Industrial Policy with Development Characteristics: Fertilizer Policy in Times of Crisis

We study time-consistent optimal policy when undiversified owner-operators face financial frictions and a planner with limited instruments. We apply it to fertilizer subsidies, one of the largest sector-specific policy instruments in the developing world and whose price is becoming increasingly volatile. We collect household-level data from 444 rural Rwandan villages from 2020 - 2024 and exploit the doubling of fertilizer prices after Russia’s invasion of Ukraine. Relative to less fertilizer-intensive villages, more fertilizer-intensive villages experience 30 percent lower fertilizer spending, 21 percent lower harvests, and 11 percent higher output prices. These patterns discipline key model elasticities. The optimal response balances a desire to reallocate production toward less fertilizer-dependent villages, while dampening the consumption loss in more fertilizer-dependent villages. Quantitatively, these forces offset and the subsidy remains at 10 percent before and after the shock.