Expert Takes: Strategic Outsourcing in an Uncertain Market

Expert Takes: Strategic Outsourcing in an Uncertain Market

By Lionel Urban, Founder and Division President, NavPros

Executive Summary

Community financial institutions have spent the past three years managing balance sheet risk. The next three will be defined by something harder to hedge: the ability to keep executing when the people with expertise who do the work are scarce, expensive, and difficult to replace. Staff retention remains a top internal risk for 68 percent of community banks, a finding that reframes workforce stability from a human resources concern into a board-level operational risk.¹

Nowhere is that exposure sharper than in residential mortgage lending, the most cyclical, most expertise-intensive, and most compliance-dependent product a community institution offers. It demands specialized underwriters, processors, closers, and compliance staff whose skills are not interchangeable with other departments. When volume falls, those salaries remain. When volume returns, the talent has often already left the industry.

The cost data makes the problem concrete. Depositories averaged $16,320 to originate a single retail loan in 2025 against $12,209 for independent mortgage banks, with fulfillment work accounting for only about 20 percent of that total for both groups.² The remainder sat in sales structures and corporate overhead. With over twenty-five years of benchmarking data, only four years have seen depositories cost more to operate than independents, and three of them were 2023, 2024, and 2025.²

2026 has added a second lesson. Nearly every forecaster entered 2026 expecting falling rates and a refinance recovery. Instead, the 30-year fixed-rate mortgage climbed to 7.40 percent in October, its highest level since November 2023, and stood at 7.32 percent on October 22, more than one percentage point above where it stood a year earlier. The Federal Reserve also raised its benchmark rate in September for the first time in more than three years.^3,4 As such, the recovery forecast has already been marked down. MBA’s October 2026 forecast now projects $2.099 trillion in total one-to-four-family mortgage originations for 2026, only about 7.5 percent above 2025 volume and well below the stronger rebound many lenders were planning around entering the year.¹³ Fannie Mae’s October housing forecast goes further, lowering expected 2026 single-family originations to $2.085 trillion, with fourth-quarter volume projected at $448 billion and the 30-year fixed rate expected to average 7.5 percent throughout 2027.¹⁴ For institutions carrying fixed mortgage staff, that difference matters. A market that grows modestly instead of sharply still requires the same compliance, processing, underwriting, closing, and technology infrastructure, but without the volume lift needed to dilute those costs.

This paper examines why home loan production costs have risen, explains how staffing volatility drives those costs, assesses what artificial intelligence does and does not change, and makes the case for a new home lending model: one with a variable-cost structure that shifts required expertise to specialized partners and allows financial institutions to focus on home loan promotion and balance sheet goals.

Operational Resilience Is a Staffing Question

Regulators and rating agencies have spent the past decade expanding the definition of operational resilience. It once meant business continuity and disaster recovery. Today, it means the demonstrated ability to deliver critical services through disruption, including disruption caused by the loss of key personnel and institutional knowledge.

Community bankers have ranked workforce issues among their highest internal concerns for years. Staff retention ranked as the third-highest internal risk in the 2023 CSBS Annual Survey of Community Banks, behind only cybersecurity and liquidity.⁵ It remained a leading internal risk in the 2024 survey, cited by 68 percent of respondents.¹ By 2025, cybersecurity and technology implementation costs had taken the top two internal slots, but the underlying labor pressure had not gone away; it had simply migrated into the cost line, with personnel expenses ranking as the second most impactful effect of inflation on community bank operations, behind only the cost of deposits.⁶

The 2025 survey captured the strategic consequence. The share of community banks that received and seriously considered an acquisition offer doubled between 2024 and 2025, rising to 12 percent, with inability to achieve economies of scale cited as the primary reason.⁷ Institutions are not selling because they lack customers or credit quality. They are selling because the fixed cost of operating a modern, compliant, technology-enabled institution has outgrown what a small balance sheet can support.

Compliance compounds the labor problem rather than sitting beside it. Community bankers reported that safety and soundness practices account for the largest share of compliance expense at 27 percent, followed by anti-money-laundering standards at 25 percent and consumer protection standards at 23 percent.⁷ Residential mortgage lending touches all three simultaneously, which is why it consumes compliance capacity out of proportion to the revenue it generates at small scale.

One community banking executive quoted in the 2025 survey described the environment as a tidal wave made up of margin pressure, efficiency challenges, technology, talent costs, and compliance.⁶ Four of those five are labor related. Resilience is mostly a staffing question wearing a strategic planning label.

A Historical Pattern: Mortgage Staffing Has Never Been Stable

Mortgage lending has always been a boom-or-bust cyclical business, and the historical employment record is clear. Employment at nonbank mortgage lenders peaked at 292,700 in June 2021 and fell to 183,100 by June 2024, a decline of roughly 37 percent that took staffing to its lowest level since 1997. Brokers lost about a third of their workforce over the same window, and together the two groups shed roughly 150,000 employees in three years.⁷

Depository mortgage divisions followed a similar curve, although at a much slower pace. The Mortgage Bankers Association tracked total mortgage employment falling to roughly 343,000 by May 2023, down 18 percent from the July 2021 peak of 420,000.⁹ Behind that aggregate sat a churn problem easy to overlook: in a single year, 14 percent of producing loan officers changed employers and nearly 12 percent left the industry entirely.⁹

This is not a one-cycle anomaly. The same pattern followed 2008, when the employment contraction in mortgage brokerage reached 57 percent, and the 2003 refinance wave before it.⁸ Rate cycles are not forecastable with precision, but the staffing response to them is entirely predictable: institutions hire into strength and cut into weakness, and both moves are expensive.

For a large independent mortgage bank, that volatility is survivable because scale and nimble hiring practices allow rapid expansion and contraction across thousands of employees. For a community bank or credit union running a mortgage department of five to twenty-five people, the same volatility is difficult to address. Losing a single underwriter is not simply a 25 percent capacity reduction. It can be the difference between a functioning program and a program that cannot clear its pipeline within a rate-lock period, and the program expertise that person carried is likely to dissipate. Many depositories responded to the post-2008 version of that problem by exiting residential lending altogether, and the market share consequences of that retreat are still felt today.

The 2026 Lesson: Planning Around a Forecast Is Not a Strategy

If the historical record establishes that mortgage staffing is volatile, the current year shows how quickly a reasonable plan can be overtaken by events. The consensus entering 2026 was close to unanimous. In September 2025, Fannie Mae projected rates ending 2026 at 5.9 percent, with single-family origination volume rising to $2.32 trillion and the refinance share climbing from 26 percent to 35 percent.¹⁰ The Mortgage Bankers Association forecast $2.2 trillion in total single-family originations, an 8 percent increase, with refinance volume up 9.2 percent.¹¹ Institutions reading those forecasts had every reason to plan for a recovery year and staff accordingly.

That is not what happened. Freddie Mac’s Primary Mortgage Market Survey put the 30-year fixed-rate mortgage at 7.40 percent on October 8, 2026, its highest reading since November 2023, and the rate remained at 7.32 percent on October 22, well above the 6.19 percent recorded a year earlier.³ Rates rose rather than fell, against renewed inflation pressure that drove the ten-year Treasury yield to multi-year highs, and in September the Federal Reserve raised its benchmark rate to a range of 3.75 to 4.00 percent, its first increase in more than three years.⁴ The MBA’s October 2026 forecast now expects the 30-year fixed to average 7.3 percent in the fourth quarter and to remain above 7 percent through 2027, a materially different picture from the one published twelve months earlier.¹²

The point is not that the forecasters were careless. It is that rate forecasts are probabilistic and staffing decisions are not. An institution that hired processors and underwriters in the first quarter on the strength of a refinance recovery is now carrying those salaries against volume that did not materialize, and will carry them through at least one more budget cycle before the decision can be unwound. An institution that cut staff in anticipation of a slower year faces the mirror-image problem, with turn times stretching past what a rate lock will tolerate.

This is the structural weakness of the fixed-cost model stated in real time. Exposing it does not require a crisis. It requires only that the future differ from the plan, which is the normal condition of the mortgage market rather than an exception to it. The executive question is not which forecast will prove correct. It is whether the institution’s cost structure can absorb being wrong in either direction without impairing service, margin, or examination readiness.

The Cost Inversion: Why Depositories Now Pay More

The clearest evidence that the traditional in-house operations model has come under strain is the cost to originate. Benchmarking through the MBA and STRATMOR Peer Group Roundtables program put depositories at $16,320 per retail loan in 2025, essentially flat against 2024 and about 4 percent better than the study high of $17,071 in 2023. Independent mortgage companies averaged $12,209, a 5 percent decline from 2024 and a 15 percent drop from their 2023 high of $14,381.² That gap of roughly $4,100 per loan exists despite depositories carrying substantially higher average loan balances, $489,451 against $342,037, which should work in their favor on a per-loan basis.²

The composition of those costs explains why. For both groups, fulfillment work, meaning processing, underwriting, and closing, accounted for only about 20 percent of the total. Sales expense consumed 42 percent of depository costs and corporate allocations another 38 percent.² The MBA analysts offered a direct explanation: depositories may be cautious about making staff adjustments in response to volume fluctuations, and corporate allocations and technology costs have escalated in recent years.² Depositories are more expensive to operate not because their people are less capable, but because their cost structure is fixed while their volume is not. Retention discipline, the right instinct both culturally and for customer service, becomes a financial liability when the revenue supporting those salaries falls by half in eighteen months.

The regulatory burden associated with financial institutions is also passed through to staffing expense. Today’s mortgage lending environment is highly regulated, and compliance training, monitoring, and reporting have become an increasing burden for bank and credit union home lending teams.

Any serious discussion of mortgage staffing in 2026 has to account for artificial intelligence, workflow automation, and an honest assessment of whether these solutions can be implemented and supported alongside competing IT objectives. The efficiency gains are real and measurable for lenders with the budgets and expertise typically afforded by scale. Freddie Mac found that lenders maximizing the digital capabilities of its automated underwriting system saved up to $1,700 per loan, roughly a 13 percent improvement over the prior year.¹⁵ Document classification, indexing, and data extraction, long the most labor-intensive and least differentiated work in a loan file, are among the clearest current applications of machine learning in lending operations.¹⁶The adoption picture, however, shows a wide gap between interest and implementation. In The Mortgage Collaborative’s June 2026 Pulse of the Network survey, which draws on independent mortgage banks, credit unions, and depositories, 83 percent of lenders said they were evaluating AI tools, but only 17 percent had deployed the technology in live production workflows, with trust the most frequently cited barrier.¹⁷ Reducing loan production costs ranked as the top operational priority for 86 percent of respondents. Yet three-quarters also reported that automated decisioning now consumes more compliance resources than any other area.¹⁷

That last finding is the part most often left out of the vendor conversation. AI does not remove regulatory obligation. It relocates it. Model validation, fair lending testing, adverse action reasoning, vendor oversight, and documentation of how a decision was reached all become new work, performed by exactly the specialized compliance and model risk personnel a community institution finds hardest to hire and retain. An institution that adopts automated decisioning without adding that oversight capacity has not reduced its risk. It has concentrated it.

There is a second constraint. Those savings accrue to lenders who have embedded the tools deeply into a production workflow, integrated them with a loan origination system, and trained staff to use them consistently. That level of integration is itself a capability requiring dedicated technologists and ongoing maintenance. Technology that requires scale to implement does not solve a problem created by a lack of scale.

AI is changing the composition of mortgage work rather than eliminating the need for a production organization. It compresses the clerical layer and raises the value of judgment, exception handling, and oversight. Notably, lenders in that same survey planning for growth said they intended to get there primarily through existing sales teams and recruiting experienced loan officers rather than through automation.¹⁷ The industry is not staffing down because of AI. It is trying to do more with the people it already has, which is the same capacity question this paper began with, arriving in a different form.

Ultimately, lenders that reduce costs through automation will have the budget to identify, implement, and support those tools across a large number of loans. Smaller, less nimble organizations will fall further behind as the foundation for new workflow practices is reserved for lenders that are well capitalized, nimble, and technology savvy.

The Competitive Consequence of Standing Still

While depositories have been managing cost and headcount, specialized lenders have been taking share. Independent mortgage banks now originate approximately 84 percent of all single-family mortgage loans in the United States, including roughly 90 percent of FHA loans, 95 percent of VA loans, and 75 percent of loans sold to the government-sponsored enterprises.¹⁸ That concentration did not happen because independents have better relationships. It happened because they built operating models designed specifically for home lending and a diverse range of programs, with the scale to absorb volume swings and the specialization to underwrite programs a generalist team cannot reasonably master.

Credit unions illustrate both the opportunity and the constraint. They originated just over one million mortgage loans in 2025, representing 14.7 percent of the U.S. market and approximately $212 billion in volume.¹⁹ Real estate lending now accounts for 57.3 percent of all credit union loans, with the industry loan-to-share ratio at 83.2 percent at year-end 2025.²⁰ First mortgage originations rose 27 percent in the first half of 2025 year over year.²¹ The demand is present and growing. The constraint is capacity.

This is the uncomfortable position many community institutions occupy. The relationship advantage is real. Local trust, deposit relationships, and knowledge of the borrower and the property are genuine assets national platforms cannot replicate. What is missing is the fulfillment engine, the product breadth, and the compliance infrastructure to convert that advantage into closed loans at a defensible cost. The risk of standing still is not dramatic failure. It is quiet erosion. A customer who finances a home elsewhere begins moving the rest of the relationship, and the institution rarely sees the decision point at which that happened.

Reframing the Build-or-Partner Decision

Bank and credit union executives evaluating their home lending strategy typically frame the choice as build, buy, or exit. Today that framing is incomplete because it treats fulfillment capacity, including staffing and the associated expertise, as something an institution must own in order to control. A modern and more useful frame separates home lending strategy into three distinct components. The first is balance sheet strategy. This includes defining the scope and pricing of programs that meet the financial institution’s risk and interest income goals. The second is customer-facing: branding, promotion, and business development. The third is the production function: processing, underwriting, closing, post-closing, compliance monitoring, and the technology stack connecting them.

Understanding this model is where a community institution’s competitive advantage lives. The first component is where the levers for financial results are defined. The second component allows the financial institution to stay focused on customer relationships and on building its role as the primary financial services provider. The third component can be outsourced, and this is where scale and economics dominate, where cost is fixed, where specialized talent is hardest to retain, and where regulatory burden compounds fastest. This means operational costs and liabilities can be restructured without disturbing the sales organization that drives revenue.²

Converting the loan production function from a fixed cost to a variable cost changes the institution’s risk profile in five measurable ways.

• Volume risk becomes proportional. Expense scales with loans closed rather than headcount hired, so a 40 percent drop in applications no longer produces a 40 percent drop in departmental profitability.
• Forecast risk is neutralized. The institution no longer needs to be right about next year’s rate path in order to be staffed correctly for it. That is the single most valuable property of a variable model in a year like 2026.
• Key-person risk is distributed. The departure of a single senior underwriter does not stall a pipeline, because underwriting capacity sits in a pool sized for a national book of business.
• Compliance risk is shared. Regulatory monitoring across a patchwork of state and federal requirements, including the model governance and fair lending testing automated decisioning now demands, is maintained by a partner whose full-time function is to maintain it.
• Technology risk is transferred. Point-of-sale, loan origination, closing technology, and the automation layered on them are licensed, integrated, and updated by the partner rather than capitalized and maintained internally.

None of this requires the institution to surrender the borrower relationship, the brand, or the credit decision. It requires only that it stop paying fixed prices for capacity it uses variably.

How NavPros Solves These Pain Points

NavPros was built to streamline community bank and credit union home lending. As a division of Land Home Financial Services, a top-fifty U.S. mortgage lender licensed in all fifty states with more than three decades of mortgage banking experience, NavPros supplies the modern production infrastructure while the institution retains everything that makes it competitive. NavPros supports portfolio and wholesale lending channels from origination through servicing as a private-label solution. This allows financial institutions to keep what is working for them and add on-demand experts and systems that complement their home lending teams.The model is best understood by analogy to a national credit card network. When a customer uses a card issued by a community bank, the bank’s name is on the card and the bank owns the relationship, the credit decision, and the economics. Behind that card sits a national network providing authorization, settlement, fraud infrastructure, and compliance at a scale no single institution could justify building. The customer never sees it, and that is precisely the point. NavPros operates the same way in home lending, cost-effectively and with superior service levels. Your institution remains the trusted name on the loan and the brand the borrower knows, while NavPros provides the backend expertise, fulfillment capacity, product access, technology, and compliance support that make the program work at scale.

Specific capabilities mapped to specific pain points

• Retention and key-person risk: On-demand access to experienced mortgage professionals across origination, processing, underwriting, closing, compliance, and secondary marketing. Capacity does not depend on any individual hire and does not disappear when that hire accepts another offer.
• Fixed-cost exposure and forecast risk: White-label fulfillment across the full loan lifecycle, from application through investor delivery, priced as a variable cost. Whether rates fall as predicted or rise as they did in 2026, capacity moves with volume rather than with a hiring plan written a year earlier.
• Product breadth without compliance risk: A wholesale product suite spanning conventional, FHA, VA, USDA, jumbo, construction and renovation, reverse, manufactured housing, HELOC, DSCR, bank statement, and non-QM programs. Your team does not need to become expert in sixty products in order to offer them.
• Technology and automation burden: A web-based platform covering point-of-sale through e-closing, with automated workflow and real-time executive reporting, deployed under your brand. Your borrowers get a modern online experience, and efficiency gains from automation arrive as part of the partnership rather than as a capital project with administrative burden.
• Compliance complexity: Established regulatory frameworks, ongoing monitoring, and audit-ready processes maintained across jurisdictions, so examination readiness does not depend on a single compliance officer’s bandwidth. Disclosures, closing documents, and compliance tests are implemented and managed with compliance warranties.
• Market visibility: Private-label rate sheet generation, borrower-facing online application, mortgage website content, and marketing collateral under your brand, closing the digital presence gap that sends local borrowers to national platforms.
• Executive oversight: Daily status reports provide insight into leads, active pipeline, and closed loans. Quarterly executive reviews translate portfolio performance, balance sheet risk, and service levels into plain-language reporting for leadership and the board.
• A full partnership deployment, from technology setup through staff training, can be completed in as little as thirty days. The real choice is not between a partnership and a hypothetical internal build. It is between a partnership that produces capacity this quarter and an internal build that, if history is any guide, will be fully staffed roughly one cycle after the volume it was designed to capture has already moved.

The same infrastructure supports portfolio lending for institutions that want to retain select loans on the balance sheet. Whether the objective is fee income, balance sheet growth, or both, the operating model is the same: local relationship in front, national-scale production behind.

The Strategic Choice

The question facing community financial institution leadership is not whether home lending matters. The financial metrics of community lenders, the relationship economics, and the competitive record all confirm that mortgage lending should be an anchor service. The question is which operating model allows the institution to sustain execution through the next cycle without betting the department on the availability of scarce specialists, the accuracy of a rate forecast, and the durability of a fixed cost structure in a variable revenue business.

Institutions that answer deliberately will be positioned to grow through the next rate cycle rather than retrench through it. Those that defer will most likely answer by default, through attrition, through a pipeline that cannot clear, or through an acquisition conversation they did not intend to have.

See What Is Actually Happening in Your Lending Footprint

Every Community Bank and Credit Union has untapped home lending opportunity inside its own market, and most are competing without a clear view of it. NavPros produces a complimentary Market Opportunity Report scoped specifically to your lending footprint. It shows which loan types are being originated in your markets, which competitors are capturing that volume, how they are winning it, which borrower segments are underserved, and where the realistic expansion opportunities sit. It is market intelligence that supports pricing committee decisions, product development, and board-level strategy rather than guesswork.

Schedule a time to walk through your Market Opportunity Report with our team. In under an hour, you will have a data-driven picture of your competitive position and the home lending opportunity already in front of you.

Contact NavPros at info@navpros.com or schedule a convenient time to go over your no-obligation Market Opportunity Report.

About the Author

Lionel Urban is a technology leader for mortgage banking and has over 100,000 hours of mortgage banking experience since 1987.  He serves as the President of Navigator Lending Solutions, a Division of Land Home Financial Services and supports lenders with mortgage Process Analysis, Technology Planning, Project Management and Advisory Board Services.

Lionel Urban – Inquiries@NavPros.com, 425.595.2012
President of Navigator Lending Solutions,
a Division of Land Home Financial Services, Inc.

Follow me on LinkedIn

Notes

1. Conference of State Bank Supervisors, 2024 Annual Survey of Community Banks (Washington, DC: Conference of State Bank Supervisors, 2024), https://www.csbs.org/survey.
2. Jenny Masoud and Marina Walsh, “Chart of the Week: Retail Production Channel–Cost to Originate a Loan,” MBA NewsLink, June 18, 2026.
3. Freddie Mac, Primary Mortgage Market Survey, weekly results for October 8 and October 22, 2026; Freddie Mac, “Mortgage Rates Decrease to 7.32%,” press release, October 22, 2026.
4. Erika Giovanetti, “Today’s Mortgage Rates Slightly Lower After Fed Hike,” U.S. News & World Report, September 18, 2026.
5. Joey Samowitz, Kathleen Navin, and Tom Siems, “2023 CSBS Annual Survey of Community Banks: Key Findings” (presentation, 11th Annual Community Banking Research Conference, October 4, 2023).
6. Joey Samowitz, Kathleen Navin, and Tom Siems, “2025 CSBS Annual Survey of Community Banks: Key Findings” (presentation, 13th Annual Community Banking Research Conference, October 7, 2025).
7. “CSBS Releases 2025 Annual Survey of Community Banks,” CommunityBanking.org, October 7, 2025.
8. Wolf Richter, “Housing Bubble & Bust #1 and #2 as Seen through Employment in Mortgage Lending,” Wolf Street, December 16, 2024, citing U.S. Bureau of Labor Statistics data.
9. Sarah Wolak, “An Exodus From The Industry,” National Mortgage Professional, August 28, 2023, quoting Joel Kan, Mortgage Bankers Association, and Dale Larson III, Modex.
10. Fannie Mae, “Mortgage Rates Expected to Move Below 6 Percent by End of 2026,” press release, September 23, 2025.
11. Mortgage Bankers Association, “MBA Forecast: Total Single-Family Mortgage Originations to Increase 8 Percent to $2.2 Trillion in 2026,” press release, October 19, 2025.
12. Mortgage Bankers Association, Mortgage Finance Forecast, October 11, 2026.
13. Mortgage Bankers Association, Mortgage Finance Forecast.
14. Fannie Mae Economic and Strategic Research Group, “Housing Forecast: October 2026” (Washington, DC: Fannie Mae, October 12, 2026).
15. Freddie Mac Single-Family, “2025 Updates to the Cost to Originate Study,” Freddie Mac Insights, November 18, 2025.
16. STRATMOR Group, 2025 Technology Insight Study (Greenwood Village, CO: STRATMOR Group, 2025).
17. The Mortgage Collaborative, June 2026 Pulse of the Network Survey, reported in “Survey Finds That Lenders Are Focused On Early-Stage AI Adoption And Reducing Cost-Per-Loan,” PROGRESS in Lending, July 20, 2026.
18. Community Home Lenders of America, CHLA 2025 Report on Independent Mortgage Banks (Arlington, VA: Community Home Lenders of America, January 2026), citing Urban Institute Housing Finance Policy Center data.
19. Polygon Research, “Credit Union Mortgage Market Intelligence,” HMDAVision 2025 HMDA Lender Rankings.
20. National Credit Union Administration, 2025 NCUA Annual Report (Alexandria, VA: National Credit Union Administration, April 1, 2026).
21. Steven Rick, Credit Union Trends Report, Q3 2025 (Madison, WI: TruStage, December 2025).