I am a big fan of the work of Dev Patel. (No, not the actor!) While he has not yet cleared the threshold of notability for Wikipedia, I can say with confidence that he will. Several of his papers have been influential upon me, and I think the world should know about them.
He works in international development, largely in India. One of his papers (and I believe his first chronologically) is “A Rosetta Stone for Human Capital”, with Justin Sandefur. We wish to know how much human capital explains economic development, and by proxy, how much we should subsidize education. Unfortunately, our measures of human capital need not be directly comparable across nations. Simply using years of schooling is inadequate if the skills gained during those years of schooling varies across countries, and it is especially inadequate if it varies in a way correlated with their economic development. Economists have tried to get around this by giving standardized tests, but not everyone takes the same tests.
What Patel and Sandefur do is administer a test to 2000 Indian school children which combines items from several standardized tests. They can use students’ performance on these items relative to each other to convert the tests scores of hundreds of thousands of students from 80 different countries into a common scale. They find that student achievement is higher in richer countries, and that using years of schooling understates the differences in human capital; that even at the same income level, students in developed countries perform better; that countries which have higher test scores export more skill-intensive goods; and that more spending on schools predicts better outcomes in poor countries but not in rich countries.
Again with Sandefur, and with Arvind Subramanian, he points out that the divergence of rich and poor countries seems to be over. One of the main motivating facts of economic development, especially during the endogenous growth revolution of the 90s, was that rich countries were growing faster than poor countries. This shouldn’t happen with a standard neoclassical production function, because capital accumulation has declining marginal returns. To explain this, economists focused on the role of ideas and technological invention, as well as substantial differences in institutions.
Well, that’s all over. Poorer countries have been growing faster, on average, than the most developed countries. Admittedly, this convergence is slow, and it will take 170 years at present rate to close half the gap. However, the volatility of growth rates has gone down, which is good. The reader is directed to Easterly et al (1993), who showed that there was little correlation decade to decade of rates of growth in the developing world. Growth was, in all likelihood, due to idiosyncratic demand shocks for their export commodity, which is hardly the basis for sustained growth.
Another theme of his work is using satellite data to estimate the effects of flooding, which he uses in several papers. “Floods” is extremely impressive in this regard. (I urge you to consult the website laying out the methodology as well). The first advance is simply being able to know exactly where flooding is going on, even when the flood itself is small or only lasts a short time. Using satellite data allows him to have an objective definition of a flood, rather than relying on the possibly varying views of government officials. This is not a trivial concern – he compares two commonly used datasets, which disagree over hundreds of flood events. Neither is river monitor data adequate. For example, the Khulna Division in Bangladesh (which accounts for 15 percent of the area of Bangladesh) has only 47 stations! He is able to validate the data against surveys and news articles, which show that the data is quite reliable.
Armed with this, he can estimate the causal effects of floods. Their effects are large, negative, and persistent. It reduces economic activity, and encourages people to move out of agriculture. It increases schooling rates and migration, and decreases agricultural productivity (again using satellite data to measure crop growth). Surveys of farmers show that those who perceive a higher risk of floods push their children to obtain jobs outside of agriculture. Interestingly, the marginal effect of a flood is less the more floods there have been. The reason is obvious – people adapt to floods the more often they happen – but he totally gratuitously extends the work yet again and uses transportation costs from road networks as an instrument for the cost of moving. People in places from which they would have an easier time moving have smaller reduction in harm from past flooding.
His CV can be found here – I am, as what might loosely be termed a practitioner, quite excited about the forthcoming database of daily flood exposure. It’s a superb pipeline to come, and I look forward to reading them when they come out.
I'm fascinated with the idea of convergence, and usually thought of it in terms of when China will surpass America. But the idea that all poor countries will catch up within 170 years strikes me as a bit too slow, and I imagine that we should be able to speed that process up.
It's so interesting!! The website for the work on floods, so cool to see researchers going beyond pdfs :)
"What Patel and Sandefur do is administer a test to 2000 Indian school children which combines items from several standardized tests. They can use students’ performance on these items relative to each other to convert the tests scores of hundreds of thousands of students from 80 different countries into a common scale." I don't completly get what they are doing there? Are they using an instrument? (sorry if that's very obvious)