Heterogeneous Effects of Weather Shocks on Food Prices in a Country with Diverse Agriculture

Abstract

We study how extreme weather shocks affect food prices in Colombia, a country with exceptional geographic and agricultural diversity. We exploit detailed information on production networks and wholesale prices of 40 crops over a nine-year period, as well as granular rainfall and temperature data, to estimate the local (crop and market-level) and global (countrywide) effects of weather shocks. Our findings reveal that global shocks drive most of the price effects, with local effects remaining small and often statistically insignificant. During El Niño episodes, global high-temperature and low-precipitation shocks are the main drivers of price increases, whereas during La Niña episodes, global high-precipitation and low-temperature shocks dominate. We also find very heterogeneous effects across crops. A machine-learning analysis reveals that this heterogeneity is driven primarily by farm-level characteristics, such as access to electricity, irrigation infrastructure, credit, and technical assistance, rather than by geographic factors. Crucially, the key predictors of price sensitivity differ substantially across shock types, suggesting that effective policies to reduce weather-induced food price volatility must be tailored to the specific risks faced by different crops.