David Snitkof: 3 Ways to Use Technology to Reduce Bias in Mortgage Underwriting

David Snitkof is Vice President of Analytics with Ocrolus, New York, responsible for the company’s analytics product and business lines. He has been with Ocrolus since April; his nearly 20 years of experience includes executive positions with Kabbage Inc., Orchard Platform and Citi. Ocrolus enables financial services companies to make high-quality decisions, with trusted data, at scale; the company’s website is https://www.ocrolus.com/.

David Snitkof

Despite federal legal protections against credit discrimination, disparities based on race, marital status, gender, and sexual orientation persist in the mortgage lending process. In 2019, 16% of Black applicants and 12% of Hispanic applicants were denied mortgages, compared to only 7% of white applicants. Given that home ownership can be a major force multiplier on wealth, these disparities are a significant contributor to wealth inequality in the U.S.

Lenders can reduce the impact of human bias on credit decisioning by building standardized, repeatable and observable processes facilitated by machines. While there’s still value in human interactions, machines are better at ensuring fairness and auditability. You can’t see inside a human mortgage underwriter’s brain, but a computer’s memory leaves a clear trail, making bias easier to measure and safeguard against. A 2019 study found that fintech algorithms discriminate 40% less than face-to-face lenders on mortgage decisions.

By adopting technology to streamline processes, empower consumers, and embrace more comprehensive credit data, lenders will build a more efficient and inclusive mortgage system.

  1. Automate and optimize processes to increase ease of use and speed of decisioning. It’s hard to accomplish fairness by fiat, because people have different ideas of what “fairness” means. But when consumers have plentiful choices and low switching costs, they’ll tell you what’s fair by voting with their feet. Empowering consumers is the fastest way to build a fairer financial services economy.

Fast decisioning and ease of use are more than just efficiency measures in this context. With a less onerous application process and faster feedback from the lender, customers can shop around with more potential lenders and compare rates, terms, and services, lowering barriers to participation and putting more power in the hands of the consumer.

  1. Make customers’ data portable. Data portability is the backbone of a financial services economy with low switching costs and strong incentives for lenders to deliver high-quality customer service. If consumers can quickly and easily share their bank transaction history with a third-party lender, they have more choice and more control over their identity and reputation within the financial system. APIs and other services providing access to consumer-permissioned data make it significantly easier to provide this level of data portability.
  1. Use data to understand more economically diverse borrowers. Consumers are changing in ways that make traditional credit data less useful. For example, studies show 10-36% of U.S. workers engage in freelance or gig work, making the determination of income, affordability and debt capacity more complex.

Lenders need a richer picture of consumers’ financial dynamics, including their assets, obligations, and cash flow, in order to understand their ability to take on additional debt. This requires more comprehensive data and different technology than what has historically been used for credit decisioning. For example, a lender could use a combination of pay stubs and APIs to ingest income data from gig economy platforms, and algorithms could be used to understand the seasonality and stability of earnings.

There are significant challenges associated with streamlining or automating credit decisioning processes or building a pipeline to access, process and extract useful data from financial records. However, current and emerging financial technology infrastructure puts these goals within reach and makes it possible to envision a more inclusive, fair and equitable mortgage industry.

(Views expressed in this article do not necessarily reflect policy of the Mortgage Bankers Association, nor do they connote an MBA endorsement of a specific company, product or service. MBA NewsLink welcomes your submissions. Inquiries can be sent to Mike Sorohan, editor, at msorohan@mba.org; or Michael Tucker, editorial manager, at mtucker@mba.org.)