Measuring the impact of hurricane incidence on agricultural production risk using insurance data
利用2002-2021年美国县级月度保险损失数据和飓风数据，采用双重差分法衡量飓风对密西西比三角洲作物生产损失的影响，发现损失在不同作物间差异显著，且飓风处理变量的构建方式对识别结果至关重要。
Abstract Hurricanes are considered among the most destructive natural disasters in the United States. The exposure of agricultural production systems to hurricanes varies between regions in contrast to global risks like commodity price volatility and international trade policies. The regional differences in hurricane exposure may lead to heterogeneity in crop insurance premium rates. This work aims to measure the impact of hurricane incidence on production losses for crops grown in the Mississippi Delta. We leverage a county‐month panel of insurance losses spanning 2002–2021 from the US Department of Agriculture Risk Management Agency, and daily data from the National Oceanic and Atmospheric Administration National Hurricane Center, to construct novel measures for hurricane treatment assignment under a difference‐in‐differences identification strategy. We find hurricane incidence results in economically significant losses that are strikingly heterogeneous across crops. We also demonstrate that measuring hurricane treatment is a critical concern for credible identification as simplistic approaches are associated with attenuation bias relative to our approach which accounts for the dynamic changes in the scope and intensity of a hurricane. We discuss implications of these results for two important policy instruments, traditional crop insurance premium rating and the more recently released Hurricane Insurance Protection—Wind Index product.