Oscar Calvo-González, Ed.
Laura Zoratto, Ed.
Publication Date
September 13, 2017

World Bank

This publication summarises five cases in which the World Bank, in collaboration with governments and partners, has applied behavioural insights to different areas of public policy in Central America. These experiences collectively show how behavioural science, when used to complement the set of traditional tools available to governments, can help explain why individuals make certain decisions in different areas of their lives, explain how collective behaviours develop and become ingrained in a society, and help design better programmes and improve public service delivery.

For example:

  • Chapter 1, "A Behavioral Approach to Water Conservation: Evidence from Costa Rica," describes behavioural interventions designed to help municipal authorities motivate homeowners in a high-income district of Costa Rica to reduce their water use. (See also Related Summaries, below.) The strategies are split between a "descriptive norm" approach (whereby homeowners are encouraged to compare their consumption with that of their neighbours and the city as a whole) and a "plan-making" approach (which pushes homeowners to set clear measures and goals for modifying their water use habits). Two of those strategies (one "descriptive norm" strategy and one "plan-making" strategy) succeeded, generating water reductions of 3.4-5.5 percent in the following month. This reflects the fact that how people act and think is often influenced by the behaviours and thinking patterns of those around them: In other words, we think socially.
  • Chapter 2, "Promoting Tax Compliance in Guatemala using Behavioral Economics: Evidence from Two Randomized Trials [RCTs]," describes a nationwide letter-based experiment to increase compliance among nonpaying individuals and firms. The RCT was structured around four treatments using behavioural design. All the letters showed positive results, with two markedly increasing the rates of income tax declaration and payment. These high-performing treatments were based on (a) a deterrent message, which framed nondeclaration as an intentional and deliberate choice; and (b) a social norms message, which showed nonpaying taxpayers that they were in the minority. Together, they helped more than triple tax receipts among the intended group, and their effects were still noticeable on taxpayers' behaviour 12 months after the pilot.
  • Chapter 3, "Enhancing Child Development through Changes to Parental Behaviors: Using Conditional Cash Transfers in Nicaragua," describes a long-term study of the Nicaragua cash transfer programme, Attention to Crisis (Atención a Crisis), that rigorously examined the mechanisms evoking behavioural change. The analysis of this programme's impact on early childhood cognitive development found that children in households randomly assigned to receive cash benefits had significantly higher development outcomes nine months after the programme began. However, the cash transfer mechanism was not the only explanation for this positive impact: nonmonetary channels such as increased social interactions and motivation also evoked behavioural change in parents. This insight has led to new efforts to make the most of these findings: An ongoing trial uses short message service (SMS) technology to send information to parents in 4,000 Nicaraguan households. The trial aims to change key behaviour patterns and thereby improve child welfare. The treatment combines variations as to who receives the messages, what type of information is provided, and how that information is presented.
  • Chapter 4, "When Winners Feel Like Losers: Evidence from an Energy Subsidy Reform," presents an analysis of El Salvador's 2011 gas subsidy reform, which proved unpopular even though it benefited most of the population, especially those on low incomes. Using ad hoc household surveys from before and after the reform's implementation, the chapter seeks to explain this apparent anomaly. Using probit/logit models, it shows that the provision of information is critical to individuals' opinions of the reform measure. Political partisanship and subsequent experience during the implementation phase also influence public opinion. The analysis suggests that, for other areas of reform, policymakers may need to look beyond a consideration of the reform's winners and losers to ascertain satisfaction levels.
  • Chapter 5, "Redistribution in Times of Fiscal Pressure: Using Games to Inform a Subsidy Reform in El Salvador," explores the insights gained from a game-based simulation of a subsidy reform. It adds to previous research mainly by considering two factors: the impact that the destination of subsidy funds has on people's willingness to forgo a subsidy, and the role of that information in influencing how willing people are to share. The chapter also considers how individual beliefs about government feed into acceptance levels for subsidy reform. Results suggest that most people are prepared to see their electricity and water subsidies decrease if the economic gains of these reductions are used for poverty reduction projects or for the delivery of public goods. Furthermore, information about the regressive nature of the current subsidy was shown to positively affect the beneficiaries' willingness to share, or forgo, their subsidy.

Chapter 6 reflects on the progress made in applying behavioural insights in a development context. In short, these cases demonstrate the possibility of using nontraditional tools, complementary to regulation, in contexts where time and resources are limited. One of the main attractions of behaviour-influenced interventions is their low cost: Unlike traditional development projects, they seek to work within established systems and use existing resources. These cases also illustrate the importance of paying close attention to both the design and the implementation details of reforms. Government actions frequently have unintended consequences that minimise or derail achievement of a policy's end goals. That small details matter implies that we first need to deliberately seek out users and understand their perspective. The World Bank finds that such an inquiry is best undertaken through mixed methods of qualitative and quantitative work. In Costa Rica, this was undertaken through focus groups among Belén citizens. The researchers learned, for example, that many residents not only understood the importance of conserving water but also were predisposed to act on this understanding. The same residents, however, were unaware of how much water they were consuming (both in overall terms and in relative terms) or how to significantly reduce their water consumption.

Other lessons learned across the pilot projects:

  • The importance of partnerships - In all the pilot projects, the Bank partnered with relevant local government agencies. The ideal scenario is for the counterpart to provide as much in-house capacity as possible. That most behavioural interventions rest on modifying standard systems should mean that existing resources will largely suffice when implementing an RCT.
  • The benefits of simplicity in both the project's design and implementation - The principle of simplicity applies as much to the project's purpose (what key behavioural change is desired?) as to the project's design (what key nudge[s] will be used?) and the participants' role (what key shift in behaviour is required?).
  • The need for flexibility - The experimental nature of behavioural economics (BE) theory, coupled with its relatively nascent application to the policy space, puts a primacy on the ability to be ready to adapt. Because behaviour-related interventions are so intimately tied up in respondents' reactions to external "nudge" factors, and because these reactions are neither homogenous nor entirely predictable, BE projects tend to be extremely dynamic.
  • The importance of replicability - A key consideration guiding the pilot coordinators was the ability to potentially replicate the pilot interventions elsewhere, particularly in other low- and middle-income country (LMIC) settings. This motive reemphasised the preference for simple interventions that did not require elaborate technology or highly specific expertise to implement.

It is noted that these cases illustrate some of the findings of World Development Report 2015: Mind, Society, and Behavior (WDR 2015 - see Related Summaries, below) in practice. WDR 2015 challenges the assumption that we are rational choice makers and presents three principles of human decision making and its interactions with policy: namely, humans think socially, automatically, and with mental models. The World Bank has since established a Mind, Behavior, and Development (eMBeD) Unit within the Poverty and Equity Global Practice to mainstream and scale up behavioural science in public policies and programmes. The hope is that these experiences will help to inform other practitioners about the potential of applying behavioural insights in a development context and will encourage them to consider such approaches as a complement to traditional policy measures - particularly in LMICs.