Disruption in ground transportation: Natural disasters and disintegration of local food markets

Spatial integration between production regions and consumer markets is essential for price efficiency, yet even a single road blockage can break that link and push prices up overnight. Climate change is expected to bring more frequent natural disasters such as flash floods and landslides, turning these small but recurring blockages into a growing threat. Leveraging a novel dataset linking 230,000 weekly food price records, a routing algorithm that identifies the roads used in the supply chain network, and 1200 landslides in Colombia from 2013 to 2020, I use an instrumental variable strategy to isolate the causal effect of transport blockages on food price differentials at the national level. A typical closure diverts roughly 2%–3% of the total quantity supplied to the affected market and raises local wholesale prices by 0.5-0.9 % within one week. The diverted quantity also lowers prices in other markets by about 0.4 %. Heterogeneity estimates using the machine learning causal forest technique reveal that the strongest effects occur in markets located far from major supply corridors, where limited connectivity amplifies the impact of road closures. The variation in results is driven by the importance of the disrupted road within the transportation network and by differences in price elasticities across cities with distinct income levels and product specializations. Using the Causal Forest results, the article identifies regions where infrastructure improvements would most effectively reduce the impact of disruptions on price differentials.