“Divide and conquer,” also to fight against poverty
Economic science has often tried to solve poverty from large macroeconomic initiatives, whose direct effect on communities is not easy to measure, and which are also usually contaminated by political and ideological biases almost inevitably. The great contribution of Banerjee, Duflo, and Kremer to the developing economy is to approach it from below, studying specific problems that can be measured at the microeconomic level in the affected communities. Thus, they discovered that low productivity problems in developing countries are mainly due to large productivity inequalities within the country itself, while in developed countries, the level of productivity of the same factor is usually similar (and close to optimal) for all agents.
The low productivity in poor countries is not a problem with a general solution but is due to large differences internally.
For example, in a developed country, practically all farmers have access to modern machinery, fertilizers, etc., making their productivity not only high on average but also there is little variation between them. In contrast, in modern developing countries, modern farms with similar productivity to those of the developed world coexist with others that are much more technologically backward, which makes the average level much lower.
The approach of the winners is not, therefore, to propose policies to ‘eliminate inequalities’ among those with better and worse performance, but identify the latter, determine a series of possible causes and design experiments to see if mitigating these causes their performance really improves. This way of breaking down the problem has not only served to identify possible solutions but also to demystify others that were considered wrongly effective.
Improving the results of poor schools is not just a matter of resources
In one of his first experiments, Kremer checked whether the poor performance in the worst schools in Kenya was due, as was believed, to the scarcity of resources of the schools and of the families themselves, for being in poor areas. They tried to provide free textbooks in some and free menus in others and compared their results with those that had received nothing. The menus only reduced absenteeism but did not improve performance, while free books also did not improve performance except for the best students (that is, those who already performed at the level of good schools).
The experiments of Banerjee, Duflo, and Kremer demystified the scarcity of resources or the ratio of students per teacher as the main cause of poor performance in poor schools.
Discarding the shortage of resources as a cause of poor performance served to focus on other problems and eventually discover that incentives for teachers (who in poor schools were less paid and much less motivated) were a really effective way to improve performance — general level among all students. Banerjee and Duflo continued Kremer’s experiments in India, discovering that another highly effective measure to improve the overall performance of a school was personalized attention to students with learning disabilities(which in the poor schools were much more numerous) thus allowing the class as a whole to advance further. Based on this evidence, India has implemented programs to assist students with difficulties in more than 100,000 schools.
The results of Banerjee and Duflo also demonstrated that a key problem in the majority of poor Indian schools was the underutilization of one of their key resources: teachers, with high levels of absenteeism and low dedication. They demystified the ratio of students per teacher as a root cause of poor school performance: they compared schools where the ratio was reduced without changing the working conditions of teachers with others where the ratio was maintained, but teachers’ contracts were linked to performance improvements of the students. In these last schools, grades improved, but not in the first.
Vaccines in exchange for lentils: how to drastically improve children’s health with minimal investments
Another important field of research for winners has been child health, and specifically, the reasons why many poor families neglect basic interventions such as vaccination or deworming, whose positive impact is brutal. In most cases, it was not due to a lack of access to these treatments, but simply to the lack of adequate incentives.
The empirical results showed that poor families were more price-sensitive than estimated. In a study in Kenya, Kremer compared schools where treatment against intestinal parasites was offered for free with others where the treatment was symbolically priced and totally affordable for affected families ($ 0.40). In the former, 75% of the families decided to treat their children, in the latter only 18%. That is, in these cases subsidizing basic medicines completely is dramatically more effective than doing it almost completely.
Banerjee and Duflo conducted similar experiments with vaccines. They discovered that mobile vaccination programs were not only cheaper but much more effective than fixed vaccination centers. Apart from absenteeism in the fixed centers, there was another problem: parents tended to postpone the decision to vaccinate their children, and in the end many ended up not doing so, while mobile campaigns forced them to make a decision with a limited time horizon (Banerjee and Duflo continued this research path to demonstrate that expired subsidies were much more effective than permanent ones). Additionally, some of the mobile vaccination units offered a symbolic incentive (a bag of lentils) for each vaccinated child.
Something so simple and cheap made the vaccination rate shoot at 31% (compared to 18% of mobile units without incentives and only 6% of fixed centers), thus reducing the unit cost of each vaccine.
In short, Banerjee, Duflo, and Kremer have demonstrated with their work that the generic concept of ‘poverty’ can be broken down into minor problems on which empirical experiments can be performed and measured, thanks to which small interventions can have dramatically positive impacts on the poorest communities.