top of page

Reimagining auto loans: The opportunity ahead for credit unions

by Scienaptic Research

Auto loans, especially those for used vehicles, make up a significant portion of the portfolio for credit unions - around 35% to 45% of their total portfolio on an average. They account for less than 5% of bank loans. Also, banks have only around 20% share of used vehicle financing, as compared to 70% in credit unions.

Let us explore why credit unions were sought after by auto loan borrowers, the new challenges owing to the COVID-19 pandemic, and how they can possibly weather the storm using AI-powered underwriting solutions.

Traditional Differentiators Offered by Credit Unions to Make their Auto Loans More Attractive

Credit unions are preferred for providing excellent member services at a higher convenience and lower cost. Vehicle buyers have always been drawn towards credit unions for their loans due to these reasons:

  • They are a customer at a bank, but an owner at a credit union since the latter is owned by its depositors. A credit union would make every effort to provide you the best financial benefits possible without any product hard-sell.

  • Credit unions provide lower auto-loan rates than other lenders. The average loan rates for credit union and bank auto loans were as follows (National Credit Union Association

Source:, June 2019
Source:, June 2019
  • Credit unions are more flexible when it comes to credit issues. You have a higher chance of getting a loan even if you have credit issues, including low credit scores. CUs have a lower credit score threshold.

  • Credit unions are locally operated and easily accessible, as compared to bigger banks whose branch need not always be near you.

  • They also provide a personalized experience with auto loans.

How did Credit Unions Underwrite Auto Loans?

Credit unions use a combination of automated and manual processes using predictors such as these to underwrite auto loans:

  • Loan to Value (LTV)

  • Debt to Income (DTI)

  • Payment to Income

  • Number of open auto loans

  • Employment history

  • Credit history

  • Tradeline balance-to-limit ratios

  • Collateral considerations like vehicle age, condition, usage (private vs commercial), and odometer reading

COVID-Induced Challenges

With the COVID-induced recession, low income levels, and high unemployment levels, credit unions are being forced to adjust their underwriting standards. One of the main challenges lies in validating employment and income considering they cannot rely only on historical data with any certainty.

As credit unions do skipped payments and offer extensions for members, and stop repossessing vehicles in some US states, there will be an impact on whether the members end up repaying the loans or not.

This has put immense pressure on credit union members. They are cautious about their lending decisions and will try to avoid any unnecessary risk. This will result in a drop in approval rates and can weaken member relationships.

Furthermore, credit unions have been slow to embrace AI in credit underwriting and other applications with concerns about data sets and robots replacing traditional underwriters. A 2018 study by Fannie Mae revealed that only 5% of CUs have adopted AI/ML solutions.

In this scenario, credit unions that have always relied on lower interest rates to attract auto loan seekers will have to make a strategic shift.

Opportunities in Underwriting Auto Loans Using Alternative Data & AI-Powered Solutions

Most credit unions lack effective systems for data, analytics, and a platform to assess borrower risk. Let’s look at what needs to done in each of these areas:

1. Alternative Data

Apart from publicly available data and loan application details, credit unions need to use data from credit bureaus and alternative data sources like LexisNexis to get a 360 degree perspective of applicants. The richer the data and features derived from the data, the better the quality of underwriting. This alternative data sources can be used to provide insightful information on auto loan applicants, as well as other sectors.

Their credit underwriting platform should be capable of capturing data from all these external sources and providing a real-time, holistic picture of the applicant. It should have custom dashboards that enable real-time tracing of risk, portfolio, and process metrics.

2. Analytics

The credit underwriting platform that CUs use should have a systematic method for feature definition, extraction, and calculation. They need explainable Machine Learning-based analytical models which are capable of assessing risks and pricing to be able to make sharp credit decisions. These algorithms should be able to transform internal and external data into meaningful insights of profitability and default trends.

3. Platform

Most importantly, the credit underwriting platform should be capable of running and managing multiple concurrent strategies and hand-offs. It should also have prebuilt predictors and models in the platform for rapid deployment and ROI.

Also, it should have a flexible architecture capable of integrating the decision systems into existing LMS and LOS workflows. The key intention should be to accelerate deployment of cutting-edge credit decisioning without disrupting existing processes and technologies.

Using Scienaptic’s AI-powered Credit Underwriting Solution

AI/ML-powered underwriting platforms like Scienaptic’s Ether platform are capable of using third-party data sources such as FCRA compliant alternate data from Lexis Nexis, etc. and raw trade line level bureau data. This can help identify traits such as rent and utilities payments, driving history, asset history, and so on.

The platform adopts a fast go to market framework that identifies and uses new data sources to bring in latest AI-ML algorithms. This helps to deliver maximum impact while maintaining the nimbleness to deploy models fast. This can help CUs make better member-friendly decisions and weather this recession by optimizing their financial resources. It will also help them differentiate clearly between potential defaulters and credit-worthy members.

It uses explainable AI-models that come with simplified and automated adverse action reasoning. They are FCRA compliant, robust, and are refreshed every three months offering 1,000+ pre-configured predictors (as opposed to the traditional 50-100 predictors.) It will integrate with existing digital applications and human oversight to provide more flexibility in loan decisions.

As we’ve seen, credit unions have a clear edge over banks when it comes to auto loans. So, with AI/ML solutions, they can save costs by making quicker credit decisions, better pricing, and exceptional member service in these trying times. With the smart AI tools on Ether, credit unions can look forward to 15-40% more approvals and 10-25% lesser losses. Ultimately, this can improve loan approval rates while bringing down default rates.

352 views0 comments


bottom of page