Гинни Мэй запускает масштабную цифровую трансформацию для стандартизации данных и автоматизации ипотечного рынка

Ginnie Mae President Joe Gormley’s message to executives at this week’s Mortgage Industry Standards Maintenance Organization (MISMO) Fall Summit acknowledged his agency’s status as an AI juggernaut but focused more on its structural data underpinnings. In a fireside conversation with MISMO President Brian Vieaux in Reston, Virginia, Gormley unpacked a Ginnie Mae operating model he said would become increasingly digital, automated and organized around standardized loan-level data. If the pieces come together around a dataset built with a single source of truth, Gormley said the changes could reduce friction for issuers, make mortgage servicing rights (MSRs) more readily transferable and create new avenues for private capital to participate in the Ginnie Mae ecosystem. “Data is at the essence of our industry,” Gormley said. “Bad data is an expense.” That vulnerability served as the connective thread across several initiatives Gormley mapped out, giving lenders and servicers a clearer window into Ginnie Mae’s direction. Ginnie Mae occupies a distinct place in the government-backed mortgage ecosystem. Unlike Fannie Mae and Freddie Mac, it does not buy or originate mortgages. Instead, the government corporation within the U.S. Department of Housing and Urban Development (HUD) guarantees the timely payment of principal and interest on mortgage-backed securities composed primarily of federally insured or guaranteed mortgages, including Federal Housing Administration (FHA), Department of Veterans Affairs (VA) and U.S. Department of Agriculture (USDA) loans. That guarantee helps connect these loans, along with the lenders and servicers that originate and manage them, to global capital markets. That makes Ginnie Mae an important part of the plumbing behind government-backed housing finance. Its standards for issuers, pooling, servicing and collateral can affect how efficiently capital moves through a market that serves borrowers who often rely most on affordable mortgage credit. Smoother path for MBS As Gormley noted in Reston, agency mortgage-backed securities collectively represent the world’s second-largest fixed-income market after U.S. Treasurys. And a significant share of MBS investment comes from overseas. Better loan-level data, friction-free servicing transfers and broader avenues for private capital participation could change the mechanics, costs and efficiencies of moving government-backed mortgages from origination into the secondary market. The modernization push Gormley described in Reston has been building for some time. At the Mortgage Bankers Association‘s Secondary and Capital Markets Conference in May, Gormley said Ginnie Mae had been accelerating work on loan-level servicing transfers over the prior year. He identified the technology needed to track certifications loan by loan and to establish a uniform cutoff date as the key remaining hurdles. At the same time, he warned that some independent mortgage bank issuers carried “risk-layered” FHA portfolios that could prove harder to finance or sell under stress. Three months later, the MISMO discussion supplied something more — an emerging execution road map. Ginnie Mae is moving from the concept of loan-level transferability toward the data architecture required to make it work, while using better data, automation and surveillance to identify problems earlier. Liquidity and visibility make up the common denominator. The idea is to make servicing assets easier to track and transfer, make issuer risk easier to see, and make the machinery connecting government-backed mortgages to private capital less dependent on manual intervention. Standardization first The immediate problem Gormley wants to address is familiar: duplicate submissions, inconsistent information, manual reconciliation and the costly cleanup that follows bad data downstream. “Bad data just has this cascading effect,” he said. With better automation, Ginnie Mae could identify a problem as data arrives rather than discovering it later in post-pooling reconciliation. “If we can tell you right up front that we believe there’s an issue, I think we gain a lot of efficiency, save a lot of time and hopefully save a lot of expense as well.” For issuers, Gormley’s strategic pivot clarifies a clearer, more direct path forward. “More digital, more automation,” Gormley said. He expects data submissions to become “more or less automated,” with AI and other tools helping identify anomalies and giving issuers near-immediate feedback. Arriving at that point would make data quality less of an IT issue and more of an operating cost issue. Standardized, unified, data-powered information reduces handoffs, exceptions, reconciliation work and the time-consuming human effort required to fix preventable errors. Automate, but stay vigilant Meanwhile, AI is being deployed primarily as an operating efficiency tool. Ginnie Mae’s automated document engine flags anomalies in issuer data and transfers information between systems, freeing employees from manual data handling. Gormley stressed that government AI deployments operate within federal and HUD frameworks, including controls designed to prevent agency data from being used to train public models. “With the use of AI comes a lot of responsibility, particularly as a government agency,” Gormley said. Ginnie Mae evaluates potential uses under federal and HUD frameworks and within its internal processes. A non-negotiable requirement, Gormley noted, is to prevent sensitive data from flowing into publicly available models that could be used to train them. He added that AI adoption and AI governance are not separate workstreams but must advance together. Widen the Ginnie Mae ecosystem In May, Gormley said Ginnie Mae was accelerating its push toward loan-level servicing transfers but still faced technology hurdles around loan-level certification and a uniform cutoff date. Three months later, he supplied an execution update. The agency’s Collateral Verification Transformation (CVT) project is in design, Ginnie Mae is working toward loan-level tracking of ownership and payment history, and Gormley is pushing to complete the CVT design phase within about six months. Another initiative Gormley discussed is running alongside the CVT push: Ginnie Mae is examining whether its traditional, all-encompassing definition of an issuer still fits the market. Possibilities include separate designations for subservicers and investing participants. Gormley said he expects more work on both ideas over the next year and called achieving such a change an “important legacy item.” Together, those two initiatives clarify a go-forward strategic priority. Modernized data infrastructure can power smoother servicing transfers, which in turn can make it easier for various forms of private capital and specialized operators to participate. The near-term message: get ready One sign of the direction of travel is eNotes. Gormley said Ginnie Mae is on pace to reach roughly three times its 2025 eNote volume this year and believes adoption has reached a “tipping point.” But his broader point was that eNotes are only one piece of the opportunity. Servicing files, transfers and collateral tracking remain ripe for standardization and digitization. “The pace of change is not slowing,” Gormley said. “We’re not trying to catch up. We’re trying to be ready for it.”