Koen Leuveld, Nathan Sivewright, Marie Mallet & Henry Cust


As intended by the organisers the CSAE 2017 conference addressed economic analysis of the broad issues relevant for economic development in Africa. The conference consisted of parallel sessions, plenary sessions and a keynote of “Reflections on evidence in age of fake news and discredited expertise” by Macartan Humphreys from Columbia University.

The keynote highlighted the need for quality data which mirrors the objectives of EDI’s work in Tanzania, Uganda and further afield. The conference took place at St. Catherine’s College, Oxford over 3 days from Sunday 19th March to Tuesday 21st March. A full programme can be found, here.

EDI’s Head of Research, Johanna Choumert-Nkolo presented a paper co-authored with researchers from the African Development Bank and CERDI: “Neighbourhood effects in the Brazilian Amazônia: protected areas and deforestation”. The paper used a strong theoretical model and a spatial econometric model to investigate empirically the interactions between deforestation decisions and the protected areas size in the Brazilian Amazonia. The study shows that “sustainable use areas”, where productive activities are allowed,canincrease deforestation. This contrasts with“integral protected areas” and indigenous lands,  which help in reducing deforestation in their vicinity.

The conference featured many sessions on a wide range of topics, many of which contained useful lessons relating to our own work. This included:

A Natural Resources session which included presentations on the ‘resource curse’ including “Resource discoveries and growth disappointments”, by James Cust from the World Bank and OxCarre, which presented evidence of developing countries consistently achieving disappointing economic growth figures following the discovery of natural resources in the country in relation to the World Economic Outlook growth forecasts from the International Monetary Fund. Stories of Ghana, Brazil and Mozambique were used as key illustrators for the findings.

Prior to this a presentation entitled: “Resource Discoveries and the Political Fortunes of National Leaders’” by Michael Keller from the University of Sussex used survival analysis to estimate the time in office of national leaders in resource rich countries with evidence suggesting the lower the institutional quality at the time of discovery in the longer the leader can remain in power after the discovery taking into account election cycles. This session also contained further presentations on the resource curse and Dutch disease. All presentations involved lively debate in the room around the general ideas being presented as well as the econometric methods undertaken by each of the authors.

Koen Leuveld, Senior Research Officer at EDI, discussing research methods during a break.

A session on Social Networks and Households presented interesting and innovative papers on women empowerment and gender inequalities, a growing research topic in developing countries. Yannick Dupraz from Paris School of Economics presented a paper on the effects of women’s education on polygamy in Cameroon and found that public education increases the probability for a woman to be in a polygamous union. A study on the effect of female early marriage on the norms related to gender roles and rights within the households were also discussed by Zaki Wahhaj from the University of Kent. The paper concludes that early marriage is the main pathway for norm transmission rather than women’s or their partner’s own education or their social network. The researchers used the timing of the onset of menarche as an instrument for the early marriage to strengthen their conclusions.

Finally, two papers presented respectively by Paulo Santos from Monash University and Uzma Afzal from University of Nottingham involved field experiments involving spouses in order to measure women empowerment, such as the bargaining power. The paper from Uzma Afzal concluded women are less involved in the decision-making as the importance of the decision increases. All these studies rely on the analysis of sensitive data and it is crucial data collecting firms put in place specific protocols (e.g. wording of the question, ensure privacy, having separate clusters by sex etc.) in order to collect high quality data. With recent projects related to domestic violence and sexual behaviour, EDI has developed its own expertise on how to deal with sensitive questions in surveys (click here to see our blogpost on the topic).

Other sessions of interest included those on health, which included four presentations from researchers all over the world including Tonny Odokonyero from the Economic Policy Research Centre in Uganda who was presenting a paper called: “Sub-national perspectives on aid effectiveness: Impact of aid on health outcomes in Uganda” which was looking at the allocation of aid within Uganda and its effectiveness on a number of health outcomes based on a populations’ proximity to the aid programme.

The other papers being presented in this session focussed on differing versions of payment for performance, a similar idea to the payment by results method implemented by the NHS in the UK. Papers focussed on case studies from Uganda, Rwanda, and Tanzania and were presented by academics from Wageningen University in the Netherlands, the World Bank and the University of Manchester respectively. After a successful pilot in Rwanda that improved allocative efficiency of health resources, a programme of incentives for co-operatives of community health workers was rolled out nationwide. In Uganda, there has been a pilot of performance based financing of health centres across the country.

A second health related session had a more varied topic of papers. The standout presentations were centred around toxic waste dumping in Ethiopia, by Caterina Gennaioli of Queen Mary University of London, resulting in counter-intuitive findings regarding health outcomes being worse in areas after completion of road construction projects. It was discovered that as road were being built thousands of tons of toxic waste were being dumped as part of the construction. Transmission into nearby water sources was the suggested mechanism for poor health outcomes.

Kiklas Heusch, from the Universitat Pompeu Fabra presented a paper about informal medical providers in developing countries. He used primary data collected by sending in ‘secret shoppers’ into informal health centres in Ghana to assess the advice and drugs that were sold to them, followed by a formal interview. Results from each visit were compared to establish what they know and how they profit finding that in response to certain conditions described, the medical centre had a greater knowledge than they revealed during the secret shopper interview and often sold unnecessary or harmful drugs when they knew that they should not.

Conferences like the CSAE provide a great opportunity for EDI to contribute to the broader research dialogue whilst both sharing and learning from current research in Africa.  The majority of projects that EDI implement are done in East Africa but our team of researchers support a range of projects across the African continent and the wider world. Ultimately the conference allows for some of the EDI team to network with other experts and academics to innovate and evolve our work to further improve research project design, field & data collection methodology and continue to drive up data quality.