Aligning automatic payments with income

Interventions
Recurrent payments form
Experiment Type
Field Experiment
Goals
Repay loan
Outcomes
Reduce debt
Focus Areas
Marketing & messaging
Behavioral Concepts
Friction Timeliness
Partner
Beneficial State Bank
Partner Type
Bank/Credit Union
Collaborator
Irrational Labs

What Happened

It worked. Sending customers a recurrent loans payment form via text message increeased their enrollment in a repayment program and total loan repayment.

Lessons Learned

Inviting loan repayment customers to ennroll in a recurrent loans repayment program can be an effective way to promote program enrollment, increase timely payment and increase loan repaymeent overall.

Background

There are many reasons someone may default on an auto loan: job loss, health emergency, an expensive car repair, or income and expense volatility. Brian Baugh and Jialan Wang found that financial shortfalls—particularly payday loans and bank overdrafts—are more common when there is a greater mismatch between the timing of someone’s income and the bills they owe.

We partnered with Beneficial State Bank, a California-based community development bank beginning in 2017, to design solutions that help make repaying car loans easier. We developed a recurring payments form designed to encourage borrowers to repay their loans automatically when they are paid.

Key Insights

We began with a behavioral diagnosis that detailed each step in the entire auto lending process to better understand that process from the perspective of both the borrower and the loan issuer. Our behavioral analysis revealed a number of insights specific to Beneficial Bank’s internal processes and barriers to repayment, as well as insights relevant to auto loan repayments broadly.

  • Monthly loan repayments are almost universally due on the day that a borrower bought their car. In some cases, this arbitrary choice does not cause any problems. If their repayment due date falls far from a payday, though, creates a disconnect between someone’s expenses and income, making it more difficult to consistently make payments on their car loan.

  • While a loan payment is due on a specific date (e.g. the 15th), many people are not paid on specific dates (e.g. every other Friday). In these cases, simply changing the date their payment is due is not sufficient and borrowers still run the risk of having income come just after their due date some months.

Experiment

We randomized over 1000 new Beneficial State Bank indirect auto loan customers into either a control or an experimental group. Beneficial State Bank conducts a welcome call that is required for all new loans or loan refinancings to confirm borrower information and to offer repayment information. During the welcome call, the experimental group was texted a Recurring Payments form designed to establish automatic loan repayment timed with when customers were paid. The control condition did not receive the form.

Results

Over 50% of customers in both the control and experimental group expressed high levels of interest in automatic, recurring payments timed with income. However, people who were texted an automatic recurring payments form were twice as likely to enroll in automatic payments (23% compared to only 12%), p < .001. Turn-off rates were equivalent across the two groups (about one third), so the proportion of members enrolled in automatic payments remained higher in the experimental group (15% vs. 8%), p < .001.

There was some evidence suggesting the form successfully encouraged individuals to time their payment with their paydays. The automatic payers in the experimental condition made 1.2 payments/ month compared the control group’s 1.05 payments/month, p < 0.001. In examining payment patterns, it seemed that about 10% of members in the experimental group had set up these smaller, more frequent payments.

We hypothesized that timing loan payment with payday would improve loan performance. On average, we find that members in both payment groups pay the appropriate amount by the monthly due date. However, members in the experimental group have fewer late payments (M = 3.83) compared to the control group (M = 5.29), p < .001, and members in the experimental group have paid more toward their loan overall (104% vs. 94.2%), p < .001.

There were 17 loan defaults during the study period, and all were in the control group, p = .128. This is not a statistically significant difference, but is a strong directional effect.