The best responses to COVID-19 involve behavior change, including social distancing, mask use, and, eventually, vaccine uptake. So why has the behavioral public policy community, encompassing governmental behavioral insights units and a variety of not-for-profit and private organizations, played a marginal role?
For the most part, our field wasn’t ready for a pandemic. We had no behavioral playbook on nudges, defaults, and other strategies for improving social distancing, mask use, remote learning, home-based work, and social transfers. A playbook would have provided a default option for policymakers, whose own bandwidth is overloaded in trying times like these. In this regard, behavioral public policy was not unique. Despite repeated warnings that we were not ready for a pandemic, governments, and society at large were caught unprepared. But like families, every community unhappy with its response to the pandemic is unhappy in its own way.
We ignored history. Behavioral public policy is rooted in the idea that biases, heuristics, and mental models determine behavior. If you reframe or alter individuals’ decision making context, you change their behavior. But what happens when individuals are responding to the behavior of other individuals and organizations, each acting under the influence, and in a novel context, of multiple biases, heuristics, and mental models? That’s what happens in a crisis. In the case of COVID-19, people and organizations have varying risk perceptions and appetites, mental models of disease, dispositions to cooperate, stereotypes of out groups, and trust in government and science. Understanding interactions like these is the realm of behavioral game theory, which unfortunately remains a long way from ready use in most policy contexts.
We had no behavioral playbook on nudges, defaults, and other strategies for improving social distancing, mask use, remote learning, home-based work, and social transfers.
So where can we turn to understand complex systems in crisis? Some of the most useful findings for COVID-19 policy have involved comparing how different jurisdictions responded to the 1918–1919 influenza pandemic. Those studies examined, at the aggregate, how organizations, society, and individuals interacted. They were able to assess how various interventions, and their timing, changed behaviors and epidemiological and economic outcomes. Those findings were widely circulated, and influential, as the early trajectory of policy responses to COVID-19 was being debated. History can be read and understood through various lenses, such as rational choice theory or class relations. We need behaviorally informed readings of history to bridge the gap between individual behavior and social systems.
We did not systematically develop intuitions. Behavioral public policy relies on randomized controlled trials (RCTs). Without an RCT to back us up, many of us are inclined to stay silent. The search for evidence is, of course, laudable, but even setting aside the problem that RCTs may tell us less than we think about how an intervention will work in a different setting, crutch-like reliance on RCTs is a problem when you need to make recommendations fast. It would be useful for the field to pre-register, systematically, our guesses for which interventions work best, and go back and learn where we got it wrong or right, so that we can refine our behavioral intuitions. The outlet FiveThirtyEight has been asking experts to estimate coronavirus outcomes in the United States in the near future. In its first survey, expert predictions, even with confidence intervals, substantially underestimated reported cases two weeks later. One could call intuition-building exercises of this sort “behavioral super forecasting.” Paul Volcker is reported to have said, “In a crisis the only asset you have is your credibility.” That’s for leaders. For policy professionals, in a crisis your main asset is your experienced intuition.
We stayed in our sandbox. The field of behavioral public policy has promoted the use of low-cost framing and related interventions to change behavior, in contrast to heavy-handed laws and incentives. In the present crisis, among the most powerful tools for promoting social distancing have been mandates from national and local governments. But the power of those mandates, it appears, has stemmed from crystallizing social rules and coordinating behavior, rather than actual enforcement and penalties. The following conjecture is worth exploring: in a crisis, the most powerful behavioral change tools consist not of laws or incentives (how credible is enforcement or reward when everyone is overstretched?) or reframing and other behavioral interventions (how to scale those immediately) but of interventions at their intersection. The playbook needs detailed ideas about the social meaning of law, and the salience and social significance of taxes and subsidies.
In a crisis, the most powerful behavioral change tools consist not of laws or incentives or reframing and other behavioral interventions but of interventions at their intersection.
The behavioral public policy community is diverse, and many individuals put forward useful ideas to promote handwashing, improve remote learning, and maintain social connection, among other challenges. But on the big questions, especially on the size and shape of requisite public health and economic policies, the field of behavioral public policy did not have much to say. When it did, we too often resorted to enumerating one of the (by now overexposed) list of cognitive biases that “explain” fear or inaction, or worse, proposed epidemiological approaches on the basis of limited evidence and historical precedent.
We can and should develop a behavioral playbook for crises. This pandemic is far from over. There will be another. Have you heard anyone say we aren’t ready for climate crises?