Some of our favorite development papers of 2020


This page in:

Development Impact will now be on break over the next couple of weeks for the holidays, resuming in early January. As we have done previously (2018, 2017), we will turn to the holidays with a gift to you of some of our favorite development papers that we came across this year. This year we thought we’d focus on papers related to impact evaluation or measurement produced by some of our colleagues at the World Bank.

·       A cash-for-work public works program in Comoros increased migration to Mayotte (Jules Gazeaud, Eric Mvukiyehe and Olivier Sterck): This paper is important for the debate of whether fostering more employment opportunities at home will deter migration, or whether it will help people overcome liquidity and risk constraints and instead enable more migration to take place. One of the challenges in identifying impacts on international migration is that it is usually such a rare event that it is hard to get enough action to see treatment effects. Here the authors use a sample of over 2,000 households and pay careful attention to measurement to document this effect.

·       Providing more information and monitoring tools for high-risk import declarations in Madagascar improved fraud detection and taxes collected, but less so for high-value items where opportunities for graft are largest (Cyril Chalendard; Alice Duhaut, Ana Fernandes, Aaditya Mattoo, Gael Raballand and Bob Rijkers): we have seen our trade colleagues trying for quite a few years to get impact evaluations underway of trade and customs reforms, and this has been a challenging area to do work. Working together with a service provider that assists the customs agency with IT and risk analysis, they are able to both gain an understanding of the process and administrative data to enable an IV, as well as to conduct an RCT in this area. And like many RCTs, working closely with the customs agency revealed a lot more problems and potential areas for improvement.

·       Paying Bangladeshi factory workers electronically instead of with cash increases formal savings, and the ability to smooth shocks (Emily Breza, Martin Kanz and Leora Klapper): There is a big push towards digital payments, which has accelerated with COVID-19. They work with over 3,000 garment factory workers and see what happens when this switch is made, and disentangle whether it is due to just getting the bank account (it is not) or also having money regularly directly paid into it. It links to the broader financial literacy debate of whether it is better to try to teach people how to do something, or make the defaults easier for them and then people can learn-by-doing.

·       Inferring sources of income inequality from labor supply elasticities (Katy Bergstrom and Will Dodds): A number of studies show that people’s preferences for redistributive policies depend on the sources of income inequality, with support declining as people believe it to be due to preferences rather than productivities. In a paper now forthcoming in AEJ-EP, Bergstrom and Dodds show that observable labor supply elasticities can be used to decompose these two sources of income inequality and then used to simulate optimal taxation taking into account not just productivities but also relative preferences for leisure vs. Consumption.

·       Does maternal depression have lasting effects on children’s cognitive development (Magdalena Bendini and Lelys Dinarte)? Depression during pregnancy and immediately after birth affects a lot of women, particularly in developing countries. Using novel data from the Young Lives longitudinal study and exogenous shocks during pregnancy as an instrument for mental health, Bendini and Dinarte find that children of mothers who experienced depression when they were very young have lower vocabulary (PPVT) scores at age 5, although these effects dissipate by age 8. The size of the effect at age 5 is large (> 0.5 SD). The authors present heterogeneity analysis that suggests that children in poorer households and whose mothers have heavy-drinking partners are more likely to be negatively affected.

·       Providing individual-focused training to mid-level bureaucrats in Ghana improves organizational culture and drives higher performance (Michel Azulai, Imran Rasul, Daniel Rogger, Martin William): increasing the efficiency of civil service workers has deep economic implications; yet, the evidence on how to nurture a culture of results in bureaucracies without creating perverse incentives is scarce. This new paper, part of a wider research agenda led by the Bureaucracy Lab, innovates not only by targeting bureaucrats with substantial discretionary powers, randomizing assignment to different trainings to separate the effects of changing norms vs direct training impacts, but also pushes the frontier on the measurement front and documents underlying mechanisms of these impacts.  

·       Supporting teachers' autonomy through a mix of grants and trainings improves education outcomes for sixth grade students in Brazil (Caio Piza, Astrid Zwager, Matteo Ruzzante, Rafael Dantas, Andre Loureiro): increasing teachers' autonomy is often cited as a lofty goal for educational system, with little evidence of actual programs that worked to promote it or to affect students' outcomes. Enters this new working paper, which shows that inciting teachers to independently develop projects on which to engage their students led to sizeable improvements in students' achievement as well as socio-emotional skills. This program was run at large scale by the State of Rio Grande do Norte, highlighting the scalability of this type of programs. This allows the authors to document underlying mechanisms, such as a reduction in teachers' turnover.


David McKenzie

Lead Economist, Development Research Group, World Bank

Kathleen Beegle

Research Manager and Lead Economist, Human Development, Development Economics

Berk Özler

Lead Economist, Development Research Group, World Bank

Markus Goldstein

Lead Economist, Africa Gender Innovation Lab and Chief Economists Office