Affording Not to Farm: Remittances after a Large Natural Disaster

with Prakash Pathak and Agnese Romiti

Working Paper · Corresponding author

Combining spatial variation in 2015 Nepal earthquake intensity with exchange-rate-induced remittance variation in a triple-difference IV design, we find that higher remittance inflows crowd out farm production in the short run while consumption remains stable.

Triple-difference instrumental variables
Pre- and post-earthquake household survey data, earthquake intensity, and exchange-rate-induced remittance variation
Nepal, around the 2015 earthquake
  • Higher remittance inflows in high-intensity earthquake areas crowd out farm production in the short run, while consumption remains stable.
  • Higher inflows also make households less likely to farm at all.
  • The decline concentrates in grains and meat, and runs through capital inputs — chiefly livestock upkeep and seed — together with households' own agricultural labour.
  • Remittances do not crowd out public aid. Non-remitting households instead lean on it in the absence of transfers from abroad.

Abstract

We study how remittance responses to a large natural disaster affect farm production, the value of crop and livestock output. We combine pre- and post-earthquake household surveys from Nepal in a triple-difference instrumental-variables design that interacts spatial differences in earthquake intensity with exchange-rate-induced shifts in the value of remittances. We find that in high-intensity earthquake areas, higher remittance inflows crowd out farm production in the short run, while consumption remains stable. Higher inflows also make households less likely to farm at all. The production decline is concentrated in grains and meat, and the adjustment runs through capital inputs, chiefly livestock inputs, handling and upkeep, and seed, together with the value of households’ own agricultural labor. Remittances do not appear to crowd out public aid. Non-remitting households instead lean on it in the absence of transfers from abroad. From a policy perspective, remittances appear to ease short-run budget constraints enough that recipient households can afford not to farm, scaling back production while smoothing consumption and without drawing on additional public support. They are therefore not a substitute for policies aimed at sustaining farm production after a large shock.

Effect on log cropland area by year, interacting earthquake intensity with the migration shift-share. Estimates are flat before the 2015 earthquake and turn negative afterwards.
Effect on log cropland area by year, interacting earthquake intensity with the migration shift-share. Estimates are flat before the 2015 earthquake and turn negative afterwards.