How to Nudge Better in Public Policy

Behavioural nudges will only improve public outcomes if the principles for their successful use–including individual choice–are followed closely.

Low Qiyan_banner-teaser

Date Posted

12 Nov 2019


Digital Issue 5, 12 Nov 2019

Many well-intended and carefully calibrated policies and programmes suffer from low enrolment rates or smaller-than-expected treatment effects. Such outcomes are surprising to policymakers and programme designers because there is often broad agreement that these interventions effect changes that are beneficial and desirable. To boost the effectiveness of interventions, policymakers have traditionally resorted to increasing incentives or intensifying education and outreach to increase participation. This is costly.

In the last decade, a seemingly low-cost alternative has rapidly become mainstream: nudges, small adjustments to the choice architecture, which promise to improve policy effectiveness by gently guiding people in the correct direction.1 The success stories of the Behavioural Insights Team within the UK government—including improved tax compliance and organ donation rates—have spurred formulaic replications as well as new applications of nudges in interventions elsewhere around the world.

The hype around nudges, however, has led to a common misunderstanding: Policymakers now think that nudges are elixirs for failing programmes, and quick fixes for undesirable behaviours. Such is the aura surrounding the nudge, that many traditional interventions are now claimed to be nudges! This misunderstanding led the originator of nudge theory, the Nobel Laureate Richard Thaler, to clarify that nudges were never supposed to be able to solve every problem.2

Policymakers now think that nudges are elixirs for failing programmes, and quick fixes for undesirable behaviours.

Principles of Successful Nudging

For nudge interventions to work, certain principles must be in play. However, these principles are often downplayed or neglected in practice, which could explain the limited success of nudges in some policies.

1. Nudges are most effective when behaviours are influenced strongly by cognitive biases.

We have all had the experience of making decisions that are not necessarily in our best interest—we eat too much and exercise too little; give in too early and give up too late. Cognitive bias plays an important role in many of these decisions. When behaviours are largely driven by cognitive biases, nudges are useful to address these biases in order to change behaviours.

One example is enrolment in retirement savings plans. Most people agree retirement savings are important, but inertia results in a failure to sign up for, and contribute to, a retirement savings plan. Automatic enrolment into a retirement savings plan has been shown to be effective in helping people overcome their inertia. Madrian and Shea found that after implementing automatic enrolment (with an opt-out option), the proportion of employees in a US company who joined the retirement savings plan increased from 49% to 86%. Only 14% opted out.3

An important strength of nudges is that people get to choose.

A nudge is effective in this context because the delay in enrolment is largely attributable to cognitive bias. When undesirable behaviours result from other causes, nudges are less likely to be effective. For example, nudges may be ineffective in raising the number of elderly persons installing grab bars in their homes or insulating their attics if the main barriers to making those changes are not cognitive but physical—for instance, the elderly may face difficulty in purchasing the required items or in moving their furniture around for the renovations!4

2. Nudges help people make choices that are in their interest, as judged by themselves.

An important strength of nudges is that people get to choose. Using the retirement savings plan example again, automatic enrolment alone is not a nudge. Automatic enrolment is in fact a shove if people must incur substantial costs—whether in time, mental effort, or money—to opt out or otherwise act contrary to the policymaker’s design intent.

Automatic enrolment turns into a nudge only when it is paired with an easy opt-out option. By making it “easy and cheap to avoid” automatic enrolment, such as one-click to opt out, those who do not want to enrol in the retirement savings plan for reasons other than inertia are free to execute their preferences; those who would like to sign up for the retirement savings plan but are slowed down by inertia will be signed up immediately.

In contrast, when automatic enrolment is forced on people without an easy opt-out, welfare is severely compromised for those who do not want to enrol in the retirement savings plan for reasons other than inertia. Such individuals may be harmed by enrolment because their preferences over time, savings, or consumption may be better met by their private financial arrangements, rather than those offered by the company or the state. This might apply for instance to lower-income earners who may have to borrow money at high interest rates to meet their expenses if part of their income were deposited in the retirement account. It could also apply to those who are more financially savvy and better able to make their own investments with higher returns.

In Singapore, most employees are required to make automatic contributions to CPF directly from their payroll, and there is no opt-out option. Hence, automatic enrolment into CPF contributions is not a nudge. Employees are also automatically enrolled to make contributions to self-help groups. While employees can opt out from the contributions, they must fill in a form which requires their employers to sign off to verify their request. In this case, even though an opt-out option is present, it is contentious whether this intervention is a nudge since employees incur not just time costs, but also reputational costs, from opting out.

Although nudges are meant to “influence choices in a way that will make choosers better off, as judged by themselves”, in practice many nudges are designed to influence choices in a way that will make choosers better off, as judged by policymakers. The presumption that policymakers know best is itself a cognitive bias which may compromise the ability of an intervention to improve social welfare. On the receiving end, people simply do not like to be told what to do all the time, even if it is in their best interest and they fully agree that is what they should be doing. Think about the times we were told to sleep early but slept late anyway. Preserving choice is beneficial because it avoids a cognitive bias where people focus on “being told what to do” instead of “what is in their best interest”.5

In practice, many nudges are designed to influence choices in a way that will make choosers better off, as judged by policymakers.

Furthermore, many people make choices that are most appropriate for their particular circumstances. For instance, low-income households are often thought to make poor decisions: they do not save enough, and they consume food that is high in calories. Working with these households can reveal that these decisions may in fact be carefully calculated ones. For some of these households, paying off bills may have to be prioritised over saving because accumulating arrears will cause anxiety and other issues.6 Some of them may choose to consume high-calorie food because such food is filling and cheap, and they have run out of money for healthier but costlier options.7,8 Observing such choices is helpful for policymakers to understand some of the unanticipated reasons that may underlie undesirable behaviours.9

3. Nudges have limited effects on matters which require changes in people’s preferences or mindsets.

The first and second principles, taken together, suggest that nudges are best used to target cognitive biases that prevent people from acting to improve their own welfare, in accordance with their own preferences. But when a broader notion of welfare requires people to consider changing their own preferences or mindsets, nudges may have limited impact.

Consider healthy eating. While nudges in cafeterias may be effective in encouraging students to eat healthily at the cafeteria, changes in food preferences may only last when nudges are present. Students may compensate for eating healthily by eating more unhealthy food later, at home or in places without those nudges. These compensating behaviours may nullify the effects of the nudges.10,11

Nudges may have limited impact.

Another example of the limited impact of nudges on preferences is foster parenting. Recruiting more foster parents is important but challenging. Understandably, policymakers are interested in using nudges and other behavioural tools to encourage participation in fostering programmes. However, reluctance to foster likely reflects personal values or parenting preferences, rather than cognitive biases (as was the case in saving for retirement). For instance, people may be averse to disruptions in their lifestyle or have concerns about their parenting capabilities. While nudges may certainly raise fostering sign-up rates by temporarily focusing parents on the upsides of fostering, it is beyond the scope of nudges, or indeed most behavioural interventions, to permanently change parenting preferences, skills, or capabilities. A successful nudge may therefore risk raising fostering rates, while lowering fostering outcomes.

Obstacles to Nudging Better

In summary, the most powerful kind of nudges are those that help people overcome cognitive biases which prevented them from acting in their own interest and according to their preferences. For issues of this nature, nudges are a good tool, since they are likely to be more effective and are less costly than other types of interventions. However, there are two obstacles that may cause effective nudges to fail: poor communication and counternudges.

Opting out may be welfare-improving for some individuals even when a nudge is otherwise effective.

Policymakers are often caught in the false dilemma of needing communicating opt-out options clearly while wanting to achieve broad impact. Fearing that communicating opt-out options might encourage opting out and undermine the success of interventions, policymakers may be inclined to keep quiet about opt-out options and when possible, make opting out harder. This is a false dilemma, because the effectiveness of a nudge intervention is measured by the size of the social good it generates and not the extent of participation. Policymakers must accept that a one-size-fits-all policy should not be the norm, and opting out may be welfare-improving for some individuals even when a nudge is otherwise effective. Instead, policymakers must believe that people would not choose to opt out if a nudge genuinely helps them to behave according to their own interests and preferences. Hence, opt-out options should be communicated clearly. Failing to do so could even be counter-productive: ‘hiding’ opt-out options invites speculation and suspicion about the intent of the intervention, breeding distrust. This not only harms the effectiveness of the existing nudge but also future nudge interventions.

‘Hiding’ opt-out options invites speculation and suspicion about the intent of the intervention, breeding distrust.

Counternudges, unlike opt-out options, are real threats to effective nudges. While nudges are implemented for social good, the objective function of commercial entities is to maximise profit. Therefore, firms may introduce counternudges to either exploit a nudge intervention or to negate its effects to increase their profits.12 For example, a fast food restaurant may reduce their standard soft drinks due to national nudges to reduce sugar intake, but counternudge by making upsizing more attractive—increasing sugar consumption overall. This suggests that policymakers using nudges as protective mechanisms for consumers must forestall counternudges and be prepared to adopt more aggressive approaches, such as incentives, mandates and bans, to improve social welfare.


Ong Qiyan is Deputy Director of Research, Social Service Research Centre, National University of Singapore. She received her PhD from the Nanyang Technological University. A behavioural economist by training, her current focus is on bringing behavioural science to social services and policies to generate new insights and stimulate innovative solutions.


  1. R. H. Thaler and C. R. Sunstein, Nudge: Improving Decisions about Health, Wealth, and Happiness (Yale University Press, 2008).
  2. R. H. Thaler, Misbehaving: The Making of Behavioral Economics (New York: WW Norton, 2015), 325.
  3. B. C. Madrian and D. F. Shea, “The Power of Suggestion: Inertia in 401 (k) Participation and Savings Behaviour”, The Quarterly Journal of Economics 116 (2001): 1149–87.
  4. Thaler (2015) detailed an initiative in the UK to encourage people to insulate their attics. Even though very few people responded to the initiative and insulated their attics, almost everyone who agreed to insulate their attics also took up the option of attic clean-up service, suggesting that cleaning up the attic is an important barrier to insulating the attic.
  5. J. M. Jachimowicz, S. Duncan, and E. U. Weber, “Default-Switching: The Hidden Cost of Defaults,” SSRN Electronic Journal (January 2016), doi: 10.2139/ssrn.2727301.
  6. Qiyan Ong, Walter Theseira, and Irene Y. H. Ng, “Reducing Debt Improves Psychological Functioning and Changes Decision-Making in the Poor”, Proceedings of the National Academy of Sciences 116 (15): 7244–7249.
  7. Y. Y. Teo, This Is What Equality Looks Like (Singapore: Ethos Books, 2018).
  8. B. K. Cheon and Y.Y. Hong, "Mere Experience of Low Subjective Socioeconomic Status Stimulates Appetite and Food Intake", Proceedings of the National Academy of Sciences 114 (1) (January 2017): 72–77.
  9. C. R. Sunstein, The Ethics of Influence: Government in the Age of Behavioral Science (New York, NY: Cambridge University Press, 2016).
  10. E. Frey, and T. Rogers, “Persistence: How Treatment Effects Persist after Interventions Stop”, Policy Insights from the Behavioral and Brain Sciences 1 (October 2014): 172–9.
  11. C. R. Sunstein, “Nudges That Fail”, Behavioural Public Policy 1 (2017): 4–25.
  12. Ibid.

Back to Ethos homepage