Insurance Desk
Cat bond desk, the cycle, modeled loss, solvency watch, protection gap, and carrier books — six voices on catastrophe-bond/ILS pricing, the reinsurance underwriting cycle, cat modeling, insurer solvency, and the coverage protection gap.
AI-generated analysis from Apprised's automated desks, synthesized from cited sources and editorially accountable to J.A. Watte. How we report · Corrections.
Chart auto-generated from this brief's structured fields. See methodology for how the underlying data is collected.
Cayman Islands' application for NAIC Qualified Jurisdiction Status could deepen its ILS and reinsurance competitive advantage just as the cat-bond market hits $18.9B in YTD issuance across 92 deals, with $65.6B outstanding and a market yield of 9.29%. Separately, the Los Angeles Department of Water and Power placed a $100M California wildfire cat bond (123 Lights Re) in July 2026, signaling institutional demand for parametric wildfire coverage as Montana burns across 47,000+ acres.
Bias-reviewed: LOW Independently rated by Kimi for political-lean, source-diversity, and framing bias before publish. Final orchestration and the published call are made by Claude, a U.S. model.
Insurance Risk Tape as of 2026-09-03
Insurance risk backdrop: elevated — catastrophe declarations rising; carrier equities leading the tape; credit spreads widening; alternative capital accessible.
-
Catastrophe Load62 active federal disaster declarations (90d)up from 34 prior 90d · led by Fire (41), Severe Storm (7), Flood (6) · 118 YTD90-day declarations: 62Prior 90 days: 34YTD: 118FEMA OpenFEMA📖 Learn more
-
Carrier Equity SignalInsurer stocks leading the marketKIE uptrend, +14.7% vs SPY (3mo) · IAK mixed, +11.7% vs SPY (3mo)KIE: 63.62 (+14.7% RS)IAK: 144.79 (+11.7% RS)Yahoo Finance (KIE/IAK vs SPY)📖 Learn more
-
ILS / Alternative Capital$18.9B cat-bond issuance YTD94 deals · $65.6B outstanding · 9.29% yield on 2.5% expected loss · avg $136M · alternative reinsurance capital remains accessibleYTD issuance: $18.90BMarket size: $65.6BMarket yield: 9.29%Expected loss: 2.5%Deals YTD: 94Avg deal: $136MArtemis.bm ILS dashboard📖 Learn more
-
Balance-Sheet Backdrop10Y 4.79% · HY 265bps10Y at 4.79% (rising) supports reinvestment income; credit spreads tight/widening on the bond book.10Y Treasury: 4.79% (rising)HY credit spread: 265bps (widening)2s10s curve: +0.4% (normal)VIX: 16.34FRED via Corvus📖 Learn more
Deterministic insurance-risk indicators — $0 LLM, computed live from public data (FEMA OpenFEMA, Yahoo Finance, Artemis ILS, FRED). Educational, not advice. Sources: FEMA OpenFEMA, Yahoo Finance (KIE/IAK vs SPY), Artemis.bm ILS dashboard, FRED via Corvus.
Today’s Snapshot
Cayman QJS bid + LADWP wildfire bond headline a busy ILS week
Two structural stories dominated the insurance desk Wednesday. First, Cayman Islands Premier Andre Ebanks publicly flagged that the territory's NAIC Qualified Jurisdiction Status application could meaningfully expand its competitiveness for ILS and reinsurance business, adding regulatory depth to a jurisdiction already dominant in collateralized structures. Second, the Artemis deal directory confirms a $100M California wildfire cat bond issued in July by the Los Angeles Department of Water and Power through the 123 Lights Re vehicle — notable for its public-utility cedent and single-peril wildfire focus. Meanwhile, RockRose Risk closed a $12.5M Series A to build an integrated wildfire risk-management and insurance platform, underscoring continued private capital interest in the hardest-to-insure residential peril. Providence Health Plan's complete closure after a failed Medicare Advantage deal collapse adds a cautionary note on regional health-plan fragility.
Synthesis
Points of Agreement
Cat Bond Desk (Vaeth) and The Cycle (Ennis) converge on a shared read: the $18.9B YTD ILS issuance pace and the Cayman QJS bid are both supply-side stories that point to a market still in capital-accumulation mode, with softening pressure building at the reinsurance layer even if the 9.29% market yield keeps investors engaged. Modeled Loss (Chandrasekar) and Protection Gap (Owusu-Reyes) agree that California wildfire is the peril where the model-to-market gap is most consequential and least resolved — they disagree on emphasis (Chandrasekar on catalog inadequacy, Owusu-Reyes on the consumer-side non-equivalence of utility hedges and homeowner coverage) but not on the underlying diagnosis. Carrier Books (Marchetti) and Protection Gap both treat Providence Health Plan's closure as a warning about what scale-deficiency does to carriers in stressed-margin lines.
Points of Disagreement
The sharpest tension is between Cat Bond Desk and Modeled Loss on the 123 Lights Re California wildfire bond. Vaeth reads continued ILS investor appetite for California wildfire as a market signal — capital is available, deals are clearing. Chandrasekar treats the same fact as a potential mispricing: if the model is calibrated to pre-2025 California loss experience and the post-2025 AAL is materially higher, investor appetite is a sanguine signal, not a reassuring one. The Cycle (Ennis) adds a second layer of tension with Cat Bond Desk: Ennis argues that some of the issuance pace is yield-seeking behavior in a compressed-HY-spread environment (2.75% OAS) rather than correctly priced risk transfer, which would imply the market-wide 2.2x multiple-on-EL is an artifact of rate environment as much as underwriting discipline. Vaeth's framework, by design, reads the spread as the honest price; Ennis reads the cycle that produced it.
Pivotal Question
What is the post-2025 LA wildfire implied AAL for California wildfire cat bonds, and does the 123 Lights Re attachment structure reflect updated model runs or pre-event pricing? If the modeled AAL has risen materially since the 2025 LA fires, the spread embedded in recent California wildfire ILS may be insufficient — and the next event would reveal whether capital-market appetite was informed or merely yield-hungry.
Bias Flags
- Cat Bond Desk: Treats spread-over-EL as sufficient price signal; underweights the possibility that the EL itself is understated due to non-stationary California wildfire hazard — a model error that no spread calculation can correct.
- The Cycle: Mean-reversion framing may miss a structural shift: if climate non-stationarity has permanently elevated California wildfire AAL, the capital flowing back is not the same cycle repeating — it's new capital mispricing a changed hazard.
- Modeled Loss: Correctly flags model inadequacy for California wildfire but does not have per-deal EL data from the corpus to quantify the specific gap for 123 Lights Re; the concern is directionally right but the magnitude is unverifiable from today's corpus.
- Protection Gap: Frames the LADWP cat bond as irrelevant to consumers — correct in immediate terms, but utility liability hedging that reduces LADWP's inverse-condemnation exposure could reduce rate suppression and benefit ratepayers indirectly; the consumer impact is not zero.
- Carrier Books: PRU's 66.8% disclosure novelty is flagged as significant, but the wording-diff data shows direction of change (sentences added/removed) not the substantive content of what changed — the inference of risk-repricing is plausible but not confirmed by corpus.
Routing
Voices seated: Cat Bond Desk, The Cycle, Modeled Loss, Protection Gap, Carrier Books
Today's corpus clusters around three insurance-relevant threads: (1) Cayman Islands QJS bid and the ILS structural-advantage story, routing to Cat Bond Desk and The Cycle; (2) RockRose Risk's wildfire-platform Series A and Montana wildfire conditions, routing to Modeled Loss and Protection Gap; (3) Providence Health Plan collapse and the 123 Lights Re LADWP wildfire cat bond, which touches Carrier Books and the broader coverage-desert question. Solvency Watch is not activated today: no rate filings, RBC actions, or insurer-of-last-resort distress appear in the corpus. The ICE liability-insurance story and ACA MarketWatch piece are noted but too thin on insurance market data to warrant primary routing.
Analyst Voices
Cat Bond Desk Soren Vaeth
The Artemis dashboard tells a clean story: $18.9B in YTD issuance across 92 deals, $65.6B outstanding, market yield at 9.29% decomposing into 5.53% insurance risk spread and 3.76% collateral yield against a market-level expected loss of 2.5%. That puts the market-wide multiple-on-EL at roughly 2.2x — not the 3x-plus prints of the 2023 hard-market peak, but still a market compensating investors meaningfully for the risk they're taking on. The pipeline remains active: Swiss Re's Matterhorn Re (Series 2026-3) at $345M and Hannover Re's 3264 Re at $200M are institutional-grade prints, and the average recent deal size of $138M tells you this is not a market of scrappy micro-deals.
The Cayman QJS story is the structural backdrop that doesn't get priced in a spread but matters to every deal in that directory. NAIC Qualified Jurisdiction Status would allow Cayman-domiciled reinsurers to post collateral on terms acceptable to U.S. ceding insurers under the credit-for-reinsurance framework, reducing friction for collateralized structures and expanding the eligible capital base. Premier Ebanks is right to flag this as an amplifier of Cayman's existing ILS appeal — it's not a pivot, it's a deepening.
The 123 Lights Re structure is the deal I want to spend time on. A $100M California wildfire bond with the Los Angeles Department of Water and Power as cedent is a public-utility using the capital markets to hedge its own liability exposure post-LA fires — not a traditional insurer layering cat protection. The peril is single-state, single-peril California wildfire, which means investors in this bond are taking on a concentrated tail that the model community has struggled to price correctly. The Harbor Crest Re deal (Porch Group, $100M, covering named storm, winter storm, severe weather, wildfire, and fire-following earthquake) is the contrasting multi-peril structure. Both are July 2026 issuances, both reflect continued appetite at the alt-capital layer even for California wildfire — which, frankly, is the peril where I most want Dr. Chandrasekar's view on model adequacy before I tell you whether the spread is sufficient.
The cat-bond market at 9.29% yield with a 2.5% EL implies a ~2.2x multiple-on-EL — healthy but below 2023 hard-market peaks — while the LADWP's 123 Lights Re wildfire bond signals public utilities are now directly accessing ILS capital for single-peril California wildfire protection.
Bias flag — Treats spread-over-EL as sufficient price signal; underweights the possibility that the EL itself is understated due to non-stationary California wildfire hazard — a model error that no spread calculation can correct.
The Cycle Margaret Ennis
Ninety-two deals, $18.9B, and we're not even through August. If you're reading the ILS issuance pace as a soft-market tell, you're not wrong — but you need to distinguish between capital coming back because the risk is repriced correctly and capital coming back because investors are chasing yield in a 3.63% fed funds environment where HY OAS has compressed to 2.75%. Both are happening simultaneously, and they have very different implications for where the cycle goes from here.
Cayman's QJS push is, from a cycle perspective, a supply-side story. More jurisdictions with streamlined collateral credit under NAIC rules means more eligible capital chasing U.S. cat risk, which is structurally softening at the reinsurance layer even if primary rates haven't fully followed. The Swiss Re and Hannover Re deals in the Artemis directory — $345M and $200M respectively, both covering U.S. and Canada named storm and earthquake — are major cedents actively shedding peak-zone exposure into the capital markets at what they presumably regard as favorable terms. When the buyers of protection are that large and that systematic, the cycle is telling you something: primary reinsurers are managing their own books by transferring to ILS investors.
Soren is right that the multiple-on-EL has come in from 2023 peaks. What I'd add is that the retrocession market — the market that reinsures the reinsurers — is where we'd see the earliest stress signal if the softening goes too far. I don't have retro-specific data from today's corpus, but the combination of strong issuance pace, Cayman QJS expansion, and benign VIX at 15.84 all point to a market that is still in the capital-accumulation phase of the cycle. Hard markets sow the seeds of the next soft market. The seeds were planted in 2023. We are watching them grow.
Record-pace ILS issuance driven partly by yield-seeking in a compressed-spread environment represents capital-cycle softening at the reinsurance layer, even as primary-market conditions remain firm — the Cayman QJS bid amplifies this supply-side dynamic.
Bias flag — Mean-reversion framing may miss a structural shift: if climate non-stationarity has permanently elevated California wildfire AAL, the capital flowing back is not the same cycle repeating — it's new capital mispricing a changed hazard.
Modeled Loss Dr. Ravi Chandrasekar
Soren asked for my view on the 123 Lights Re California wildfire bond, and I want to answer that carefully. California wildfire is, at present, the peril where the gap between modeled and actual loss is widest — not because the physics are mysterious but because the exposure base has changed faster than the event catalog can price it. The WUI expansion, the defensible-space compliance gap, and post-fire debris flow have all moved the loss surface in ways that historical event sets underrepresent. When LADWP places a $100M wildfire bond, the investors buying it are relying on a model calibrated partly to pre-2017 California loss experience. The 2018 Camp Fire, the 2021 Caldor and Dixie fires, and the 2025 LA fires each pushed the modeled AAL upward after the fact. The question is whether the spread embedded in 123 Lights Re reflects the post-2025 view of California wildfire frequency and severity, or whether it's still anchored to a model that hasn't fully digested the last event.
The Montana wildfire story in today's corpus is a secondary-peril reminder. The Sand Creek Fire and its siblings have burned more than 47,000 acres across southwestern Montana, with containment on the Sand Creek Fire starting from just 2 percent before recent rains. The corpus notes conditions are deteriorating again — warmer temperatures, lower humidity, breezy winds. These southwestern Montana fires are not the insured-loss event that a Los Angeles urban wildfire would be, but they are the kind of event that accumulates into a bad aggregate year for wildfire-exposed cat bonds and ILS structures with aggregate triggers.
RockRose Risk's $12.5M Series A for an integrated wildfire risk-management and insurance platform is the private-market response to the modeling gap. The bet is that property-level mitigation data — defensible space assessments, ember-resistant construction verification — can improve the underwriting signal beyond what regional cat models provide. That's the right hypothesis. Whether $12.5M is enough to build a data moat in a peril as geographically and meteorologically complex as California wildfire is a question the next capital raise will answer.
California wildfire cat-bond pricing likely still relies on pre-2025 event catalogs that understate current AAL, while accumulating Montana secondary-peril losses and the RockRose model-improvement bet both signal that the underwriting community knows the gap exists.
Bias flag — Correctly flags model inadequacy for California wildfire but does not have per-deal EL data from the corpus to quantify the specific gap for 123 Lights Re; the concern is directionally right but the magnitude is unverifiable from today's corpus.
Protection Gap Daniela Owusu-Reyes
Let me say clearly what the 123 Lights Re bond does and does not do for California homeowners. The Los Angeles Department of Water and Power is hedging its own liability exposure — the kind that arises when utility infrastructure ignites a wildfire and the utility faces inverse condemnation claims. That is a legitimate and important use of the cat-bond market. But it is not the same as homeowners in Altadena or Pacific Palisades getting their coverage back. The insured-vs-economic-loss gap in California wildfire is structural: the FAIR Plan is the insurer of last resort for hundreds of thousands of properties that the voluntary market has non-renewed, and FAIR Plan's capacity constraints have been well-documented in prior coverage. A $100M utility liability hedge does not close that gap.
RockRose Risk's Series A is more directly relevant to the protection-gap question — if the platform can produce property-level wildfire risk scores that allow standard-market carriers to re-underwrite properties they've been non-renewing on a blanket geographic basis, that's a genuine contribution to coverage availability. The pitch — brokerage plus risk-mitigation data plus placement — is the right model. But at $12.5M and Series A stage, this is a years-away solution for a now-problem.
Dr. Chandrasekar's point about Montana is worth connecting to the consumer side: those 47,000+ acres burning in southwestern Montana are in counties where the E&S market is already the only option for many rural landowners. Secondary-peril accumulation in the aggregate year erodes ILS collateral and tightens retrocession supply, which flows through to primary market tightening in exactly the communities that can least absorb it. The protection gap isn't just a California story. It runs from Los Angeles to Lolo, Montana.
The LADWP wildfire cat bond hedges utility liability, not homeowner coverage availability — the California protection gap remains structurally unaddressed while RockRose Risk's platform solution is still years from scale.
Bias flag — Frames the LADWP cat bond as irrelevant to consumers — correct in immediate terms, but utility liability hedging that reduces LADWP's inverse-condemnation exposure could reduce rate suppression and benefit ratepayers indirectly; the consumer impact is not zero.
Carrier Books Theo Marchetti
Today's corpus is light on primary-carrier earnings data, so I'll anchor on what I have. The SEC filing wording-diff data shows the Insurance sector averaging 30.3% Item 1A novelty across 8 leaders — below the cross-sector average and notably below Energy Majors (55.4%) and Defense (54.5%). That low average masks two outliers worth flagging: PRU at 66.8% novelty with 304 sentences added and 148 deleted, and TRV at 47.2% with 246 added and 251 deleted. Prudential's 66.8% risk-factor rewrite is the largest in the insurance sector by a wide margin — that level of disclosure turnover in a life-insurance and financial-services context typically signals management is materially repricing or restructuring risk exposure, not just updating boilerplate. TRV's near-equivalent adds and deletions at 47.2% reads more like active risk-factor rebalancing than a directional shift.
Against the macro backdrop: the 10-year/2-year spread at 0.46pp is flat but positive — mildly favorable for the reinvestment-rate story that life carriers like PRU depend on. Effective fed funds at 3.63% and HY OAS at 2.75% (tight, risk-on) suggest the investment portfolio environment remains constructive. The broad dollar index at 118.9028, down 1.67 points over 30 days, is a modest headwind for carriers with significant international reinsurance placements priced in dollars.
Providence Health Plan's complete closure after a failed Medicare Advantage deal is the carrier-books cautionary note of the day. The corpus cites a Providence spokesperson acknowledging the agreement collapsed 'despite significant effort on all sides.' Regional health plans attempting to survive by selling or partnering their MA books are in a structurally weak position: Medicare Advantage margins have compressed industry-wide, and a regional plan without the scale to absorb MLR volatility is a stranded carrier. Providence's closure is not a systemic event, but it is a case study in what happens when a regional carrier can't find a buyer willing to absorb deteriorating MA economics.
PRU's 66.8% risk-factor novelty — largest in the insurance sector — warrants close reading for directional disclosure shifts, while Providence Health Plan's closure illustrates the margin-compression endgame for sub-scale Medicare Advantage plans.
Bias flag — PRU's 66.8% disclosure novelty is flagged as significant, but the wording-diff data shows direction of change (sentences added/removed) not the substantive content of what changed — the inference of risk-repricing is plausible but not confirmed by corpus.
Simulated Opinion
If you had to form a single opinion having heard the roundtable, weighted for known biases, it would be: the ILS market's $18.9B YTD issuance and the Cayman QJS push are genuine structural positives for alternative capital supply, but the enthusiasm is partly rate-environment-driven and the California wildfire segment — exemplified by the 123 Lights Re LADWP bond — carries model risk that the market-level 2.5% EL figure cannot fully price without post-2025 catalog updates. The more important story for U.S. policyholders is that ILS capital flowing into utility-liability hedges and reinsurance treaties does not translate into restored homeowner coverage in the California voluntary market; RockRose Risk's Series A is the right idea at insufficient scale. Providence Health Plan's collapse is a contained but instructive case study in what regional carrier fragility looks like in a compressed-margin, scale-dependent line — watch for similar closures in Medicare Advantage as the MA margin cycle continues to tighten.
Independent Cross-Check — Kimi
Consensus 12 Contested 2 Developing 1
Bitcoin briefly reached $70,000 for first time since June, rallying over 7% Consensus
SK Hynix shares surged over 12% in Seoul after announcing massive stock buyback Consensus
Federal Reserve released July 28-29, 2026 FOMC meeting minutes Consensus
APM Terminals and Hapag-Lloyd partner, with Hapag-Lloyd to acquire 25% stake in Maasvlakte II terminal Consensus
Home Depot rolls out 3-hour express delivery across US using 2,000+ stores as fulfillment hubs Consensus
RockRose Risk secured $12.5 million Series A funding for wildfire insurance platform Consensus
ICE pitching legal insurance plan for local police officers making immigration arrests Consensus
Providence Health Plan to close completely after Medicare Advantage deal collapses Consensus
Georgia Interior Ministry denied reports of 20,000 Russian citizens entering through Larsi border in one day, stated actual figure was 7,880 'citizens of various countries' Contested
Hormuz shipping transits remain low amid conflicting US-Iran claims Contested
Trump urged Senate to pass crypto CLARITY Act and teased more Bitcoin purchases Consensus
Bitcoin.com integrates UAE-registered USD stablecoin (USDU) into self-custodial wallet Consensus
Axle Logistics signed five-year jersey patch deal with Tennessee Basketball Consensus
Enugu State Governor Peter Mbah appointed 23 new special advisers and assistants Developing
ZIM reported Q2 net profit of $64 million, up from $24 million year-over-year Consensus
Watch Next
- Cayman Islands NAIC QJS application status: any formal NAIC committee action or comment period opening would be the next data point on whether the regulatory-access story becomes real supply-side capacity.
- 123 Lights Re (LADWP) deal terms disclosure: watch for Artemis or Bloomberg confirmation of attachment point, exhaustion level, and trigger type (indemnity vs. parametric) to allow Modeled Loss to assess pricing adequacy against post-2025 California wildfire AAL estimates.
- Montana wildfire containment reports: Sand Creek Fire and related fires began at 2% containment; any deterioration driven by returning heat and low humidity could push aggregate secondary-peril ILS loss estimates higher for 2026.
- PRU 10-K risk-factor content review: with 66.8% novelty and 304 sentences added, the substantive direction of Prudential's disclosure rewrite warrants analyst scrutiny — earnings call or investor day would be the forum where management explains the shift.
- Providence Health Plan Oregon DOI filings: regulatory run-off filings will determine whether policyholders have gap coverage and whether any MA plan assumes the membership — watch Oregon DOI for emergency orders in the next 72 hours.
Historical Power Lenses
J.P. Morgan 1837-1913
Morgan's defining move during the Panic of 1907 was to concentrate fragmented private capital into a single coordinated intervention, preventing systemic collapse by making the terms of liquidity supply explicit and credible. The Cayman QJS bid reads similarly: a jurisdiction consolidating its regulatory credibility to become the preferred conduit for dispersed ILS capital into U.S. risk markets. Morgan understood that the value of a clearinghouse is not just operational efficiency but the trust signal it sends to counterparties — NAIC QJS is precisely that trust signal for Cayman, reducing the haircut that U.S. ceding insurers must apply to Cayman-domiciled collateral and thereby lowering the cost of capital transfer. Morgan would recognize the move: it is not about adding capital but about removing friction that prevents existing capital from being deployed at scale.
Andrew Carnegie 1835-1919
Carnegie's vertical integration thesis — control the ore, the furnaces, the rails, and the ships — finds a direct analog in RockRose Risk's platform ambition: own the property inspection data, the mitigation verification, the brokerage relationship, and the insurance placement. Carnegie recognized that whoever controls the input to the production process controls the margin at every downstream step. In California wildfire insurance, the scarce input is defensible-space and ember-resistance data at the parcel level — data that standard cat models do not have. RockRose's $12.5M Series A is a bet that the company that builds that data moat first will control the underwriting margin for the hardest-to-place residential wildfire risk. Carnegie sold his steel empire at the top of the cycle; the question for RockRose is whether $12.5M is enough to reach the scale at which the data asset is defensible before a larger player acquires or replicates it.
Napoleon Bonaparte 1799-1815
Napoleon's doctrine of the central position — placing his army between divided enemy forces to defeat each in detail before they could combine — applies to the jurisdictional competition now playing out between Bermuda, Cayman, and Singapore for ILS domicile supremacy. Cayman's QJS application is an attempt to occupy the central position in U.S. cat-risk capital markets by achieving regulatory equivalence with Bermuda's existing NAIC recognition, preventing U.S. ceding insurers from being forced to choose one jurisdiction over the other purely on credit terms. Napoleon's campaigns also illustrate the risk of overextension: the same administrative reforms that made his empire efficient created a system too centralized to adapt when conditions changed rapidly. A Cayman ILS market whose competitive advantage rests on regulatory arbitrage rather than genuine underwriting expertise faces the same brittleness if NAIC standards tighten or if a major ILS loss event in a Cayman-domiciled vehicle triggers collateral disputes.
Thomas Edison 1847-1931
Edison's approach to invention was systematic and patent-portfolio-first: build the lab, generate the IP, then use the portfolio to create switching costs for competitors and dependency for customers. The 123 Lights Re transaction is the LADWP doing something Edison would recognize — using a new instrument (the parametric cat bond) to create institutional infrastructure around a risk that the traditional insurance market could not price or absorb. Edison's Menlo Park was not a single invention but a system for generating inventions; the LADWP wildfire bond, if replicated by other California public utilities facing inverse condemnation exposure, becomes an infrastructure for sovereign-adjacent entities to self-hedge catastrophe risk through capital markets rather than the insurance system. Edison lost the AC/DC war to Westinghouse partly because he underestimated the scalability of the competing platform — the risk for traditional property insurers is that public-utility cat bonds scale into a competing coverage architecture for the highest-value California wildfire risks, leaving the voluntary market with adverse selection in the residential layer.