Tom McDermott: Testing our Methodologies and Ideologies

Dear Kiari and Niloufar,


Thanks for the articles you contributed:
Kiari for the excellent article you authored.
Niloufar for passing on the letter from the 15 development economists.


We have posted both on the website and will cover them in the next email update.


While not necessarily in conflict, the two articles present a good contrast of views on the three winners of the Nobel for Economics. This is exactly the sort of exploration of an issue that I would like to see more often expressed on our website. Too often we receive an interesting article sent in by a member, but without any opinion expressed. While it is great that we share interesting articles, I like to see people express opinions, including contrasting or conflicting opinions.


So thanks to you both.


I should mention my own prejudice, formed largely by an online EDX-MIT course I took a few years ago on 'the economics of poverty' and taught jointly by Esther Duflo and Abhijit Banerjee. I found parts of the course challenging, mainly due to my only half remembered statistics courses from university days. Banerjee in particular is a 'stickler' for sound statistical analysis. Overall though, I thought the course was an excellent introduction to their 'epidemiological' approach to poverty issues.


They challenged some of the orthodoxy and ideology with which we former field workers often approached issues. On the other hand, much of what they argued would sound familiar to any of us on questions like :
- How do you best popularize ORS? Give out free vs. small price? Cheap wrapper or glossy?
- How do you popularize bednets? Free or charge?
- Do you seek to replace village money lenders with microcredit groups?
- What economic / social issues drive people to use village healers and quack doctors vs. the health center?


My criticism of the course would center mostly on the fact that so much of the field research was for South Asia with only a smattering of work on East Africa. Also that they didn't pay enough attention to the impacts of conflict.


The positives however far outweighed the negatives. They preach a scientific approach which follows any hypothesis with rigorous testing and a clear understanding that history and culture will lead to differing results from place to place. They push for quick and frequent evaluation of results from interventions in place of the costly and time-consuming 'end of project' or end of the 5 year programme cycles so beloved of the UN.


As to the letter by the 15 development economists, I felt they are arguing a rather different point. They seem to say that randomized testing should not become an excuse for constant shifting of funding streams by development agencies. As the authors put it, using such trials "to guid aid spending will lead to short-term, superficial, and misplaced policies." That is not at all what Duflo and Banerjee intend. They just ask that development agencies and NGOs stop trying to push interventions without testing whether they work.


Like you Kiari, I am happily surprised to see economists who focus on issues of poverty at the micro-level are finally winning recognition from the wider economics community, just as a few years ago those working on behavioral economics finally gained respectability.


Best