Science - December 16, 2010

Experimenting in Africa

Development economist Erwin Bulte bagged a Vici grant from Dutch research organization NWO last week. He is to receive 1.5 million euros to find out which agriculture projects in Africa work, and why. 'Without learning aid will dry up.'

Erwin Bulte: 'You have to have a story for the people who are paying for the aid.'
We spend billions of euros on development projects without knowing whether it is money well spent or not. We do not know enough about which projects succeed (and which do not) and why this is. It is high time we spent our development funds effectively.
Is this the new Dutch government speaking? Or the World Bank? Or Erwin Bulte, winner of the Vici grant from the NWO? The correct answer is: all three. 'The zeitgeist is on our side', says Bulte. 'Even development organizations are approaching us to ask if we can find out what impact their projects are having. You have to have a story for the people who are paying for the aid. Show them: see, this is what has been achieved. Without a story and lessons learnt from previous projects, aid will dry up at some point. What is more, it is difficult to accept that a lot of money is spent in ways that do not have much impact.'
Agriculture plays a crucial role in economic development in Africa, Bulte is convinced. Over the next couple of years, the economist is going to find out what sort of impact institutions have on agrarian development.

Institutions?
'Institutions are the rules that enable people to make good investments. It is a broad concept: they are the rules of the game. The legal framework is important, for example. If the land I am tilling is not my own and can be reclaimed next year, that makes it difficult to invest and I will want to make a quick profit. If I am a small farmer in a rural area, artificial fertilizer may well be too expensive. If I link up with other farmers and purchase jointly, it becomes feasible to use it. Or is it collaboration with buyers in the chain that will lead to investments in development? There are numerous different theories about this. I want to know which rules are important for agricultural development and find out how that relates to development projects and government policy.'

How do you find that out?
'One of the ways is to do experimental research. In Liberia we are going to carry out a project with the aid organization CARE, in 130 villages. In half the villages, CARE will implement an intervention and in the other half we shall do nothing. Normally, an aid organization waits to see which villages want to participate. But we are going to allocate the villages on a random basis - that is necessary for the experiment. The intervention will involve things like CARE organizing farmer cooperatives in order to create advantages of scale and empower farmers in relation to traders and suppliers. We will then look at whether an intervention of this kind works.'

Development aid has been studied before though, hasn't it?
'Yes, of course, but development projects have never been set up systematically as an experiment. You need to do that in order to be able to evaluate rigorously. Which villages participate in our experiment will be decided by chance. That provides a far more reliable picture of the average success than you get when a development organization makes its own plan and villages can sign up for it on a voluntary basis.'

Do you measure development by income or by behaviour?
'We measure the impact using questionnaires but also by playing games with real money. In Burundi we are conducting an experiment to measure the levels of mutual trust among the villagers - trust is a very important condition for economic development. We give people an envelope with money. They can keep it, but they can also give the money, or some of it, to other people in the village. We do that for them and they don't know who gets the money. When we give away money we triple the amount. The recipient is also given the option - anonymously again - to give away some of the money. So people have to consider: if I give away money, will I then get some back from my fellow villagers?'

What is achieved by a game like that?
'When markets develop and people start thinking more commercially, trust increases. At first you see it mainly within the family, and then it spreads to strangers. That is important if, for example, you want to buy something on credit. Without trust a society stays stuck at the flea market level - fair exchange. We research the factors that explain mutual trust and could therefore contribute to more effective aid.'

Improving development aid with five researchers: isn't that a bit ambitious?
'We cannot achieve anything alone. We dovetail this research with research on institutions and agricultural development done by the International Food Policy Research Institute (IFPRI). That project is much bigger. We focus on African countries that have recently had a civil war. We also link up with ongoing projects by development organizations, including some of the World Bank's. I will soon have four PhD students in Africa researching development at village level. And then I will look for a Postdoc who will analyse the relation between agricultural development and institutions at the macro level, and work with me on the links with the micro-studies. That should generate broader findings on which interventions in agricultural development raise a farmer's income.'


Erwin Bulte has been in the top ten for publications by Dutch economists for years. In 2007, he came to Wageningen from Tilburg as a 'high potential'. He became professor of Development Economics and already had a Vidi grant from the NWO in the bag. He used those funds to research the 'resources curse': the idea that development countries suffer more than they benefit from their mineral resources, because of the civil war and corruption they provoke. When he began the research, he believed in this curse. Five years and fifteen articles later, he is a critic of it. He wrote about this in Science. The Vidi grant ended this year. Time for a Vici, the last of the trio of NWO grants, Veni, Vidi, Vici.
Last year, his research proposal for a Vici grant was rejected by a narrow margin. This year's proposal was assessed as excellent and he won the grant. 'I worked on it fulltime for three weeks. The research had more focus and coherence.' The money will enable Bulte to appoint three PhD researchers and a postdoc, which will help him develop a clear research line in the chair group. 'We have about thirty PhD students in the group. But because of the external funding, there is no real research agenda to speak of. The Vici grant will enable us to get a group of researchers to work together intensively.' 
 
 

Re:act