Borrowing constraints and (seasonal) famines

Imagine that local food crops have very strong and predictable seasonal fluctuations. The obvious way to smooth consumption and avoid famines before new crops is to store food. Suppose now that this is for some reason not possible. Then, you should trade food with another region that has countercyclical crops, or at least storable crops. What if even that is not possible? Then it should be seasonal migration that should bring people where there is food. And what if that does not happen? You have the rural northwestern districts of Bangladesh.

Shyamal Chowdhury, Ahmed Mushfiq Mobarak and Gharad Bryan performed there a randomized intervention by giving a monetary incentive for seasonal migration. While this would obviously improve outcomes, they varied the conditions for the payout to see what would work best: cash or credit, mandating group or individual migration, changing group size, imposing destinations, etc. For once a randomized intervention study tackles the efficiency of the intervention, a welcome change.

While seasonal migration in the control group was 13%, it was 40% in the intervention group, a significant difference, and much more efficient than just informing about wage opportunities elsewhere, which only triggers a 2% increase in migration. The fact that these payments have such a large effect highlights that the true problem is the liquidity constraint these people are facing.

That said, all this depends on the size of this cash payout. With a very large payout, it would not be surprising to see a large response. The financial incentive corresponded to US$11.50, or about 4 days worth of wages in the destination regions. That does not seem very high, and one could hope that with the normal improvement of conditions over time, this problem should be relatively easy to overcome.