Syndicate content

Weekly links April 13: militant randomistas, show them the germs, should your next paper not be a paper? and more...

David McKenzie's picture
  • A sociology of J-PAL (in French) – I didn’t feel like I learned anything much new from it (J-PAL was started by elite researchers, attracts other elite researchers, has received a lot of funding, and has pushed randomized experiments is about the whole story here), but the author did manage to get an RA to give the following quote: “« En fait, c’est pas vraiment des congrès, c’est vraiment “on va vous apprendre ce que c’est qu’une vraie évaluation”. Ils le disent pas vraiment, mais y a un vrai côté militant derrière, y a un vrai côté “on va vous convaincre de faire des expérimentations aléatoires”. […] Pour le coup c’est vraiment : “faites des expérimentations dans vos bureaux, à la Banque mondiale, à ce que vous voulez”. Il y a un côté “témoins de Jéhovah”, on va aller répandre la bonne parole. Et en interne, y a pas vraiment de discussion... » (Nadine, assistante de recherche au J-PAL).”
  • Econlog summarizes Lant Pritchett’s six bitter pills – politically incorrect research findings that you wished more people cared about e.g. “the variance of economic growth rates is much lower among democracies than non-democracies and hence (descriptively) nearly all episodes of rapid, sustained, (and hence poverty reducing) economic growth were initiated by non-democratic regimes (e.g. Indonesia, China, Korea, Taiwan, Vietnam).” And “my view in development today there is too much attention to inequality within poor countries and not enough to the very low levels such that things are pretty bad even for the, say, 80th percentile.”
  • Cyrus Samii provides R and Stata code for inverse covariance weighting
  • The Economist on how little economists understand about the causes of growth and the importance of going beyond accounting decompositions: “An economist might explain China’s rapid growth in the 1980s by saying that it began to deploy more capital per worker and to adopt foreign technologies. Yet it was very clearly the result of a political decision to loosen state control over economic activity. It would similarly be accurate to say that China’s future growth will depend on how well it develops and deploys new technologies. But that depends on decisions about economic governance taken by its leaders, which will in turn be influenced by social and geopolitical forces that economists scarcely understand and generally ignore. “
  • Jessica Creighton and Wayan Suriastini compare two approaches to identifying women in Indonesia who had given birth in the past year – and find a cheaper key informant method had greater coverage than a door-to-door listing exercise.

Add new comment