Abstract:
Floods are among the most frequent and destructive natural hazards worldwide, yet their economic impacts remain poorly understood. This paper examines the effects of two cyclone-induced floods on firms' economic performance in New Zealand. To identify flood damage, we propose a novel method using high-resolution satellite imagery to track changes in ground vegetation before and after the cyclones. By integrating this information with detailed business administrative records and enterprise survey data, we construct a unique plant-level dataset to estimate the flooding causal effects. Using a Difference-in-Differences (DID) approach, we find that firm premises located in flood-affected areas experienced significant declines in gross output and sales, alongside increased losses in profit and value-added. Contrary to common belief, our analysis reveals that capital damage and labour displacement, rather than productivity losses, are the primary channels through which flooding affects firms. Plants responded to these events by liquidating inventories and cashable assets, though no evidence of relocation is found. Our results are robust to various econometric specifications and alternative estimation methods, including the Synthetic Difference-in-Differences (SDID) estimator.