<img src="https://ws.zoominfo.com/pixel/pIUYSip8PKsGpxhxzC1V" width="1" height="1" style="display: none;">

Educational Redlining Report Raises Questions About Fair Lending

author Harrison Shaw
3 min read
Feb 12, 2020

Imagine a 24-year-old analyst applying to refinance a $30,000 student loan. He lives in New York City, makes $50,000 a year, and has held his job for five months.

Will where he went to school impact the rate and terms of his refinance?

It might, according to a new report by the Student Borrower Protection Center (SBPC).

The SBPC submitted nearly identical applications to fintech lender Upstart Network changing only one factor: where the analyst went to school. Borrowers who graduated from historically black colleges or universities (HBCU) or Hispanic-serving institutions (HSI) were charged higher interest rates and origination fees, SBPC found.

                                                                               

                                     Loan interest rate   Origination fee   Total cost

New York University                                      

                                        16.34% APR             $1,231                 $42,288

Howard University (HBCU)                          

                                        21.29% APR             $1,960                 $45,785

New Mexico State U (HSI)                            

                                        19.23% APR             $1,862                 $44,011

That’s not all.

The SBPC also reached out to Wells Fargo, making inquiries for a $10,000 student loans using similar applicants. It found that community colleges students would pay, on average, $1,134 more for a student loan than those who attended a 4-year college. They also have shorter repayment periods: 12 years vs. 15.

SBPC called out both Wells Fargo and Upstart for “educational redlining,” the practice of discriminating against borrowers based on where they attend or attended school. Both companies denied engaging in discrimination, but only one had readily available and public data to combat the allegation.

Using Educational Data for Underwriting

As lenders seek out new ways to determine creditworthiness some have turned to education data such as SAT scores, college majors, college selectivity and sector (for-profit, non-profit, public). For example, they can use cohort default rates at a college to predict likelihood of repayment.

In theory, it seems like a novel idea. In practice, there are serious fair lending implications. When it comes to college, 56 percent of Latinx students and 44 percent of black students attend a two-year public school vs. 39 percent of white students. That means making a lending decision based on school attendance can inadvertently lead to discrimination against borrowers of color.

It’s another pitfall to consider when looking at alternative data for underwriting.

How Accurate Was This Study?

Before legislators, regulators, and consumer advocates come for Upstart and Wells Fargo, it’s important to consider the thoroughness of the study. It’s based on two case studies, a very small sample. It doesn’t mention how many borrowers were simulated. 

As Upstart co-founder and CEO Dave Girouard noted to NPR, the study used "hypothetical, contrived applicants for a loan who aren't real people was both anecdotal and not reflective of the real world." 

That’s not to say there isn’t truth in the report. It’s possible that lenders may be inadvertently discriminating against borrowers of color by taking where they go to school into account. Previous reports have found issues with educational data in underwriting. And HBCU’s themselves pay more to borrow money.

Fair Lending Analysis Prepared Upstart for the Challenge

The good news for Upstart is that its proactive fair lending analysis left them in a strong position to defend against the claims.

Upstart, which claims to assess more than 1,000 factors—including alternative data like education— when making underwriting decisions, received the first-ever No-Action Letter from the Consumer Financial Protection Bureau in 2017. In it the lending platform promised to “regularly report lending and compliance information to the CFPB to mitigate risk to consumers and aid the Bureau’s understanding of the real-world impact of alternative data on lending decision-making.”

In a follow up letter last year, Upstart shared the result of its analysis. It reported “access-to-credit comparisons show that the tested model approves 27 percent more applicants than the traditional model, and yields 16 percent lower average APRs for approved loans.” Testing across race, ethnicity, and sex segments saw acceptance rates increase by 23 to 29 percent and APRs decrease by 15 to 17 percent, Upstart says. 

It was helpful that Upstart had solid data to counter the more anecdotal report. Not only did the company have readily available data to demonstrate its commitment to fair lending, it also demonstrated a proactive approach to fair lending and a willingness to work with a regulatory agency to uphold fair lending laws.

Compare that to Wells Fargo and its history of fair lending violations. It’s much harder for Wells Fargo to fight off even anecdotal accusations with its well-publicized compliance failures.

Read also: How to Make the Most of Expanded CRA Modernization Comment Period

Proactive Fair Lending Analysis

This is yet another example of the importance of fair lending data analysis. Analyzing consumer, commercial, small business, student, and small farm lending to uncover disparities and identify potential risks is more than just good compliance. It’s good risk management, particularly when it comes to reputation.

Don’t wait for someone else to start asking questions about your underwriting data and practices. Make sure your institution is proactively analyzing loans for potential discriminatory patterns and is able to defend itself with data if necessary.

Regular monitoring and analysis of your portfolio demonstrates a commitment to both fair lending and the laws that enforce it.

 

Related: How to Build a Strong Fair Lending & Redlining Compliance Management System

Subscribe to the Nsight Blog

All Topics Risk & Compliance Product Insight Banks Lending Compliance Credit Unions Risk Management Fair Lending Nfairlending Cluster: Risk Management Compliance Lending Compliance Management Integrated Risk Blog Regulatory Compliance Management Nrisk News & Updates HMDA Risk Mortgage Lenders Vendor Management Ncomply Business Resiliency CRA Lending Compliance Blog Vendor CFPB Ncontinuity Business Continuity Third-Party Vendor Management Cybersecurity Fintech NVendor Strategic Planning Ncyber Regulatory Updates Ntransmittal Audits & Findings OCC Regulatory Compliance Company Culture Audit Nverify Compliance Management Exams FDIC Nfindings Findings NCUA Contract Management Employee Intranet Business Continuity Management Contract FRB Findings Management Enterprise Risk Management FFIEC Ncontracts manager Third-Party Vendors Vendor Risk Vendor Risk Management Wealth Management 1071 ABA Bank Access Control Artificial Intelligence Audit Findings Board Portal Call Report Compliance Risk Cybersecurity Assessment Digital Tansformation Due Diligence Efficiency Employee Engagement Exam findings FIEC Governance and Risk Management Inc. 5000 Internal Audit Management Mortgage Lending Nboardportal Verify integrated risk management
How Regression Analysis Can Help Your Fair Lending Program
Fair Lending

How Regression Analysis Can Help Your Fair Lending Program

Mar 11, 2020 2 min read
[Survey] Are You Tuned Into the Compliance Pulse?
CRA

[Survey] Are You Tuned Into the Compliance Pulse?

Dec 5, 2018 1 min read
All the Fair Lending News That’s Fit to Print
CRA

All the Fair Lending News That’s Fit to Print

Oct 30, 2019 3 min read