Research

Agglomeration Beyond Productivity: Economic Organization and Resilience to Natural Disasters

Job Market Paper

Agglomeration generates economic benefits by concentrating economic activity, but whether these same urban structures amplify or mitigate vulnerability to natural disasters remains unclear. I examine how industrial composition, road networks, and trade linkages shape disaster resilience across U.S. metropolitan areas.

Natural disasters reduce local economic activity by approximately 3 percent on average. Valuing the associated loss in electricity service, the median estimated economic loss is equivalent to approximately 14 percent of monthly MSA GDP. The effect varies substantially with urban structure. Greater industrial concentration more than doubles the decline, whereas more centralized road networks and stronger intra-city goods sourcing reduce it by around half. The effects of industrial concentration also depend on where that concentration occurs, with concentration in Manufacturing mitigating disaster impacts and concentration in Accommodation, Wholesale Trade, and Transportation and in Health and Education amplifying them. These findings show that urban structures associated with agglomeration can either amplify or mitigate disaster impacts, making resilience an important dimension of how the benefits and costs of agglomeration are evaluated.

Download Job Market Paper

Working Papers

Who Bears the Cost of Staying Urban? Skill-Differentiated Migration Responses to Natural Disasters

Abstract

This paper examines how natural disasters shape migration responses of high-skilled and low-skilled workers across metropolitan and rural areas in the United States. I measure within commuting zone migration using a binary location choice framework that accounts for disaster conditions in both the current place of residence and potential next destinations. Using disaster driven monetary losses, I find that a one percent increase in four year cumulative disaster damage growth raises out migration of low-skilled workers from metropolitan areas by 1.44 percentage points, while high-skilled inflows rise by 0.63 percentage points. I also consider a measure of relative spatial disaster exposure, defined as the likelihood that surrounding areas experience greater damage than the focal location. A one percentage point increase in this measure raises low-skilled population growth in rural areas by 2.58 percentage points over four years and 1.79 percentage points over five years, while generating no comparable response in metropolitan areas.

Overall, the results show that natural disasters reinforce skill based spatial sorting, shifting low-skilled workers away from urban areas, with implications for unequal access to the opportunities provided by cities and longer run inequality across skill groups.