次优世界中的最优洪水保险: 财政溢出、重新分类风险与道德风险
Optimal Flood Insurance in a Second-Best World: Fiscal Spillovers, Reclassification Risk and Moral Hazard
NBER Working Papers · 2026 · [{"name": "Jonathan Gruber", "affiliation": []}, {"name": "Adam Solomon", "affiliation": []}]
中文摘要
气候变化加剧使得针对自然灾害风险的防护——无论是通过事前保险还是事后援助——成为首要的公共政策议题。美国的洪水风险防护传统上依赖联邦紧急事务管理局（FEMA）提供的事后援助，以及国家洪水保险计划（NFIP）提供的高度补贴保险。为纠正补贴导致的洪水易发地区过度建设与减灾不足问题，NFIP 近期转向按精算公平原则厘定保险保费。但这一变化产生了两个非预期后果：随着洪水易发地区保险覆盖率下降，财政溢出效应波及 FEMA 灾害援助；以及家庭因气候预测的不确定性而面临不可保险的重新分类风险。我们构建了一个最优洪水保险模型，纳入了补贴保费的静态与动态权衡。随后，我们运用多种识别策略和数据，估计出实施该模型所需的五个关键参数。我们发现，财政溢出效应与重新分类风险保护收益显著超过道德风险成本，且洪水保险的最优补贴率为 52%，与近期改革前的补贴水平相当。
Abstract
Intensifying climate change makes protection against natural disaster risk – either through ex ante insurance or ex post aid – a first-order public policy issue. Flood risk protection in the U.S. traditionally featured ex post aid through FEMA and heavily subsidized insurance through the National Flood Insurance Program (NFIP). In an effort to correct subsidy-induced overbuilding and under-mitigation in flood-prone areas, the NFIP recently moved to actuarially fair insurance premiums. But this change has two unintended consequences: a fiscal spillover onto FEMA disaster aid as insurance coverage declines in flood-prone areas, and household exposure to uninsurable reclassification risk from uncertain climate projections. We develop a model of optimal flood insurance that incorporates the static and dynamic tradeoffs of subsidizing premiums. We then use a variety of identification strategies and data to estimate the five key parameters needed to implement this model. We find that the fiscal spillover and reclassification risk protection benefits significantly outweigh the moral hazard costs, and that the optimal subsidy to flood insurance is 52%, comparable to the level of subsidization before the recent reform.