Published on Development Impact

Weekly links February 25: what is causing inflation, why some types of causal inference was slow to take off in tech companies, Chinese policy experimentation, and more…

This page in:

·       Noah Smith interviews Emi Nakamura about what is causing inflation at present, the role of micro to macro evidence in macro, and her approach to research and advice for young researchers: “Regarding the path forward, I often find myself an advocate for the merits of "boring" work chipping away at basic questions. Just because a question has been studied before, doesn't mean that we are all convinced of the answer. Multiple studies coming to the same conclusion, using different, and hopefully increasingly convincing,  methods can have a lot of value in cementing our views on a subject”.

·       Scott Cunningham has an interesting interview with Steve Tadelis: Steve is an applied theorist, who was one of the first generation of economists to go and work for tech companies in the Bay Area – he gives an interesting background of how chief economists always used to be economic forecasters, and then how tech companies started down different paths/approaches with using I/O and applied econometrics. He also discusses what types of causal inference have been faster and slower to take off in tech companies, and how to make a convincing case for more than simple A/B testing.

·       Gangsters want to be good people too argues Chris Blattman – with discussion on special interests in general “Everyone gets up in the morning and rationalizes how they are a good person. And if we want to combat that special interest, we need to remember that.”

·       On the Brookings Future Development blog, Misha Lokshin and Nithin Umapathi discuss reasons to be cautious about the potential for using AI in social protection, with different issues to consider such as  “The accountability and “explainability” problem: Public officials are often required to explain their decisions—such as why someone was denied benefits—to citizens. However, many AI-based outcomes are opaque and not fully explainable because they incorporate many factors in multistage algorithmic processes”.

·       On VoxDev, Shaoda Wang and David Yang study the process and politics of large scale policy experimentation in China, looking at how sites are chosen for piloting, how politicians may exaggerate the effectiveness of a policy by allocating more resources to that area while on trial, and how national decisions to scale-up policies do not take these factors into account.

·       If instrumental variables were advertised by big Pharma – a great graduate student skit on Youtube (h/t Peter Hull).


David McKenzie

Lead Economist, Development Research Group, World Bank

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000