Over the past few weeks, the BI4GOV team in the Western Cape Government connected with leading behavioural insights (BI) practitioners from across the globe, hailing from research organisations, government, academia, and multilateral organisations alike. We set out on a journey to explore the application of BI in “low-resource contexts”, to understand the challenges that practitioners in these spaces face and provide guidelines and practical tips to overcome them.
Although the objective seemed straightforward in our minds, subsequent conversations quickly revealed the challenges of developing a systematic, one-size-fits-all framework. As it turns out, even the term “low-resource contexts” itself is not well-defined, taking on different definitions amongst different people. For example, a lack of resources may refer to resources in the broader, systemic environment, such as poor physical infrastructure or low technological penetration. Alternatively, a lack of resources may refer to organisational limitations, in terms of operating budget, team capacity, or human capital. The term “low-resource” might also seem analogous to developing countries. But in reality, is not necessarily so; in fact, practitioners operating in low-income or rural neighbourhoods in the developed world might face similar challenges.
Given the ambiguity of the term “low-resource” and the diversity of behavioural insights work that qualifies as taking place in a “low-resource context”, we bid adieu to that term for the time being and focussed on BI work in developing countries. Below are some key themes that emerged from our global conversations.
What we learned
Although developing countries are similar in that they are all challenging settings, the specific challenges they each pose significantly vary.
For example, recent social cohesion projects in Pakistan conducted by the World Bank needed to be gender-sensitive, ensuring the separation of male and female fieldworkers and participants where necessary. While gender may not be as critical a consideration in every developing country, other factors will arise, and practitioners need to juggle a myriad of idiosyncrasies in each context. Sometimes, similarities do emerge across societies; for example, two completely unrelated education projects in Peru and South Africa both cited a need to overcome the lack of physical infrastructure in remote and under-resourced schools. However, more often than not, developing countries present various permutations of the same challenges. This also makes sense after considering the sheer social, economic, political, and religious diversity of countries broadly categorised as “developing”, from Indonesia to Nigeria to Colombia and to Bangladesh, to mention just a few examples.
Although the above seems like a fairly intuitive takeaway, especially since contextualisation is one of the foundational principles of behavioural insights work, it highlights the challenge of creating a framework for BI work in developing countries without over-generalising things. What practitioners have identified to be possible and helpful, however, is more rigorous documentation of individual contexts as uncovered through each completed project. James Drummond from the OECD shared with us that while his team spent a lot of time understanding local contexts in a project on safety regulation involving Mexico and Oman, they didn’t spend much time documenting those cultural differences or exploring their underlying drivers. A deeper understanding, and a record of that understanding, would have helped with longer-term backtracking as well as any subsequent future projects. As BI work in developing countries continues to grow, thorough documentation of different contexts and their nuances could save practitioners valuable time in the planning phase, eliminating the onerous task of starting from scratch.
The diffusion of BI work and innovation, more broadly speaking, still flows in one direction: from the West to the rest.
Although the field of behavioural insights, especially as it pertains to public policy, emerged from Western countries such as the UK and US, it is now being deployed across the globe. Developing countries have been quick to adopt and even lead the field as pioneers. For example, the Western Cape Government in South Africa began with behavioural insights projects in as early as 2012. However, the agenda for BI work today – tools, methodologies, areas of focus; what is popular now, what will be popular in the future – this is still commonly set by those in the West. Practitioners across the globe still largely look to projects emanating from the West as inspiration for their own initiatives, but this is not necessarily the most beneficial or useful approach. There are significant differences between Global North and Global South countries in how society is configured and society’s most pressing challenges. Rearranging produce displays in supermarkets to encourage healthy eating, for example, loses impact and relevance in a nation where supermarkets are not used by the masses.
As Ammaarah Martinus from the Western Cape Government shared, “There are pockets of excellence [in BI work across developing countries], but there isn’t a deliberate effort [to share and collaborate with each other]. Right now it’s more of a push towards being part of the West versus being part of your own.” There needs to be greater effort to strengthen the network of BI practitioners in developing countries, whether that involves academics with significant fieldwork experience or local governments just looking to start. A stronger community would foster a greater sense of shared identity and encourage knowledge dissemination and open communication. It would also allow practitioners in developing countries to take inspiration directly from each other and encourage innovation to flow more openly, in multiple directions.
A variety of players and operational models exist in the BI space in developing countries.
Deploying BI projects in government requires a multidisciplinary team – policymakers, behavioural scientists, data specialists, project managers, and field staff. Each member is a critical piece of the puzzle; policymakers understand the regulatory and societal context, behavioural scientists bring a deep understanding of human behaviour and empirical research methods, and the remaining players coordinate tactical details on the ground including logistics, impact assessment, and partnerships. There are also a variety of ways to build up a team, whether that is hiring for a complete, full-time internal nudge unit or simply crowding-in expertise and talent from partners such as ideas42, BIT, the World Bank, OECD, or Busara, to name a few players in the space. Collaboration with universities, research centres, or even individual academics is also common.
There is no one right answer; across the landscape we’ve seen a variety of operational models play out depending on organisational requirements, capacity, and comfort level with BI. Regardless of the model however, one common theme is the importance of partnerships. BI projects especially in developing country contexts depend on multiple stakeholders to pool together skillsets, perspectives, and resources to truly achieve success.
Despite the diversity of hurdles faced in every individual context, data emerged as a common challenge.
From Peru to Lebanon to South Africa to India, a consistent struggle faced by practitioners from developing countries was data. Data would be inaccessible, outdated, unsuitably formatted, or in many cases, simply unavailable. The unavailability of data also knows no bounds; from employment to school enrolment to even something as fundamental as citizen identification. As Dilip Soman of BEAR shared, “In India, you can develop interventions but then are unable to deliver them. A large portion of your target population have no identity cards, no home address… They are essentially invisible to the government. Ironically, the people who need the most help are often those that are the hardest to reach.”
Data collection itself also becomes a formidable challenge. In many developing countries, you can’t just, for example, run a Qualtrics survey and distribute it electronically as one might in the UK. Data may need to be collected in the field by a large number of field staff using good old-fashioned pen and paper. As a result, data accuracy and completeness become a concern. Staff may make recording errors, records may get lost, important details may slip through the cracks as coordination becomes increasingly complex. Mobile solutions such as tablets have emerged as a promising solution, but these challenges still persist.
What is the way forward?
BI is more than just nudges and randomised control trials (RCTs); it should be seen as a policymaking toolkit.
Although this movement has universal application, it has especially profound potential in developing countries, where nudges may not be the most suitable intervention and RCTs are simply too expensive to feasibly conduct. BI work makes a general case for rigorous, empirical impact assessment, which has potential to improve monitoring and evaluation in governments even if not deployed in a format as extensive as a RCT. BI also encourages a comprehensive methodology that has potential to influence all stages in the policy cycle, including policy development. BI also encourages a deeper understanding of human behaviour, which has the potential to not only change external service delivery but also attitudes and culture internally within organisations. All these possibilities are exciting development areas as the field of behavioural insights continues to grow.
Behavioural insights needs to be mainstreamed into organisations rather than being seen as a separate add-on.
We’ve heard from several practitioners that without sufficient buy-in and commitment, projects that once began with lofty ambitions and promising results fall flat of their potential. Pilots never scale, partnerships fizzle out, and BI becomes a back burner “nice-to-have”. As a result, there is an absolute need to mainstream behavioural insights into an organisation to achieve longer-term sustainability. Again, this phenomenon is not unique to developing countries but rather universal. No behavioural insights project should be fully reliant on a third-party affiliate or entity regardless of how new the work is to the organisation. Internal capacity building should be top-of-mind throughout the entire process. In government in particular, every civil servant designing and implementing programmes, should have some knowledge of BI in their toolkit. The future of BI is to see it as part and parcel of the policy and planning cycle.
BI work in developing countries will grow with a change in how and where information is communicated.
Developing countries could do a better job in how they communicate their projects. Specifically, more could be done to synthesise discoveries and learnings into broader concepts that have relevance and reach across the world including in the Global North. For example, the City of Cape Town in South Africa recently narrowly averted a water crisis, and behavioural interventions played an important role in steering the city away from Day Zero where taps would completely run dry. The lessons learned from that initiative are not only relevant to drought-prone zones, but also to communities around the globe as water conservation is an important element of environmental sustainability. By changing the narrative and generalising outcomes to be more globally applicable, BI work in developing countries would not be merely seen as “interesting” by the Global North but also as highly relevant and useful. A change in the narrative – the how – of communication would go a long way.
At the same time, we also need to think about the “where”. A significant proportion of BI work is still being disseminated through academic journals, especially work that is being taken seriously and respected amongst the BI community. Academics, who play an integral part of the BI puzzle, are often under pressure to work on projects that seem highly promising and highly publishable. As a result, interventions which may have huge societal promise but are not publication-friendly for whatever reason may not receive the support and attention they deserve. As incentive structures in academia evolve and behavioural science talent outside of academia grows, BI work in developing countries too will proliferate.
Over the next few months, we will be releasing more articles detailing our conversations with global behavioural insights practitioners and shedding further light on these themes. Stay tuned!
Icon made by Free icons from www.freeicons.io