The OECD’s six principles for spreading behavioural insights across government


Public sector decision making and regulated organisations are the next frontiers of behavioural economics, says the OECD. Its new report, featuring over 100 case studies, contains a guide to using BI in new ways.

Behavioural insights is not merely a trend — it has the potential to improve how governments work across many areas.

But while it is increasingly being used by policymakers across the world, BI is still mainly being applied in only a few of its potential guises. The most common are policy tweaks targeted towards influencing individual citizens, typically when they are making financial, health and safety or consumer decisions.

The next frontier is applying behavioural insights to regulated organisations and even government itself, argues a new report from the OECD. The report is based on surveys with 60 public bodies in 23 countries and two international organisations, and is the first big comparative survey of behavioural insights units around the world.

“Soft” incentives nudging public servants towards coordinating with other departments could assist in achieving whole-of-government goals such as increasing diversity or good regulatory practice, for example.

“As with all policy interventions, results of behaviourally informed interventions should be monitored and evaluated over a period of time.”

Applying it to regulation could also increase compliance in some areas better than a purely prescriptive approach, the report argues, though this has not been properly tested yet. Capital markets and banks, energy consumption for large firms, transport used by big business are all potential targets, the OECD suggests.

Applying behavioural insights earlier in the policy cycle has the potential to bring new gains too. At present BI tends to be introduced fairly late in the process, typically during the implementation phase, to fine tune a pre-decided policy. Incorporating it during ex post evaluation and from the start of policy design to provide evidence on where rational actor and perfect information assumptions might fail can help help governments attain their policy goals and reduce the likelihood of needing to go back and fix initiatives that are not working.

The report includes more than 110 case studies, covering issues related to consumer protection, education, energy, environment, finance, health and safety, labour markets, public service delivery, taxes and telecommunications. The three-page case studies include an HIV risk education game in South Africa, a Canadian trial using air miles to tackle obesity, and a British experiment using reminders to encourage consumers to check whether their insurance premium was too high.

Six principles to spread behavioural insights

The report boils down the lessons from these cases into six principles for the use of behavioural insights in public policy.

  1. Strategy: engage in a multi-staged strategy for implementing behavioural insights with a menu of types of assistance for all appetites.
  2. Data and evidence: undertake robust calculations before engaging in testing and experimentation to ensure the use of sufficiently large sample sizes, so that effects can be detected. Use of data should be based on the recognition that good and reliable data is key to applying behavioural insights and upholding its integrity. It should also build on the recognition that data are not the same as evidence; it is also necessary to understand the limitations of data for public policy.
  3. Validity of results: replicate to ensure the observed results are correct in the same context and setting (internal validity) and also test the application of the same approach to other contexts and settings (external validity).
  4. Segmentation: consider applications that could work for a part of the population but not for the entire population and whether these applications can implemented given the legal and cultural context.
  5. Evaluation: conduct ongoing monitoring to identify short-term and long-term effects: as with all policy interventions, results of behaviourally informed interventions should be monitored and evaluated over a period of time.
  6. Transparency and accountability: publish work (successful and unsuccessful trials) for transparency and accountability: a number of countries already publish in journals or produce annual reports of their activities. This level of transparency is good practice that should be adopted by all behavioural practitioners in public policy. Transparency should also inform the actual practice of applying behavioural insights by disclosing and understanding more about the actual costs (to be compared against the benefits) of applying behavioural insights to deploy it appropriately.

Apart from recommending application of behavioural insights to the rules and practices governing public service agencies and influencing regulated organisations, the report suggests policy makers:

  • Develop some consistency in the organisation, methodologies, quality controls and capacity support functions of the different behavioural initiatives.
  • Encourage the development of knowledge and capacity among public officials to understand and apply behavioural insights within public bodies, for example through mainstreaming, training and information.
  • Take into consideration behavioural insights when designing and evaluating policy implementation.
  • Develop processes to determine when there is a behavioural issue that can be successfully addressed through the application of behavioural sciences. Equally, determine when a behavioural intervention may not be appropriate and understand why.
  • Exploit the full potential of behavioural insights as a tool for engaging with stakeholders and collect feedback on what works.