# How to lower your QC defect rate and repurchase risk

Updated September 26, 2026

Measure the defect rate the way the Guide requires, trace each defect category to the step that produced it, fix that step, and run the same test at prefunding so the defect is caught before closing. Repurchase risk falls as fewer loans reach delivery with a significant defect, the only category that can lead to a repurchase demand.

## Key numbers

- **30 days** from the date of confirmation to self-report a loan ineligible as delivered (D1-1-01)
- **60 days** to file a written Fannie Mae appeal, unless the demand says otherwise (A2-3.2-03)
- **$500** cap on a de minimis correction, unless a higher amount is agreed (D2-1-04)

## Measure it the way the Guide does

You define your own severity levels, but Fannie Mae requires the highest one to be loans not eligible as delivered. Under D1-1-01 you set a target defect rate for at least that level, drawn from the post-closing random sample and "as reasonably low as possible." You measure it at least quarterly, review the target at least annually, and document the rationale. Fannie Mae "may provide input on a more appropriate target."

Reporting is at least monthly and in writing, with at least three months of trending and the highest-severity rate against target at least quarterly. D1-1-03 also requires reports to "use consistent methodology and terminology, including defect categories and severity levels." If you rename or regroup categories every few months, your trend line stops meaning anything.

When a trend shows up, you owe a written corrective action plan with the expected resolution and a time frame.

## Know which defects can cost you a loan

After a full-file review, Fannie Mae puts each defect into one of three buckets (D2-1-03). A finding doesn't change the price or make the loan unacceptable, so no remedy is required, though a data update may be. A price-adjusted loan (PAL) would have been eligible had the correct data been delivered and the loan-level price adjustment paid; you pay the adjustment, and Fannie Mae "may not demand repurchase of a PAL." A significant defect would have changed the price or made the loan "unacceptable for purchase" had the true facts been known. That's the one where Fannie Mae will require repurchase or may offer a repurchase alternative.

A significant defect has to relate to a listed area: the borrower's income, credit, liabilities or assets, property eligibility and the appraisal, loan terms such as LTV and occupancy, required insurance, or required loan documents. Freddie Mac uses the same three designations and a closely matching list (3401.1). Fannie's framework covers loans purchased or pooled on or after January 1, 2016.

Your "not eligible as delivered" level and the agency's "significant defect" are separate labels aimed at the same loans. Push one down and you push down exposure to the other.

## Trace each defect to where it entered

The Guide says what reporting is for: finding "loan-level issues and broad-based systemic, procedural, or operational issues, and address and remedy those issues to reduce the lender's defect rate" (D1-1-01). It prescribes no root-cause method.

Common practice is to tag each defect with its category, the origination step where it entered the file, and the channel or branch, then rank categories by highest-severity count. Loan-level findings have to go to the responsible business units, so every tag should name an owner.

## Move the test before closing

A defect cured before closing never reaches the post-closing sample or delivery. Under D1-2-01, prefunding results let you catch defects "such as analysis or calculation errors, inaccurate data, or inadequate documentation" before closing or acquisition, which "prevents the lender from selling ineligible loans to Fannie Mae." Prefunding reviews have to happen early enough for corrections to be made, with a monthly selection from every production channel.

Alongside the required full-file reviews, you can add component reviews aimed at one element, such as income or assets, "to confirm that a particular control or process is working as intended, such as newly added controls or processes." That's the natural way to prove a corrective action worked. Prefunding results are reported separately from the post-closing target rate.

## Cure what gets through, and report what you can't

A significant defect can be corrected during the appeal process, within limits. The correction has to show the defect didn't exist at purchase, or that it was corrected in the time and manner the Lender Contract specifies. Correction documents must be based on information available at the time of underwriting (no later than the note date), or that covers the time of underwriting (D2-1-04).

One form is a de minimis correction: an amount "not to exceed $500" (or a higher amount if agreed) refunded or otherwise provided to resolve the defect. It can't be used if it would leave a required minimum borrower contribution unmet, or to correct a Charter violation. Freddie Mac applies the same limits. Fannie Mae appeals must be filed in writing within 60 days of the demand unless the demand says otherwise (A2-3.2-03). Because cures depend on what the file held at underwriting, a defect your own QC finds early is easier to fix than one the agency finds later.

Self-reporting isn't optional. A loan with a defect that makes it ineligible as delivered must be reported within 30 days of the date of confirmation, which is the date the monthly post-closing report that includes it is published to management. It goes through Loan Quality Connect with a summary and supporting documents (D1-1-01). A repurchase alternative offered on a self-reported loan doesn't bring representation and warranty relief.

If a vendor does your QC, you stay "fully accountable" and fold its results into your own reporting and corrective action. Each month you must re-review at least 10% of the vendor's post-closing sample yourself, loans with and without defects, and report concurrence rates (D1-1-02). Low concurrence means your measured defect rate may be wrong.

## What this looks like over a quarter

The numbers are illustrative. A lender closes about 600 loans a month and samples 10% at random. Its documented highest-severity target is 2%, the lender's own choice rather than a Guide number.

In July, 6 of 60 random-sample loans are not eligible as delivered, a 10.0% rate: 3 income (variable income averaged over a shorter history than documented), 2 assets (large deposits not sourced) and 1 liabilities (a debt on the credit report left out of the DTI).

Income is the biggest category, and all three loans came from one channel using an outdated calculation worksheet. The written action plan retires the worksheet, retrains that channel's underwriters, and adds a prefunding income component review for 90 days, with a due date and an owner for each step. The 6 loans are self-reported within 30 days of the July report's publication.

| Month | Sample | Highest-severity defects | Rate | Income | Assets | Liabilities |
|---|---|---|---|---|---|---|
| July | 60 | 6 | 10.0% | 3 | 2 | 1 |
| August | 64 | 5 | 7.8% | 2 | 2 | 1 |
| September | 58 | 3 | 5.2% | 0 | 2 | 1 |
| Quarter | 182 | 14 | 7.7% | 5 | 6 | 3 |

The quarterly rate is 14 ÷ 182 = 7.7%, still above the 2% target, and the quarterly report says so. Income defects fell from 3 to 0. Assets didn't move, so asset sourcing is the next corrective action.

The lender remains responsible for its targets, corrective action and self-reporting.

## FAQ

**What defect rate should a lender target?**
Fannie Mae sets no number. The target must be "as reasonably low as possible," set at least for the highest severity level from the post-closing random sample, measured at least quarterly and reviewed at least annually (D1-1-01).

**Which QC defects can lead to a repurchase?**
Only significant defects. Findings need no remedy, and price-adjusted loans are settled with a price adjustment and can't be repurchased (D2-1-04).

**Do prefunding catches lower the post-closing defect rate?**
Not directly. The target is measured on the post-closing random sample. Prefunding catches reduce what reaches it.

## Sources

- Fannie Mae Selling Guide, published September 2, 2026: https://selling-guide.fanniemae.com/
- D1-1-01, Lender Quality Control Programs, Plans, and Processes (04/01/2026): https://selling-guide.fanniemae.com/sel/d1-1-01/lender-quality-control-programs-plans-and-processes
- D1-1-02, Lender Quality Control Staffing and Outsourcing of the Quality Control Process (04/01/2026): https://selling-guide.fanniemae.com/sel/d1-1-02/lender-quality-control-staffing-and-outsourcing-quality-control-process
- D1-1-03, Lender Quality Control Reporting (04/01/2026): https://selling-guide.fanniemae.com/sel/d1-1-03/lender-quality-control-reporting
- D1-2-01, Lender Prefunding Quality Control Review Process (04/01/2026): https://selling-guide.fanniemae.com/sel/d1-2-01/lender-prefunding-quality-control-review-process
- D2-1-03, Outcomes of Fannie Mae QC Reviews (11/03/2015): https://selling-guide.fanniemae.com/sel/d2-1-03/outcomes-fannie-mae-qc-reviews
- D2-1-04, Identifying and Remedying Origination Defects Under the Remedies Framework (08/30/2016): https://selling-guide.fanniemae.com/sel/d2-1-04/identifying-and-remedying-origination-defects-under-remedies-framework
- A2-3.2-03, Remedies Framework (08/30/2016): https://selling-guide.fanniemae.com/sel/a2-3.2-03/remedies-framework
- Freddie Mac Seller/Servicer Guide Section 3401.1, Postfunding quality control processes (effective 12/17/2025): https://guide.freddiemac.com/app/guide/section/3401.1

This guide summarizes agency requirements as of the date above. It is not legal advice.