Our Perspectives

Solving last mile challenges: The potential of behavioural insights for the 2030 Agenda


Behavioural insight draws from research findings from psychology and neuroscience. These insights about how people make decisions matter for development. UNDP photo

Across the globe, all people – poor or rich – sometimes make choices that are not conducive to their own well-being. Saving enough for retirement, eating healthy, investing in education – all too often we humans postpone intended actions to ‘tomorrow’, succumb to inertia or get stuck in habits.

In light of the extensive research on the cognitive biases that influence human decision-making, there is a broad consensus that traditional economic models are insufficient for effective policy-making. Behind every policy lie assumptions about how humans will behave in light of new regulations and why we act the way we do.

Nonetheless, behavioural insights are only being leveraged by a relatively small, but growing number of policy-makers around the globe. Now, United Nations agencies and funds are catching up. Behavioural insight draws from research findings from psychology and neuroscience. These insights about how people make decisions matter for development. They matter for policy-formulation and addressing last-mile problems.

In Busia, Western Kenya, many smallholder farmers do not use fertilizer. Research shows that it significantly increases yields and a standard policy might assume that farmers aware of these benefits would invest in fertilizer and therefore seek to raise awareness amongst farmers while providing subsidies for it. Here, insufficient awareness was not the problem. Farmers expressed their commitment to purchase and use fertilizer in interviews - but most did not translate intent into action. A behaviourally-informed intervention reframed the problem. What barriers prevent farmers from taking action? How can we redesign the choice environment to help farmers act in accordance with their long-term goals?

This intervention entailed parallel experiments, designed based on what we know about factors that influence decision-making such as mental accounting and friction costs. Most farmers have cash readily available immediately after harvest, but not later in the year when it is time to buy fertilizer. The first experiment helped farmers to set aside money in different mental accounts, by offering the option of purchasing vouchers for fertilizer at harvest time to solve mismatches in timing. The second experiment removed friction costs - the extra costs associated with traveling to purchase fertilizer that represent a seemingly small barrier, yet deter action - by offering free delivery. The results of the randomized control trial showed that 14 percent of the farmers in the control group purchased and used fertilizer, compared to 36 percent in the treatment group that was offered vouchers and free delivery.

The farmers were successfully nudged to act in accordance with their long-term goals. Nudges change the choice architectures that surround us, while maintaining freedom of choice. An example of a nudge is placing healthy food at prominent places in school cantinas. Banning fast food in the same cantinas is not a nudge. Nudges are usually based on careful empirical testing, very often through randomized control trials. Good nudges have high benefits and low costs. In the UK, simple changes to reminder letters on taxes resulted in approximately £210 million additional tax revenue. In Guatemala the most successful variant of a reminder letter increased payments by 43 percent.

Applying behavioural insights and its rigorous monitoring and evaluation frameworks to achieving the Sustainable Development Goals has the potential to better address last mile problems. Many development challenges are not primarily policy-problems, but rather, implementation challenges.

Last-mile problems in this context are the challenges created when users interact with policies and services. The term was coined from the early days of the telegraph. The new technology enabled words to travel much faster than by mail, but the telegraphed messages only reached the end of the telegraph line. They still had to be taken to their respective households – they had to go the last mile. In the context of public policy, last mile problems refer to the failure of adoption of government programmes. For example, scientific innovations that can combat diarrhoea and the distribution of the drugs are first mile problems whereas the proper adoption and use of the drugs are examples of last mile challenges.

In Moldova, many adults treated for tuberculosis discontinue medical treatment. Relapses have significant impacts on the individual’s health and on the economy. To address this last mile problem, UNDP partnered with the UK Behavioural Insights Team and explored the behavioural barriers preventing medical adherence. We found that a major barrier were the mandatory regular visits to a clinic to take the drugs in the presence of a doctor. To test whether more patients would follow-through if allowed to take medication at home while connected through their phone-cameras to a doctor or nurse, a randomized control trial was set up. Early results indicate a 20 percent increase in the treatment group, compared to the control group. An important premise of these experiments is: if you want to encourage people to do something, make it easy!

Another one is: let’s move from ‘raising awareness’ to facilitating action.

In January 2016, the UN Secretary-General appointed two Behavioural Insights Advisors for initially six months. They worked with the UNDP Innovation Facility to improve uptake of an e-waste recycling solution in China, crowdfunding efforts for green energy in Ecuador, the anti-corruption initiative ‘Phones Against Corruption’ in Papua New-Guinea, and more.

Insights from behavioural science that informed the behavioural advice on these initiatives include:

  • Network nudges: people are influenced by the behaviour of friends and members of our extended social network
  • Prompts: people are more likely to take action when prompted at the right moment.
  • Commitment contracts: people are more likely to execute actions when they have committed to achieving a goal, for example by signing a document that states the intention.
  • Descriptive norm: people observe other people’s behaviour as guidelines for what’s acceptable and desirable.

The initial results will soon be shared by the UN Secretary-General’s Office. The World Bank, as well as governments in Australia, Denmark, Germany, the Netherlands, Singapore, the United Kingdom, and the United States have all launched dedicated initiatives to leverage behavioural insights to address policy and last-mile challenges. The work of these teams shows that approaching development challenges with behavioural insights leads to better diagnoses of problems and to better designed solutions. It is time to scale this approach across the SDG achievement efforts.

This article was originally published in Devex.

Benjamin Kumpf Blog post Innovation Development Effectiveness Knowledge management Sustainable development