· Mariajose Silva-Vargas (the development economist and photographer who took the pictures we use in the headers on this blog), has a photo exhibit on “The Journey of an Impact Evaluation” which I saw on display in Paris this week. This chronicles the different phases of an evaluation of a refugee integration project in Uganda. She has also posted some of the pictures and story on her photography page – and here is an older blogpost I did on using photos and videos to communicate your research for others thinking about doing this.
· For your summer reading, Clément de Chaisemartin and Xavier D'Haultfœuille have posted a draft in progress of a textbook they are writing on TWFE and DiD with heterogeneous treatment effects and imperfect parallel trends. They note how common these designs are, with 26 of the 100 most cited papers published in the AER between 2015 and 2019 having at least one TWFE regression to estimate the effect of a treatment on an outcome. They summarize the recent literature and provide Stata and R guidance. There is some guidance of how to choose between all the new heterogeneity-robust DiD estimators, but it is currently still a pretty technical read, I think targeted at advanced graduate students or researchers already familiar with these methods and trying to understand how they all fit together. Looking forward to seeing this evolve. I’m also intrigued by the “coming soon” sections titled at the end of the manuscript, such as panel Bartik designs and sequential randomized experiments.
· Do you not have a panel but still want to do difference-in-differences, so are tempted to ask participants to recall living conditions from several years before? Rob Fuller summarizes a recent paper of his which looks at how accurate such recall was in an evaluation in rural Ethiopia, finding “that the information that survey respondents gave when we asked them to recall their baseline situation was much more closely correlated with their situation at the time of the survey than it was with their actual baseline situation. In other words, they tended to underestimate how much had changed since the time of the baseline.” – which can badly bias results if you control for these retrospective measures. It reminded me of work by Martin Ravallion which showed similar concerns with what he dubbed “shoestring evaluations” for an evaluation in China.
· On the CGD blog, Jacobus Cilliers and Nompumelelo Mohohlwane summarize their experiments in South Africa which look at the impacts of teaching incoming primary school students in their home language versus in English. Teaching in the mother tongue boosts both mother tongue and English literacy, but teaching in English actually had a negative impact on literacy in the mother tongue, and not a significantly bigger impact on English literacy than teaching in the mother tongue.
· Orley Ashenfelter is doing a podcast series called “The Work goes on” which interviews leading labor economists. Here’s the transcript of his interview with Jim Heckman, talking about Heckman’s college days and how he turned down Cornell, Stanford, Yale etc to go to Colorado College, and this great quote from Orley about something Al Rees, who taught Heckman at Princeton in grad school, is meant to have said “And I once heard him say, I thought I heard him say, but this could be apocryphal in my memory now, that the best class he ever taught in labor economics with the highest quality students was the class you were in, and in fact, he said you were the only one in it.”
· Lant Pritchett on what types of policies/work he is actively for as a development economist: few would argue about being “for” rapid and sustained growth in broad based labor productivity,..higher levels of state capability…effective education for all” and I also am a big supporter of his fourth item, “labor mobility”. But making progress/knowing what to do on each of these without having a magic wand is tough – would be good to see what he is for in terms of actions/policies, rather than just desirable outcomes.
· NEUDC 2023 will be at Harvard on November 4-5. Paper submissions are now open and due by August 17.
Join the Conversation