Insurance and Reinsurance Withdrawal Patterns - What the Actuaries Know

Dossier 056 Date: 2026-04-04 Status: PRIVATE - intelligence analysis Analyst: por. Zbigniew Method: PARDES Cross-references: Dossier 050 (SAFE Aerie), 048 (Billionaire Prepping), 047 (2028 Signal Sweep)

SEED

Insurance withdrawal is the canary - when actuaries price risk as uninsurable, they are revealing a future the public hasn’t been told about.

PARAGRAPH

Six consecutive years of $100B+ insured catastrophe losses (2020-2025) have triggered a systemic retreat by the insurance industry from climate-exposed, politically unstable, and conflict-adjacent markets. This is not a business cycle - it is a one-way exit. State Farm, Allstate, Farmers, and 20+ carriers have abandoned Florida, California, Louisiana, Texas, and Colorado. Lloyd’s of London has mandated new war and cyber exclusions effective 2025 that quietly redefine what conflicts are “insurable.” Reinsurers Swiss Re and Munich Re are posting record profits while catastrophe exposure growth outpaces premium growth - they are harvesting the last profitable years before the math breaks. The billionaires receiving the same actuarial intelligence (Thiel, Zuckerberg, Bezos, Ellison) are buying property in insurable, defensible locations - Hawaii, New Zealand, Indian Creek Island, rural compounds - while the regions they abandon become uninsurable. The NFIP carries $22.5B in debt and faces reauthorization by September 2026. Swiss Re models project insured losses could hit $320B in a peak year by 2026. At 7% annual growth, losses double every decade. The insurance industry is not retreating from risk - it is retreating from a future it can see and we cannot.

Confidence: 0.85 - Primary data from Swiss Re sigma, Munich Re NatCat, Lloyd’s market bulletins, FEMA, state insurance departments. Projection confidence lower (0.6) for 2028 clustering thesis.

PESHAT (Facts - The Data)

1. US Homeowners Insurance Collapse - State by State

California

• State Farm: Stopped writing new home insurance policies in May 2023. Dropped 30,000 existing policyholders in high wildfire-risk areas. Nonrenewed 1,600+ policies in Pacific Palisades specifically, 2,000+ in Brentwood/Calabasas/Hidden Hills ZIP codes. In May 2025, approved 17% emergency rate hike on ~1 million remaining policies after $7.6B projected Palisades/Eaton fire payouts.
• Allstate: Stopped writing new policies in California. Among carriers that have “backed away” from the market.
• Farmers: Pulled back from California market, later partially re-entered for condo and renters.
• AIG, Chubb: Also retreated from California market.
• Scale: 150,000+ uninsured households in highest fire-risk areas. 1 in 5 homes in extreme fire risk areas lost coverage since 2019. Over 100,000 homeowners collectively dropped by major carriers between 2019-2024.
• Palisades/Eaton fires (Jan 2025): 11,800+ acres destroyed, 1,000+ structures burned, insured losses approaching $10B, total losses exceeding $50B. Combined LA wildfire insured losses: ~$40B.
• Moratorium: Commissioner Lara issued one-year moratorium on non-renewals/cancellations in fire-affected ZIP codes. Expires January 2026. When it expires, mass cancellations expected.

Sources:

• State Farm California overview
• California insurance crisis carrier list
• LA wildfire losses NBC
• State Farm rate hike agreement

Florida

• Scope of collapse: Florida home insurance market collapsed 78% according to Yahoo Finance report. Private insurers largely abandoned Florida homeowners, concluding large areas are “essentially uninsurable.”
• Citizens Property Insurance: Peak of 1.42 million policies in October 2023. Fell to 936,182 by start of 2025. Plummeted to 395,144 by start of 2026 - a 73% decrease from peak. This was driven by depopulation program transferring 546,000+ policies to new private carriers entering after 2022 tort reform.
• Stabilization signals: 17 new insurance companies entered Florida market after 2022 reforms. Citizens announced rate decrease of 8.7% for 2026 renewals. But this masks continued affordability crisis.
• Companies that left/reduced: Farmers, Progressive, AAA, and numerous smaller carriers. At least 6 carriers went insolvent between 2020-2023.
• Mortgage risk: National banks hold hundreds of billions in Florida mortgages that could become distressed if insurance remains unavailable. Without affordable insurance, property transactions freeze and home values crater.

Sources:

• Florida home insurance market collapse
• Citizens policies plummet 2025
• Citizens rate decrease 2026

Louisiana

• Scale: 20+ insurance companies stopped doing business in Louisiana. At least 11 went insolvent. ~120,000 property owners left scrambling for coverage in 2022-2023.
• Drivers: Hurricanes Laura (2020) and Ida (2021) caused billions in damage, bankrupted multiple carriers.
• Costs: Average homeowners insurance $4,031/year - $1,608 above national average.
• Partial recovery: 10 new insurers began offering products since 2024. Rate increase requests dropped from 80 (2023) to 50 (2024). Average rate increases dropped from 14% to 6.6%.

Sources:

• Louisiana insurance crisis
• Louisiana insurers gone broke or left

Texas

• Scale: Over 400 insurance companies stopped writing new business in Texas since 2021 (per Texas Department of Insurance).
• Drivers: Severe convective storms caused $52B in insured losses nationally in 2025 - Texas at the epicenter. Hail and tornado exposure.
• Premiums: Average homeowners premiums increased 40-60% since 2019. Some coastal/hail-prone areas saw 100%+ increases.
• Nonrenewals: Complaints more than doubled in 2024. Homeowners receiving notices citing “high wind and hail exposure” even without prior claims.
• TWIA deficit: Texas Windstorm Insurance Association started 2025 with $413.5 million deficit.

Sources:

• Texas insurance crisis
• Texas homeowners losing coverage

Colorado

• Scale: Premiums rose nearly 60% over five years. Rates more than doubled between 2020-2025. Ranked 8th most expensive state for home insurance.
• Dual catastrophe: Both wildfire and hail. Marshall Fire (2021) burned 1,000+ houses in suburban Boulder County.
• Underinsurance: 74% of Marshall Fire-affected homeowners were underinsured. 36% so severely underinsured their policy covered less than 75% of rebuild cost.
• Market withdrawal: Insurers making decisions to not write throughout the entire state.

Sources:

• Colorado insurance crisis CPR
• Colorado communities priced out

National Picture

• 1 in 13 homeowners (7.7%) is uninsured nationally - Consumer Federation of America 2024 analysis.
• $1.6 trillion in unprotected market value.
• 14% of owner-occupied homes uninsured nationwide, jumping 6%+ from 2023 to 2024.

Source: Center for American Progress

2. Reinsurance Pricing Signals

The Big Three

• Swiss Re: Top reinsurer by 2024 GPW at $43.1B gross reinsurance premiums. Record group net income. P&C Re profit $2.8B.
• Munich Re: Targeting EUR 6.3B profit in 2026. ROE above 18% by 2030. Slashed retrocession, scrapped sidecars - retaining more profit in-house.
• Hannover Re: Third largest global reinsurer.

Pricing Trajectory

• January 2026 renewals: Swiss Re achieved only 0.3% price increase, but loss assumptions increased 4.6%, yielding a net price decrease of 4.3%. This means reinsurers are accepting less money relative to expected losses. The cycle is softening at exactly the wrong time.
• Soft market expected in personal and commercial lines 2025-2026. Strong industry returns on equity driving competition for market share and putting pressure on rates.
• The paradox: Reinsurers are posting record profits while the math of catastrophe losses grows exponentially. They are harvesting the late-cycle gains before the market breaks.

Catastrophe Exposure Growth Outpacing Premiums

Swiss Re’s own analysis shows catastrophe exposure growth continues to outpace insurance premiums - meaning the gap between what’s at risk and what’s covered keeps widening.

Sources:

• Swiss Re overtakes Munich Re
• Munich Re Ambition 2030
• Catastrophe exposure outpacing premiums

3. Nuclear/War Risk Exclusions - Lloyd’s of London

Cyber War Exclusions (Effective January 2025)

• Market Bulletin Y5381 (August 2022): All standalone cyber-attack policies must include exclusions for state-backed cyber-attacks.
• Market Bulletin Y5433: Updated wordings for state-backed cyber-attack exclusions.
• January 1, 2025: Type 4 clauses (which provided cover for state-backed cyber-attacks during conventional war) eliminated for new and renewal policies. Syndicates wanting to cover war-related cyber must do so through separate, affirmative products.
• Attestation requirements: From January 31, 2025, syndicates must attest compliance with cyber war exclusion requirements twice yearly.

Translation: Lloyd’s is systematically removing coverage for the exact scenarios that geopolitical tensions make more likely - state-sponsored cyber-attacks on infrastructure. They are not adding coverage for new risks; they are removing it. This is the actuarial equivalent of quietly moving to higher ground.

Nuclear Risk Exclusions

• LMA5621 (2023 revision): Retains broad restrictions on ALL “nuclear” activities including fusion. The Lloyd’s Market Association’s nuclear exclusion clause has not been loosened - it has been maintained at maximum breadth even as nuclear energy investment surges globally.

Sources:

• Lloyd’s cyber war exclusion developments
• Lloyd’s state-backed cyber-attack mandate
• Lloyd’s updated cyber war wordings
• Nuclear risk exclusion modernization case

4. Climate Insurance Retreat - Geographic Map

The Withdrawal Pattern

| Region | Primary Peril | Insurer Status | Population Density | |——–|————–|—————-|——————-| | Southern Florida | Hurricane/flood | Mass withdrawal, partial recovery post-reform | Very high (6M+ Miami metro) | | California coast + WUI | Wildfire | Major carriers exited, moratorium expiring Jan 2026 | Very high (LA 13M metro) | | Louisiana Gulf Coast | Hurricane | 20+ carriers exited, 11 insolvent | High (NOLA 1.3M metro) | | Texas Gulf Coast + Tornado Alley | Hurricane/hail/tornado | 400+ carriers stopped new business since 2021 | Very high (Houston 7M metro) | | Colorado Front Range | Wildfire + hail | Carriers pulling back statewide | High (Denver 3M metro) | | Outer Banks / Carolinas | Hurricane | Rising premiums, reduced availability | Moderate | | Pacific Northwest | Wildfire/earthquake | Tightening underwriting | High (Portland/Seattle metros) |

The Numbers That Matter

• 2024 global economic losses: $320B (Munich Re) - highest adjusted total on record.
• 2024 insured losses: $140B (Munich Re). $141B (Swiss Re). Sixth consecutive year above $100B.
• 2025 insured losses: $107B. Down from 2024, but $40B from LA wildfires alone.
• 2026 projections: $148B on trend. $320B in peak-loss scenario (Swiss Re).
• Annual growth rate: 7% real terms - equivalent to doubling every decade.
• Protection gap: $1.83 trillion globally (Swiss Re, premium equivalent terms, 2023). ~49% of 2025 economic losses were insured - meaning ~51% was uninsured.

Sources:

• Munich Re 2024 natural disaster figures
• Swiss Re 2025 marks sixth year above $100B
• Swiss Re sigma natural catastrophe trend

5. Political Risk Insurance

Market Status

• Global capacity: Nearly $4B in 2025 ($2.23B private insurers, $1.75B Lloyd’s syndicates). Up from $3.71B in 2024.
• Demand surge: Howden’s 2025 survey of multinationals with $1B+ revenue estimates PRI demand to rise 33% due to trading environment instability and tariff uncertainty.
• Standard pricing: ~1% of limit (e.g., $1M annual premium for $100M coverage).
• DFC/MIGA: Continue offering standard political risk coverage (political violence, expropriation, currency inconvertibility, breach of contract).

What the PRI Market Signals

The 33% demand surge is the signal. Multinational corporations are not buying more political risk insurance because the world is getting safer. They are buying it because their internal risk teams - which have access to the same actuarial and geopolitical intelligence as insurers - see elevated instability ahead.

Sources:

• DFC Political Risk Insurance
• MIGA Political Risk Insurance

6. Health/Pandemic Insurance

Post-COVID Reality

• Pandemic risk is “largely uninsurable” - consensus of actuaries, insurers, and reinsurers. Fails key insurability criteria: losses are too large, too widespread, too correlated, and too difficult to predict.
• Business interruption exposure: Nearly all BI exposure units had highly correlated risk under COVID-19, violating the independence assumption fundamental to insurance.
• Model resolution problem: Existing pandemic models cannot differentiate high from low risks at sufficient resolution for actuarial pricing.
• Moral hazard: Government responses (lockdowns) determined the size of BI losses, creating a moral hazard actuaries cannot price.

Regulatory Response

• Standardized pandemic exclusions now widespread across commercial lines.
• Mandatory BI coverage options being developed in some jurisdictions.
• GAO analysis concluded federal pandemic insurance approaches would entail significant costs to taxpayers, and businesses might not participate.
• No federal pandemic backstop has been created - unlike terrorism (TRIA), there is no mechanism to absorb pandemic losses.

The Signal

The insurance industry has effectively declared pandemics uninsurable - meaning the next pandemic’s economic losses will be entirely absorbed by businesses, governments, and individuals. There is no private-sector safety net. The actuaries looked at the numbers and walked away.

Sources:

• COVID-19 implications for insurer risk management
• GAO pandemic risk federal insurance
• Pandemic protection gaps - International Actuarial Association

7. The “Uninsurable” Threshold - Florida Case Study

The Cascade

1. Insurers withdraw - carriers exit or raise premiums beyond affordability.
2. Insurer of last resort absorbs load - Citizens Property Insurance peaked at 1.42M policies.
3. State intervenes - tort reform (2022), new carrier incentives, depopulation programs.
4. Property values respond - without affordable insurance, mortgages cannot be written. No mortgages = no transactions = price collapse.
5. Mortgage portfolio risk - national banks holding hundreds of billions in Florida mortgages face distressed asset scenario.
6. Tax base erodes - property tax revenue falls as values decline.
7. Infrastructure investment stops - why build in an uninsurable zone?

NFIP: The Federal Time Bomb

• Debt: $22.525B borrowed from Treasury as of February 2025. $7.9B remaining borrowing authority.
• Reauthorization deadline: September 30, 2026. If not reauthorized, borrowing authority drops from $30.425B to $1B.
• Risk Rating 2.0: FEMA’s updated pricing methodology (fully implemented April 2023) has caused rate increases of up to 183% in some areas. 18% annual cap limits the speed, but the trajectory is clear.
• Participation decline: Premium increases under Risk Rating 2.0 have substantially reduced NFIP uptake, particularly in lower-income communities.
• The math: Only current and future NFIP participants are responsible for repaying the $22.5B debt through premiums. As participants drop out, the remaining pool bears more debt per capita. This is a death spiral.

Sources:

• NFIP reauthorization status
• NFIP borrowing authority
• Risk Rating 2.0 leaving households vulnerable

8. Lloyd’s Realistic Disaster Scenarios

Lloyd’s requires all syndicates to stress-test against specific catastrophe scenarios. The 2025 RDS framework includes:

Compulsory Natural Catastrophe Scenarios

• Gulf of Mexico windstorm
• European windstorm
• Japanese typhoon
• California Earthquake: Los Angeles
• California Earthquake: San Francisco
• New Madrid earthquake
• Japanese earthquake
• UK flood

Compulsory Terrorism Scenarios

• Terrorism: Rockefeller Center
• Terrorism: One World Trade Center
• Note: CBRN (Chemical, Biological, Radiological, Nuclear) hazard exposures are explicitly excluded from certain terrorism scenarios, suggesting separate classified modeling.

Compulsory Cyber Scenarios

• Cyber power outage
• Widespread malware event
• Cloud outage event

Extreme Disaster Scenario (New 2025)

• Strike, Riot and Civil Commotion (SRCC) triggered by Cascadia Subduction Zone earthquake. Modeling shows follow-on civil unrest could exceed $20B in insured losses - mid-sized cities and “overlooked exposures” posing unexpected threats.

The signal: Lloyd’s is now modeling civil unrest as a compounding factor on natural disasters. They are pricing the scenario where an earthquake triggers not just physical damage but social breakdown.

Sources:

• Lloyd’s Realistic Disaster Scenarios
• SRCC losses could top $20B

REMEZ (Connections - What the Patterns Reveal)

Connection 1: The Actuarial Intelligence Advantage

Insurers now use:

• Daily satellite imagery at 3-meter resolution covering the entire Earth’s land surface
• Machine learning classifying individual property risk (roof condition, vegetation proximity, building materials, defensible space)
• AI catastrophe models simulating “black swan” scenarios with 20% greater accuracy than traditional models
• Real-time flood monitoring fusing satellite observation with ground data (Floodbase)
• Wildfire prediction: Companies like Kettle analyze billions of lines of weather, fuel, satellite, anthropogenic, and ground truth data to predict wildfire likelihood per square mile

Translation: The insurance industry has access to a surveillance and prediction infrastructure that rivals military intelligence. They can see, at the individual property level, what risks are coming. When they withdraw from a market, they are acting on information the public does not have.

Source: AI catastrophe modeling with satellite imagery

Connection 2: Billionaire Geography Correlates with Insurance Geography

The pattern: Not one of these locations is in the insurance withdrawal zones (inland Florida Gulf, California WUI, Louisiana coast, Tornado Alley). They are buying where insurance remains available - meaning where actuaries still see the risk as manageable. The billionaires are not buying survival bunkers in random locations. They are buying in the places the insurance industry’s models say will still be functional.

The Fortune article from April 2, 2026 is explicit: billionaires are “bolting for Florida from the West Coast” - but specifically to Indian Creek Island and Palm Beach, not to Tampa or Jacksonville. They are fleeing California (uninsurable) for the one part of Florida (insurable, sea-walled, privately secured) that still works.

Sources:

• Zuckerberg Hawaii bunker
• Bezos Indian Creek purchases
• Zuckerberg Indian Creek purchase
• Billionaire tax migration to Florida
• New Zealand “apocalypse insurance”

Connection 3: The Reinsurance Paradox

Swiss Re and Munich Re are posting record profits. Munich Re targets EUR 6.3B profit in 2026. Swiss Re delivered record net income in 2024. Yet:

• Catastrophe exposure is outpacing premiums (Swiss Re’s own analysis)
• Insured losses have grown at 7% annually = doubling every decade
• Reinsurance pricing softened at January 2026 renewals (net 4.3% price decrease)
• Munich Re is slashing retrocession and scrapping sidecars - retaining profits rather than sharing risk

Translation: The Big Three reinsurers are behaving like someone who knows the party ends soon. They are maximizing current extraction (record profits), reducing exposure to external claims (slashing retrocession), and positioning for a future where the current model no longer works. They are not investing in expansion. They are harvesting.

Connection 4: Lloyd’s Models What Governments Won’t Say

Lloyd’s compulsory scenarios include:

• Nuclear terrorism at specific NYC targets
• Cascadia Subduction Zone earthquake triggering civil unrest
• State-backed cyber-attacks on power infrastructure (modeled on Ukraine 2015-2016 attacks)
• Massive simultaneous hurricane scenarios

These are not hypothetical exercises. They are mandatory stress tests that every Lloyd’s syndicate must price into their reserves. Lloyd’s is requiring the entire London insurance market to hold capital against scenarios that political leaders insist are unlikely. The gap between what Lloyd’s models and what politicians say is the gap between actuarial truth and political comfort.

DRASH (Mechanism - How Insurance Withdrawal Cascades)

The Feedback Loop

Catastrophe losses increase (7% annual, doubling/decade)
    -> Insurers raise premiums or withdraw
        -> Reinsurance becomes more expensive
            -> Primary insurers cannot afford reinsurance
                -> More primary insurers withdraw
                    -> State insurer-of-last-resort absorbs load
                        -> State insurer becomes insolvent or taxpayer-funded
                            -> Property values decline (no mortgage without insurance)
                                -> Tax base erodes
                                    -> Infrastructure investment stops
                                        -> Vulnerability increases
                                            -> Catastrophe losses increase further

This is a positive feedback loop with no natural stabilizer. The only circuit breaker is government intervention (NFIP, Citizens, TWIA, FAIR Plans), but these entities are already insolvent or approaching insolvency:

• NFIP: $22.5B in debt
• TWIA (Texas): $413.5M deficit
• Citizens (Florida): Depopulating by policy count but this just transfers risk to new, untested private carriers

The “Uninsurable” Cascade Timeline

Based on current trajectory (7% annual loss growth, carrier withdrawals, reinsurance softening):

Strongest Counter-Argument (Adversary)

The insurance market is cyclical. Florida’s market is already stabilizing - 17 new carriers entered, Citizens is depopulating, rates are decreasing. The 2022 tort reform worked. Reinsurers are profitable. This could be a temporary adjustment, not a structural collapse.

Response: The cycle argument held when catastrophe losses were stationary. They are not. The 7% annual growth in insured losses is driven by climate change (increasing hazard frequency and severity) and asset concentration (more expensive things in harm’s way). The new Florida carriers are small, untested, and undercapitalized. One major hurricane season will reveal whether the “stabilization” was real or a mirage built on subsidized wind pools and litigation reform.

SOD (Emergent Pattern - What the Actuaries Know)

The Core Insight

Insurance companies are the only institutions in the global economy that must, by regulatory requirement, price the future accurately or go bankrupt. They cannot afford ideology, optimism, or political expedience. Their models must be right, or they die.

When the entire insurance industry simultaneously:

1. Withdraws from climate-exposed regions
2. Excludes state-sponsored cyber warfare from coverage
3. Declares pandemics uninsurable
4. Models civil unrest as a compounding factor on natural disasters
5. Posts record profits while reducing long-term exposure

…it is telling us something about the next decade that no government report, media analysis, or think tank paper will say plainly:

The actuarial models show a convergence of climate, geopolitical, and pandemic risks that makes significant portions of the built environment functionally uninsurable by the late 2020s to early 2030s.

The 2028 Cluster - What Exists

The 2028 cluster in insurance is not a single regulatory deadline. It is the convergence point where multiple deteriorating trends reach critical mass simultaneously. The trajectory lines are clear. The exact year is uncertain. But the direction is not.

TZELEM (When This Truth Is Weaponized)

How This Intelligence Gets Corrupted

1. Disaster capitalism: Entities with actuarial intelligence buy distressed properties in uninsurable zones at deep discounts, then lobby for government insurance backstops that socialize the risk while privatizing the recovery.

2. Insurance as population control: When insurance withdrawal makes regions uninhabitable for the middle class but not the wealthy (who can self-insure), it becomes a mechanism for economic displacement - a market-driven form of forced migration.

3. Actuarial intelligence as insider trading: Those with access to proprietary catastrophe models (reinsurer executives, tech billionaires who fund climate research) can position their real estate portfolios before the public models catch up. The billionaire bunker geography is circumstantial evidence this is already happening.

4. Climate nihilism: The insurance withdrawal narrative can be weaponized to argue that climate adaptation is futile, suppressing support for mitigation measures that would actually reduce losses.

5. NFIP as political weapon: The September 2026 reauthorization becomes a hostage situation - flood-prone states’ representatives forced to accept policy concessions in exchange for continued insurance availability.

KEY FINDINGS SUMMARY

1. The insurance industry is conducting an orderly retreat from the future. Six consecutive years of $100B+ losses, 7% annual growth, and carrier withdrawals from five major US states represent a structural, not cyclical, shift.

2. The actuaries have better data than the public. Daily satellite imagery, AI catastrophe modeling, and proprietary risk analytics give insurers a 5-10 year visibility advantage over public discourse.

3. Billionaire property geography tracks insurable geography. Thiel (NZ), Zuckerberg (Hawaii + Indian Creek), Bezos (Indian Creek + Hawaii), Ellison (Hawaii + Palm Beach) - all in areas that remain insurable and defensible.

4. Lloyd’s is modeling scenarios that governments deny. Nuclear terrorism, civil unrest cascades, state-sponsored infrastructure attacks - all compulsory stress tests for every Lloyd’s syndicate.

5. The NFIP is a $22.5B time bomb expiring September 2026, with a death spiral dynamic as participants exit and remaining policyholders absorb more debt per capita.

6. Pandemics have been declared actuarially uninsurable. No private-sector safety net exists for the next pandemic. No federal backstop has been created.

7. The 2028 convergence is plausible but not proven. Multiple insurance-sector timelines (loss trajectory, NFIP, Solvency II, California moratorium cascade, reinsurance cycle) point to late 2020s as a structural crisis window. Confidence: 0.6.

Classification: PRIVATE - Intelligence analysis for Zbigniew Brief subscribers Next steps: Cross-reference with Dossier 050 (SAFE Aerie geography) and 048 (billionaire prepping patterns). Map insurance withdrawal zones against Dossier 044 (opposition infrastructure) to identify which political organizing regions lose economic viability first.