2026 MBA Compliance & Risk Management Conference Recap 

The sessions at the Mortgage Bankers Association’s Compliance & Risk Management Conference in Washington, D.C. this past month covered a narrowed federal reading of ECOA, increasingly coordinated state regulators, and fraud losses climbing at their fastest pace in years. The common thread was not necessarily new rules abounding, but that the structure of a product, a program, or a control carries new regulatory lens based on outlined administrative priorities.  

The Economic Outlook and a High-7% Forecast 

In the opening remarks, MBA President & CEO Bob Broeksmit described a sluggish housing market despite unemployment below 4.5%, with hiring characterized as “low hire, low fire.” The forecast of upward trending mortgage rates was driven by aspects of heavy debt financing, reduced foreign Treasury demand, and a Federal Reserve that is no longer buying bonds. As of October 1^st, mortgage rates for 30 year FRM were reported at an average 7.28% according to Freddie Mac. For the MBA, advocacy recommendations included lower GSE guarantee fees and LLPAs, lower FHA premiums, and lower credit report costs. The call from the pulpit was for those in power at these organizations to make concessions for the good of the consumer.  

RESPA Section 8: Payment Structure Matters as Much as the Label 

First a reminder: as outlined in the CFPB’s RESPA FAQs, an agreement to refer does not necessarily need to be written; a pattern of conduct can establish a disallowed referral relationship. What was once a compliant scenario could have a dramatic compliance decision change based on the tweaking of seemingly minor fact points. The conference sessions included RESPA example scenario testing that highlighted those very principles: 

• Gifts: A broadly distributed $20 calendar-and-pen lender-only branded set was permissible as a marketing initiative. However, a $150 year-end gift tied to customer satisfaction scores failed because only agents with closings could earn those scores, making the criteria a proxy for referrals. 

• Dual roles: Panelists nearly unanimously rejected an agent who obtained an MLO license and was paid per closed loan while originating only on the agent’s own transactions. A salaried, non-producing branch manager role was potentially defensible, but only if the person actually performs it. 

• Affiliates: the typically outlined consumer disclosures of the affiliated business relationship and “no required use” are necessary but are potentially not sufficient depending on the nature of the business arrangement. A title company affiliate whose sole employee forwards files to an outside vendor, with revenue only from parent-company referrals, strongly suggests a sham. A $3,000 credit for using three affiliates can work if the discount is real and use is not required. 

• Marketing Services Agreements (MSAs): structuring an agreement where compensation is partially adjusted on performance, such as the number of rate checks or prequalification requests, tracks referrals too closely; adjustments tied to scope, footprint, or independent fair market revaluation are more defensible. 

Regulation B: A Narrower Rule Is Not a Reason to Stop Testing 

In April of 2026, the CFPB removed the “effects test” under Reg B and narrowed discouragement coverage by outlining that that ECOA does not recognize disparate-impact liability. However, this federal change was cited in sessions as at times in direct conflict with state laws with private actions still remaining. The advice from the panel for institutions stuck between a rock & a hard place? Fair lending testing should still be a paramount priority and any recalibration of testing should be clearly documented with defensible business reasons.  

For Special Purpose Credit Programs (SPCPs), the use of protected characteristics is foundational to program structures. These types of programs drew additional commentary from panelists. If a financial institution changes an existing SPCP’s criteria to better match the recent Reg B changes, could those new criteria constitute a proxy for a protected characteristic? And would that garner scrutiny from the federal regulatory bodies? The ultimate risk decision lies with the financial institution, but panelists advised that lenders be purposeful with their business decisions and track those changes closely, as regulatory lookback periods can span years. 

State Supervision Trends & MBA Driven NJ Suit 

With changes at the CFPB firmly solidified at this point in the administration, readers may have heard premonitions in the recent past such as the expectation of “50 mini CFPBs” who would rise up through a reinvigorated state level enforcement. Panelists throughout various sessions reminded conference attendees that beyond state regulators that examine entities to identify risk of consumer harm, there is also oversight from state attorneys general (AGs), who often drive their investigations from complaints. Cybersecurity exams now expect more substantive risk assessments, current incident response plans, and documented third-party oversight. AI was framed as an amplifier of existing fair lending, UDAAP, and privacy obligations, with a Massachusetts AG student loan lender settlement over untested lending models as a key example. Another unique item on the state front: on September 3^rd, 2026, the Mortgage Bankers Association filed a lawsuit in the U.S. District Court for the District of New Jersey challenging the disparate impact rules adopted by the state. The assertion is that in the law’s current form, disparate impact may be established through national, state, or local statistics without a challenger needing to show disparity among an entity’s own applicants or customers. 

Risk QA Closing Super Session: Mortgage Fraud Is an Insider Problem 

The conference closed with the Risk QA Super Session, where I had the opportunity to represent ActiveComply on the panel to discuss emerging mortgage fraud statistics. Real estate fraud losses reported to the FBI reached $275.1 million in 2025, a 59% increase from $173 million in 2024, on 12,368 complaints. Cotality found that 1 in 129 mortgage applications showed signs of fraud risk in Q1 2026. We also saw the fraud-related cases below presented by state regulators at the AARMR (American Association of Residential Mortgage Regulators) conference in Seattle this year:  

• Falsified documents: An originator who worked for seven mortgage companies over three years created false loan documents. The RMLO license was revoked on April 29, 2025. 

• Unperformed credit repair (pending): An individual used a personal phone, 400+ text messages, and Apple Pay and Venmo accounts to collect roughly $11,540, promising 80 to 100-point score increases. 

• Likely AI-generated documents (pending): Eleven originators submitted fabricated tax returns, P&Ls, and bank statements on 81 closed loans totaling more than $23 million. Post-close QC caught it through IRS transcripts, identical P&L formatting, and verifications of deposit. 

None of these schemes stopped at origination. Each surfaced through a referral or a post-close review, at times, self-reported by the lending organization themselves.  

Key Recommendations 

• Audit Payment Criteria for Referral Proxies: Review gifts, MSAs, and co-marketing for criteria that stand in for referrals, and monitor the social posts where those arrangements become visible. 

• Document Fair Lending Recalibration: Any change to disparate impact testing, SPCP eligibility, or credit cutoffs needs a written rationale and approval trail. 

• Inventory State Requirements Separately: Map state CRA, servicing, and AI obligations, and regression-test that federal updates do not disable state configurations. 

• Treat Post-Close QC as a Fraud Control: Build transcript checks, verifications of deposit, and document forensics into QC, and extend monitoring to the communication channels where insider schemes start. 

Conclusion 

Institutions that read the federal changes as a reduction in risk may face exposure in state exams and private litigation. A gift tied to the wrong metric, a system update that silently removes a state setting, or a fabricated tax return that clears origination each produces a costly finding that documentation and monitoring could have prevented. These findings will impact future examinations not only at the state level, but also exams with future federal regulatory bodies should we see a pendulum swing in the next administration’s priorities. Compliance programs built on continuous monitoring and organized records will absorb that shift with far less disruption. 

Key Terms

• RESPA Section 8: The part of the Real Estate Settlement Procedures Act that prohibits giving or accepting anything of value in exchange for referrals of settlement service business, and prohibits splitting fees for services that weren’t actually performed.
• SPCP (Special Purpose Credit Program): A program permitted under Regulation B that lets a lender extend credit to a defined group of economically disadvantaged people, using criteria that would otherwise be restricted, under a written plan.
• MSA (Marketing Services Agreement): An agreement in which one party pays another for marketing services. Under RESPA, scrutiny focuses on whether payments reflect the actual services provided or track referrals.
• LLPA (Loan-Level Price Adjustment): A pricing adjustment applied to a loan sold to the GSEs, based on risk characteristics such as credit score and loan-to-value ratio.

Catch referral riskeariler.

Request Demo