Can we build savings through coin deposits?

Background

Consumers have a number of options for converting loose change into bills. Like many retail stores and other financial institutions, Digital Federal Credit Union (DCU) offers access to coin machines in many of their branches. The coin machines that DCU offers remain popular among members.

Collected baseline data shows that the average coin deposit amount was $91, with a median of $45. While 3% of members split their slip amounts between cash, savings, and their checking accounts, more than 80% of members either cash the entire amount or deposit it directly into their checking account. Only a handful of members save their coin deposits.

Key Insights

Our baseline data collection indicated that a large portion of members cashed their coin machine slip, likely in order to spend the money from the coin deposit. This makes sense – spare change is rarely taken into account when creating budgets.

Similar to tax refunds and other kinds of windfalls, people think of the money from spare change as extra money that exists outside of someone’s normal mental accounting. The money can be spent without the same constraints that people place on income from other sources, but that also means coin deposits represent an interesting opportunity to nudge individuals to deposit all or a portion of the coin slip amounts into savings. We felt that a few small nudges could increase the number of members who saved their coin deposit windfalls.

To determine how we might go about developing an intervention, we started by understanding the process in detail. We quickly identified a few touch points with members. The machines convert spare change into a slip with the total amount printed on it. Members then take the slip directly to a teller, at which point they must decide whether to cash the slip or deposit the amount into one of their accounts.

Experiment

Together with DCU, we worked to change the choice architecture around the coin deposit process. The new process included a salient reminder to save, as well as an active choice with implied default that encouraged members to deposit their coin amounts into a savings account.

We randomly assigned DCU branches into two conditions. In the first set of branches, we posted a behaviorally-designed sign by the coin machines. The sign was intended to be a reminder for members using the machines to consider saving when they may not have previously.

In the second set of branches, the savings message communicated by the sign was reinforced by tellers. When a member took their coin machine slip to the teller, the tellers said, “Can I go ahead and deposit this into your primary savings account today?” This script intended to imply that saving a portion of the slip was the expected behavior.

Results

The study was launched in December of 2019. In total, DCU tellers tracked 3,302 coin machine transactions – those in the treatment group were more likely to agree to save a portion of their coin slip into their DCU primary savings account (23.1% in control, 26.6% in treatment, p = 0.006).

Interestingly, when we look at just those who decided to save, those in the control group saved a higher percentage of their coin slip (62% vs. 54%, p= 0.049). So, while the implied default appears to have led to more people in the treatment to agree to save, those who agreed to save tended to save a smaller percentage of their slip, on average. We think that this is because the implied default convinced members to at least save a portion, bringing down the average amount saved. Indeed, while an equal number of members in both groups saved their whole slip, members in the treatment group were more likely to save a portion of their slip compared to the control (p=0.001).

Taken together, we believe that this suggests the implied default in the treatment group successfully nudged members who otherwise would not have saved at all to save, even if it was just a small portion of their coin slip.