Should we Flee? The Labor Market Impacts of Hurricane Evacuation Orders

• In Event: US Crisis Response

Abstract

Hurricane evacuation orders are one of the few policy levers available to save lives during extreme

storms. They are, however, costly for governments and households, heeded only partially by residents,

and their net economic effect remains largely unknown. Policymakers therefore lack credible evidence

when deciding whether and when to evacuate beyond the most exposed areas.

Do evacuation orders alter county labor markets after landfall?1 Exploiting quasi-random gaps between

forecasted and realized storm intensity, I compare counties that issued orders with equally damaged

counties that did not. By holding storm damage constant – and verifying robustness by instrumenting

with forecasted landfall – the study isolates the order from hurricane damages, delivering the

first causal estimates of evacuation economic impacts.

The analysis covers the 14 Atlantic hurricanes (2012–2022) that triggered at least one order in AL, FL,

GA, LA, MS, NC, SC or TX. Treatment is mandatory evacuation order from the Hurricane Evacuation Order

Database (Anand et al., 2023). Outcomes are county-month employment from BLS LAUS; deaths,

injuries and property loss come from SHELDUS (Center for Emergency Management and Homeland

Security, Sheldus Database, 2024); peak wind speed from National Hurricane Center’s HURDAT2.

Two orthogonal shocks drive identification: (i) unanticipated deviations in storm path, timing and

intensity, and (ii) local officials’ heterogeneous propensity to order evacuation. The main specification is

a 24-month event study. In future models, instrumenting Ec with forecasted intensity addresses

remaining endogeneity and identifies a local-average treatment effect.

Early estimates show employment falling 1–2% for the year following an order, fully recovering by

month 24 shows estimates for heterogeneity by severity of the storm. This suggests that

there may be little to no cost of evacuating unnecessarily if the storm is mild and residents can easily

return.

Ongoing work will incorporate mortality, migration, housing prices and sectoral heterogeneity. If

employment losses are modest relative to lives saved, blanket evacuations are justified; if not, targeted

shelter-in-place guidance may dominate. Credible cost–benefit metrics will equip FEMA and state emergency

managers to refine hurricane response protocols and to allocate resilience funding.

Author

• Lucas Kaplan, Syracuse University
  Presenting Author