Ensuring Fair Access to Credit in the Gig Economy Era

In a time when multiple income streams complicate income verification, AI tools can help CUs practice fair lending.

In recent years, auto dealers and credit union lenders have faced growing challenges in promoting financial inclusivity and fairness for today’s car buyers. The issue stems less from intentional discrimination and more from the tendency of applicants to unintentionally misrepresent their income on loan applications. This is particularly common with income that is difficult to document, such as freelance work or multiple jobs, and is often compounded by the applicant’s or lender’s lack of understanding in calculating earnings from gig economy jobs.

How pervasive is this situation today? Consider that 37% of respondents in a January 2023 poll of approximately 1,000 workers from the job site Monster reported having more than one full-time job. What’s more, freelancers make up 36% of the workforce in the U.S. today – according to Upwork’s 2021 Freelance Forward Economist Report, freelancers comprise more than one-third of the U.S. labor market, with 59 million freelancers nationwide.

If that’s not enough to cause confusion over what to include in a car loan application, also consider that the gig economy has made it easier to take a business overseas. According to MBO Partners, in 2018, 19% of full-time independent workers in the U.S. had clients in other countries, and 16% of Americans say they’ve used a gig website or app to make money.

Growing Complexities in Calculating Income
The challenge arises from increasing complexities like variable earnings, duplicate documents, multiple employers, and W-2 wage earners with additional fixed income requiring gross-ups. Frequent staff turnover at credit unions compounds the issue, as inexperienced employees often make mistakes and introduce biases during manual reviews. They may inconsistently account for overtime, shift differentials and per diem pay, leading to unequal treatment of applicants. This inconsistency not only affects fairness but also exposes lenders to potential fair lending violations.

A lender and major U.S. financial institution that we serve actually “tested” this theory and provided the same income documents and funding guidelines to 10 different employees ranging in seniority from executives and funders to underwriters. Shockingly, no two people arrived at the exact same calculation.

To further complicate matters, consumer-permissioned sources often generate new barriers for low-income earners since many lenders underwrite based on net income rather than gross income. Every American has a different tax situation leading to unpredictable withholdings, tax deductions, garnishments, 401(k) contributions, etc.

When credit unions underwrite based on net income, they can inadvertently increase the borrower’s payment-to-income ratio, leading to higher interest rates. This process adds extra hurdles for low-income borrowers, such as the need to remember account usernames and passwords, while lacking comprehensive support. Significant variable pay types like per diem, shift differentials, overtime and per-mile payments are often overlooked, further disadvantaging these borrowers.

Traditional Formulas and Tools Reduce the Chance of a Fair Loan
Today’s auto lenders often rely on intricate formulas to calculate applicant incomes, which can fluctuate throughout the year and are difficult to explain. This complexity frequently leaves car shoppers and the dealers assisting them puzzled about how to accurately calculate or report income, potentially leading to unwanted repossessions. Additionally, lenders miss opportunities to secure more loans without increasing credit risk due to these confusing income calculations.

Dealers and lenders are increasingly turning to new AI tools to help bring more fairness to the process. Many of today’s leading OCR solutions read data off paystubs, but they’re only pulling text from the documents. They also rely on changes in the metadata to “detect fraud.”

OCR tools read data off of documents, but they aren’t an end-to-end solution helping lenders comply with ECOA and Fair Lending. Their mission isn’t to lower the cost of credit using AI for financial inclusivity, real-time transparency, and improved compliance and market share. To accomplish this, an organization must enable higher function decisioning and unlock data to ensure credit policies are unbiased and effective.

Click here to read the entire article from Justin Wickett, Credit Union Times