Every quarter, thousands of closed loans get pulled for post-closing quality-control review, and the results keep pointing the same direction. Not appraisals. Not assets. Not closing docs. The category that keeps finding the top of the critical-defect table — first place in all but two quarters since late 2024 — is the one this industry considers its bread and butter: income and employment.1,2
We've written about where income review costs hours — the income calculation bottleneck. This brief is about where it costs money: the defects that survive underwriting, surface in QC or an agency audit, and turn into remediation, insurability problems, or a repurchase demand averaging north of $32,000 per loan.3
The defect scoreboard: what post-closing QC actually finds
The cleanest industry-wide window into loan quality is ACES' quarterly benchmarking of lender post-closing QC audits. Across calendar 2025 the critical defect rate averaged 1.50% — statistically flat against 2024's 1.52%, and roughly one defective file in every 67 reviewed.2 The headline number hides the drama underneath: quarter to quarter, the mix of what goes wrong swings hard. Step through 2025 below.
Eligibility findings were also elevated (15.87%) as borrowers stretched to qualify in a tight affordability market.
Income sits at the top of the table — again
In Q3 2025, income/employment findings spiked 47.6% in a single quarter — from 18.45% to 27.24% of all critical defects — making it the largest category by a wide margin.1 Q4 improved: compliance overtook it at the top, which ACES notes was only the second time income has ceded first place since Q4 2024 — and it still held over a fifth of all critical findings.2 And this isn't one vendor's dataset talking: Fannie Mae's own quality-control reporting lists DTI-related issues — missing documentation, incorrect income calculation, improper payment calculation — among its top defect drivers, with self-employment and rental income historically the most defect-prone calculations in its random QC sample.5,7
"The increase was driven primarily by concentrated deterioration in income- and compliance-related findings, reinforcing the importance of documentation integrity."
— Nick Volpe, EVP, ACES Quality Management, on the Q3 2025 results1
The takeaway
Income defects aren't an edge case. In post-closing QC benchmarking, income/employment has been the No. 1 critical-defect category in all but two quarters since late 2024 — and it's the same calculation your team performs on every file.
What one defect costs when it's caught late
A defect caught in underwriting costs a condition. A defect caught in post-closing QC costs remediation. A defect caught by the agency costs real money: a study of GSE repurchase activity by STRATMOR and Reggora put the average cost of a buyback at more than $32,000, with typical instances near $40,000, at an average repurchase rate of 49 basis points — and found income- and appraisal-related issues drove 57% of the buybacks in the period.3 Slide your annual volume to see what that exposure looks like at your shop.
That's before the costs no study captures: the underwriter hours burned on remediation, the agency scrutiny that follows a pattern of findings, and the investor relationships that quietly reprice your paper. Freddie Mac's data points the other direction too — loans run through its digital tools are on average four times less likely to carry defects.4 Quality has become the measurable differentiator.
Anatomy of an income defect
Fannie Mae's Quality Insider series is unusually candid about where income and DTI defects come from. The pattern isn't exotic fraud — it's ordinary files failing quiet checks. Five failure modes recur throughout its reporting:5,6,7
The missing document
The calculation was fine — the file can't prove it. A paystub short of the required period, a missing schedule, an expired VOE. Lack of documentation has become a leading defect driver in Fannie Mae's QC results.
The wrong method
Incorrect income calculation — the top of Fannie's DTI defect list. Semi-monthly booked as biweekly, overtime annualized without a trend check, a bonus averaged over the wrong window.
The debt that didn't make it
Improper monthly payment calculation on the other side of the DTI — a liability missed, misread, or entered once and never reconciled against the credit report and statements.
The rental that wasn't
Rental income and rental-loss miscalculation is a perennial top finding in random QC reviews — Schedule E, Fair Rental Days, and lease treatment reconciled wrong or not at all.
The self-employed maze
Complex calculations from self-employment have historically topped Fannie Mae's random QC sample — returns, add-backs, distributions, and business trend judged across documents that rarely agree on the first pass.
Notice what most of these have in common: they aren't judgment failures — they're reconciliation failures. A figure that didn't get traced to a document; a document that didn't get checked against its siblings. Even the genuinely judgment-heavy cases, like self-employed stability analysis, begin with documents that were never fully reconciled — which is exactly why so much of this is preventable in-flight and expensive after the fact.
The refi stress test
The refi comeback offered a preview of what happens to quality when volume returns. Year over year, refinances nearly doubled their share of QC reviews in CY 2025 — from 11.14% to 21.04% — but their share of critical defects more than doubled, from 15.30% to 32.20%.2 Defects concentrate wherever teams move fastest on files they assume are easy. FHA files told a similar story, carrying about 31% of critical defects — persistently above their share of reviews.1,2
The lesson for 2026 planning is uncomfortable but useful: your defect rate is not a fixed property of your team. It's a property of your throughput. Any process that only stays accurate at low volume isn't accurate — it's idle.
First-pass yield: the fix is upstream
Post-closing QC is a smoke detector, not a sprinkler — by the time it finds the income defect, the loan is closed, sold, and priced. The economic answer is to move the same checks upstream, where a defect is still just a condition. That's a workflow argument, not a staffing argument:
- Make the documentation set provable, not assumed. Completeness checks by income type — before the calculation, not after the finding.
- Calculate by the agency method, visibly. Show the method, the window, and the trend test next to the number, so QC re-derives the same answer you did.
- Reconcile every figure across the file. Paystub vs. W-2 vs. transcript vs. VOE — the five failure modes above are all cross-document misses. It's the same discipline we described in the income calculation bottleneck, now with a QC price tag attached.
- Flag contradictions while the borrower is still responsive. A discrepancy costs an email in processing and a condition in underwriting; let it reach a completed buyback and the study-average cost is $32,288.3
- Audit your prevention like the GSEs audit your production. Track first-pass yield and defect mix monthly — the same categories ACES benchmarks — and ask every tool to demonstrate measurable defect reduction, the way Freddie Mac's data does for its own digital offerings.4
This is also where AI belongs in the quality story — not deciding loans, but reading every page, tracing every figure, and refusing to let two documents disagree in silence. A cited, auditable income calculation survives QC by construction — and documented, auditable AI use is exactly what the new GSE AI-governance frameworks now require lenders to be able to demonstrate.
Frequently asked questions
Sources
Defect statistics are drawn from the published benchmarking and agency materials below. The repurchase-exposure figures in this article are an illustrative model built on the cited study averages — your shop's rate varies with product mix, channels, and QC rigor.
- ACES Quality Management — Q3 2025 Mortgage QC Industry Trends ReportCritical defect rate 1.79% (up 18.5% from 1.51%); income/employment findings 27.24% of critical defects (up 47.6% from 18.45%); compliance 18.97%; FHA 31.10% of defects; Nick Volpe commentary.
- ACES Quality Management — Q4 & CY 2025 Mortgage QC Industry Trends ReportQ4 rate 1.38% (annual low); CY 2025 average 1.50% vs. 1.52% in 2024; income/employment 21.52% in Q4 — off the top spot for only the second time since Q4 2024; refi review share 11.14%→21.04% and refi defect share 15.30%→32.20% year over year (CY 2024→CY 2025); eligibility +291.6% YoY.
- National Mortgage News — The leading drivers and costs of GSE repurchase requests (STRATMOR / Reggora study)Average repurchase cost $32,288 (typical instances near $40,000); average repurchase rate 49 bps (Apr 2023–Oct 2024); income- and appraisal-related issues drove 57% of buybacks.
- Freddie Mac — How Freddie Mac Is Powering Efficiency and Cost SavingsLoans using Freddie Mac digital tools are on average 4× less likely to carry defects.
- Fannie Mae — Quality Insider: Understand Top Defects to Help Strengthen Loan Quality (Sept 2025)Top defect reasons relate to DTI: lack of documentation, incorrect income calculation, and improper monthly payment calculation.
- Fannie Mae — Quality Insider: Minimize Defects With Fannie Mae's Income Calculator (April 2025)Complex self-employment calculations are historically top defects in Fannie Mae's random QC sample; the free Income Calculator targets exactly these errors.
- Fannie Mae — Quality Insider: Solving Rental Income Challenges (Aug 2025)Miscalculation of rental income and rental loss continues to be a top issue in random loan-quality reviews; documentation gaps are a growing driver.
- Fannie Mae — Selling Guide, B3-3 Income AssessmentThe agency methods the calculations must follow: pay frequency, YTD trend analysis, minimum history, declining-income treatment, rental and self-employed requirements.