Why Self-Employed Borrowers Still Struggle to Get a Mortgage – and What's Changing

Written by:

Chris Caggiano

Date:

03 Jul 2026

Share

Share via Facebook

Share via Linkedin

Share via X

Copy to clipboard

You earn a good living. Your bank account reflects it. But when you sit down with a mortgage lender, suddenly your income doesn’t quite add up – at least not on paper.

For self-employed borrowers, this is a familiar frustration. The mortgage system was built around a simple assumption: that income arrives in regular paychecks from a single employer and gets reported on a W-2. That assumption is increasingly out of step with how Americans actually work. Roughly 16.8 million Americans – about 10 percent of the total workforce – are self-employed today, and workers receiving a short-term W-2 or 1099 accounted for 27 percent of all jobs in 2024. Freelancers, independent contractors, small business owners, and gig workers don’t fit the traditional lending model. And the documentation requirements that follow can make qualifying for a home loan feel nearly impossible, even when the money is clearly there.

The mortgage industry has a name for loans that fall outside conventional qualification standards: non-qualified mortgages, or non-QM. The category has grown substantially as the workforce has shifted, from roughly $40–50 billion annually to an estimated $239 billion in 2025, or about 9 percent of the total U.S. mortgage market. That growth reflects real demand.

But the lenders processing those loans are still working through largely manual systems that weren’t designed for the complexity non-QM borrowers bring. For self-employed borrowers, that gap means longer waits, more paperwork, and – in some cases – loans that fall through entirely. This is the problem that Tradata Inc. is trying to solve. The San Francisco-based startup is building AI-powered document processing tools specifically for non-QM underwriters, with a focus on reducing the manual work involved in gathering, reading, and organizing loan files, without attempting to replace the human judgment at the center of the underwriting process.

The Document Bottleneck

Non-QM underwriting is fundamentally different from conventional mortgage processing. Where a standard qualified mortgage relies on straightforward income verification – W-2s, pay stubs, tax returns – non-QM loans require underwriters to construct a financial narrative from complex, often voluminous documentation. A self-employed borrower might submit thousands of pages of bank statements. A real estate investor’s cash flow may require careful parsing to determine what counts as qualifying income and what does not.

Danny Tang, Founder and CEO of Tradata, says the core inefficiency is clear: underwriters are paid for assessing loan risk, but the documentation demands of non-QM loans mean they routinely spend more than an hour per loan simply gathering and reading files before any actual underwriting begins. “Underwriters are paid for assessing the risk of the loan, not for reading documents,” Tang says.

That imbalance is at the core of what Tradata is addressing. Through interviews with working non-QM underwriters, the company found a consistent pattern: document handling, not decision-making, is where productivity is lost. The judgment calls that experienced underwriters are trained and paid to make can often be resolved in minutes. It’s the hours of document work surrounding those decisions that drives inefficiency across the industry.

Targeted Automation

The push to automate mortgage underwriting has been building for years, but non-QM presents a particular challenge. The complexity and volume of documentation involved makes it difficult to streamline without also displacing the experienced judgment that non-QM lending requires. A self-employed borrower’s financial picture rarely fits a template. The underwriter’s role is precisely to evaluate situations that don’t reduce to a formula.

That distinction matters because not all automation in this space is designed with it in mind. Many AI underwriting tools are positioned as end-to-end solutions – handling not just document processing but the approval decision itself. Tang sees that as a category error. “We are solving the document problem,” he says. “We want underwriters to make decisions faster, not disappear.”

The conditions for that more targeted approach have recently fallen into place. Document AI accuracy has improved to the point where it can function reliably in a regulated environment, a threshold that wasn’t consistently met even a few years ago. And as non-QM lending has grown, the cost of the document bottleneck has grown with it. An hour of underwriter time per loan is a manageable inefficiency at low volume. At the scale the market has reached, it’s a structural problem.

Hidden Costs and Widening Gaps

The cost pressures facing non-QM lenders extend beyond obvious line items. Industry figures suggest the cost per loan still exceeds $10,000 on average for many lenders, even after some improvement from peak levels. Tang points to two underappreciated cost drivers in the non-QM space. The first, again, is time. An hour or more of underwriter time per loan, multiplied across growing loan volume, adds up quickly.

The second cost driver is less obvious: the rules governing who will buy the loan. Non-QM loans are typically sold to institutional investors on the secondary market, and each investor has its own standards for what constitutes an acceptable loan file. If an underwriter misses a documentation requirement or fails to present the right narrative, the investor may refuse to purchase the loan or require a buyback after the fact. The financial exposure from buyback risk can be substantial and is not always visible in standard lender dashboards.

The gap between large and small lenders is also becoming more pronounced. Major mortgage operations have the resources to build proprietary AI systems in-house, giving them a processing speed advantage that smaller independent lenders struggle to match. In wholesale lending, where brokers can route loans to whichever lender closes fastest, that speed gap translates directly into lost volume. A lender that cannot approve a loan within a soft deadline risks losing the deal entirely, as brokers simply move to a faster competitor. “The gap is widening, and we want to narrow it for small and medium lenders,” Tang says.

What’s Next

The growth of non-QM lending is not a temporary adjustment. It reflects a fundamental and ongoing shift in how Americans earn a living – away from traditional salaried employment and toward independent, contract, and gig-based work. As that shift continues, the demand for mortgage products that can accommodate non-traditional income will only increase.

For lenders, that creates both an opportunity and a pressure point. The volume is there. The question is whether the operational infrastructure can handle it. Manual document processing made sense when non-QM was a small corner of the market. At current scale – and with further growth likely – it becomes a ceiling. Lenders that can move loan files through the system accurately and efficiently will be better positioned to capture that volume. Those that can’t will find themselves losing business not because the borrowers aren’t qualified, but because the paperwork takes too long.

For self-employed borrowers, the implications are direct. The mortgage system is slowly catching up to the way they work. Whether it catches up fast enough to matter depends largely on whether the lenders serving them can clear the document bottleneck that has defined non-QM processing from the start.

About the Expert: Danny Tang is the Founder and CEO of Tradata Inc., a San Francisco-based startup building AI-powered document processing tools for non-qualified mortgage underwriters.

This article is intended for informational purposes only and does not constitute legal, financial, or investment advice. The views and opinions expressed herein reflect those of the individuals quoted and do not represent an endorsement of any company, product, or service mentioned. Readers should conduct their own due diligence and consult qualified professionals before making any investment decisions.