This post was first published on my corporate blog
Financial institutions have a wealth of information available to them from consumers. Due to manual and antiquated models, residential lending processes so far have had several negative experiences for both the lender and the borrower. Banks are plagued with application limitations, transaction complexities and data collection and processing challenges.
The ‘one-size-fits-all’ loan application simply does not work anymore. The newly implemented and redesigned URLA (Uniform Residential Loan Application), aims to simplify, organize and streamline the entire consumer journey – from loan request, to the underwriting and approval process.
The new URLA document, implemented from July this year (2019), (with a mandatory use date of February 1, 2020), comes with a completely new layout and content. The all so familiar four-page document used for so long is now replaced by a system of five critical components that have been mixed and matched to correlate the requirements of a specific transaction. Digitization of traditional workflows along with the more user-friendly approach of the new residential loan model can solve many of the existing issues. Let us understand what the changes mean for banking compliance, as well as the opportunities they can create.
The need for this change in the data communication format
Fannie Mae and Freddie Mac understood the need for a critical change in the data communication format – relevance of the data collected. For today’s underwriting to be accurate while being predictive, new uniform data sets needed to be created. Hence, the GSEs have both removed and added information from the URLA.
How will AI help reshape mortgage lending
The promise of Artificial Intelligence (AI) or Machine Learning (ML) towards transforming the lending industry is now beginning to unfold as several key players across the consumer lending frontiers are already leveraging the technology to streamline processes, improve productivity, efficiency, borrowers experience, risk alignment and loan quality. Lending by itself is a data-rich environment, and the benefits are there for everyone to see.
For example, in its quarterly Mortgage Lender Sentiment Survey®, Fannie Mae’s ESR (Economic & Strategic Research) Group found that almost 63% of lenders are acquainted with AI/ML, but only about 27% have tried or used AI tools for mortgage functions. Even almost 3/5th of lenders (~58%) said that they expect to implement AI tools in two years.
In the coming months, the additional data provided by the new URLA document enables several new use cases for deployment of AI/ML tools. We will see AI and ML come to frequently influence the end-to-end borrower journey from shopping for homes to owning said homes. AI and ML can be leveraged to prioritize and score leads, match borrowers to assets, manage risks, predict propensity to close, convert and so much more!
To reduce the negative experiences, lenders can accurately detect and highlight anomalies for underwriters, and essentially predict borrower default propensities within servicing to either further engage or incentivize borrowers for timely payments.
AI/ML tools through enterprise tools like SAP or Microsoft are already opening up new lines of revenues for financial institutions in other segments like consumer lending. Servicers working on portfolio retention can use AI or ML to not only predict customer behaviour 3-6 months in advance, but also combine it with attractive offers or discounts that turns them into customers-for-life. Better use of relevant data also includes additional benefits like fewer losses from frauds, accurate automated valuation models or AVMs for the properties, and reduced credit losses through accurate EPD & general loan default predictions.
Preparation is key
The revised URLA is probably one of the single largest compliance changes in a long time as it creates an effective pipeline management system, which can be tailored to the lender’s required business outcomes.
Preparing is a key component and it must always begin with a strategic implementation plan, as any bank must first understand which areas of its functions will be impacted by this change. For instance, whether your existing loan origination system supports the dynamic version of the URLA or not, how you can use the voluntary information from borrowers to see its benefits early on. Regardless of where you stand, there are benefits of testing the waters early as this major redesign in the URLA can cause headaches later, if caught off-guard.
If lenders are interested in exploring AI and ML, we recommend starting in small areas where you can use legacy or historical data, can help implementation teams gain confidence with AI on a practical level to further inform decision-making/investment decisions.