Two pricing decisions separated the lenders that hit their return-on-equity targets in 2023 and 2024 from the ones that struggled to break even. Both still apply.
After two of the leanest years on record in 2023 and 2024, one pattern is hard to miss. The mortgage lenders that consistently make money are not just the ones running circles around capital markets. They are the ones being deliberate in two places at once: how they price to borrowers and who they sell to in the secondary market, with an honest view of what those decisions are worth two, five and 10 years from now.
That may sound obvious, but in practice, it’s not what most shops do, mostly out of habit. In richer-margin cycles, the consequences of imprecise primary market pricing are blunted by stronger overall economics. That’s not the cycle the industry is in today.
The Mortgage Bankers Association’s first-quarter 2026 Quarterly Mortgage Bankers Performance Report shows independent mortgage banks earned a pre-tax production profit of $727 per loan, with average revenue of $12,626, against all-in production costs of $11,898. That 16-basis-point margin is less than half the long-term industry average of 39 basis points, which goes back to 2008. In an environment that thin, every pricing and execution decision either protects that $727 or erodes it. There is little room for imprecision.
Primary market pricing is where most of the profit leaks
A common practice across the industry is to pull a dozen competitor rate sheets, apply a fixed basis-point margin and load it into the pricing engine, which then shows the loan officer the highest price. It is easy and feels disciplined. It is neither.
Consider a 250-basis-point margin applied uniformly across the book. On a $200,000 loan, that produces $5,000 in revenue. Against an $11,898 manufacturing cost, the lender has closed that loan at nearly a $7,000 loss. Even with a smaller-dollar commission, it’s still a significant loss.
On an $800,000 loan, the same 250-basis-point margin generates $20,000 in revenue, but a competitor that has sharpened its pencil for that size of deal will win it instead. The engine fills the pipeline with low-balance loans that fail to cover their costs and cedes the high-balance deals to others. Multiplied across a year of originations, the pattern explains a significant share of broken P&Ls.
The above examples illustrate a mindset issue. Basis points drive pricing, loan sales and commissions, so most lenders use that metric to measure margins. However, what actually shows up on the P&L is dollars, not basis points. Shifting the internal conversation from basis points per loan to dollars per loan is where a thoughtful pricing strategy begins, and from there, the fix becomes straightforward: pricing by loan size, segment and borrower economics, rather than defaulting to a stack of competitor rate sheets.
Across Vice Capital Markets’ client base in 2023 and 2024, the single biggest differentiator between lenders that hit their return-on-equity targets and those that struggled to break even was not volume. It was whether they priced thoughtfully to the primary market or simply mirrored competitor rate sheets. The pricing models Vice Capital Markets provides to its hedging clients are part of why that gap appears so clearly in the data, but the underlying point applies regardless of the tool used: outsourcing primary pricing policy is the same as outsourcing margin.
The long view in the secondary market
A parallel issue lives on the secondary side. Every shop is hunting for basis points on the bid, and most are hunting them on a single-day horizon. The buyer paying up today, though, is often the one with the strongest recapture engine, which means today’s basis-point win often becomes tomorrow’s forfeited refinance. That is where the second profitability leak hides.
ICE Mortgage Technology’s May 2026 Mortgage Monitor reports that the industry-wide servicer retention rate for refinances was 32% in the first quarter of 2026, down from 35% in the fourth quarter of 2025, even as refinance volumes rose. That average sits inside a much wider range when measured shop by shop. Some of the largest aggregators that actively market to their portfolios operate in the 60%-70%+ range. Many large depositories sit in the mid-teens. Some independent mortgage banks recapture only 2% to 3% of their borrowers. This spread is visible in the public company disclosures tracked by the Mortgage Bankers Association.
From the seller’s perspective, the buyer’s recapture profile is silently embedded in the price being offered. The strongest aggregators bid more aggressively because they expect to win the borrower back. Selling to one of them, winning two basis points on today’s cover bid and forfeiting the next refinance (along with the one after that and the referral the borrower might have provided) is not best execution. It is short-run optimization mistaken for discipline.
The same dollars-versus-basis-points lens belongs here. A basis point on today’s cover is a defined dollar amount. A lost borrower relationship, at a 65% recapture rate over the next decade, is a much larger one.
Most desks already apply this kind of math in another corner of their lives. Many lenders informally adjust an investor’s effective price by a few basis points to reflect the real cost of dealing with a buyer who is slow to purchase or constantly issuing stipulations. The same logic applies to recapture. Translating each buyer’s recapture profile into basis points of long-run revenue at risk and incorporating that figure into the best execution calculation meaningfully changes the picture. When that math is applied, the investor who “wins” on price by 2 basis points today at a 65% recapture rate becomes the worst bid on the seller’s desk, not the best.
The mega-servicer effect changes the math, retroactively
This consideration matters more now because the buyer landscape is consolidating. A wave of servicing-related mergers and acquisitions is concentrating the market into a smaller set of very high-recapture aggregators. When a high-recapture one acquires a mid-recapture servicer, borrowers previously sold to the former are suddenly being marketed by a sharper recapture operation than the original sale contemplated.
There is no clean way to predict acquisition activity on the day of a sale, but there is a reasonable way to handicap it. Assume more consolidation is coming and that the buyers most likely to be acquired are those with under-monetized servicing portfolios, and price execution decisions against that future, rather than against today’s bid alone.
Alternative structures are also worth understanding. Some buyers will pay less for the servicing asset in exchange for an agreement not to market to the borrower and, in some cases, to refer inbound inquiries from the borrower back to the originator. The trade is upfront dollars for downstream relationship protection. That arrangement is not appropriate for every lender, but for those that prefer not to service in-house and do not want to deliver their borrower to a recapture juggernaut, it merits serious consideration.
Sharp secondary execution still matters
None of this is an argument against running a sharp secondary market desk. Thoughtful pricing does not replace tight execution but instead compounds with it. Loan-level pull-through measurement, disciplined use of assignments of trade to compress bid-ask spreads and intraday repricing so that bids reflect market moves between issuance and acceptance remain the places where basis points are won or lost. The ViceEx(TM) platform from Vice Capital Markets exists to enforce exactly that kind of discipline.
The takeaway: the lenders that came through the lean years intact did so by treating the primary and secondary markets as two halves of the same strategy. Thoughtful primary market pricing protects the unit economics of every loan originated, while long-view secondary market execution preserves the lifetime value of every borrower acquired. Optimize for both, and that per-loan profit begins compounding into something worth defending. Optimize for one and ignore the other, and the next downturn will reveal the cost.
