by Saugato Datta, Vice President, ideas42
Behavioural economics studies human behaviour in all its messy complexity. Its practitioners pay attention to all manner of things that standard neoclassical economics ignores (or waves away as unimportant: the context in which decisions are made, visual, aural or social cues, salience, social or group norms, and the subtleties of the way options are framed, to name but a few). As applications to a variety of fields – ranging from healthcare to agriculture to education to finance – shows, this is, generally speaking, a good idea. Paying attention to the sorts of details that economists (and many policymakers) have grown used to dismissing as unimportant – such as whether program benefits are framed as gains or losses, or what aspects of a person’s identity a policy foregrounds – or paying closer attention to the details of tools economists do use – such as the timing and frequency of payments, not just their quantum – can lead to insightful ideas about how to make progress on difficult policy problems.
Yet, of course, economists – behavioural or otherwise – do not make policy, even if they think they ought to. Inasmuch as this is the case, it behoves them to pay close attention to what policymakers think and believe and what they care about when it comes to policy choices. As we know from the public choice literature, policy is not made or implemented by disinterested technocrats but by self-interested individuals whose motivations and objectives are as complex and as messy as those of any of the people their policies hope to influence.
So for insights to have policy impact, it is important not just that they be insightful or that their foundations be sound; it matters perhaps just as much what those who might potentially use them on a large scale think about them. And this includes not just what they think about direct costs and benefits, but indeed a whole host of other more “behavioural” features – their resonance with policymakers’ sense of purpose; their impacts on fairness and equity, what policymakers think they are actually about, and so on.
Yet for a discipline that takes so much interest in how things are perceived and about the importance of context, behavioural economists have (by and large) not really tackled the question of how policymakers think about their discipline – if indeed they think about it at all. The recently published paper by Trujillo et al. takes a stab at addressing some of these questions using a survey of practitioners, policymakers and academics. It tries to gauge their interest in and knowledge of behavioural economics concepts in relation to health policies in developing countries, as well as their receptiveness to the kinds of policy recommendations that the field is known for. To the best of my knowledge, this is the first survey of this kind to focus on the reception of and receptiveness to behavioral economics and its applications. And while the paper focuses specifically on health policy, its findings should be informative – and provocative – for those interested in the application of behavioural economics more broadly.
A few things seem particularly noteworthy. First, respondents seem quite familiar with basic concepts in behavioural economics, although it is not quite clear whether we have any way to know if they know what they do not know. In particular, certain terms whose knowledge the paper probes have fairly precisely delineated and specific meanings within the behavioural economics literature, which may not be widely comprehended. On the other hand, one might argue that it hardly matters whether a policymaker knows what hyperbolic discounting is, as long as s/he grasps the implications of hyperbolic discounting on behaviours of policy interest.
In that sense, the most interesting aspect of this work is the vignettes the authors use to understand the level of acceptability of various “behavioural” policy levers. Here, I was struck by something that may be worth thinking about more. The authors find that policies around framing, communication, information campaigns, etc. are pretty widely acceptable. This is not surprising, given that these are all quite widely used by governments and agencies. What was more surprising was the relatively high degree of consensus around reminders and small incentives to overcome immediate costs. To the extent that these findings are generalisable, this is important news for applied behavioural economics. There remain many areas where timely reminders or micro-incentives for “good” immediate behaviours could be helpfully deployed. Thinking about the best ways to scale these relatively simple ideas, as well as ways to refine their design so that they become more effective – for example, by working out the optimal frequency and timing of reminders or small incentives – should be a larger part of the research agenda than it is at the moment.
Indeed, the areas where there was a lower degree of consensus about desirability should also provide applied behavioural economists considerable food for thought. On the one hand, the paper finds that policymakers are not fans of things like pay for performance or introducing uncertainty into payment schedules. It’s worth noting that some of these areas are ones where there’s considerable debate within the (behavioural) economics community to begin with, so that it’s not surprising that policymakers remain sceptical. (Indeed, it’s not clear to me whether we should think of some of this as discomfort with behavioural economics’ preferred policies, or those of economics without “behavioural” as a qualifier. It’s also worth pointing out that this finding may be one of those that doesn’t carry over, or carry over quite as strongly, to other domains, inasmuch as health policymakers and officials are particularly sensitive to issues around intrinsic motivation and are thus considerably more resistant to and skeptical of monetary incentives or payments in any form to patients in particular than policymakers in other domains.) On the other hand, the effectiveness of commitment devices has been demonstrated in so many settings that it seems disappointing that policymakers do not much like it. Finally, it’s not surprising at all that respondents do not seem to like the idea of paying people for reducing risky sexual behaviours. Here, too, policymakers are probably right to be worried about general equilibrium and long-term effects, which applied behavioural economists sometimes pay less attention to than they ought. Of course, the caveat above continues to apply, because these are precisely the areas where financial or monetary incentives have made the least headway, even when they are not explicitly behaviourally motivated.
At some level, all this provides a good basis for thinking about where research should focus. For instance, research around how to structure and time reminders or small incentives optimally, as well as that around the long-term and general equilibrium effects of some of the more exciting behavioural economics interventions is clearly needed. Perhaps similar work might also help to reduce policymakers’ discomfort with things like commitment devices or performance pay. And yet it would be odd (and perhaps naive) for behaviouralists to assume that more evidence is all that’s going to be needed to change minds – because, as they know only too well, information and evidence are only a small part of what determines how people – policymakers included – think about things.