The economic impact of the 2008 floods in Santa Catarina, Brazil: A municipality-level Difference-in-Differences analysis

Details

Supervisors

Parienté, William

Faculty

Faculté des sciences économiques, sociales, politiques et de communication

Degree label

Master [120] en sciences économiques, orientation générale, à finalité spécialisée

Abstract

enThis thesis examines how official disaster classifications influence municipal economic outcomes using the 2008 floods in Santa Catarina, Brazil, as a natural experiment. I employ a difference-in-differences framework comparing 75 affected municipalities against 218 unaffected controls over 2005-2013, analysing seven economic indicators including GDP, sectoral value added, and fiscal measures. The analysis distinguishes between graduated disaster classifications: 14 municipalities received "public calamity" status (most severe) and 61 received "emergency status" (moderate severity). The results challenge conventional assumptions about disaster impacts on local economies. Rather than persistent negative effects, affected municipalities experienced substantial positive economic impacts that strengthened over time. By 2013, all affected municipalities showed GDP gains of R$504.4 million annually (significant at 1% level), with public calamity municipalities achieving even larger effects of R$769.4 million (significant at the 10% level). Emergency status municipalities demonstrated more modest but positive gains of R$369.1 million (significant at 5% level), confirming a graduated institutional response pattern. Sectorally, agriculture experienced persistent negative effects (-R$3.5 million by 2013), while industry (+R$129.6 million) and services (+R$167.9 million) showed strong positive responses, suggesting disaster classifications accelerated structural economic transformation. Robustness checks, including fixed effects specifications, logarithmic transformations, event studies, and matching estimators, generally confirm these findings. However, placebo tests reveal significant pre-treatment differences between affected and unaffected municipalities, raising concerns about parallel trends assumptions. The event study shows convergence patterns in pre-disaster years but cannot fully address identification concerns. Matching estimators produce consistent results for the full sample but fail for the smaller public calamity group due to sample size limitations. These findings provide evidence that effective institutional disaster responses can transform crises into development opportunities, with response intensity directly correlating with economic recovery trajectories. The results have important implications for disaster management policy design, particularly highlighting concerning distributional effects favouring metropolitan areas over rural communities. While methodological limitations require cautious interpretation of causal claims, the analysis documents strong correlations between institutional disaster classifications and subsequent economic outcomes that merit serious consideration for policy design in developing countries facing increasing climate risks.