The post-closing framework in plain language
The lender needs a documented process that selects loans every month, combines a representative random sample with risk-based discretionary reviews, completes the defined file work, reports results to management, and drives timely corrective action. The evidence must be strong enough for Fannie Mae to audit both the process and individual conclusions.
Random selection: ten percent or a valid statistical method
D1-3-01 allows the random component to use ten percent of the applicable production or a statistically valid sample designed at a 95% confidence level with a 2% precision rate. The method is not merely a target count: retain the population, channel classification, calculation inputs, selection process, and selected membership.
The current guide calls for separate random samples for retail and third-party-originated loans. That distinction should exist in the source population and remain visible in cycle reporting.
Discretionary selection: target the risks the random sample may miss
Discretionary reviews complement the random sample. Selection criteria should reflect the lender’s actual risk—new products, higher-risk characteristics, production sources, personnel, concentrations, unusual defects, fraud indicators, or other findings from monitoring. Keep the targeted reason attached to the selected loan.
Timing: manage one monthly control cycle
Fannie Mae’s current process is monthly and uses an overall 90-day cycle. Build due dates from the production month through selection, file review, reverification, reporting, management response, and corrective action. A dashboard is not enough; retain timestamps and evidence of completion.
File review and reverification
The review needs to address the accuracy and integrity of the underwriting and closing information applicable to the loan. Reverification is a separate evidence-producing process: record when a request was sent, what response was received, any variance, and how the reviewer resolved it. Do not treat an outbound request as a successful result.
Reporting, defect analysis, and corrective action
Management reporting should show the population and sample, completion status, defect types and severity, channel or source trends, root causes, and required action. Corrective action should name an owner, due date, expected evidence, and closure approval. Preserve the loan-level record behind every aggregate.
Controls to keep in the system
- Controlled QC plan version and effective date
- Frozen monthly population with channel classification
- Random method, parameters, seed, and selected membership
- Discretionary selection reason
- Document inventory and missing-evidence status
- Reverification requests and actual responses
- Finding source, rule, severity, response, and disposition
- Cycle counts, defect analysis, management review, and corrective action
Frequently asked questions
How much of production must be selected for Fannie Mae post-closing QC?+
The current guide permits a ten-percent random sample or a statistically valid random sample using a 95% confidence level and 2% precision rate, subject to the detailed requirements and separate channel treatment in D1-3-01.
Are discretionary reviews part of the random defect rate?+
Discretionary selections are targeted for additional risk and should remain identifiable from the random sample. Follow the current guide and your QC plan for reporting and trending.
How quickly must the review cycle be completed?+
Fannie Mae’s current guide establishes monthly selection and an overall 90-day QC cycle. Confirm the precise selection, review, reporting, and corrective-action timelines in the current D1-3 sections.
Primary sources
Requirements change. Confirm the current guide and your approved QC plan before changing a production process.