Written by Economics is commonly seen as a theoretical branch of the social sciences – one of its own branches, however, is not. Development economics, which deals with growth of lower income countries, provides, in contrast, a very hands-on approach to the application of its models. Central to this branch of economics is the last mile problem.
The Last Mile Problem
Many fail to understand the obstacles towards providing supplies to remote areas in underdeveloped countries. Medical supplies, schoolbooks, and transportation of essential goods (all of which are the basis of human development) cannot be sent to their destinations due to a lack of supporting infrastructure. To illustrate this, let us take the example of a shipment of vaccines to a village in Africa: while it certainly is possible to get the vaccines to the closest regional capital, many times, there is no infrastructure that can safely transport from the capital to the village. This last stretch is the crux of the last mile problem.
Theory vs Experience
Older development economics theories have sought to fight the last mile problem (decades before the term was even coined), but in a standoffish, figure-it-out-yourselves way, post-WWII economists such as Ragnar Nurske saw that development was to be overseen by the nation’s government, investing in many different industries, rejecting trade with other countries. Others preferred neoclassical models, where a laissez-faire market approach would of most benefit, the government interferes the minimum in economic affairs of both individuals and society. However, what most economists from this era failed to understand were the underlying issues that were at the heart of these countries’ underdevelopment: disease prevention and education. These were crucial points that not just theory could solve. They had to be tested on-field.
A revolutionizing and experimental approach to Development Economics was introduced to us in 2019 by Michael Kremer, Esther Duflo and Abhijit Banerjee, awarding them the Nobel Prize in Economics. This innovative approach implemented Randomized Control Trials (RCT) experiments, a reliable way to infer and confirm causality relationships, and so much was achieved. When faced with a certain question/concern, through randomised allocated regions or groups of people, experiments are assigned, tackling the problem in many ways. Afterwards, the outcomes between the different trials with or without interventions are compared. With this experimental process one can assess the impact of social policies on poor and middle-income countries.
So, if in the past, most development economists created their work tucked away in the comfort of their offices, many have now started to complement and base their ideas and deductions in experimental on-field work. In special focus on the work of these three pioneers, their trials were mainly conducted in India and Africa, striving to find new and innovative ways to fight poverty and improve people’s living standards. Some of the problems tackled regarded important old concerns such as health and education.
The impact of randomized controlled trials
When it comes to health, immunization rates are the most predominant issue. According to UNICEF, to this day, around 27 million children and 40 million pregnant women are not provided with a basic package of immunizations, yearly. This leads to many deaths that could have been prevented by more efficient practice and supply of incentives for vaccination.
In India, although vaccination services are offered exempt of any charge, in public health facilities, the immunization rate still remains low in some areas, specifically the more rural and poorer. As a response to this, Duflo and Banerjee randomly grouped 134 villages in rural India into 3 groups: one in which there was a monthly reliable mobile immunization camp, where a nurse and an assistant went to the villages, providing easier access to vaccination services in those villages; another with the same immunization camp, but for each completed immunization, the children’s households received incentives (raw lentils and metal plates); and finally the control group (i.e. with no intervention done).
The results showed that the second group had the most positive outcome, increasing vaccination rates by six times more than before. This was a result of people having easier access to getting vaccinated and incentives that gave a reason to not postpone their immunization. In fact, it ended up being less costly per vaccination to provide these incentives, since the medical professionals needed to be paid regardless of the services being provided or not.
Moreover, efficiency in education, in terms of incentivizing children to attend school, was also empirically studied by these three economists. They started by discarding previous objectives/solutions that came from the assumption that the problem was centred around the poor management and inefficiency of teachers, curriculums, and pedagogy programs, and instead focused on the design of the educational system itself. Thus, in contrast, Pathram and J-PAL, the research organisation founded by Banerjee and Duflo, studied a practical and innovative way to teach primary grade children which prioritized and assessed their actual learning levels (not the specific class year they were in). Instead of constantly flunking children that lacked specific knowledge from previous school years, they undertook another methodology, which was evaluated and found effective, known as “Teaching at The Right Level”. Such method grouped students into the same learning levels and provided them with targeted resources to achieve higher levels. This ended up showing that children were capable of learning fundamental skills such as how to read and basic math in just 30-45 days.
At a 2010 Ted Talk event, Esther Duflo also shed light on how to subsidize education by observing in which interventions a 100$ grant would be more effective in incentivizing extra years of education. Surprisingly, it turned out that just by explaining to people the benefits of education, something inexpensive to do, it increased school attendance by 40 extra years in that community. Furthermore, based on a past field study by Kremer, in regions where it is common to have worm infections, offering deworming treatments to children in schools added up to 28.6 more years, being the second most effective intervention.
A step in the right direction
All in all, the world of development economics was forever changed with the introduction of experimental research. For years, most economic reasoning and foreign aid has derived from a neo-classic economic approach in which it was believed that by increasing GDP per Capita through improvements in market structure, technology and public goods would make people automatically behave in a more socially beneficial way. As a result, the impact of this help was never capable of being accountable or fully efficient and most poor countries remained in the same situation.
However, as one can imagine, these small and geographically exclusive experiments are also restricted by some issues. The broader is the problem we want to solve, the harder it is to draw a specific and effective resolution for it, so this field of study needs to be continuously evolving. Nevertheless, a new collaboration with behavioural economics and experimental approach, through Randomized Control Trials, opened the doors to new ways of unconventional thinking, providing life changing solutions and helped development economists improve substantially causality reasoning.
Team’s note
Dear reader, as we have reached the end of the article and being RCT a very broad topic, if you would like to further explore RCT in the context of economics, then take a look at this article. If you also wish to know what Nudging and RCTs have in common, check this article.
Sources: Brookings, ResearchGate, Journal of African Economies, Habitat for Humanity, Great Britain, J-Pal, MIT News, The Conversation, TED Talk Ideas Worth Spreading, Firstpost, Thebmj, Hindustan Times
Benedita Elias
Guilherme Barroca
theawarenessnews.com 