When a Government Nudge Goes Wrong, the Disadvantaged Pay
New York City parking ticket scofflaws teach us a lesson about the limits of behavioral economics
Over the past decade, if you feel you’ve been played like a puppet — nudged to pay overdue taxes or limit water consumption — it’s not your imagination. It’s just some government department dabbling in the dark arts of behavioral economics to encourage (arguably) better choices for society.
Those nudges often work. Manipulating the way choices are presented to people can indeed have an impressive impact on outcomes. A reminder note based on a social norm (“Nine out of 10 people pay their taxes on time!”) or a personalized letter comparing our energy consumption to that of a neighbour’s gets our attention. At heart, we’d rather not stick out for the wrong reasons.
There are now at least 200 behavioral science units in governments around the world dedicated to these principles — encouraging COVID-19 vaccination and organ donations, collecting on corporate tax payments, and shaping many other behaviours essential to a well-functioning society. Governments see nudging as a “soft” tool that’s inexpensive to implement and administer and not overtly bothersome. Plus, it’s politically safer than “hard” tools such as fines.
Nudging is still a relatively new idea. It was popularized in 2008 with the publication of Nudge: Improving Decisions About Health, Wealth, and Happiness. Applications to date have been largely based on first-generation research. Recent studies, however, reveal a much more nuanced picture. Some people respond to nudges as intended, others don’t respond at all, and still others respond in unintended ways.
That’s what researchers learned in a study they conducted on New York City parking ticket payment. Ori Heffetz and Ted O’Donoghue of Cornell University and Henry Schneider of Queen’s University analyzed how people responded to notification letters after being ticketed for a parking infraction in New York City between 2011 and 2013. In that time, 6.6 million tickets were issued to two million vehicle licence plates, totaling $424 million in fines and $85 million in late penalties (lesson: when in the Big Apple, leave your car at home).
The research team exploited a quirk in the timing of the notification letters. Before June 2012, the first reminder letter was sent out 40 days after the infraction, alerting people ticketed to a missed payment deadline. After that date, the first letter was sent on day 20, reminding them of the impending day 30 deadline.
That one change alone had a significant and positive impact on the aggregate response. The number of people who paid their parking tickets after receiving the letter sent on day 20 was 10 percentage points higher than those who paid after getting the letter sent on day 40 (46 percent versus 36 percent).
It was an impressive result, more so considering the content of the letter wasn’t altered much, primarily just its timing. But the result masked a much more intriguing finding. The researchers noticed that, over the two-year study period, there were quite a few people who received multiple parking tickets. And their behaviour fit a pattern: they were either consistently slow or remiss in paying their parking tickets, or they consistently took care of business without much delay. They were consistent in their response to reminder letters as well. Those with a low baseline propensity to respond to tickets — a natural nudge target — reacted least to letters.
It was a big gap between the two groups. When the first notification letters were sent out on day 20 rather than day 40, the response from the “taking care of business” group increased by 15 percentage points. But ticket payments by the “low-response group” only grew by one percentage point.
The researchers then matched the plate owners’ address to their census block group and related socioeconomic data. They discovered that the people least likely to respond to nudge letters, and who incur significant late penalties, disproportionately come from neighbourhoods that have lower income, less education, and higher proportions of Black or other racial groups.
While the low-response group largely ignored the nudge letters, they finally did respond to a third notification that communicated a heavy-handed “incentive” — a threat that their vehicle would be towed or booted within weeks.
Why would the hammer work where nudges failed? One reason, says one of the researchers, Henry Schneider, is that “the pressures of life may get in the way of handling parking tickets and other tasks like that.” The second may be financial constraints. “You may get a reminder, but if you don’t have the funds you won’t be able to pay the ticket,” he explains.
If you’re a policy-maker in a behavioral insights unit, this study may prompt you to revisit your playbook. For one, you don’t necessarily need to manipulate the content of the nudge letter to get results. As part of their study with the City of New York, the researchers ran a five-week experiment in which they tested four versions of the first notification letter: some with or without additional information and varying degrees of forcefulness. The results showed that the content of the letters hardly mattered.
The study also calls into question the one-size-fits-all approach to nudging. Targeting individuals based on past behaviour looks to be a more efficient and equitable approach to reminders and other forms of nudging. Otherwise, most of the resources to support an intervention go to those who don’t really need the nudge and not to the people least likely to respond and who pay most of the late fees.
“It’s not like piling on the penalties will really improve payment rates or help the community,” says Schneider. “It seems to be mostly harming the folks who are already in a disadvantaged position.”
The numbers back him up. In the case of New York City parking tickets, three notification letters were sent out, each with a deadline and a late payment penalty that could add up to $60 to the original infraction. Ninety-one per cent of the extra spending on notification letters and 87 percent of savings in late fees went to the people who responded in a timely manner.
Changing the Nudging Playbook
Schneider says that identifying people as low- or high-response types can be fairly simple. A municipality wanting to collect late parking ticket fees can determine how quickly people paid their previous parking tickets, library fees, or utility bills. People who consistently paid fees in the past are unlikely to need a nudge in the future. Instead, attention should go to the small group of chronic delayers. For these individuals, municipalities may need to get more creative with their nudges or think outside the box to avoid simply levying late payment penalties.
In the short term, this study and others like it will give a more realistic view of behavioral science’s true value. That couldn’t come soon enough. For some, nudge campaigns too often fall short of high expectations. This partly relates to overly optimistic results from randomized control trials (RCT) that don’t transfer to the real world. One recent study found that academic RCT nudge experiments had an 8.7 percentage point effect compared to 1.4 percentage points from nudges implemented by real-world behavioral insight units.
In the end, it’s in the interest of both government officials and ordinary citizens that nudging, responsibly deployed, continues to improve. The alternative levers that governments use to encourage pro-social behaviour — financial penalties and taxation (often regressive) or legislation (usually leaky with loopholes) — are hardly more appetizing.
