Article Summary
With 30-year rates near 7%, lenders cannot rely on refinance volume. RETR data on 17,438 originator moves shows brokers and non-delegated channels were the only net gainer in 2026, underscoring the need to choose and fund a defensible moat.
From the vantage of the independent mortgage bank, every moat on this list is a direct threat: Each one takes market share, erodes relationships with past customers and interferes with existing purchase relationships that the IMB spent years building.
A moat, by definition, protects one competitor by imposing a cost on another, and right now those costs land disproportionately on the IMB. So, the strategic task is really two tasks: know which moats defend your channel and which are aimed at it, then decide where you intend to stand. What the movement data actually says The best evidence of where the moats are is where the originators go. RETR tracked 17,438 originator moves from January 1 through September 9 of this year, involving more than 3,000 companies that gained people and more than 3,500 that lost them. The channel totals are revealing. Independent mortgage banks gained 7,485 originators and lost 7,582, a net loss of 97, with $75.7 billion of production arriving and $77.5 billion departing. Banks and credit unions gained 5,012 and lost 5,242, a net loss of 230 and $5.0 billion of volume. The broker and non-delegated channel gained 4,941 and lost 4,614, a net gain of 327 originators and $6.8 billion, the only channel with positive net movement in 2026. Net originator and volume movement by channel, January 1 through September 9, 2026. Source: RETR Data Intelligence Platform. The flow matrix tells you even more. Of the originators who left an IMB, 4,871 went to another IMB, carrying $52.8 billion in production. That is the IMB channel paying transition bonuses to trade its own people in a circle. Another 1,897 left IMBs for the broker channel with $17.4 billion, while 1,572 came back the other way with $10.6 billion. The broker channel is attracting the more productive IMB originator and returning the less productive one. Banks, by contrast, mostly kept their movers in-channel: 3,922 of the 5,242 bank originators who moved went to another bank. Channel flow matrix: originator count and production volume moving from origin channel (rows) to destination channel (columns). Source: RETR Data Intelligence Platform. Read positively, the data says three things. First, every channel is gaining thousands of originators, so every channel has a proposition that works for someone. Second, the companies at the top of the gainers list, CrossCountry, New American Funding, Guaranteed Rate, Huntington and NEXA among them, are winning recruits in a year of extraordinary churn, which means a clear proposition still wins. Top ten companies by originator headcount gained, January 1 through September 9, 2026. Source: RETR Data Intelligence Platform. And third, the broker channel’s net gain is not an accident. It is the market telling us that renting a manufacturing floor is now a moat. Six moats, six owners Consider each attacker we named in our first article, this time as the owner of a moat rather than as a threat.1. Wholesale Lenders
The wholesaler owns manufacturing productivity. UWM originated $39.7 billion in the second quarter with 40.5% of the wholesale channel, and NEXA, now with more than 4,000 originators after acquiring UMortgage, funds more than half its volume through correspondent channels. Their moat is offering an originator a full manufacturing floor without a single underwriter or back office employee on payroll. Their exposure is that they own neither the originator’s loyalty nor the customer. 2. Banks The bank owns the balance sheet and the relationship. Seven large banks grew mortgage volume 20.8% in the second quarter, with Wells Fargo up nearly 43%, and proposed capital rules would cut risk weights on low-LTV first liens to as low as 20%. A bank underwriting the whole household does not need the loan itself to earn a monoline return. Its exposure is commitment. The same movement data shows Synovus, Cadence and U.S. Bank shedding originators while Huntington, Pinnacle, Fifth Third and JPMorgan added them. The moat is real for the banks that decide to use it. 3. Builders The builder owns the purchase opportunity. Lennar captured 83% of its buyers’ mortgages last quarter, D.R. Horton 81%, and Green Brick moved from 53% to 66% in a year. Their exposure is that the moat is only as wide as the new home cycle. 4. Servicers The servicer owns the prior customer. Rocket now closes 54% of its refinances from loans it already services, and the Homebuyers Privacy Protection Act preserves trigger access for the current servicer. The exposure, as we showed last week, is permission without engagement leaks: industry retention of refinancing borrowers was just 32% in the first quarter. 5. Real Estate Platforms The real estate platform owns intent. Zillow Home Loans originated $2.2 billion in the second quarter, up 95%, and reports that its originators close roughly twice the industry average of purchase loans because pre-approval lives inside the home search. Compass closed its combination with Anywhere with a mortgage venture attached, and eXp is replacing one mortgage partner with another. The platforms’ exposure is that someone still must manufacture the loan. 6. MLOs The productive originator owns trust. The Realtor relationship travels in the originator’s phone, and 17,438 moves prove it travels freely. The exposure is that trust is a personal asset, and a personal asset needs a platform that makes it more valuable. What about IMBs? The integration moat is gone. That is why it is time to choose. The IMB integrator’s dilemma, and the way through For forty years the independent mortgage banker bundled all four functions: distribution, manufacturing, capital and customer. The bundle was the moat, because no one else combined them as well. That is what has changed. Each specialist above now takes one component with a scale or a subsidy the integrator cannot match, and none of them needs the whole bundle.Integration by itself is no longer defensible. Of every position in the ecosystem, the IMB that is average at all four functions is the most exposed.
Here the authors part ways, at least in emphasis. The moat framework above is Jim’s. Rick’s view is that the bundle is still the moat; what has changed is who holds it and how it is constructed. UWM and Rocket made the unbundling of the IMB possible by rebuilding the bundle themselves. Their technology delivers a vertically integrated experience comparable to an IMB’s, but at far greater scale, and the resulting cost advantage is passed through to the broker and non-delegated correspondent partner’s pricing. That pass-through is the vital competitive differentiator. The IMB holds the bundle vertically integrated under one licensing umbrella; Rocket and UWM hold the same bundle horizontally integrated across thousands of partner firms, an inherently more scalable structure that drives the delivered cost to originate lower every year. The boundaries of the bundle have been redrawn around the channel rather than the company, and that is why a bundle the IMB still legally owns no longer protects it. But the same data that describes the dilemma also describes the way through. On this, the authors agree entirely: The bundle was never really the moat. Excellence in a component was. The IMBs at the top of the gainers list are not winning because they integrate. They are winning because originators believe they will close more loans there than anywhere else. The IMBs that manufacture at about $6,000 per loan, against an industry average near $11,900, are not threatened by wholesale scale. They are wholesale scale. And IMBs collectively hold roughly 95% of Ginnie Mae issuance because they built a competence in government lending that banks chose not to keep. This is exactly the lesson of cost to produce from our second article. The performance gap between top- and bottom-quintile lenders was never about size. It was about the decision to manage cost as a strategy. Defensibility works the same way. There is a moat available to every channel and to every lender within it. The lender simply must identify it and commit to it. Identify and commit For an IMB, the moats are specific, and each is achievable. Own the manufacturing frontier. A lender at $6,000 per loan can defend against the wholesaler and, beyond that, become the manufacturing floor that brokers and non-delegated originators rent. The channel that is gaining originators needs partners. Own the permissioned customer. Retain servicing where capital allows, maintain the consent record and build the engagement that turns a 32% retention rate into Rocket’s 54%. Own a segment no one subsidizes. Government, renovation, manufactured housing and non-QM lending reward competence rather than a subsidy from elsewhere in the transaction, and the private credit platforms acquiring specialty lenders this month have noticed. Own originator productivity.If your originators produce more here than they could anywhere else because of your leads, your turn times and your recapture, they are not a rental. They are a moat, and the recruiting treadmill stops on its own.
Own a permissioned vertical. The Costco precedent we described last week required a trusted brand and deliberate consent, not a real estate brokerage. Affinity partnerships outside the housing transaction remain almost entirely unclaimed. No lender needs all five. Most need one or two, chosen deliberately and funded fully. That choice is the strategy. The CEO’s questions Three questions can be answered from your own systems before the next planning cycle. Which of your top twenty originators would produce as much next year if they left, and which would not? The second group is your moat. Make it larger. The first is a rental workforce. Where does your cost per funded loan sit against the price a wholesaler charges a broker for the same loan? If you are below it, you own manufacturing. If you are above it, that is the first moat to build. Which of the moats above is closest to your reach today, and what would it take to commit to it fully within twelve months? On October 1, at the HousingWire Mortgage Banking Summit in Dallas, these questions go before the industry in a live survey, with the results displayed as the room responds. This week’s entry is on the ballot: Which moat have you committed to, and which is your biggest exposure? Next week, article five turns to the lever that decides whether manufacturing can reach the frontier at all: where AI is changing labor and unit economics, how operational and customer data become assets and how to tell real productivity from another shiny object. This is the fourth article in a six-part HousingWire Mortgage Banking Summit series. Jim Deitch is the CEO and Founder of Teraverde. Dr. Rick Roque is the Sr. VP of Strategic Growth and M&A at NFM Lending and founder and Managing Director of Menlo Company. Originator movement data provided by RETR, the industry’s largest repository of housing and origination data. This column does not necessarily reflect the opinion of HousingWire’s editorial department and its owners. To contact the editor responsible for this piece: zeb@hwmedia.com. Find your moat. Register for the HousingWire Mortgage Banking Summit. Click Here Related Trump threatens trade halt if Fed does not cut ratesSeptember 4, 2026In “Mortgage Rates”Chase offers up to 25-bps mortgage rate discount through Oct. 4September 14, 2026In “Mortgage”Mortgage applications rise 0.8%, ARM share now at 8%September 2, 2026In “Mortgage”
