Climate Risk, Insurance and Housing
Paper Session
Sunday, Jan. 3, 2027 8:00 AM - 10:00 AM (EST)
- Chair: Shan Ge, New York University
Natural Disasters, Property Insurance, and Housing Markets
Abstract
We study how local climate risk affects house prices and location choices through insurance and migration channels in spatial equilibrium. We develop a dynamic spatial housing model in which households choose location, tenure, and property insurance coverage in the presence of heterogeneous climate risk across regions. Insurance decisions and spatial mobility interact to determine equilibrium house prices and population distributions across regions. The model generates realistic patterns of insurance demand, including underinsurance and strong heterogeneity across wealth, age, and regional risk exposure. Counterfactual increases in local climate risk induce selective migration out of exposed regions, leading to lower house prices in equilibrium as higher insurance costs are capitalized into asset values. Our results underscore the importance of jointly modeling climate risk, property insurance, and spatial sorting when evaluating the housing market impacts of climate change.Better Early than Late: Insurance Payments and Mortgage-Market Stress after a Disaster
Abstract
This paper studies how the timing of homeowners' insurance claim payments affects post-disaster mortgage performance. Using Freddie Mac Single-Family loan-level data, I document that mortgage-market stress after Hurricane Irma is short-lived and that delinquency responds to only the loss-year insurance payments—the share of incurred losses paid within one year—rather than to payment progress measured over longer horizons. Using an instrumented difference-in-differences (IV-DID) design, I estimate that a one-percentage-point increase in the county-level share of unpaid losses in the year of losses raises a loan’s probability of delinquency by 0.34 percentage points. The evidence points to a liquidity mechanism: the effect is larger for loans in higher-damage areas, for borrowers with tighter liquidity constraints (higher debt-to-income ratios), and for borrowers with weaker access to external credit (lower credit scores). Delayed insurance payments impose costs on both borrowers and Freddie Mac, but do not materially affect mortgage investors, and the combined losses to borrowers and Freddie Mac exceed insurers’ financial gains from delaying payments. On the insurer side, lower loss-year payment is associated with higher opportunity costs of liquidating assets and with limited liquid asset holdings during disaster periods.Model Risk in Physical Risk Models
Abstract
Model risk captures the uncertainty in model-derived point estimates of unobserved risks and has implications for financial markets. This paper studies model risk in the setting of residential physical risk assessment by comparing outputs of two prominent risk modelers, CoreLogic (CL) and First Street Foundation (FSF). We document five facts about model risk in physical risk models, defined as the absolute difference in the two modelers’ estimates of average annual loss (AAL) for the same property and set of perils. (i) Model risk costs average single-family residences (SFR) $300 per year, which is roughly equal to the mean of each modeler’s AAL estimates. (ii) Aggregating the estimates up to the tract or the county does not substantially reduce the disagreement. (iii) Model risk is disproportionately borne by economically vulnerable households. (iv) There is little association between model risk and insurance premiums, as insurers mostly use data from one modeler to price insurance contracts and ignore model risk. (v) Net in-migration rates are highest in places with the highest levels of model risk, implying that aggregate migration decisions ignore model risk. These facts imply that uncertainty about modeled estimates of physical risk is not appropriately priced by insurers and insurees.Discussant(s)
Shan Ge
,
New York University
Derek Wenning
,
Indiana University
Stephanie Johnson
,
Rice University
Tobias Huber
,
University of Georgia
JEL Classifications
- R3 - Real Estate Markets, Spatial Production Analysis, and Firm Location
- G2 - Financial Institutions and Services