New Episode: The U.S. Housing Supply Crisis: Homeownership Demand Challenges
Hi! I’m Pete Carroll, executive of public policy at CoreLogic, and in this final episode of the season, we’re going to talk about how technology like artificial intelligence, machine learning and more are paving the way for the future of mortgage underwriting—and how these innovations are working to identify new opportunities to create responsible, yet affordable mortgage financing.
Until recently it was very difficult for lenders to accumulate consistent and thorough data outside of standard metrics such as debt-to-income ratios and credit scores. But with new technology, lenders now have the ability to screen candidates based on a more holistic view of their payment behavior. These are more subtle, responsive and accurate for low-to-moderate income, or LMI, people.
One great example of this is looking at prospective borrowers’ rental and utility payment histories. This kind of data can augment traditional sources of data to provide an even more reliable picture of an LMI person’s ability and willingness to repay their mortgage loan.
Additionally, lenders are becoming increasingly savvy at electronically validating borrower financial information. Increasingly, lenders are relying on digital tools that can verify the accuracy and authenticity of a borrower’s financial information, as opposed to paper documents such as W-2s and paystubs.
For example, digital checking account and direct deposit information can provide a comprehensive picture of the prospective homeowner’s monthly income and expenses.
In addition, payroll databases can instantly verify a prospective homeowner’s employment status, and financial institutions can provide digital investment and other account statements that describe borrower assets.
All of these methods create a very comprehensive picture of a household’s financial statement, including their balance sheet, monthly cash flow, and monthly residual income after expenses. When compared to a cruder debt-to-income ratio, it is easy to see how sophisticated, electronically-enabled methods can be more responsive to more LMI homebuyers and facilitate their access to affordable mortgage loans.
We’re also continuously improving new mortgage products designed to meet the needs of LMI people. Mortgage products such as rehabilitation loans that allow LMI people to obtain mortgage financing and rehabilitate an existing structure, without having to come up with a large down payment can be a very effective tool.
Additionally, clever financing solutions that address “appraisal gaps” or scenarios where an independent appraiser values a home lower than the market sales price of the home will go a long way to facilitating access to affordable mortgage lending. Such solutions can include direct subsidies credit enhancements, or second mortgage loans from federal, state, and local agencies, mortgage lenders and investors, as well as other non-profit and private sector contributors.
Of course, with innovation comes risk. Algorithmic underwriting is only as unbiased as the people who design it. Public/private partnerships can work to rigorously test and evaluate new sources of data, artificial intelligence and machine learning models as they work to create novel mortgage underwriting methods. In doing so, they will ensure that we are accomplishing our goals of expanding access to affordable mortgage lending—and doing so responsibly and equitably.
This includes creating sand boxes where mortgage lenders can test novel approaches in partnership with their regulator and investor counterparts.
This testing is particularly important for a market that relies on a steady flow of capital to support $11T in mortgage loans outstanding. Formal testing processes validate transparency and predictability to the investors who supply this capital.
This testing also ensures that novel approaches are consistent with the spirit of important consumer protections and prudential requirements. This includes validating that affordable lending regulations and other policies that ensure equitable lending are being applied consistently, such as the Community Reinvestment Act, Fair Lending laws and the Affordable Housing Goals and Duty to Serve mandates of both Fannie Mae and Freddie Mac.
Like all big challenges, the Affordable Lending Gap is complicated, and there is no simple solution. It will take a combination of doing the hard work of building trust, utilizing existing lower down payment options, and exploring the technology that can paint a more accurate picture of someone’s reliability as a borrower to move the needle meaningfully in closing this gap—and getting people into the homes that will enable them to build wealth for years to come.
