Everyone wants more approvals, better rates and fewer losses – but is it all possible? 🥇

Specialty auto lenders are unheralded champions of the economy – they make auto loans available to people with below average credit who would otherwise be unable to buy a car. These cars help people get to work, travel and elevate themselves in life, driving short-term and long-term economic benefits.

However, it’s also a very challenging industry. Specialty auto lenders face heavy regulatory scrutiny, significant capital requirements, high business cyclicality and difficult relationships with the media, which is often quick to label these businesses as predatory. The industry’s challenges have been frequently highlighted in news coverage, including articles from Bloomberg, Business Insider and the Wall Street Journal.

Despite the challenges, specialty auto lenders have an imperative to approve as many creditworthy applicants as possible. High approval rates increase overall efficiency and strengthen the lender’s relationships with auto dealerships, which typically leads to additional volume (dealers don’t like when their customers are rejected for financing!). As a result, auto lenders are frequently searching for ways to approve more customers, offer better rates and mitigate losses – a tricky balance to achieve.

Fortunately, advances in technology and data science have helped create a powerful new tool that can assist lenders – Automated Machine Learning (or “AutoML”). AutoML leverages historical data and uses it to quickly train highly accurate predictive models. In the data intensive specialty auto finance industry, it can transform multiple functions to drastically improve business and financial performance.

Here are a few critical areas where machine learning can assist specialty auto lenders. If you’re not familiar with automated machine learning, see this post for an introduction to the key concepts.

1. Loan Underwriting

Underwriting, including credit review and pricing, drives financial performance for specialty auto lenders. Accurate underwriting mitigates loan losses and the operational costs associated with collecting on defaulted loans. The goal of auto underwriting is to accurately assess the borrower and collateral (i.e. the vehicle) to determine the associated risk level, decide whether to approve the application and, if so, price the loan at a level that’s attractive to the applicant while generating an appropriate expected return.

Machine learning analyzes historical data, “learns” from it and generates a predictive model that can make future decisions. It’s most useful when many data points inform a given decision, which is common in specialty auto lending (applicant-provided information, credit reports, vehicle information, etc.). As the volume of data increases and the number of potential relationships expands, machine learning discovers sources of predictive power not possible with rules-based underwriting.

AutoML can be used in numerous ways within the underwriting process, including:

  • Providing an accurate assessment of the borrower’s risk level (i.e. likelihood of default)
  • Determining the likely recovery rate on the collateral (i.e. resale value of the repossessed car)
  • Recommending the most appropriate pricing (i.e. the APR to the consumer)

Using machine learning models in the underwriting process has numerous benefits, including more consistent risk assessment, additional predictive accuracy and cost reduction driven by eliminating the manual underwriting effort that remains common at many lenders.

2. Internal Workflow Prioritization

Auto lending involves back and forth between the lender, auto dealer and borrower, and loan officers need to carefully manage this process to close the financing transaction. This often involves manually processing applications in a simple queue, which results in sub-optimal allocation of the loan officer’s time.

As auto lenders gather data, they can use automated machine learning to train models that recommend the “next best action” to loan officers. This lets them work through their queues dynamically, based on the opportunities that are most in need of attention. The prioritization can be based on a variety of predicted factors, including the likelihood that the deal will close, expected profitability of the loan, time-based factors and more.

Using machine learning to prioritize workflow lets loan officers focus exclusively on processing, which accelerates response rates, reduces costs and improves customer experiences. Further, by centralizing the decision-making, auto lenders can ensure that staff are always focused on the right applications, creating more consistent workflows.

3. Income Verification

Auto lenders often rely on customer-stated income information when making lending decisions, and then need to verify this information. Income verification is critical to the overall process, because a customer’s income (and related metrics such as debt-to-income and payment-to-income) plays a major role in determining financing eligibility and maximum loan amounts.

Income verification often involves collecting supporting evidence such as employment information, proof of stated income (such as pay stubs) and income estimates from third parties. This takes time, often results in imperfect information and requires judgmental human decisions.

Machine learning can greatly help with verification to the benefit of both lenders and applicants. Auto lenders can use historical data from verified applications to train predictive models that accurately estimate the applicant’s income on future applications. This estimate can then be compared to the applicant’s stated income level to help confirm that it was reasonable and accurate, either to assist with prequalification or the final verification process. Automated income assessment can also flag high-risk cases for manual review.

Machine learning models can accelerate the time to close a deal by being used as either decision-making tool or decision-support tools. This is critical as the dealer often has multiple financing offers from lenders and prioritizes lenders that can close quickly.

4. Loan Servicing

Machine learning can also be used within auto loan servicing to predict which customers are likely to miss their payments. Predictive models analyze large volumes of data regarding historical payments and then accurately determine which customers are unlikely to make their next payment. These models are then run against data from outstanding loans on a frequent basis (typically daily or weekly), and high-risk loans are added to a queue for additional customer contact, such as extra emails, text messages or phone calls.

NSFs and no-pays are the first step towards a default, so being able to identify and target these loans before they occur has the potential to greatly reduce overall default rates. Further, this can provide an early-indicator for identifying potential issues with the underwriting process, which can be used to create a rapid feedback loop between the servicing and credit functions (rather than waiting months to learn if the loan actually defaulted!).

Final Thoughts

Specialty auto lending is an important and challenging industry that is always searching for tools to improve performance. Fueled by exponentially increasing data and compute power, automated machine learning can assist auto lenders by making it easier to assess creditworthiness, prioritize workflows and issue profitable loans.

About DigiFi

DigiFi is a technology company that helps businesses make better automated decisions.

Our platform lets businesses easily use automated machine learning and rules management to optimize critical decisions with no coding or technical expertise required. Repetitive work that used to take hours can now be completed in minutes, letting your team focus on what matters most.

Learn more at digifi.io