Are Repeat Nudges Effective? For Tardy Tax Filers, It Seems So

While behavioral scientists sometimes aim to nudge one-time actions, such as registering as an organ donor or signing up for a 401K, there are many other behaviors—making healthy food choices, paying bills, filing taxes, getting a flu shot—that are repeated on a daily, monthly, or annual basis. If you want to target these recurrent behaviors, can introducing a nudge once lead to consistent changes in behavior? What if you presented the same nudge several times—would seeing it over and over make its effects stronger, or just the opposite?

Decades of research from behavioral science has taught us a lot about nudges, but the field as a whole still doesn’t have a great understanding of the temporal dimensions of most interventions, including how long nudge effects last and whether or not they remain effective when repeated.

If you want an intervention to lead to lasting behavior change, prior research argues that it should target people’s beliefs, habits or the future costs of engaging in the behavior. Many nudges, however, focus instead on manipulating relatively small factors in the immediate choice environment to influence behavior, such as changing the order in which options are presented. In addition, relatively few field experiments have been able to administer and measure an intervention’s effects more than once, making it hard to know how long the effects of nudges are likely to persist.

What if you presented the same nudge several times—would seeing it over and over make its effects stronger, or just the opposite?

While there is some research on what to expect when repeating nudges, the results are mixed. On the one hand, there is an extensive body of research in psychology on habituation, finding that, over time, people show decreased responses to the same stimuli. It wouldn’t be a giant leap to presume that seeing the same nudge again might decrease how much attention we pay to it, and thus hinder its ability to change our behavior. On the other hand, being exposed to the same nudge multiple times might help strengthen desired associations. Research on the mere exposure effect, for example, illustrates how the more times we see something, the more easily it is processed and the more we like it. It is also possible that being nudged multiple times could help foster enduring change, such as through new habit formation. Behavioral nudges aren’t going away, and their use will likely grow among policymakers and practitioners. It is critical to understand the temporal dimensions of these interventions, including how long one-off effects will last and if they will continue to be effective when seen multiple times.

We studied these questions through a multiwave field experiment conducted in partnership with the Ministry of Finance in Ontario. Our research was focused on testing the effectiveness of planning prompts on organizations’ tax compliance. Planning prompts—step-by-step instructions detailing when, where, and how to complete a task—have been shown to be effective at motivating individuals’ behavior changes across a variety of domains, but they haven’t been tested as a tool for motivating tax compliance, nor has their effectiveness within organizations been tested.

Our field experiment focused on all organizations that failed to file timely annual returns for a payroll tax during the 2013 and 2014 tax years, around 6,300 and 6,200 organizations, respectively. Each year, late filing organizations were randomly assigned to receive one of two letters: the government’s standard late notice or a revised experimental late notice with the information from the standard late notice reorganized into planning prompts.

We found that our planning prompt intervention increased organizations’ tax compliance, and these effects replicated across both tax years. Organizations that received our experimental letter were significantly faster filing their taxes, filing an average of 4.1 days and 5.3 days sooner, respectively. As a result, across these two tax years our intervention saved the government more than $11,000 in collection costs and brought in approximately $600,000 more in remittances before additional collections efforts began. It also saved organizations money in terms of penalties and fines. In fact, the positive effect of our experimental late notice persisted over 10 and 16 weeks after the letters were sent, even weeks after additional collection efforts began.

Seeing the same nudge again might decrease how much attention we pay to it, and thus hinder its ability to change our behavior. On the other hand, being exposed to the same nudge multiple times might help strengthen desired associations.

As designers of the intervention, these were all encouraging findings, but would just one letter lead to sustained behavior change the next year? Unfortunately, no. Despite our nudge’s initial success, when it came time for organizations to file their taxes the following year, it made no difference. Late filing organizations that received our experimental letter in the first year were just as likely as those that received the standard late notice to file their taxes late again the following year.

The effects of our intervention did persist, but only to a point. We were able to nudge organizations to file faster and remit more money than those who received a standard late notice. For those organizations that received the standard notice, it took weeks of costly sustained collections efforts for the government to match the effectiveness of our intervention.

While the behavior change from our nudge was only temporary, an unintended opportunity arose when many organizations were late filing both years. Of the almost 6,200 organizations who participated in the second year of our field experiment, approximately one quarter of them (1,500) had also been late in the first wave. Their persistent tardy filing enabled us to test how effective our nudges were when they were deployed more than once.

In our field experiment, all organizations that were late filing were randomly assigned to receive one late notice, each year, independently. Therefore, some organizations received our nudge once (for example, only in the first or second year), some organizations twice, and some organizations simply received the standard late notice each year. This random assignment allowed us to compare how effective our nudge was when only seen once versus twice.

Perhaps being exposed to the same simple plan more than once makes it even easier to implement the second time around.

Was the nudge effective if received a second time? Yes—in fact, we found that there was a trend toward the experimental letter being more effective for those organizations previously exposed to it. We know from prior research that simple concrete plans can be effective at getting individuals to act, and it is encouraging to see that these plans can be effective when repeated. Perhaps being exposed to the same simple plan more than once makes it even easier to implement the second time around. We found this to be true for both small and large organizations, as well as for firms with more or less experience filing. And as a result of these findings, Ontario’s Ministry of Finance has since implemented our planning prompt intervention as its new standard late filing notice.

There are some key factors that may have contributed to the success of our repeated intervention that might not transfer to other repeated nudges. In the case of tax returns, there is a year between exposures, which is likely enough time to forget the intervention and mitigate any habituation. Also, organizations—the target of our experiment—are prone to procedural changes and staff turnover that could have caused our nudge to be new to the actual receiver each year. It is possible that if the late notices were sent more often, organizations may eventually habituate to the changes, and as a consequence we would observe decreased attention toward our intervention. Our research also only studied nudges repeated once, so it’s not clear whether the same nudge repeated two, three, or more times would continue to be as effective as they were the second time. Future research should continue to explore the effects of nudges over time, as these dynamics are critical for ensuring long-term success.

While there’s clearly more work to be done on the temporal effects of behavioral interventions, our work provides some key initial insights. The effects of our tax filing nudge did not persist over time for those organizations that only received it during the first year. Timing is critical and planning prompts must be provided at key points during the decision-making process to be maximally effective, not just one-off interventions. Our results also provide some early evidence that when interventions are deployed multiple times, it may be possible to utilize the same nudge without diminishing returns on efficacy. In fact, our planning prompt trended toward being more effective when seen again. It seems you can, in fact, nudge twice.