In my undergrad thesis at Harvard, I tried to do two things:
- Throw a simple model at some ideas suggested by Carl Shulman’s post on upward and downward biases in the estimates suggesting that opening the wrold’s borders would double global GDP.
- Pass an ideological Turing Test on a controversial idea that I believed in quite confidently (way too confidently in hindsight). It definitely served the purpose well.
My take-away from the exercise is that the models underlying huge gains from open borders are extremely non-robust to reasonable considerations. My current view is that promoting relatively large increases in temporary migration from poor to rich countries seems like a promising and under-explored avenue
The economic literature consistently finds that current international border restrictions on the movement of people impose a large cost on the world economy. Lifting these restrictions could thus plausibly increase the world’s GDP by more than 50 per cent. However, the static models underlying these estimates imply flows of at least hundreds of millions of migrants without specifying how long this process takes. This paper introduces a dynamic model to test the sensitivity of the gains from open borders to time-bound effects, calibrating its parameters using multiple empirical strategies. A key finding is that economic convergence between rich and poor countries under business-as-usual growth significantly reduces estimated gains compared to a static model with otherwise identical parameters – from 57 percent to less than 20 percent of global GDP. The exact magnitude of this effect crucially depends on speed of migration and discount rate. The paper concludes by considering welfare implications and dynamic effects of migration externalities.