What E053 Learned About Wildfire and Insurance

What E053 Learned About Wildfire, Insurance, and Blame

E053 reached its most durable conclusion early: California's wildfire insurance crisis was not the work of one villain. Fire behavior, vulnerable structures, concentrated exposure, insurance contracts, reinsurance, catastrophe models, capital, regulation, mitigation, and affordability were acting on one another.

Tom Mercer and I recorded the conversation while the January 2025 Los Angeles emergency was still changing. The episode was distributed on January 28. That timing matters because the opening containment, damage, and loss figures were preliminary.

This revision keeps the systems argument and corrects the chronology. It is not a current market report, policy recommendation, or final account of either fire.

The numbers belong to different moments

The recording opens with Eaton at zero containment and Palisades at six percent. By the January 28 release, the incident picture had moved. Later records moved again.

The current CAL FIRE Eaton record shows a January 7 start, 14,021 acres, 9,419 destroyed structures, and 19 civilian fatalities. It lists the cause as undetermined and identifies January 31 as the containment date.

The current CAL FIRE Palisades record shows a January 7 start, 23,448 acres, 6,845 destroyed structures, and 12 civilian fatalities. It identifies January 31 as the containment date and currently lists the cause as arson.

Those are later official records with their own update dates. They should not be projected backward as facts available to Tom and me during the recording.

flowchart LR
    A["Fire behavior and exposure"] --> B["Structure and neighborhood loss"]
    B --> C["Insurance portfolio loss"]
    C --> D["Reinsurance and capital"]
    D --> E["Underwriting, rate, and availability choices"]
    E --> F["Regulation and residual markets"]
    F --> G["Mitigation and rebuilding incentives"]
    G --> A

Tom brought a program builder's view

Tom had previously supervised me. In the episode, he described a career building commercial-property programs and catastrophe-risk capabilities.

He said he began in catastrophe-exposed insurance at ICAT, helped develop a California residential earthquake program, then joined Velocity Risk and helped form catastrophe-modeling, analytics, and commercial-program functions. He later worked on commercial-property programs at MSI and SafeLease.

Tom also said he had served in the Navy in San Diego and attended business school at UCLA. Those are his source-era statements. His current role, preferred profile, exact chronology, and contact link still require his confirmation before this revised page can be released.

What mattered in the conversation was his operating perspective. An insurer cannot discuss one exposed property without asking how thousands of properties aggregate, how much loss it retains, what protection it buys, what capital supports the portfolio, and which rates and forms the regulator permits.

The fire can reach the building by more than flame

Our conversation spent time on embers because the popular image of wildfire is a wall of flame touching a house.

The U.S. Fire Administration describes four exposures: ember deposition, radiant heat, direct flame contact, and convection. Embers can land on a structure, enter through openings, ignite nearby landscape, or ignite attached and neighboring fuels.

That changes the unit of analysis. The roof matters, but so do vents, gaps, decks, fences, vegetation, maintenance, adjacent structures, wind, and the event. No single material or cleared distance makes a building fireproof.

Tom's point was not that one product would solve wildfire. It was that physical risk has to enter underwriting, modeling, building, and incentive decisions more clearly.

California did not have one insurance market

The episode used a loose contrast between admitted and non-admitted insurance. The more accurate map includes the voluntary admitted market, licensed agents and brokers, surplus-lines placement, the California FAIR Plan, and supplemental coverage.

The current California Department of Insurance residential guide explains producer roles, policy review, shopping, availability problems, surplus lines, and the FAIR Plan. The Department states that surplus-lines insurers are not backed by the California Insurance Guarantee Association.

The FAIR Plan is not simply a taxpayer-funded state insurer. The Department's FAIR Plan page describes an insurance-company-run association operating under California law and regulatory oversight as a residual option.

Each route has different forms, limits, oversight, intermediaries, protections, and eligibility. None can be reduced to "the same policy through another door."

Reinsurance connected the local fire to global capital

Tom explained reinsurance as protection insurers purchase for defined parts of large loss.

The NAIC reinsurance overview lists capacity, stabilization, financing, catastrophe protection, risk spreading, and expertise among common purposes. The insurer still owes its policyholder under the original policy.

A catastrophe treaty may respond only after a retention and only up to a limit. Covered perils, portfolios, periods, occurrence definitions, exclusions, reinstatements, and counterparty terms matter. One annual price increase or event clause is not an industry standard.

California finalized a net-cost-of-reinsurance regulation at the end of 2024. The rule and related coverage commitments became part of a larger reform program. It did not turn reinsurance into the only rate input or guarantee new consumer coverage.

Models are distributions, not prophecies

The episode argued that forward-looking catastrophe models could help distinguish risk and inform capital.

The NAIC catastrophe-model overview describes hazard, vulnerability, exposure, and financial modules. A model simulates many plausible events, applies damage relationships to a portfolio, then applies policy and reinsurance terms to estimate a loss distribution.

It does not predict the exact date, ignition, path, or loss of the next fire. Data, event sets, construction attributes, valuations, financial terms, demand surge, assumptions, and model version all affect the output.

California completed review of its first wildfire catastrophe model in July 2025, months after the episode. Regulatory review permits defined use under a process. It does not certify every output for every portfolio or property.

Later events tested the system

In February 2025, the Commissioner approved a $1 billion FAIR Plan member-insurer assessment after the January wildfire and wind losses. That later order replaced our source-era speculation about what the residual market might need.

The Department's Sustainable Insurance Strategy now tracks model use, reinsurance-cost treatment, coverage commitments, filings, and market snapshots. In July 2026, the Department reported additional insurers committing to expand writing. A commitment, filing, approval, issued policy, and sustained market outcome remain different facts.

California also authorized a public wildfire-model program in 2025. As of July 2026, the Department was selecting university expertise and preparing development, not operating a completed public model.

The market was still moving. That is why a reform article must function as a dated tracker, not a declaration that the crisis is solved.

A systems view changes the question

Blame can be politically satisfying because it makes a complex failure feel legible. It is a weak operating model.

If fire exposure changes while buildings remain vulnerable, losses rise. If concentrated losses rise, capital and reinsurance matter more. If rates and forms lag risk, carriers change appetite or capacity. If availability falls, residual markets accumulate exposure. If mitigation is not measured or maintained, incentives may not change physical outcomes.

None of that excuses an insurer, regulator, builder, owner, utility, government, or another actor from a specific documented failure. It means the failure should be named with evidence and placed in the system it affects.

The full conversation remains available on Spotify and YouTube. It preserves an informed discussion recorded during a moving emergency, including claims that require the corrections and boundaries above.

This page was developed with AI assistance from the E053 transcript and linked primary sources, then structured for human guest, incident, insurance, legal, actuarial, catastrophe-model, fire-science, media, founder, and editorial review. It does not provide insurance, legal, actuarial, building, or safety advice.