Editor’s Note: This feature originally appeared in the February issue of DS News, out now.
In the early months of 2019, default servicing continues on much the same path it’s traveled for some time now, with all corners of the industry continuing to try and find ways to thrive in an ongoing low-volume environment. With the cost of servicing non-performing loans increasing, it behooves servicers, vendors, and their partners to explore their systems and processes in order to ensure they’re working in a smart, streamlined, and efficient manner.
“Fewer loans are going to full foreclosure, and therefore infrastructure costs previously set in place from the 2008 crisis are now being spread over fewer loans, making the per-loan cost of foreclosure much more significant,” said Steven Mowers, President, Claims Recovery Financial Services (CRFS).
However, Yvette Gilmore, Freddie Mac’s SVP, Servicer Relationship and Performance Management told DS News that evolving and innovative technology can help offset those costs. “We understand that, in servicing, we all need to innovate. We must keep up with constant innovation and stay ahead of the technological curve,” Gilmore said. With a dizzying array of new technologies continuing to evolve, businesses have plenty of new opportunities to get the most bang for their buck—but only if approach innovation in the right way.
For this month’s edition, DS News spoke to a collection of servicers, vendors, and government representatives to see what lessons they’re learning in the trenches, and what takeaways default servicing professionals can apply to their own day-to-day processes and longer-term plans.
MORE WITH LESS
In December, Altisource released a “State of the Servicer Industry” report based on their 2018 survey of 200 professionals working in the mortgage default servicing industry. The report touches on several important trends at play in the current servicing landscape—a trend toward consolidation, the importance of compliance management, the ways in which technology can improve efficiencies, but also the crucial nature of interpersonal communication and collaboration.
Kristen Estrella, VP, Division Operations, First American Mortgage Solutions, said that one way the industry is adapting is through service providers introducing multi-functional products and technology solutions that enable servicers to do more with less. For example, Estrella said, “Technology solutions now allow servicers to leverage prior title work, which can shorten timelines and lead to quicker decisions for borrowers, as well as potentially increase data integrity.”
For servicers looking to take ownership of these issues and improve their processes, there’s a host of financial technology in varying stages of development and applicability. Just one question: where to begin? “You have to invest to support the work that you’re doing, and the default side of servicing is the most complex, labor-intensive piece of the industry,” said Denis Brosnan, CEO, DIMONT. “It’s challenging to maintain the level of investment needed to handle that work with the volume levels that we have.” Brosnan added that these costs were also compounded by compliance issues.
Whether it’s a case of choosing single vendors who can provide end-to-end services or simply working to ensure that all points of contact with partners are as efficient as possible, one thread runs through nearly every corner of the default industry today: doing more with less. “Mortgage service companies need to choose a field services partner that has a proven record of providing quality results in a timely manner, and technology plays a key role in this,” said Alan Jaffa, CEO, Safeguard Properties. “Those field services companies that make a major investment in technologies like mobile, geolocation, and automation are more likely to deliver property data more quickly.”
Anuj Jain, Head of Loss Mitigation and Default Servicing, Chase Home Lending, told DS News that Chase is working toward several priorities in 2019, including “working alongside HUD and other agencies to make the retention/loss mitigation, property auction, and conveyance process easier.” Jain also emphasized the leveraging of technology such as artificial intelligence (AI), optical character recognition (OCR), and robotics to help reduce the need for manual work.
NEW TECH, NEW TOOLS
Technology offers the promise of new solutions to even the most stubborn problems. But all that shiny potential demands a measured, deliberate, and strategic approach in order to provide the best results. “Seeking a ‘technology solution’ is similar to developing a ‘better mousetrap,’” Mowers said. “Any new system is only as good as the integrity and structure of the underlying data that supports it. There is a more pressing need for the integration of transactional data and source documentation with the goal of an industry-accepted naming convention, standard, and interface.”
Once those pieces are in place, Mowers said, technology gains will become more effective. “From a technology and workflow perspective, we’re only beginning to scratch the surface of what we can do with data in the foreclosure and default process,” Estrella said. “Servicers who recognize the transformative power of data are actively engaging in different integrations and partnerships with software companies.” Estrella cited First American’s own Digital Gateway, which, she explained “enables servicers to flexibly access only the data assets and services they need through APIs.”
Steve Butler, General Manager and Founder, AI Foundry, predicts that 2019 will see one long-used technology headed toward increasing obsolescence: optical-character recognition (OCR), the conversion of written or typed characters into a digital medium via scanners or other means. “OCR was invented a century ago and is losing its utility in modern times because it can’t do anything intelligent with the text it scans,” Butler said. He anticipates that OCR will increasingly give way to AI technology that enables machines to ‘read and react’ to content. “This will enable a boom in ‘white collar automation’ where manual document processes (such as mortgage processing) are replaced by software-based robotics processing.”
Susan Connally, VP, Surveillance Operations, Radian, said she believes robotic process automation could be hugely helpful in improving the claim-filing process. She told DS News, “There is a push to try to automate that process as best we can in order to pull all of the supporting documents that need to be filed with the claim and automate that process.”
One common response among the industry players DS News spoke to for this piece is a need for more efficient ways to handle the volumes of data inherent in default servicing and its related practices—ways to track it, ways to keep it up-to-date, and ways to more effectively share and communicate it between different entities. Technologies such as blockchain, AI, machine learning (ML), and robotic assistants all have enormous potential to help, whether on the data--management front or in streamlining or automating interactions with borrowers or industry partners.
As Chase’s Jain told us, it all amounts to improved efficiencies and overall cost savings—in theory, at least.
“Intuitively, we can all see and understand the benefits, but servicers must be careful not to fall into a trap,” said Sean Ryan, CEO, Aspen Grove Solutions. “If you don’t invest in gathering the data and creating effective data models, your investment in AI/ML tools will be much more difficult and expensive and may not have the desired outcomes.”
Brosnan spotlighted another common industry buzzword in the fintech realm: robotic process automation (RPA). “That’s a fancy term for ‘better workflow,’” Brosnan said. “There’s not going to be an end-to-end workflow system that handles everything. What servicers need to do is to look to their vendors that have their own systems and enable those systems to work with each other.” That can be a challenge, but some industry players are already seeing key benefits from implementing RPA solutions.
“We recently took a process in our bankruptcy department that used to take a person about 30 minutes to do,” said Gagan Sharma, President & CEO, BSI Financial. “Through an RPA pilot, we were able to reduce that to about two minutes that effectively requires a human being just to do one click. It saves time, but it also improves compliance. We are minimizing the likelihood of human error in the process.”
Debbie Hoffman, CEO and Founder, Symmetry Blockchain Solutions, said that advances within other industries in the areas of AI and data aggregation are beginning to seep into the worlds of mortgage lending and servicing, but there is still a long way to go yet. “AI can be used to assist with the underwriting and loan-processing actions by extracting borrower information,” Hoffman said. “It can be used to verify income, insurance, and assets that would otherwise need to be manually checked.
The biggest issue in this area may be the inherent challenge of addressing bias in AI.” First American’s Estrella told DS News that “improvements in workflow software options to include and leverage data, title work, and new document generation functionalities are all innovations that will significantly impact servicing in the next few years, as the industry continues to move towards digital mortgages.”
“Machine learning, combined with RPA, could be transformative in our business,” Sharma said. “However, it will be a continuum and an evolution. There’s no one silver bullet.”
KEEPING CHANNELS OPEN
Navigating the modern default servicing landscape requires constant, efficient communication between one’s own team, other partner vendors, and the various government agencies with which the industry interacts, including the GSEs. Keeping one’s bearings in a twisting maze of regulatory requirements from federal, state, and local levels can be overwhelming.
Where technology has perhaps the greatest potential to assist the industry is in not just automating more processes, but in helping ensure that the massive amounts of data inherent in mortgage servicing can pass from one point of contact to another as seamlessly as possible. “Workflow and documentation technologies are not, in and of themselves, all that transformative,” Brosnan said. “What’s transformative is moving data between various systems so that everybody can interact. That continues to be a spotty and frustrating challenge for the industry.”
Lori Eshoo, President and CEO, National Tax Search, told DS News that she’s already seen robotic technology giving her company a huge advantage in its ability to keep track of tax payments and impending penalties for their clients. In the past, National Tax Search would have to buy current and delinquent tax files from the various respective states on a monthly basis in order to stay abreast of penalties.
Now, with robotic technology that allows the company to access digital versions of those files for states that have put them online, NTS can access that same data on a weekly or even a daily basis. However, that level of automation and technological advance is not consistent across the board, with some states jumping aboard the digitization bandwagon sooner than others. Nevertheless, taking advantage of automating these processes and data pulls as much as possible can result in significant cost savings.
Eshoo also emphasized the importance of building and maintaining automated program interfaces that allow external vendors such as NTS to interact with servicer’s systems to share and update data on both sides in a seamless manner. “That’s the future of updating servicing systems on a nightly basis,” Eshoo said, “being able to transfer data from our system to their system without having somebody in the middle.” Aspen Grove’s Sean Ryan told DS News that ensuring more positive outcomes for borrowers will require the facilitation of more “multidirectional interactions,” rather than focusing only on the loan itself.
Ryan explained that he has often seen instances where loss mitigation teams are not connected to the inspection/ preservation oversight teams, or there are different teams for different aspects of the default process. “If the loss mitigation team has no access to what is happening with the property at a critical point in time when a decision is required, how can they make the best-informed decision for the borrower?” Ryan asked. “What if the property has just flipped from occupied to vacant as evidenced by the most recent inspection? What if work is ordered on a property and the borrower reinstates the loan but may not actually be in the house? How does the repair work order get canceled to prevent leakage?”
Ryan said that these issues can be addressed with a connected property servicing platform, integrated in real time to loan servicers, vendors, and other systems used across default servicing. “The dollars to drive technology initiatives can be found in the savings to be realized by the investment itself,” Ryan said.
However, knowing the general shape of a solution doesn’t necessarily make it easy to implement. DIMONT’s Denis Brosnan told DS News that one roadblock is, ironically, that the traditional servicing systems already in place are good at what they do. “Their job is to account for a tremendous number of financial transactions across a huge volume of loans,” Brosnan said. “However, the legacy nature of servicers’ technology frameworks frustrates what we call the ‘ilities’—things like usability, extensibility, and interoperability. That’s a huge roadblock to improving efficiency in a complex supply chain.” The solution to many of these disconnected problems, DIMONT’s Brosnan suggests, is to take a holistic approach—to look at ways to both improve internal processes and experiment with technological solutions in a parallel manner.
Brosnan cites as an example work DIMONT has done to help streamline the FHA conveyance process. “By working with the adjacent vendors in the supply chain, we’ve come up with an end-to-end solution that combines many processes related to property preservation, hazard claims, and the FHA claims process,” Brosnan said. “It’s more than just technology. We need ways that we can leverage the commonality of data and try to be multiple parallel processing while we have the opportunity.” As servicers and service providers continue to innovate, so too do the GSEs.
Tracy Stephan, VP - Enterprise Innovation, Fannie Mae, pointed to Fannie Mae’s Servicing Marketplace as a key enabler of Fannie’s work toward building greater efficiency within the housing finance sector. “In the past, matchmaking between a seller and servicer has taken several months or more for a seller to identify a servicer, agree on the price, and finalize the transaction,” Stephan said. Servicing Marketplace was developed to help decrease that timeline, providing all parties with transparent pricing, a standardized process, and standardized data requirements.
CRFS’ Mowers told DS News that they have seen an abundance of what he called “avoidable errors” that occur during servicers’ loss mitigation actions—often early in the process and which aren’t identified until much later. “Once realized, the cost of these errors can include curtailment of advances or the rejection of the full claim amount within an audit,” Mowers said. “These ‘errors,’ which often result in monetary losses, become more difficult to identify as loan populations are transferred from servicer to servicer.”
Ferreting out these mistakes that are lying dormant in a servicer’s inventory can be an enormous challenge, and are something that won’t easily be solved by technological innovation alone. It requires an ongoing commitment to examining those internal processes to identify breakdowns and inefficiencies. Radian’s Susan Connally also hit upon this lack of visibility as an issue that stands out for her. “Many servicers have built robust exception reporting that is available but not necessarily shared,” Connally said. “Making sure you’re getting all of the information that you can so you have a complete picture of what the default activities and challenges of your specific portfolio look like is critical.”
Fannie Mae’s Stephan explained that one way the GSE is working to help servicers zero in on these problems sooner is through a strong focus on standardizing data and providing better integration through Application Program Interfaces (APIs). “This allows all mortgage industry participants to identify risks earlier in the loan lifecycle, and helps servicers better identify default risks,” Stephan said. This focus has included the launch of the Fannie Mae Developer Portal, which Stephan said “promises to streamline processes for lenders at a time when they need to improve efficiencies and cut costs within their businesses.”
Last year, Fannie also created The Exchange, a free online community designed to allow Fannie to collaborate “with housing innovators across the nation to develop and adopt breakthrough ideas and solutions.” Freddie Mac’s Gilmore told DS News that it was important to recognize that the landscape of default servicing is shifting fundamentally. “Many institutions no longer want to service default loans, resulting in our defaulted loans being serviced by a smaller group of customers, most of them being non-banks,” Gilmore said. How are these new players changing the day-to-day way that default servicing is handled? Gilmore said, “You have people who think differently. They’re faster, more agile, and easier to adapt to change.”
Gilmore said this change in the landscape is impacting how Freddie Mac pilots its initiatives. “Freddie Mac is focused on updating our processes and technology, based on direct feedback from our servicers, to provide cutting-edge solutions that are faster, simpler, and more user-friendly.” “The end goal is to be more thoughtful in how Freddie Mac works with servicers in order to implement new tech while disrupting existing processes as little as possible, and Freddie Mac is committed to making the overall mortgage experience easier, more cost effective, and more efficient. We are looking at ways to improve the mortgage servicing experience by leading the way through collaboration with our servicers,” Gilmore said. “The only way we can move the mortgage servicing industry forward is together.” As an example, Gilmore pointed to the area of title. “We knew that we had an issue in being able to market REO properties subsequent to a default, and a lot of it had to do with title,” Gilmore said.
As such, Freddie is now piloting a direct-title program. “We’re looking at procuring title differently, working with our servicers, so that we ensure a better outcome,” Gilmore said. “We’re on a collaborative journey with our servicers to improve their overall experience. Our servicing partners are critical to bringing about the positive change that we all will benefit from.”
THE NEXT STEPS
While technology is presenting ways for servicers and their partners to build better, stronger, and more efficient systems, those solutions will only succeed if the industry invests in that technology while also closely examining what other aspects of their processes need to be rethought as well. Aspen Grove’s Sean Ryan cautions that the pressure of day-to-day operations can often take priority over addressing underlying issues in order to reduce expense and increase efficiency. “Servicers and subservicers are challenged from an expense and overhead perspective, and asset owners struggle to ask the right questions and apply the appropriate pressure,” Ryan said. “Default servicing often operates on a postmortem model and tends to look at the problems that have already happened.
At Aspen, we implement solutions that ingest and normalize real-time data from various departments, vendors, and service providers on the property timeline to help manage issues in real time. I’d like to see those concepts take root in the larger industry as a whole, as it will help to change the business, drive efficiency, provide for better outcomes for all stakeholders, and enable implementation of data models that could lead to the next level of business understanding through artificial intelligence and machine learning.”
Ryan told DS News that one way to move in the right direction is for the industry to adopt standards around data modeling. “If we can increase the standardization of forms and processes across the industry, then data will be more useful, integrations will become easier, and we can cut down expenses.” “Communication is key to our efforts,” said Freddie Mac’s Gilmore. “Making sure that we all understand who the critical partners are along the entire chain—not just the servicers, but the vendors that they utilize and the different ways they obtain their data.” As Gilmore explained, “More and more, the servicers, the GSEs, and their vendor partners are ultimately pulling data from the same sources.
So, it’s really about understanding how our partners are managing their business and plugging into that, versus building something brand new that they may not need.” DIMONT’S Denis Bronsan also emphasized the critical nature of an ongoing dialogue between the various parts of the industry, whether directly on a day-to-day basis or via groups such as the National Mortgage Servicing Association or various industry events that bring different organizations together in one place. “The traditional ‘vendor-vendee’ mentality requires the vendee to say, ‘I don’t want to commit to certain contract terms or volume levels.’ I get that, but that’s illogical, and it’s not how other industries work,” Brosnan said. “Amazon and its vendors, they talk all the time about volumes and delivery and standards because they’re dependent upon each other. It’s time for our industry to recognize those interdependencies.”