The pros and cons of ‘behavioural policymaking’ as an art and science

This is the final and third part of a series of blog posts on Evidence-based Policymaking in the Public Sector. Part I focused on The Rise of the Social Experiment in Policymaking and Part II looked at Big Data for Governance. You can follow the links to read them

Behavioural Economics — a relatively new sub-branch within the discipline — concerns itself with addressing behavioural patterns, drives and (dis) incentives among economic players. Developed steadily over the past century, the field has not as much exploded into the economic and policymaking field as it has gained a loyal following among certain public administration practitioner groups who consider it the solution to a host of social issues that range from addressing the climate and resource crisis to public health and community cohesion.

Very broadly put, behavioural economics accommodates and accounts for ‘unideal’ (and even irrational) decision making on the part of economic agents. This approach differs from earlier neoclassical readings of economic behaviour that held individuals as perfectly rational decision-making agents and maintained that only under information asymmetry could they be forced to make less-than-ideal choices.

As the discipline of behavioural economics grew and gained its rightly deserved space in the world of social sciences, it grew to be embraced subtly — and sometimes not as subtly — by decision makers across sectors. One element of this science, ‘choice architecture’ has, for instance, been very commonly used to drive human-centric design of institutions across the world. Choice architecture refers simply to the fact that things such as “the order of traffic lights, the display of food in a cafeteria, the design of a web page, all shape the salience of options and the likelihood of certain choices… [This is because] the social environment in which people make decisions may trigger some of these heuristics and inhibit others.”.

These subconscious heuristics lie at the heart of behavioural ‘nudging’, which is define as “any aspect of choice architecture that alters people’s behaviour in a predictable way, without forbidding any options or significantly changing their economic incentives.”

This definition covers three distinct subcategories, “each tapping into a different psychological human failing that leads people to fall back on automatic rather than systematic information-processing strategies and decision making: Budges that tap into people’s propensity to choose options that demand the least (physical or intellectual) effort (i.e., the path of least resistance); nudges that tap into people’s propensity to conform or succumb to prevailing group norms and peer pressure; and interventions that harness people’s eagerness to identify with peer groups or valued groups that provide them with positive self-esteem”.

For profit-maximising private sector players, this can take the form of ‘nudging’ buyers and potential consumers into altering purchasing their behaviour. Interestingly, when some commercial companies in the European Union began to exploit this behavioural tendency to influence shopping and spending behaviour by including “pre-ticked” default options during the check-out process as a design feature on their online platforms, the EU Directive on Consumer Rights had in June 2014 step in to categorically ban these companies from utilising these exploitative design choices.

How do Nudges Work?

At least one of the psychological roots behind the predictability of behaviours and choices when presented with uncertainty or an excess of information is the human tendency to use heuristics, biases, mental shortcuts and simple solutions “even if this means acting against their own interest.” This is, in essence, what lies behind the irrationality propounded in behaviour economics sector.

The rules of the game have been somewhat clearly delineated for the public sector in several parts of the world to prevent consumer exploitation. However, these rules remain fuzzier in the public sector, where policymakers and sector experts in different parts of the world frequently assume the role of choice architects to drive desired public behaviours. Indeed, the use of behavioural nudges has slowly but survey grown in the public administration sector, especially over the past decade, as evident not only by a spate of behaviourally focused public policy decisions being rolled out globally but also mainstream acceptance of the science as evidenced by the creation of specialist journals such as Behavioural Public Policy and Journal of Behavioural Public Administration and the fact that behavioural policy researchers — such as Richard Thaler, Esther Duflo, Abhijit Banerjee & Michael Kremer — have been recognised and feted for their work in using behavioural insights to trial and introduce lasting solutions for intractable social problems.

However, the use of behavioural science for public policy has not been without its fair share of criticism. At its most ‘innocuous’, behaviour governance policies make use of simple nudges to direct the public towards making socially ‘preferable’ or healthier choices. Some researchers have compared behavioural nudges to harmless but useful “signposts”, noting that, “simplifying messages and reducing complexity may have large effects on people’s behaviour. For example, making registration forms or information letters clearer, and options more salient, can move decisions in desired directions, e.g. by encouraging people to enrol into pension schemes or change energy providers”.

Notably, such approaches have yielded also fruit in some economic contexts. For instance, in the United Kingdom, Her Majesty’s Revenue and Customs (HMRC) “managed to increase the number of people paying their income tax on time by 15% through using a letter informing the addressees that most people in their postcode had already paid… [and this scheme] could generate an extra £30 million if it were rolled out nationwide. In another example, the media reported that Her Majesty’s Court and Tribunal Services (HMCTS) [had] conducted a successful trial using personalised text messages to remind people to pay their court fines. In this trial, the rate of people paying their fines increased from 5-33%, and it is estimated that the intervention could generate £30million in extra revenue if rolled out nationally.” In the same country, evidence from randomised controlled trials has showed that exposing people to SMS reminders to save money increased their savings balances by 6% and similar emails to homeowners informing them of their neighbours’ electricity bills led to reductions in power consumption equivalent to energy price reductions of 11–20%.

Similar efforts have been undertaken elsewhere in the world too. While in the UK, most contemporary work in this area is driven by the Behavioural Insights Team (BIT) (sometimes called the ‘the nudge unit’), which conducts “experimental trials, designs behavioural interventions and advises other organisations on how to apply behavioural insight in the public domain,” the US has its own Social and Behavioural Sciences Team at the White House’s Office of Science and Technology Policy (OSTP) and the German Chancellery too has started up a project group, the Wirksam Regieren (“governing effectively”) with a similar mandate.

Such efforts in the field behaviour public administration make use of commonly identified behavioural and decision-making mechanisms as heuristics or “organising devices” to base policy designs on. Table 1 below lists some of these main mechanisms and the corresponding policy design phases where they are used.

Table 1: Major Causal Mechanisms in Behavioural Governance

As a still-developing knowledge area, ongoing research in behavioural governance has the advantage of enhancing our collective “understanding of the psychological causal mechanisms that influence the development of public policies and link those policies to implementation and to actual behaviours at the micro-level.” Consequently, the practice may well serve to increase the effectiveness of policy design, delivery and evaluation, to more effectively target policy recipients and to enhance cost-effectiveness of addressing “wicked” governance problems. Viewed from this angle, behavioural policymaking sounds like the manna to almost every governance challenge everywhere, and yet, these facts are the precise reason why critics of behavioural governance raise an alarm.

So What is The Problem?

The key concerns around the use of such approaches hinge partly on the philosophical question of free will: Should(n’t) citizens have the choice to make their own decisions, instead of being coerced, covertly or overtly, by their elected governments? A related point: Who is to say that administrations are best suited to make decisions on driving individual decisions? This is the classic policing-the-police question, and surely, critics of the use of behavioural science in public administration have focused on studying biases among policymakers themselves and have, alarmingly, found evidence of it as well. Yet another strand of behavioural policymaking thus has been driven by these concerns and is centred around understanding issues of bureaucratic stereotyping, discrimination, as well as methods to mitigate cognitive and decision biases amongst actors in the bureaucracy including civil servants and public managers.

Further, at what point does innocuous government messaging and simple ‘nudging’ evolve into more egregious forms of choice control? This is an ethical question that is relevant for every single political context, but particularly so for fragile and unstable democracies that suffer from low levels of trust in their governments and often, for good reason.

As the frontiers of behavioural science expand, it is also becoming increasingly clear that individual-level decision making is almost always a function of surrounding environments. Indeed, “humans and their decisions are grounded in, and influenced by, the social context in which they live” and these factors are so inextricably tied with each other that they cannot be treated or targeted separately. Some researchers thus recommend a multifaceted approach to policymaking that addresses both social structures and systemic practices to encourage social change instead of merely targeting individual actions. For example, “attempting to encourage people to change the way that they travel by focusing on each individual’s choices is likely to be ineffective as it does not challenge the systems and processes giving rise to social practices of (perhaps increasingly) unsustainable travel.”

Further, instances of behavioural public administration — if leaked to the public — has been linked to instances of behavioural backlash as well. Indeed, as “supermarket designers know, ‘pure’ choice architecture nudges (such as leading the customer onto a particular path which takes them past particular product groups in a particular order) can be effective over long periods of time as long as the customer remains unaware of the manipulation. However, findings in the ‘reactance’ literature suggest that, once they are aware, customers are likely to feel an urge to defy the system. The same can be expected to apply to pure choice architecture nudges used by policy makers.”

“Reactance” refers to the tendency of backlash if people perceive that a message has originated from an untrusted or unlikeable source, or if the message is perceived to be against ‘our identity’ or ‘way of life’, or if it violates the recipient’s sense of agency and autonomy. An excellent example of such reactance was the contrasting behaviour of ideologically different American households who were given information on their own and  peers’ home electricity usage. Here, researchers found that liberals “started saving electricity (as expected), but that conservatives responded defiantly and actually increased electricity consumption... This example illustrates that ‘social nudges’ are not only more likely to be ‘found out’, but that they have considerable potential to trigger reactance – especially among those who are politically opposed to the intended outcomes.”

In another example of an unexpected result, a Swiss–US study on energy conservation found that “giving people feedback on their water use [did] successfully [reduce] water consumption, but coincided with an overall increase in electricity use. An explanation for this unforeseen outcome is what researchers call the ‘moral licensing effect’: the people who saved water felt entitled to be wasteful in another area.”

Indeed, this is an increasingly oft-observed cognitive problem associated with behavioural interventions (e.g., around obesity policies or fitness campaigns), especially when the word of the government’s efforts in these areas gets out, which in turn leads people to feel concerned by the elitist assumption that they as “ordinary people” know less than experts on how to live a good life.

Despite these apparent setbacks, a positive result of all this push and pull within the knowledge area has been the development of set of interventions that include both “cognitive and decision processes through which decision-makers, implementing actors and target populations… shape and react to public policies and to each other, as well as the impacts of these processes on individual and group behaviour.”

The Last Word

Another point of contention for critics of behavioural public administration is the relatively undersized effects that such efforts have had in actually triggering lasting social change. Some researchers have noted that the work of governmental ‘nudge units’ have actually led to either extremely small public effects or even no effects at all. It thus appears that nudges need to be accompanied by additional strategies to spur citizens into desirable actions.

One widely offered recommendation by critics has been for policymakers to treat stop individual-level decision-making as an isolated phenomenon and instead study it within the rich and detailed cultural contexts that people live in. Indeed, “mental models, the interrelated schemes of meaning that people use when they act and make choices, are shaped by economic relationships, religious affiliations and national contexts. This needs to be recognised when designing nudges, so that they are communicable in different communities.”

Further, “cultural contexts have an effect on the formation and expression of heuristics and biases. Mental models, the interrelated schemes of meaning that people use when they act and make choices, are shaped by economic relationships, religious affiliations and national contexts. This needs to be recognised when designing nudges, so that they are communicable in different communities.”

Today, researchers admit that there is no one-size-fits-all approach to behavioural policymaking success. However, our shared understanding of the main drivers and motivators for individual- and community-level decision making is evolving as well. For instance, as an extension of the preceding point, it is now becoming increasingly evident that the way to trigger lasting group-level change is to treat “individuals as social beings; [to] recognise that people are active agents who are aware of what others around them value; [to] understand the importance of norm internalisation; and [to] ensure interventions ‘work with’ existing social norms (rather than against them).”

This is a clear shift from previous neoclassical models of understanding that treated individuals as “fallible decision makers” who benefitted from nudges designed by an all-knowing administrator. Social researchers now instead propose an approach “which recognises that enduring behaviour change involves an identity-change process whereby people proactively choose to engage in behaviour that is perceived as identity-consistent” leading to transformations in their “belief systems in such a way that the desired behaviour arises from new, deeply held convictions” which is more likely to trigger lasting behavioural change.

It is clear that there is a long way to go for both researchers and policymakers to fully understand and formalise the “rules” of the behavioural game. Some things, however, are clear from the evidence gathered thus far: The public does not appreciate being patronised by its administrators. People are not the unthinking hordes that early behavioural science theories would have us believe, and are in fact frequently deeply cognisant of the contexts and circumstances they live in. They do not respond well to nudges that stigmatise behaviour or trigger shame or guilt, and their use of cognitive biases is almost always to simplify choice-making in an uncertain and complex world. People are also broadly responsive to effective information, which may in the long run be more useful in encouraging behavioural change than unconscious choice architecture. Shared communal identity is also a significant force in driving behaviours, and people thus tend to respond well to socialisation treatments such as ‘peer education’, which reduces friction among communities, tends to be more effective in “converting” opinions within homogenous groups of people and promotes shared identities as in these programmes, “people are not being nudged into particular behaviours but persuaded to adopt new norms and behaviours as an integral part of ‘who we are’, ‘what we stand for’ and therefore ‘what we do’”.

Interestingly, some of the most effective behavioural interventions in the world have relied on non-coercive, non-unconscious nudging that took into account these aforementioned social phenomena, even if these tend to be more expensive than implementing simple choice architecture choices. Perhaps this is indeed the way to go.