Economic Losses Associated With Tropical Cyclones in Mexico: A Two-stage Model

Tropical cyclones (TCs) are among the primary drivers of disaster declarations in Mexico, causing substantial economic losses. This study assesses reported TC-related economic losses in Mexican states from (Funk 1980 ) to 2023 and identifies the hazard and socioeconomic determinants of both the probability of loss occurrence and loss magnitude. A two-stage econometric framework was employed: a binary logistic regression model to predict the probability of reported economic losses ( n = 304), and a Generalized Linear Model with Gamma distribution, state and year fixed effects, and clustered standard errors to estimate conditional loss magnitude ( n = 262). Two alternative hazard exposure measures were compared, state-level spatial maxima and population-weighted averages, to assess the robustness of results to spatial aggregation choice. Hazard variables included maximum sustained wind speed and maximum 5-day cumulative precipitation. Socioeconomic variables included prior-year state economic activity and population density. Mexico experienced US$40.54 billion in TC-related economic losses during the study period. The two-stage model identifies distinct determinants at each stage: wind speed, economic activity, and population density significantly predict loss occurrence, while wind speed and precipitation are the key determinants of loss magnitude. Notably, prior economic activity, when hazard variables are population weighted, is negatively associated with loss magnitude once unobserved state heterogeneity is controlled for, suggesting a role for economic resilience in moderating TC impacts. Findings on wind speed and precipitation are consistent across both hazard exposure measures, the ITAEE lag effect, however, is only significant under the population-weighted specification, indicating some sensitivity to how hazard exposure is measured.