Menu
Three numbers explain more about the next decade of credit union sustainability than any strategic plan currently sitting in a board packet.
The average credit union member is 47 years old. That places the typical member past the peak borrowing years and squarely in the savings years, the stage of life where a member’s relationship with your balance sheet shifts from the loan side to the share side.
The median first-time homebuyer is now 40 years old, an all-time high, according to the National Association of Realtors’ 2025 Profile of Home Buyers and Sellers. The median repeat buyer is 62. The median buyer overall, across the entire market, is 59.
Read those three numbers together and the problem stops being abstract.
The three ages
The typical credit union member is seven years older than the median first-time homebuyer, and the household-formation event that starts a lending relationship now happens at 40.
| Group | Median age |
|---|---|
| Credit union member | 47 |
| First-time homebuyer | 40 |
| Repeat homebuyer | 62 |
Sources: CUInsight, "Aging credit union membership"; National Association of Realtors, 2025 Profile of Home Buyers and Sellers.
Homeownership is the event that starts a financial household. It is where a person stops being a customer of financial products and becomes a household with a balance sheet: a mortgage, an escrow account, an insurance relationship, a payment that clears every month, and within a few years a home equity line, an auto loan, a card, and eventually a wealth conversation. That entire sequence has one front door, and the front door is now opening at age 40.
Your existing membership base has largely walked through it already, somewhere else. It is aging steadily into the repeat-buyer and deposit-holder profile: fewer originations, more shares, less spread, and a slow erosion of the loan side of the balance sheet.
You can see it in the aggregate data. Credit unions closed the first quarter of 2026 with 145.8 million members, up 2.5 million year over year, and $2.48 trillion in assets, up 4.9 percent. Healthy numbers on the surface. But the loan-to-share ratio slipped to 81.5 percent, down from 81.8 percent a year earlier, and the industry contracted by 161 institutions through consolidation.
Deposits are growing faster than loans. Institutions are disappearing. And the average member is seven years older than the median first-time buyer your growth plan depends on capturing.
Here is what makes this a strategy problem rather than a demographic inevitability. Most credit unions answer aging membership with a member growth target: more accounts, more checking, more rate-led deposit campaigns aimed at whoever will open one. That response treats the symptom and worsens the disease. Adding more 47-year-old deposit relationships to a book of 47-year-old deposit relationships does not change the age curve. It raises the cost of maintaining it.
The alternative is to stop thinking of the mortgage as a product line and start treating it as the acquisition engine for everything else. A single purchase mortgage does not produce one relationship. It produces a lead into a household that will need a sequence of products over the next twenty years, and it produces that lead at the exact moment the household is deciding who its primary financial institution will be. No other product in the credit union lineup arrives at that moment. Nothing else attaches for seven to thirty years, pulls direct deposit with it, creates a servicing asset worth more than the loan’s own origination margin, and reaches the only cohort in the market forming genuinely new financial households.
This is not a marketing question. It is a sustainability question. An institution that cannot originate purchase mortgages profitably at its own average loan size has no mechanism for replacing the households it is losing to age, and every year it waits, the replacement cost rises and the remaining pool of first-time buyers gets more contested.
Credit unions currently hold a first mortgage with 2.4 percent of their members.
The rest of this article is about why that number has not moved, what it is costing, and what it would take to move it.
The industry reports member growth every quarter, and every quarter it looks fine. That is the problem.
Credit unions added 2.5 million members in the year ending Q1 2026, roughly 1.7 percent growth. Assets grew 4.9 percent. Read the press release and the trajectory looks stable.
Now look at what those additions actually are.
J.D. Power’s October 2025 data finds that 52 percent of new checking relationships are secondary rather than primary. The figure is 48 percent for investment accounts and 65 percent for credit cards. Roughly 34 percent of new checking accounts go inactive within their first year. And a meaningful share of the industry’s reported growth is not organic acquisition at all. It is absorption, the arithmetic result of 161 credit unions disappearing through consolidation in a single year. Those members were already counted. They just moved.
So the headline number is measuring account openings. It is not measuring households, and it is certainly not measuring households in the borrowing years.
The distribution underneath the average is worse than most boards assume, and three independent datasets find the same shape. Raddon’s household profitability work puts 31.7 percent of credit union households in profitable territory, with the top 12 percent averaging $2,101 in annual profit and contributing between 182 percent and 360 percent of total institutional profit. They do not merely carry the book. They fully fund everyone else’s losses. StrategyCorps, looking only at checking, finds the top 10 percent of relationships hold 63 percent of relationship dollars, while 34.9 percent of checking accounts are outright unprofitable against a cost to service of $250 to $400 a year.
Run those distributions against NCUA aggregates and the spread is roughly $4,270 in annual net revenue for a top-decile member against about $62 for the bottom 35 percent. That is a gap of nearly 70 times. Any acquisition budget priced against an average member value is wrong by an order of magnitude at both ends of the curve.
Which makes the cost side hard to defend. Acquiring a credit union member now costs $565, up 15 percent year over year, according to the Clutch 2026 Member Acquisition Cost Report. Average pre-provision contribution, derived from NCUA aggregates, is $216 per member per year. That is a payback of two and a half to three years for an average member, and on a net income basis, no payback at all for the bottom 40 percent.
Spending $565 to open an account that arrives secondary, goes dormant inside twelve months, and never generates a loan is not growth. It is cost of acquisition booked as a marketing win.
The number that would actually tell you whether the institution is sustainable is one almost nobody reports to the board: net new households in the borrowing years, anchored by a loan. That is a different metric, it produces a different budget, and it points at a different product.
There are two kinds of mortgage borrower, and credit unions have spent a decade competing for the wrong one.
Repeat buyers are 79 percent of the market. Their median age is 62. Thirty percent of them pay cash, and those who finance put down a median of 23 percent. They are, almost without exception, already somebody’s member. They have a primary institution, an existing mortgage relationship, a card, a deposit history, and in many cases an advisor. Winning one is a share-of-wallet fight against an incumbent who has held the relationship for years and who will see the payoff request before you see the application.
First-time buyers are 21 percent of the market, the lowest share the National Association of Realtors has ever recorded. Their median age is 40.
They are also the only cohort in the entire market forming a genuinely new financial household.
That distinction is the whole strategic argument. A first-time buyer has no incumbent mortgage servicer, no established primacy, no thirty-year relationship to displace, and no switching cost working against you. They are choosing a primary financial institution for the first time as a household, and the mortgage is the decision that anchors every choice that follows it. This is the single moment where a credit union competes on level ground, and it is the only moment that produces a member who will still be generating revenue three decades from now.
Consider what one of those relationships actually opens. The mortgage brings an escrow account, which brings a recurring monthly obligation, which is the most reliable path to direct deposit and checking primacy there is. Title and settlement attaches at closing. Debt protection attaches at closing. A home equity line follows as equity builds. Auto, card, and eventually a wealth conversation follow the household as it matures. One origination is not one product. It is the entry point to a sequence of them, sold to a household that has already told you it trusts you with the largest transaction of its life.
The economics of that household are not the economics of an average member. A relationship that reaches genuine primary financial institution status generates eight times the fee revenue and ten times the deposits of a non-primary relationship, according to Curinos, and it attrits at under 2 to 3 percent annually against an industry baseline near 15 percent, according to BCG. At those retention levels, member lifetime value moves substantially:
Member lifetime value by annual retention rate
Small movements in retention produce large movements in lifetime value. The bottom row reflects the attrition rate BCG observes for true primary financial institution relationships.
| Annual retention | Member lifetime value | Implied tenure |
|---|---|---|
| 85 percent | $582 | 6.7 years |
| 90 percent | $836 | 10 years |
| 95 percent | $1,372 | 20 years |
| 97.5 percent | $1,950 | 40 years |
Derived figure. Calculated as m × [r ÷ (1 + i − r)] at a $130 annual margin and a 4 percent cost of capital, with the margin derived from NCUA aggregate net income divided by published membership for calendar year 2025.
And a Raddon A-tier household, held at the primary-institution attrition rate BCG observes, is worth roughly $31,500 in lifetime profit, about 38 times an average member at 90 percent retention.
Now the objection, because a CFO will raise it before the end of this paragraph. If first-time buyers are only 21 percent of the market and that is an all-time low, is this not a shrinking pool?
It is a smaller pool, and it is more contested. It is also the only pool that produces net new households, which means the scarcity raises the value of the capability rather than lowering it. Institutions that can originate profitably for this borrower capture a disproportionate share of a decade of household formation. Institutions that cannot are left competing for 62-year-old repeat buyers against the incumbent who already holds them, while their own membership ages another year.
There is also a reason to think this borrower is winnable specifically by credit unions. The binding constraint for first-time buyers is capital, not credit. Their median down payment is 10 percent, the highest since 1989. Fifty-nine percent funded it from personal savings, 26 percent from financial assets such as 401(k) balances, stock, or cryptocurrency, and 22 percent from a family gift or loan. Those are files that need judgment, documentation work, and a human conversation. They are exactly the files that high-volume independent mortgage banks handle worst, and exactly the files that credit union manual underwriting, operational flexibility, and lower rate sensitivity handle best.
The capability gap is not desire, and it is not underwriting skill. It is cost per loan, and that is the subject of the next two sections.
If the argument above is right, you would expect to see it in market share data. You do.
Credit unions claimed roughly 7 percent of the purchase mortgage market in 2024, a figure that has remained essentially flat over the preceding five years. Independent mortgage banks took 68 percent of purchase volume. Traditional depositories took 25 percent. Credit unions originate about 15 percent of all mortgage loans despite representing roughly one third of U.S. mortgage lending institutions.
Now look at refinance. Credit union refinance share ran 8 to 9 percent in 2020 and 2021, rose to 14 percent in 2022, peaked at 20 percent in 2023, and settled at 16 percent in 2024. And a substantial portion of that share came from smaller-dollar loans, including home equity and renovation products.
Put those two trends side by side and the diagnosis writes itself. A refinance is a transaction with a borrower who is already yours, or who is at least already a homeowner shopping on rate. A purchase is an acquisition. Credit unions are demonstrably competent at serving an existing book and demonstrably absent from the event that creates new households.
That is the aging membership problem restated as a market share statistic.
The penetration data underneath it tells the same story. As of mid-2025, product penetration across the industry looked like this:
Credit union product penetration
Checking reaches nearly two thirds of members. The first mortgage, the product with the longest duration and the highest switching cost, reaches 2.4 percent.
| Product | Share of members |
|---|---|
| Share draft / checking | 63.2 percent |
| Credit card | 18.8 percent |
| Used auto | 18.8 percent |
| Other unsecured | 10.9 percent |
| Share certificates | 10.7 percent |
| Money market | 8.0 percent |
| New auto | 6.7 percent |
| IRA | 3.1 percent |
| First mortgage | 2.4 percent |
| Home equity or second lien | 2.3 percent |
Source: America's Credit Unions / Michigan Credit Union League, U.S. Credit Union Profile, data as of June 30, 2025. Auto lending reaches 25.5 percent of members when new and used are combined. Bars are scaled to the highest value in the table and are illustrative.
Auto combined reaches 25.5 percent. First mortgage reaches 2.4 percent. And the clearest single measure of the resulting primacy gap is direct deposit: penetration sits under 50 percent at credit unions against 77 to 80 percent at the three largest banks.
The consequence shows up in attrition. Industry-wide attrition runs near 15 percent annually, and roughly 10 percent for checking. For relationships that reach true primary financial institution status, BCG observes attrition under 2 to 3 percent. Primacy is not a satisfaction score. It is a retention multiplier, and the mortgage is the most reliable way to earn it.
So the strategic conclusion is not in dispute inside most credit unions. Everyone in the room already agrees they need younger, borrowing-age households and that mortgage is how you get them. The share number has not moved in five years anyway. That is not a strategy failure. It is an economics failure, and Section 6 is where the economics live.
Before accepting that conclusion, it is worth ruling out the alternatives, because most credit unions have already tried them.
Rate-led deposit acquisition. High-yield savings runs from zero to negative on a spread basis. Share certificates run zero to roughly $110 a year. These are wholesale funding in a retail wrapper, justified by liquidity and balance sheet growth rather than by revenue. Any campaign that leads with rate is buying wholesale funding at retail acquisition cost, and the member it buys is a rate shopper who will leave for the next offer.
Auto lending. Auto is the strongest loan-side penetration credit unions have at 25.5 percent combined. But it is increasingly captive-financed at the dealership, its duration is short, and it does not pull direct deposit or escrow behind it.
Checking-first acquisition. Fifty-two percent of new checking relationships arrive secondary, and roughly 34 percent go dormant within a year.
Youth and club accounts. Worth doing, and worth being honest about: no public per-account balance or revenue data exists for these products anywhere. They are behavioral and retention products whose value is real but unmeasured. No growth business case should rest on them.
What the mortgage does that none of these do is attach a household for seven to thirty years, at the highest switching cost of any retail financial product, with a mandatory monthly obligation that pulls the payment relationship and direct deposit along with it.
It also concentrates the highest-value cross-sell moment a credit union gets. Ranked by revenue per participating member, the most valuable products in the entire credit union lineup all attach at a loan closing: title and settlement at $600 to $1,400 per closing, mechanical breakdown coverage at $250 to $400, debt protection at $150 to $400 a year, and GAP at $150 to $250. Missing the attach at a mortgage closing is expensive in a way that missing a checking cross-sell is not, and at 2.4 percent penetration, most credit unions are missing it by simply never being at the closing table.
One more point deserves emphasis, because it changes who you market to. Credit unions under $10 billion in assets are exempt from the Durbin interchange cap and earn roughly $0.51 per debit transaction against $0.23 for covered issuers. Because that revenue is largely flat-rate per transaction, high-frequency low-balance households are disproportionately profitable. Bank On certified accounts average a $1,273 balance but 33 debit transactions per month, more than double the portfolio average of 13.6, producing roughly $249 a year in revenue with no overdraft income at all. That is more than a typical free checking account generates.
The implication is uncomfortable and useful: the younger, lower-balance households that most credit unions screen out of mortgage marketing on balance criteria are frequently the households worth acquiring.
The constraint is cost per loan. Here is the arithmetic.
The Mortgage Bankers Association reported total loan production revenue of 333 basis points and total production expense of 308 basis points in the second quarter of 2026, netting $973 per loan on an average loan balance of $386,359.
The average credit union origination is $198,167, according to 2025 HMDA data.
Revenue scales with basis points. A large share of production expense does not; it is fixed per loan, driven by headcount, compliance, and touch count rather than by loan size. Apply the industry cost structure to a credit union loan and the result is this:
Per-loan production economics at the average credit union loan size
Revenue scales with basis points. A large share of production expense does not. Applying the industry cost structure to a $198,167 loan produces a negative result.
| Line | Per loan |
|---|---|
| Revenue at 333 basis points on $198,167 | $6,599 |
| Production expense, discounted 20 percent from the industry figure of roughly $10,900 | $8,750 |
| Implied result per loan | −$2,151 |
Derived figure. Revenue basis points and per-loan production expense from the Mortgage Bankers Association Quarterly Mortgage Bankers Performance Report, Q2 2026, which reports 333 basis points of production revenue and 308 basis points of production expense on an average loan balance of $386,359. Average credit union origination of $198,167 from 2025 HMDA data. The 20 percent expense discount is an assumption intended to credit lower marketing and commission structures, not a published figure. A first-lien-only average nearer $235,000 improves every basis-point line by roughly 18 percent.
A credit union originating $198,167 loans on that cost structure loses money on production. Not thin margin. Negative.
This is where the demographic argument and the economic argument collide, and it is the reason purchase share has not moved in five years. First-time buyers borrow less than repeat buyers. The exact segment that solves the aging membership problem is the segment the in-house cost structure punishes hardest. Every credit union that has concluded “we need younger members” and then quietly discovered that purchase lending does not pay for itself has run into this arithmetic without necessarily naming it.
Two honest qualifications, because a CFO should stress-test this before acting on it.
First, the $198,167 HMDA average includes home improvement loans, which account for roughly 26 percent of credit union HMDA volume. A first-lien-only average is more likely in the $220,000 to $250,000 range. Run the case at $235,000 as well, which improves every basis-point line by about 18 percent and narrows but does not close the gap.
Second, the 20 percent expense discount applied above is an assumption, not a published figure. It is intended to credit credit unions for lower marketing spend and lower commission structures than the independent mortgage bank cost base MBA measures. Institutions with heavier branch overhead will look worse than this; lean shops will look better.
What survives both qualifications is the threshold. To make purchase lending pay at credit union loan sizes, per-loan production expense has to come in under roughly 330 basis points of a $198,167 loan, which is about $6,500. That single number is the entire economic case for changing how origination is delivered, and it is the number every option in Section 10 should be measured against.
Even at breakeven production, the origination is not where the value sits.
At a 25 basis point servicing fee and the bulk servicing multiple of 4.50 to 5.25 times reported by MCT in July 2026, retaining servicing on a $198,167 loan creates a mortgage servicing right worth roughly $2,358. That is about 4.8 times the net production income the MBA reports on a much larger average loan. The asset created at closing is worth several times the margin earned at closing.
Selling servicing released surrenders 12 to 17 basis points of that fair value, roughly $240 to $340 per loan. It also surrenders every dollar of future servicing income, and it surrenders the member relationship. Read that against Sections 1 and 3: the institution spends $565 and considerable effort to acquire a net-new borrowing-age household, then hands that household’s monthly statement to a competitor for the next thirty years.
Retaining servicing and sub-servicing the operational work nets roughly $314 per loan per year. The components are a 25 basis point servicing fee on an average balance near $185,000, producing $462 gross, less approximately $108 in sub-servicing cost and approximately $40 in oversight. On a 5,000-loan portfolio that is about $1.57 million a year in recurring, rate-insensitive revenue, against zero if servicing is released.
Escrow deserves separate attention because it is the cleanest revenue line in the lineup. A typical escrowed loan carries roughly $3,100 across the escrow cycle. At a 3.25 percent funds transfer price credit against roughly 0.30 percent blended interest paid, that is $91 to $102 per loan per year with no acquisition cost, no rate competition, and no fee disclosure. Ten thousand escrowed loans is approximately $1 million a year. Worth noting for institutions considering third-party arrangements: a non-depository servicer placing escrow at a custodial bank typically receives an earnings credit of 60 to 90 percent of fed funds, which is a direct argument for keeping escrow on the credit union’s own balance sheet.
There is a strategic dimension beyond the revenue. A retained servicing portfolio is a first-party dataset covering every borrower’s rate, equity position, payment history, escrow change, and life stage. It is the only asset that reliably lets an institution win the next transaction from a household it has already acquired. Release servicing and that dataset, and the recapture opportunity it represents, belongs to somebody else.
One caution on sourcing. The commonly quoted price of $6 to $10 per loan per month for sub-servicing performing loans is industry convention with no public document behind it. What can be defended are the bounds from MBA’s own servicing cost data: large servicers operate at roughly $127 to $132 per loan per year, mid-size servicers at $227 to $232, with performing loans costing about $176 a year against $1,573 for non-performing, a difference of roughly nine times. Sub-servicers price somewhere between those bounds plus margin. Anyone presenting a precise per-loan-per-month figure as sourced should be asked for the source.
Strategy is the easy part. Here is what has to be true operationally.
Product readiness. The binding constraint on first-time buyers is the down payment, now at a median of 10 percent and the highest since 1989. Low-down-payment structures, down payment assistance layering, and the operational competence to actually close those files are the price of entry, not a differentiator.
Underwriting posture. Twenty-six percent of first-time buyers funded down payments from financial assets such as 401(k) balances, stock, or cryptocurrency, and 22 percent used a family gift or loan. Those files require documentation work and human judgment. Manual underwriting capability is a genuine credit union advantage, but only if the cost structure in Section 6 makes it affordable to use.
A real data foundation. This deserves to be said plainly: there is no published dataset of revenue per product per member for credit unions. Not from NCUA, not from the trade associations, not from any vendor. Most institutions are targeting on penetration reports and instinct. The highest-value fix is free. NCUA collects account counts by share and loan type on the 5300 call report but does not publish them in the quarterly summaries. Pulling the raw 5300 files replaces roughly a dozen derived denominators with actual counts and makes it possible to tier the membership by contribution. If you cannot identify your A-tier households, you cannot build a lookalike model of them, and every acquisition dollar is spent against an average that Section 2 showed does not exist.
Where AI earns its keep. Specifically, and without hype: propensity modeling against 5300-grounded segments rather than demographic guesses; document intake and conditions clearing, which attacks the fixed per-loan cost in Section 6 directly; automated disclosure and status communication; and recapture triggers running against the servicing portfolio described in Section 7. Where it does not help is equally worth stating. AI does not fix a per-loan cost structure driven by headcount and channel design, and it does not create primacy.
Measurement discipline. Published member acquisition cost figures vary by roughly ten times depending on whether they count paid media only or fully loaded cost including branch staff, incentives, and onboarding. The $565 figure cited here and a much lower figure can both be accurate for the same institution, but not in the same board deck. Pick one definition, hold it, and then measure the outcome that matters: net new households in the borrowing years, anchored by a loan.
Rate sensitivity as planning hygiene. Every spread figure in this article assumes a 3.25 percent funds transfer price credit and a 1.90 percent blended cost of funds, consistent with the August 2026 environment of a 3.75 percent upper fed funds target and SOFR at 3.66 percent. In an easing scenario at a 2.00 percent funds transfer price, every spread figure falls by roughly 38 percent, and rate-paying deposit products go decisively negative. That scenario strengthens rather than weakens the case for loan-side household acquisition, but it should be run before any multi-year plan is approved.
For the CFO, the case reduces to a comparison.
Acquiring an average member costs $565 and returns $216 a year in pre-provision contribution. Payback lands between two and a half and three years, and for the bottom 40 percent of members it never arrives on a net income basis. Two ceilings from NCUA aggregates constrain any attempt to improve that through fees: fee income runs $69 per member per year and total non-interest income runs $171. Operating expense per member is $511 a year, which is the hurdle any new product or channel has to clear at the margin before it contributes anything.
Now run the same acquisition dollar through a mortgage-anchored household and count what attaches:
The retention effect compounds the rest. Moving a relationship from 85 percent to 97.5 percent annual retention moves member lifetime value from $582 to $1,950. And the Raddon A-tier household, the top 12 percent that funds the institution, generates roughly $2,101 a year in profit and approximately $31,500 in lifetime profit at primary-institution attrition levels, about 38 times an average member.
That is the whole argument in one line. Spending $565 to acquire an average member is a marginal decision. Spending $565 to acquire a 40-year-old net-new household and anchoring it with a retained-servicing mortgage is not close.
The aging membership problem and the per-loan cost problem are not two problems. They are the same problem, and they have the same solution.
Which brings the decision to a single question, the one established in Section 6.
Can your institution originate at under roughly $6,500 per loan at your actual average loan size, while retaining servicing and escrow?
If the answer is yes, build. If the answer is no, there are three honest paths and they should be evaluated on their merits.
Build scale in-house. Fixed per-loan cost falls with volume. But generating the purchase volume required to get there is the same capability the institution does not currently have, which makes the problem circular for most credit unions under $1 billion in assets. This path is real, and it is slow, and it requires sustained board patience through a period of negative contribution.
Exit or broker it out. This solves the cost problem immediately. It also surrenders the acquisition wedge described in Section 3, the primacy asset in Section 4, and the servicing annuity in Section 7. It is a defensible decision, but it should be made explicitly rather than by drift, because it is functionally a decision to let the membership age.
White-labeled operations through a credit union service organization. Fixed cost is shared across institutions and converted toward variable cost. The credit union keeps its brand on the disclosures and the monthly statement, keeps the member data, keeps retained servicing and escrow, and keeps the recapture rights that make the portfolio a growth asset rather than a run-off asset.
Whichever path is chosen, these are the questions worth asking any partner, and they are worth asking even of an internal build:
A piece like this is only as good as its willingness to say what it does not know.
Several figures here are derived rather than published. Revenue, contribution, and expense per member are calculated from NCUA aggregates divided by published membership, not reported directly by anyone. The per-loan credit union economics in Section 6 apply an industry cost structure to a credit union loan size with an assumed 20 percent expense discount. The member value decile spread is reconstructed from Raddon and StrategyCorps distributions run against NCUA totals.
Several sources are older than ideal. Raddon’s publicly available household tier dollar figures date from 2016 to 2021 and predate the entire rate cycle; the tier structure is stable, the dollar amounts are not. StrategyCorps’ checking relationship data is 2016 vintage.
And several things simply are not published. There is no public sub-servicing price per loan per month. There is no credit union specific overdraft revenue figure that can be trusted, with major estimates disagreeing by a factor of five, and NCUA stopped collecting overdraft and NSF data in March 2025, making the last vintage final. There is no public title insurance revenue per loan, because it typically flows through a service organization rather than the call report. Raddon and Curinos values sit behind subscription.
None of that changes the shape of the argument. The three ages are published. The market share data is published. The penetration rates are published. The MBA cost data is published. The conclusion those sources support is that credit unions are aging out of the borrowing years and are absent from the one event that creates replacement households, and that the obstacle is a per-loan cost structure rather than a strategic disagreement.
If you want to know where your own institution sits against these numbers, the inputs are your average loan size, your annual origination volume, and your current servicing disposition. That is enough to run your per-loan economics against the MBA benchmarks and to model what a retained-servicing portfolio would be worth on your balance sheet in ten years.
America’s Credit Unions / Michigan Credit Union League, U.S. Credit Union Profile, data as of June 30, 2025. Product penetration rates by share and loan type, including first mortgage penetration of 2.4 percent of members.
BCG. Attrition rates for primary financial institution relationships (under 2 to 3 percent).
Clutch, 2026 Member Acquisition Cost Report. Cost to acquire a credit union member of $565, up 15 percent year over year.
CUInsight, “Aging credit union membership.” Average credit union member age of 47. https://www.cuinsight.com/aging-credit-union-membership/
Curinos. Primary financial institution relationships versus non-primary: 8 times fee revenue, 10 times average deposits.
Customer Service Profiles. Share of new checking accounts inactive within year one (approximately 34 percent); alternate acquisition cost figures for credit unions and retail banks.
Federal Reserve. Durbin interchange amendment exemption: approximately $0.51 per debit transaction for issuers under $10 billion in assets against $0.23 for covered issuers.
HMDA, 2025 data. Average credit union mortgage origination of $198,167; home improvement loans as approximately 26 percent of credit union HMDA volume.
J.D. Power, October 2025. Share of new accounts opened as secondary rather than primary relationships: 52 percent checking, 48 percent investment, 65 percent cards.
Mortgage Bankers Association, Quarterly Mortgage Bankers Performance Report, Q2 2026. Total loan production revenue of 333 basis points, total production expense of 308 basis points, net production income of $973 per loan on an average loan balance of $386,359. Servicing cost data: $127 to $132 per loan per year for large servicers, $227 to $232 for mid-size servicers; $176 per year for performing loans against $1,573 for non-performing.
Mortgage Capital Trading (MCT), July 2026. Bulk mortgage servicing rights multiples of 4.50 to 5.25 times.
National Association of Realtors, 2025 Profile of Home Buyers and Sellers. First-time buyer share of 21 percent (historic low); median first-time buyer age of 40 (all-time high); median repeat buyer age of 62; median buyer age of 59; first-time buyer median down payment of 10 percent (highest since 1989); repeat buyer median down payment of 23 percent; 30 percent of repeat buyers purchasing all cash; down payment sources of 59 percent personal savings, 26 percent financial assets, 22 percent family gift or loan. https://www.nar.realtor/press-releases/first-time-home-buyer-share-falls-to-historic-low-of-21-median-age-rises-to-40
NCUA, Quarterly Credit Union Data Summary and system performance data, Q1 2026 and Q4 2025. Total membership of 145.8 million, up 2.5 million year over year; total assets of $2.48 trillion, up 4.9 percent; 4,250 federally insured credit unions, down 161 year over year; loan-to-share ratio of 81.5 percent against 81.8 percent a year earlier. Aggregate interest income, fee income, other operating income, and operating expense used to derive per-member figures. https://ncua.gov/newsroom/press-release/2026/ncua-releases-first-quarter-2026-credit-union-system-performance-data
Raddon. Household profitability tiering: 31.7 percent of households profitable; top tier averaging $2,101 in annual profit and contributing 182 to 360 percent of total institutional profit.
Scotsman Guide, “Can credit unions figure out purchase mortgages?” Credit union purchase mortgage market share of approximately 7 percent in 2024, flat over five years; independent mortgage banks at 68 percent and traditional depositories at 25 percent of purchase volume; credit unions originating approximately 15 percent of all mortgage loans while representing about one third of mortgage lending institutions; credit union refinance share of 8 to 9 percent in 2020 and 2021, 14 percent in 2022, 20 percent in 2023, and 16 percent in 2024. https://www.scotsmanguide.com/news/can-credit-unions-figure-out-purchase-mortgages/
StrategyCorps. Checking relationship distribution: top 10 percent holding 63 percent of relationship dollars; 34.9 percent of checking accounts unprofitable; cost to service of $250 to $400 per year.
Bank On / Cities for Financial Empowerment Fund account standards data. Certified account averages of a $1,273 balance and 33 debit transactions per month against a portfolio average of 13.6.
The following are calculated rather than published and should be treated as defensible estimates rather than citations.
Per-member revenue and expense. Gross revenue per member of $1,026, net revenue per member of $727, pre-provision contribution per member of $216, net income per member of $130, and operating expense per member of $511 are derived from NCUA aggregate income and expense divided by published membership for calendar year 2025. Fee income per member of $69 and total non-interest income per member of $171 are the corresponding published ceilings.
Member value distribution. Top-decile net revenue of approximately $4,270 per member per year and bottom-35-percent net revenue of approximately $62 are reconstructed by applying Raddon and StrategyCorps distributions to NCUA aggregate totals.
Member lifetime value. Calculated using m × [r ÷ (1 + i − r)] at a $130 annual margin and a 4 percent cost of capital, producing $582 at 85 percent retention, $836 at 90 percent, $1,372 at 95 percent, and $1,950 at 97.5 percent. The A-tier lifetime profit figure of approximately $31,500 applies the same method to Raddon’s $2,101 annual profit at BCG’s observed primary-relationship attrition.
Credit union per-loan production economics. Section 6 applies MBA’s Q2 2026 revenue basis points to the 2025 HMDA average credit union origination and applies a 20 percent discount to MBA’s per-loan production expense to credit lower marketing and commission structures. The 20 percent discount is an assumption, not a published figure.
Mortgage servicing right value. Approximately $2,358 per loan, calculated at a 25 basis point servicing fee on a $198,167 balance at MCT’s July 2026 bulk multiple range.
Retained servicing net income. Approximately $314 per loan per year, calculated as a 25 basis point fee on an approximately $185,000 average balance ($462 gross) less approximately $108 in sub-servicing cost and approximately $40 in oversight.
Escrow float. $91 to $102 per loan per year, calculated on an approximately $3,100 average escrow balance at a 3.25 percent funds transfer price credit against approximately 0.30 percent blended interest paid.
Interchange per account. Debit interchange of approximately $68 to $80 per checking account per year and $61 to $68 per total member, derived from NCUA aggregates.
Every spread and float figure in this article assumes a 3.25 percent funds transfer price credit and a 1.90 percent blended cost of funds, consistent with the August 2026 environment of a 3.75 percent upper fed funds target and SOFR at 3.66 percent. At a 2.00 percent funds transfer price, every spread figure falls by roughly 38 percent.
Raddon’s publicly available household tier dollar figures date from 2016 to 2021 and predate the current rate cycle. The tier boundaries are structural and stable; the dollar amounts are not. StrategyCorps’ checking relationship data is 2016 vintage and has not been publicly restated. Scotsman Guide market share figures reflect 2024 data. Where these sources are used, the directional finding is the load-bearing claim, not the precise dollar amount.
The commonly quoted sub-servicing price of $6 to $10 per loan per month for performing loans is industry convention with no public source and is not used in this article. Credit union specific overdraft revenue is not cited because major estimates disagree by roughly a factor of five and NCUA discontinued collection of overdraft and NSF data in March 2025. Title insurance revenue per loan is not available publicly because it typically flows through a credit union service organization rather than the 5300 call report. Per-account revenue data for youth and club accounts does not exist in any public source. Raddon and Curinos values are held behind subscription; findings are cited without links.
To become a member 888-900-8409 Opt. 1
Already a member? 888-900-8409 Opt. 2
contactus@hrcuso.com
4210, 50th Street, Suite C
Lubbock, TX, 79413
NCUA CUSO Registry
#111-0007899
Have Any Questions?