How risk models and data can help insurers after LA wildfires

These modern risk models not only improve underwriting accuracy but also provide a pathway for homeowners to mitigate their exposure—and potentially lower their insurance costs.

Why traditional underwriting falls short

Historically, underwriting in wildfire-prone areas has relied on broad categorizations, often leading to inaccurate risk assessments and mispriced policies. Modern wildfire risk models offer a more precise and data-driven approach by evaluating multiple factors at a granular level.

One major consideration being evaluated in wildfire risk models is vegetation density and fuel load, which assesses the presence of flammable materials near a property that could contribute to fire spread. Topography also plays a significant role, as steep slopes can accelerate the speed and intensity of a wildfire.

Additionally, climate conditions, including drought levels, temperature trends, and historical fire activity, can help determine the likelihood of future wildfires. Another critical factor that should be considered is fire suppression capacity, which accounts for proximity to fire stations, water availability, and the strength of local emergency response infrastructure.

Finally, wind patterns and ember travel must be considered, as embers can ignite structures miles away from the main fire-front. Unlike traditional hazard mapping, modern wildfire risk models integrate real-time data, climate analytics, and AI-driven predictive capabilities to offer dynamic risk assessments from the community level down to the exact property address.

This allows insurers to set premiums, define coverage terms, and manage risk at the individual property level rather than applying broad exclusions or blanket rate hikes across entire ZIP codes.

Risk models vs. catastrophe models: Understanding the difference

A common misconception in wildfire risk management is that risk models and catastrophe models serve the same purpose. While both are essential, they play distinct roles.

Wildfire risk models focus on property-level underwriting and pricing, evaluating individual exposure based on factors like those discussed above, as well as mitigation efforts. In contrast, catastrophe models estimate portfolio-level losses, helping insurers and reinsurers plan for large-scale wildfire events by calculating metrics such as Probable Maximum Loss (PML) and Average Annual Loss (AAL).

To effectively manage wildfire risk, insurers must use both models in tandem—risk models to guide policy-level decisions and catastrophe models to determine capital reserves and reinsurance needs. This integrated approach enables insurers to price coverage more accurately, maintain market stability, and better protect both their portfolios and policyholders.

The next frontier: Mitigation-based wildfire risk modeling

One of the most significant recent advancements in wildfire risk modeling is the ability to incorporate mitigation efforts. Research from Guidewire HazardHub, based on 91,800 home inspections in California over a decade, has revealed compelling insights:

• Even basic mitigation measures, such as clearing vegetation around a home, can reduce wildfire risk by 20% or more.
• Comprehensive strategies, including all mitigation recommendations from the National Fire Protection Association—such as installing ember-resistant vents, using fire-resistant materials, and maintaining defensible space—can reduce risk by up to 74%.

This presents a major opportunity for insurers to incentivize risk reduction by offering premium discounts, supplemental coverage, or improved policy terms for policyholders who invest in fire-resiliency.

Is California still insurable? Data-driven risk modeling holds the key

With major insurers exiting or restricting coverage in California due to the wildfire risks, a pressing industry question remains: Is California still insurable?
The answer depends largely on statewide rate adequacy and how accurately wildfire risk is assessed.

Recent data from HazardHub found that 98% of wildfire-related property losses occur in just 10% of the state—those areas receiving an “F” wildfire risk grade in the HazardHub model.

This suggests that while certain areas are high-risk, the majority of homes remain insurable if underwriting can effectively distinguish between extreme-risk and lower-risk properties. However, if statewide base rates are too low, even properties with moderate to low risk can destabilize the insurer.

Rather than resorting to broad market withdrawals, insurers can maintain coverage in California by collaborating with the California Department of Insurance to establish statewide rate adequacy and leveraging advanced wildfire risk models. This approach enables insurers to:

• Encourage mitigation efforts – Incentivize homeowners to reduce risk throughout the policy lifecycle: At policy acquisition, ensure underwriting guidelines are met and rates are applied appropriately. Within 60 days, identify high-risk locations requiring inspection for a thorough underwriting analysis. Before renewal, remind customers to maintain initial mitigation improvements and assess any additional risk reduction efforts. Inspect high-risk properties, those with new mitigation work, and properties with significant mitigation potential.
• Enhance portfolio management – Maintain market presence while avoiding overexposure to high-risk properties in close proximity.
• Precisely price risk – Establish sustainable rate levels that reflect future expected claims costs for each location.

The future of wildfire risk modeling: AI, IoT, and smarter data

The next evolution of wildfire risk modeling is already underway, with advancements in AI and Internet of Things (IoT) sensors enabling even more precise risk assessments. Key innovations include:

• AI-Powered Predictive Analytics – Machine learning models continuously update risk scores based on new data inputs.
• Real-Time Weather and Fuel Monitoring – IoT sensors track ignitions, smoke, wind speeds and direction, humidity levels, and vegetation conditions to refine risk assessments dynamically. This real-time data can protect lives and property by enabling early wildfire detection, increasing the chances of extinguishing fires before they grow beyond 10 acres.
• Geospatial AI & Satellite Imaging – AI-driven satellite analysis detects high-risk property changes in real-time, allowing insurers to adjust coverage accordingly.

A smarter path forward for insurers

With wildfire losses reaching historic levels, insurers face a critical decision: retreat from high-risk areas and shrink the available market or embrace modern wildfire risk models to navigate the increasingly complex landscape.

By adopting modern risk modeling, mitigation incentives, and AI-driven analytics, insurers can move toward a more resilient, profitable, and insurable future—even in wildfire-prone regions.

Wildfires aren’t going away. But with smarter underwriting, better mitigation strategies, and advanced risk models, insurers can adapt to this new era of climate-driven risk—protecting their bottom lines, the communities they serve, and the planet we all share.

Tammy Nichols Schwartz, CPCU, is the Senior Director of Analytics at Guidewire HazardHub, where she leads wildfire risk modeling and hazard data solutions. With 27 years of experience in insurance underwriting and risk modeling—including leadership roles at the California FAIR Plan and Black Swan Analytics—she is a recognized expert in leveraging data to improve insurance resilience.

See also:
AI vs. human touch in workers’ comp claims management
7 Ways financial professionals are using AI to streamline workflows
Which states see the largest losses from online fraud?