Writing

Academic

We all have to start somewhere.

📜 Islam and Its Science Deficit (abridged version)

Muslim-majority countries are, on average, much less democratic than the rest of the world, even for their level of income. But a paper by Eric Chaney shows that this the democratic deficit is entirely explained by the institutional legacy of medieval Arab caliphates: once you add a variable for that captures how much of a country's territory was conquered by Arab forces in the Middle Ages, other countries (think Bangladesh, Indonesia) are normal for their income levels. I wondered if the same thing was true for the "Science Deficit" and ran the same regressions, combining Chaney's replication dataset and measures of scientific output from Scimago's country ranking. It found the opposite: Arab-conquered countries that do somewhat better than the rest (though the difference is not large enough to be statistically significant in a small sample). Effect of a Muslim majority hovers around statistical significance depending on the controls included but is consistently negative.

Timeline appendix with some of the relevant history here.

(I would be surprised if no one ran these regressions before but I haven't found any paper along these lines, let me know if you find something better. If anyone knows a skilled scientometrician, I have a couple of ideas for extensions.)

📜 Open Borders in a Converging World (heavily abridged, full version with lots of rambling here)

Economists modelling international migration estimate that we're leaving "Trillion-dollar bills on the sidewalk" by having overly restrictive migration policy. Opening all borders would, depending on the model, lead to somewhere between 50 to 130 % of a GDP boost because of the so-called place premium: migrants are much more productive in developed countries than in poor ones.. My thesis was inspired by Carl Shulman's blog post on potential robustness checks for these models and I decided to test one of them: "If it takes decades, even more than a century, for migration flows to take their course, there will be plenty of time for economic development in poor countries to reduce wage gaps and place premium." I ran a couple of simulations with a simple two-region model previously developed by my supervisor Borjas, extended it into a dynamic setting. I looked at speed of migration in the EU, Puerto Rico-US migration, rural-to-urban migration in China and Gallup world polls to calibrate the speed of migration in the model. I found that, for plausible parameters of catchup growth in developing countries, the discounted benefits are reduced by about two thirds, from 60% to c. 20%.

In reality, I think the benefits are probably higher and remain but it was a useful "ideological Turing Test". I am very bullish on much more liberal rules for short-term migration, which has much less of the downsides that people fear (institutional effects, brain drain).