Model uncertainty raises Florida hurricane insurance premiums
A Federal Reserve paper by Erik Heitfeld finds that greater uncertainty among hurricane risk models leads to higher homeowners insurance premiums in Florida. The study quantifies model dispersion using regulatory data and links it empirically to insurance pricing across ZIP codes.
Quantifying model disagreement
Erik Heitfeld's paper investigates how model uncertainty, distinct from inherent aleatoric uncertainty, influences catastrophe risk insurance pricing.
Using unique data from seven hurricane risk models approved by Florida regulators, the study measures the dispersion in loss projections across ZIP codes and over time.
The analysis finds strong empirical support for the hypothesis that greater dispersion among model forecasts correlates with higher homeowners insurance premiums.
For instance, a ten percent reduction in model dispersion in 2021 could have saved a typical Florida homeowner $50 to $90 on their annual premium.
The research highlights that significant dispersion exists due to differing assumptions and judgments by model developers, rather than just sampling error.
Ambiguity and strategic pricing
The paper explores two mechanisms through which model uncertainty can lead to higher premiums.
First, insurers may be averse to ambiguity, preferring risks with known probabilities over those with unknown ones, leading them to charge higher prices.
Second, under rate regulation, firms might strategically "cherry-pick" more pessimistic models to justify aggressive rate increases when forecasts diverge.
This behavior creates incentives for firms to manipulate costs, as observed in other regulated markets.
The social benefits of catastrophe insurance are substantial, providing community resilience and capital for rebuilding, making understanding its pricing drivers crucial.
A new lens for regulators
This study offers a critical empirical link between model disagreement and consumer costs, moving beyond theoretical discussions.
It provides concrete evidence that regulatory-approved model diversity, while intended for robustness, can inadvertently contribute to higher premiums.
For policymakers, this suggests a need to scrutinize not just the models themselves, but also how their collective uncertainty impacts market behavior and consumer burden.