· Dan Björkegren on how could AI impact developing economies? 5 years ago I wrote a post on the use of AI and ML in development interventions and impact evaluations which noted that many of the uses were for improving measurement and improving econometric estimation, with far fewer on improving the interventions themselves. Dan’s post looks at how recent advances in LLMs and other AI technologies could impact economies through avenues such as cheap, tailored expertise (e.g. targeted advisors and chatbots); through changing which sectors and industries have comparative advantage (e.g. perhaps lowering costs of exporting and learning about foreign consumers); improving education quality with AI tutors; etc. He also notes the current limitations when it comes to using these technologies in poor countries with more limited internet access and less training data.
· A series of three e-books on Thriving in Economics, with volumes for aspiring PhD students, PhD students, and assistant professors, is now out with all proceeds going to Ukraine relief. The table of contents looks great for each one, with lots of bite-sized pieces of advice. Here’s a little sample, with organizer Tatyana Deryugina offering great advice on getting started as a new assistant professor: “Adopt the mindset of “my learning is just beginning”… The best way to learn is as you go. Remember opportunity cost. Reading an entire textbook on econometric methods to brush up on the latest techniques will probably not be the best use of your time. Reading a few chapters on techniques relevant to your research may well be worth it. Don’t try to anticipate what you will need to learn, just do it on an as-needed basis. If you come across resources that look like they could be useful in the future but currently aren’t, save them to a folder/document for the occasion and move on…There should be limits to what you’re willing to learn. If you’re a theorist with an idea for a paper that has a substantial empirical component and you’ve never touched data, you are probably better off finding an empirical coauthor rather than trying to learn data analysis. If your project straddles two literatures, one of which you are clueless about, consider finding a coauthor who is active in that literature. Academia is a team sport”
· The second part of the Ideas Untrapped interview with Lant Pritchett has him discussing school systems and exams and the need for political commitment to really improving things: “This big research project, RISE…we included Vietnam as one of our focused countries because it was a success case….we really wanted to answer the question, how did Vietnam do this? Why did they succeed? …And five years into the research effort…they had produced a bunch of empirical research of the econometric type. Is Vietnam success associated with this or that measurable input? Nothing really explains Vietnam at the approximate determinant input level. And finally, one of the researchers said to me, Lant, we’re trying to get around the fundamental fact that Vietnam succeeded because they wanted to….I can’t go back and tell the British taxpayers that they spend a million dollars for a research project on Vietnam, and the conclusion to why Vietnam succeeded was because they wanted to…on another level, it’s a deep and ignored truth….Everybody wants to assume it’s a technocratic issue, it’s a design issue. I think the fundamental problem of these failing and dysfunctional education systems, it’s a purpose problem.”
And it wouldn’t be Lant without some strong opinions: “conditional cash transfers are just stupid in a trivial way…Charter cities are wrong in a very deep and sophisticated way….people promoting tech in education are promoting the pornography of education rather than real education”.
· Replicability and the use of macro variables: in a new working paper, Iasmin Goes notes that macro variables like GDP, inflation, unemployment etc get revised – so the same analysis conducted on the same time series can end up producing different results depending on what vintage/version of the WDI/Penn World Tables/etc you use. For example, running the exact same cross-country regression to look at the association between the genuine savings rate and trade/GDP ratio results in different magnitudes and even different signs depending on which version of the data are used. “These findings have two practical implications. First, researchers should always be transparent about their data sources and vintages. Second, researchers should be more modest about the precision and accuracy of their point estimates, since these estimates can mask large measurement errors.”
· The NYTimes makes a video and gets Alexander Skarsgard to explain the 610-page Dasgupta report on the Economics of Biodiversity (full report) and summarizes the main message in 5 words “pay for what we use”.
· More detailed videos that would be great for teaching or getting students interested in how history can be combined with development economists are in the CAGE research centre series on “why isn’t the whole world developed” – e.g. Are Africa’s colonial borders holding it back? and is colonialism responsible for medical mistrust in Africa? The main pages also have teaching notes for using this in class with 14-18 year olds – e.g. here is the page for “Why is Africa so poor?” based around Nathan Nunn’s work on the slave trade.
· And for your next causal inference class discussion – the Washington Post headline “Watching live sports in person may be good for you, researchers say” “New research connects viewing live sporting events with higher levels of life satisfaction and lower levels of loneliness — and researchers say live sporting events could be used to improve public health” – and the having it both ways of communicating these results. First “The researchers were careful to point out that the data doesn’t mean watching live sports actually causes those gains. But the association is worth exploring further, they say” and then a sentence later ““Our findings could be useful for shaping future public health strategies, such as offering reduced ticket prices for certain groups,” says …the study’s lead author”
Join the Conversation