· GiveWell has a nuanced discussion of the evidence for giving cash directly to poor households and how this compares to deworming and ant-malaria bednets as interventions they recommend charitable donors should support.
· Lant Pritchett and Shrayana Bhattacharya discuss what cash transfers are and are not good for in the Indian Express in discussing the pros and cons of India’s move towards direct cash transfers: “If the goal of transferring resources to citizens is simply to attain a socially desirable distribution of money and ability to buy things, cash works very well. However, if the idea is to tackle market failures and attain a socially desirable form of behaviour, where administrators allocate benefits to the poorest and the poorest are able to use the subsidy amounts for good nutrition and health outcomes, the idea of cash as a cure-all is problematic.” (h/t @franciscome).
· The IGC has a policy note highlighting the strong early impacts of BRAC’s ultra-poor program in Bangladesh. Nice findings, although the failure to mention the lack of any net impact in a similar program in India leads to overselling the policy implications.
· Lawrence Haddad asks whether the evidence pendulum is swinging back towards the center after swinging too much towards RCTs: an example he gives is a recent paper asking for “Realist RCTs: that is, RCTs which have many arms, which focus on mid level programme variables, operate in several sites, and draw on complementary qualitative analysis.” (h/t @fp2p).
· Marginal Revolution University has an intro guide to RCTs in two parts. Nice example of why we need RCTs through the case of a medical drug (premarin – hormone replacement therapy) and how observational studies led to the wrong conclusions.
· The latest From Evidence from Policy note overviews the program I worked on to get young women into jobs in Jordan.
We will be off until the new year, so happy holidays to all our readers and good luck to those of you doing job market interviews. You can follow me on twitter (@dmckenzie001) for occasional links in the interim.
Join the Conversation