IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.



Happy 100th links everybody. Thanks for the wonderful feedback and to Chris for loaning out his space, and to Jenn Cowman and Cara Vu. On to the links:


  • An interesting investigation from one of the more interesting development writers, Francisco Toro. He used a reporting fellowship to go to Uganda and investigate why farmers there keep using old, less productive seeds and technology rather than more productive hybrid seeds used in many parts of the world. He finds both metaphorical and literal market failures. There are newer seeds and better fertilizer readily available, but the markets are flooded with adulterated and counterfeit hybrid seeds, and farmers quickly learn not to risk it. One farmer bet his son’s education money on “better” soybean seeds to find that only 20 percent germinated.
    • Why doesn’t the government monitor the seed quality? Toro (with an assist from Lant Prtichett) concludes the Ministry of Agriculture has evolved to be dependent on foreign aid groups, and ends up being not terribly connected to the farmers. (It is worth noting the irony that the private firm he mentions at the end as having the tech solution seems to have been caught faking soil tests in the U.S.)
    • Bold, Kaizzi, Svensson, & Yanagizawa-Drott confirm the Uganda counterfeiting problem in a forthcoming paper in QJE. They tested seed and fertilizer purchased in Ugandan markets and found 30% of fertilizer nutrients missing, and 50% of seeds counterfeit. They propose that diluting batches, rather than selling 100% fake ones, confuses the signals enough that farmers’ yields are inconsistent. The result is that farmers stay hungry but the markets don’t collapse. (Some of the same authors had a paper on fighting counterfeit anti-malarial meds by introducing cheaper real ones.)
  • A look at the “Replicability Index” of studies Daniel Kahneman cited in the bestseller Thinking, Fast and Slow about priming (changing behavior based on subtle or non-conscious messages). Calculations based on effects and sample sizes of the studies show they’re all a little too good to be true, and sure enough many of them didn’t replicate. (h/t David Batcheck)
    • This finding isn’t a surprise, but it’s worth reading Kahneman’s eloquent response, admitting he fell victim to a phenomenon he himself studied:

      As pointed out in the blog, and earlier by Andrew Gelman, there is a special irony in my mistake because the first paper that Amos Tversky and I published was about the belief in the “law of small numbers,” which allows researchers to trust the results of underpowered studies with unreasonably small samples. We also cited Overall (1969) for showing “that the prevalence of studies deficient in statistical power is not only wasteful but actually pernicious: it results in a large proportion of invalid rejections of the null hypothesis among published results.” Our article was written in 1969 and published in 1971, but I failed to internalize its message.

He points out that he still believes that one should believe the preponderance of published studies, but that he didn’t understand the extent to which only positive results were being published. (He should also get credit for being one of the forces behind the replication push).

  • Bill & Melinda Gates’ letter is structured as a reply to Warren Buffett and is very well written and communicated. One message is that the world is getting better, with fewer children dying every year. They calculate 122 million children have been saved since 1990.
  • Economist Jishnu Das wonders if econ researchers are really the best people to be in the policy communication/recommendation business. Among the points he brings up is if their comparative advantage is in communication. He also wonders about  the focus on finding effective things and push to scale them up, compared to stopping things we’re not sure about.

Which still brings us to wonder how economists ended up near the top of the Y axis here (h/t Charles Kenny):