of the $1.27 billion that went into new certificates in October, $1.04 billion, or 82%, went into terms of one year or less. The five-year bucket took 240 accounts and $9.0 million.
That is a funding profile that reprices almost entirely within twelve months. If rates fall, it reprices in your favor. If they do not, you are renegotiating 82% of your new certificate book inside a year, against a member who has by then had twelve more months of practice comparing your rate to a Treasury bill.
The rate on new one-year certificates moved from 3.44% in January to 3.91% in October. On terms over one year and under five, it moved from 3.72% to 4.18%. The direction is already set.
What Your Credit Union Should Do Now
- Find your six-figure accounts paying 0.00%. This is a single query against your share file and it will take under an hour. Whatever the number is, it is a list of names, not a statistic.
- Size the concentration before you set next year’s deposit strategy. If your distribution resembles this sample, a single-digit percentage of accounts is carrying most of your funding, and your retention plan for those accounts should not be the same as your retention plan for the other 90%.
- Stop managing to blended cost of funds alone. Report it by size band. The 1.65% total in this sample describes none of the four bands underneath it.
- Check your certificate ladder against the Treasury curve, not against peers. Your competition for a $100,000 balance is no longer only the institution across town.
- Look at who holds the sub-$1,000 accounts earning 4.00%. If a promotional rate is landing on your smallest balances and not your largest, that is a product configuration decision somebody made, and it may not be the one they intended.
Every one of these comes from the share side of the AIRES file you already produce.
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 deposit tables here come from the share side of AIRES files that credit unions already produce. The Data Collaborative pools them so you can see your own balance concentration and rate distribution next to credit unions like yours, by size band, by product, and by percentile. A call report gives you a single blended cost of funds, which is the one number in this article that turned out to describe nothing.
We can run these cuts against your file. Book a 30-minute walkthrough and we will find your zero-rate six-figure accounts and show you where your concentration sits against your peers.



