Housing Order Directs FHFA Toward Proactive Mortgage-Insurance Cancellation

A White House order directs the Federal Housing Finance Agency toward a pilot that would periodically screen eligible Fannie Mae and Freddie Mac loans for proactive mortgage-insurance cancellation. For credit unions that originate or service enterprise-backed mortgages, the consequential work is not an immediate cancellation campaign. It is preparing valuation, payment-history, notice and exception controls for rules that FHFA and the enterprises still must establish.

The October 22 executive order says the pilot would apply only to mortgages made after the program is established. It calls for automated valuation models and other available data to identify borrowers who appear to satisfy equity and other cancellation requirements. It also calls for safety-and-soundness standards, payment-history requirements and prompt borrower notices when cancellation may be available.

That boundary is important: the order does not cancel mortgage insurance on existing loans, establish a launch date or define a universal equity threshold. FHFA first must determine that the pilot is in the interests of the enterprises, the agency and the public, and then direct Fannie Mae and Freddie Mac to establish procedures.

The valuation is a trigger, not the decision

An automated valuation can identify a loan for review, but the order keeps several other conditions in the decision path. Cancellation must be consistent with safety and soundness, enterprise contracts and all applicable statutory, regulatory and enterprise requirements. A credit union should therefore avoid treating a rising value estimate as an automatic member entitlement.

A controlled workflow should preserve the model and version used, valuation date, property-data inputs, confidence or exception indicators, payment-history check, investor and contractual eligibility, final decision and member communication. Low-confidence valuations, recent property changes, disputed data and loans with unclear contractual treatment need a defined human-review route.

The FHFA automated-valuation-model rule resources already frame AVM governance around confidence, data integrity, conflict avoidance, random-sample testing and nondiscrimination. The pilot's eventual operating instructions will determine how those controls attach to cancellation screening and servicer evidence.

Member communication needs a two-step design

The order asks servicers, where legally permitted and operationally feasible, to notify borrowers promptly when mortgage insurance may be canceled and to reduce unnecessary appraisals, documentation and administrative barriers. That creates a useful distinction between an eligibility signal and a completed cancellation.

The first communication should say that available data indicates the loan may qualify and explain any remaining review. The second should document the final decision, effective date and payment effect. If the loan does not qualify, the record should preserve the reason and the member's route to correct property or loan data. Marketing language should not imply that appreciation alone guarantees a lower payment.

The order also targets refinance title costs

A separate section directs FHFA toward expanding the enterprises' title-acceptance pilot into an ongoing nationwide program for eligible refinances. The order asks FHFA to consider higher loan-to-value eligibility, more title-assessment providers, waivers of certain documentation for payment-reducing limited cash-out refinances and lower enterprise fees. It also requires resulting savings to be passed through and disclosed to borrowers.

Those provisions are a roadmap, not a current product rule. Mortgage leaders should inventory which loans are sold or serviced for the enterprises, identify where AVM and title-risk decisions enter the workflow, and assign an owner to monitor FHFA and enterprise implementation. Existing borrower rights, contracts and procedures remain the baseline until official guides and system instructions change.

Our coverage of enterprise model choice shows why a data option must remain consistent through underwriting, pricing and delivery. The mortgage AI go-live tests provide a parallel control set for valuation, explanation, exceptions and post-launch monitoring.