It's hard to overstate just how much Duflo, Kremer, and Banerjee have transformed the scientific rigor and strength of development economics, and econ in general. Extremely well-deserved prize.
Other fields which have been slower to reform take note.
Other fields which have been slower to reform take note.
.... and yes, I mean you, epidemiology.
Duflo, Kremer, and Banerjee are known as the founding "randomistas," for putting useful randomized experiments into the forefront of development.
But that's really only part of the story. They are also brilliant econometricians (read: causal inference statisticians).
But that's really only part of the story. They are also brilliant econometricians (read: causal inference statisticians).
Much of their work involves what we might call meta-science now: calling into question and testing the basic assumptions required for useful and rigorous science.
Take this paper: nber.org/papers/w8841 (note: author order is always alphabetical in econ)
Take this paper: nber.org/papers/w8841 (note: author order is always alphabetical in econ)
This paper systematically evaluates the extent of bias in standard errors due to failing to account for clustering/serial correlation for difference-in-difference papers, a favorite method for policy analysts.
It's a BRILLIANT paper, should be required reading.
It's a BRILLIANT paper, should be required reading.
I won't go through their whole list of works, but it's like a fireworks show that's gone on for decades, showing variously the failings in how many methods and ideas in econ are flawed or limited, as well as really incredibly creative and useful new ways of using them.
The randomista part is a consequence of those decades of work, not the primary driver.
Some of the most important questions we want answers to just can't be answered with any reasonable degree of reliability, unless we have extraordinary circumstance.
Sometimes, that's an RCT.
Some of the most important questions we want answers to just can't be answered with any reasonable degree of reliability, unless we have extraordinary circumstance.
Sometimes, that's an RCT.
But of course, this isn't just a methods story; it's a REFORM story. Being right isn't enough. Reform involves people. And people are the worst.
I don't know that there are simple, tweetable lessons here for how they managed to do it, but it's clear that they did.
I don't know that there are simple, tweetable lessons here for how they managed to do it, but it's clear that they did.
That's especially impressive, given that Duflo is a woman in a field dominated by a classic boy's club, which was even more extreme when she was getting started.
Fun mini story: I once taught a climbing class with her in it. Didn't realize who she was until after the class.
Fun mini story: I once taught a climbing class with her in it. Didn't realize who she was until after the class.
