Most lenders and servicers likely may pay no mind to the work of the Mortgage Industry Standards Maintenance Organization’s contingent of dedicated volunteers. They don’t care how the technology works; they just want it to work. But the latest incarnation of MISMO standards, Version 3.1, and its predecessor Version 3.0—collectively called the Version 3.X architecture—represent a major shift in how disparate software products crucial to mortgage industry interact with each other.
And if industry participants aren’t paying attention to whether their software follows the Version 3.X standards, they risk missing out on an opportunity to inject transparency and business intelligence into virtually every aspect of their operation.
MISMO is a wholly owned subsidiary of the Mortgage Bankers Association and its volunteers come from more than 150 subscriber organizations. MISMO released the Version 3.1 Reference Model as a candidate recommendation in May. As a candidate recommended reference model, MISMO believes the specifications laid out in Version 3.1 can be implemented and used by the mortgage industry.
“To the best of our ability to tell within the several hundred people who have reviewed it, we believe it is a stable-enough environment to be implemented and used in production,” said Greg Alvord, the chairman of MISMO’s architecture committee and a senior data architect at RealEC Technologies, a subsidiary of Lender Processing Services.
MISMO’s volunteer corps is organized into workgroups, each focusing on a different component of the mortgage industry. In the Version 2.X of the MISMO architecture, each workgroup created a standard for how software products that serve a specific component of the industry transmit and receive data.
With each MISMO workgroup working semi-autonomously, software that needed to communicate with multiple types of transactions in the mortgage industry—like a loan origination system or servicing system of record—had to implement multiple instances of the same data transmission.
“We found, for example, 14 different definitions of property,” Alvord said.
If MISMO was to succeed in its mission of promoting an open standard for the entire industry, something clearly had to change. MISMO began the process of reconciling each workgroup’s differing data definitions into a cohesive system that meets the needs of all segments of the industry.
The imminent, real-world implications of reconciling the data standards is that software across all tasks in the mortgage industry will not only be able to communicate more efficiently, new technology and integrations between pieces of software can be introduced to lender and servicer users faster.
Harry Gardner is president of document preparation vendor SigniaDocs. He’s also a longtime MISMO volunteer, previous serving as president and currently as chair of the residential governance committee. He likens the Version 2.X architecture to a forest of data standards. Each workgroup had its own tree, with standards and definitions emanating from its branches and leaves.
“If you were trying to use Version 2, you had to understand the branches on each individual workgroup’s tree,” said Gardner. “What we did with Version 3 was to create this single uniform overarching reference model.
“So it’s one big tree and all the branches, no matter which workgroup is using a branch on that tree, have the same interrelationships,” he added.
Once the Version 3.X architecture is implemented by software developers, they can employ different workgroups’ transactions easier because the structure doesn’t differ from one workgroup to another.
For example, LOS vendors can build integrations to the software of third-party underwriting services with all sides using a common set of specifications.
That means existing integrations work better and new ones are easier to build. It also creates the opportunity for vendors to abandon the myriad proprietary standards previously used.
Many mortgage technology vendors have built proprietary data transmission standards to support integrations. It’s expensive and time-consuming and takes resources away from developing new innovations. It can also be a point of contention when two developers need to integrate their products and both want to use their own data standards.
Version 3.0 was released as a candidate recommendation reference model in December 2009 and was the first step in the reconciliation process of the MISMO standards. But the reference model was only a partial reconciliation of the MISMO standards because it did not encompass all of the residential workgroups.
The Version 3.0 reference model was developed on an accelerated pace to accommodate the government-sponsored enterprises’ work on the Uniform Mortgage Data Program, and focused primarily on the two workgroups whose standards support that effort. The flood workgroup was the only one not involved in the UMDP that made the release.
The Federal Housing Finance Agency announced the UMDP in May 2010. It is a sweeping overhaul of the way Fannie Mae and Freddie Mac accept electronic data on loan files. The program also includes a new initiative for the GSEs to collect full electronic appraisal reports on four commonly used forms, a first for the two mortgage giants.
In order for the loan delivery component of the UMDP to take advantage of the consistency designed for the Version 3.X architecture, MISMO focused on reconciling just three workgroups so work could begin on the Uniform Loan Delivery Dataset. The ULDD is the file format and set of definitions that makeup the new file that lenders will begin submitting for all loans sold to the GSEs. The remaining nine workgroups have reconciled their data standards to the new specifications in the Version 3.1 Reference Model.
“The original plan was to do it all at once,” Alvord said. “But because of this very important project requested by FHFA to Freddie and Fannie, we partitioned the work in order to have a publication that met their needs and a following publication that brought everyone else in.”
The 3.X architecture represents a major advance in the quest for data standards that are shared by the entire mortgage industry. For the individual industry participant, that opportunity will promote efficient internal communication, data accuracy and transparency in dealings between lenders, investors and servicers.
“In any previous format, whether it was a private one or in the Version 2 architectures, there was never enough transparency so that everybody could know what was going on,” Alvord said.
With software that follows the Version 3.1 standard, all participants can view verification and underwriting documents like credit reports and appraisals using a single model for data transmission.
“With the Version 3 model, the seller of the loan can communicate exactly what data they used in their analysis and the receiver—the potential buyer of the loan—can put that data in their own process and make their own estimate of value in order agree to purchase it or evaluate it as it’s being delivered,” Alvord said.
“We’re going from a reps-and-warrants environment to a trust-but-verify environment, where the data and facts that are necessary for doing your own evaluation, your own process as a purchaser of the asset are available for you to use and open up the transparency,” he added.
Those abilities are bolstered by the data quality the Version 3.X architecture introduces into the origination process.
“The ability to include the source data used to make a decision and it can be separately verified by some other process, is what improves data quality,” Alvord said. “Data quality comes from having agreed-upon definitions of what the data points mean and the ability to compare what it is supposed to be and what you want it to be, with what it should be.”
