Causal analysis of trade loss from pathogens: A global study of foot and mouth disease impacts on meat exports

Abstract

Our general interest is in global trade loss from livestock pathogens, specifically exports. We adopt a causal inference approach that considers animal disease outbreaks over time as non‐staggered binary treatments with the potential for switching in (infection) and out of treatment (recovery) within the sample period. The outcome evolution of switchers and non‐switchers identifies the treatment effects. Using a recently proposed dynamic difference‐in‐differences estimator, we estimate the treatment effects that decompose into “infected” (switch in) and “recovered” (switch out) effects. As a case study, we investigate the global impacts of foot and mouth disease outbreaks on meat exports using panel data across 178 countries from 1996 to 2016. We find that the outbreak decreases meat export by 19,000 to 48,000 metric tons per year (14% ‐ 35% of the mean annual meat export) in the 5 years following an outbreak, which is brought about by a decline in the infected countries' meat exports, while the disease‐recovered countries do not fully regain the export losses even after 5 years. The average total effect is estimated at about 54,000 metric tons per outbreak over 5 years following an outbreak, resulting in an export revenue loss of $141 million. Asymmetric post‐infection and post‐recovery trade losses imply a significant disease burden on the endemic regions. Regional decomposition of treatment effects reveals that while the magnitude of export loss is greater for the non‐endemic regions, the relative magnitude (compared to the mean annual export per country) is many times higher in the endemic regions.