Skip to content
Aerial view of the Munich Re headquarters in Munich

VII Reinsurance & Specialty Risk·Analysis·Europe

Reinsurance capital is plentiful, but unmodellable risk waits at the state's door

The catastrophe bond multiple has fallen from 5.15 in 2023 to 2.4; private capital competes over modellable catastrophe risk. Non-catastrophe issuance in the third quarter was zero. War, cyber and uninsured disaster losses move to the public balance sheet, yet public pools do not work everywhere.

Talay Insight Editorial · 31 October 2026 · 10 min read · 21 sources

Talay Insight Analysis DeskAI-assisted analysis, reviewed by an editorPublisher of record: Lu Apps LLCTrack recordRequest a correction

Munich Re headquarters, Munich, May 2022 — archive photo, representativePhoto: Carsten Steger / Wikimedia Commons · CC BY-SA 4.0 · resized · Source

Why it matters

Noise: reinsurance capacity is at a record, so no risk is left uninsurable. Signal: that capacity flows into a single channel. While the catastrophe bond multiple fell to 2.4, Artemis counted no third-quarter issuance carrying non-catastrophe risk. Aon puts the nine-month protection gap at 55%, and roughly 25% of losses in the EU are insured. Washington's 40 billion dollar Hormuz facility had written no voyage as of 3 September, while India's 1.5 billion dollar pool issued policies. The question is not how much capacity exists, but which balance sheet the risk lands on.

Implications

  • The average catastrophe bond multiple (the ratio of the interest spread to expected loss) stood at 5.15 in 2023 and fell to 2.4 in the first nine months of 2026; over the same period expected loss rose from 1.7% to 2.85%, so investors are paid less for carrying more risk.
  • The US DFC's 40 billion dollar marine reinsurance facility had insured no voyage as of 3 September, while India's 1.5 billion dollar pool issued policies on the back of a 1.4 billion dollar sovereign guarantee. Public pools solve the access problem, not the price problem.
  • Compulsory earthquake insurance in Türkiye covers 58% of homes, or 11.7 million dwellings; of the 104 billion dollar loss from the Kahramanmaraş earthquakes, roughly 5.5 billion dollars was met by insurance. Abundant global capacity does not close that gap on its own.

Noise

Record capacity leaves no risk uninsurable

Signal

Capital is chasing the tower downwards, not spreading sideways

Signal vs Noise ›

Map: Reinsurance capital is plentiful, but unmodellable risk waits at the state's door

Where the abundance flows

In yesterday's analysis of the January renewal we wrote that pricing is moving at two speeds. This piece asks a different question: which risk is abundant capital refusing to carry, and where does that risk go? The Artemis data gives the first answer. In catastrophe bonds — instruments whose principal passes to the insurer once a disaster is triggered — the average multiple was 5.15 in the first nine months of 2023 and 2.4 over the same period of 2026. Expected loss rose from 1.7% to 2.85% in that time, while the average interest spread fell from 8.76% to 6.83%.

A single transaction shows the shift more sharply. The Hartford's Foundation Re IV bond paid 6.25% against expected loss of 1.36% in 2023; the 2027-1 series, which came to market on 30 October, offers 3 to 3.5% for expected loss of 1.56%. In the same week the California Earthquake Authority opened a tranche with expected loss of 6.35% at a spread range of 8.5 to 9.5%. Capital is chasing the tower downwards.

That same capital is not spreading sideways. According to Artemis, none of the 948 million dollars issued in the third quarter carried non-catastrophe risk; the last third-quarter transaction in that class was Beazley's 210 million dollar cyber bond in 2024. Jordan Nickerson, a co-founder of Icosa, said on 30 October that how cyber behaves in a systemic event remains untested. The nine-month record of 18.9 billion dollars is a record for modellable catastrophe risk.

Yields are not closing the door, they are sorting it

The standard reading runs as follows: when risk-free yields rise, reinsurers trim their risk appetite. The 2026 data supports that in part, but the mechanism is finer. The US 10-year yield stood at 5.05% on 28 October; SCOR's reinvestment yield rose from 4.3% on 30 June to 5% on 30 September. High yields are not driving capital away from catastrophe risk. Because catastrophe bonds carry a floating coupon, the investor pockets the collateral yield and protects the total return even as the spread narrows. Nickerson attributes insurance-linked securities returns holding above their historical average to precisely that collateral yield.

This is where yields do the sorting. Because the total return comes from collateral, the multiple paid for modellable risk can compress as far as 2.4. For unmodellable risk, the same investor sees no measurable excess return that would beat a risk-free yield of around 5% over a near horizon. That last sentence is our inference; no source establishes it with figures. The contradiction should also be noted: comments from Chubb's chief executive Evan Greenberg, reported by Artemis on 21 October, attribute the money flowing into insurance to low interest rates, which does not square with SCOR's 5% figure.

War: the pool solves access, and only access

Two state pools produced two different outcomes in the same year. The Chubb-led facility of the US Development Finance Corporation rose from 20 billion to 40 billion dollars, yet Insurance Business reported on 3 September that it had insured no voyage. Current usage data for 30 October could not be verified. According to the May report, cover was conditional on a US Navy escort, and that convoy system was never established. Howden Re data puts the Strait of Hormuz transit premium at 0.10 to 0.125% of hull value, rising to 2 to 3% in March, yet capacity remained open. In a Lloyd's survey, 88% of marine war underwriters maintained their appetite for hull risk. Brokers say the problem was price, not access; the pool was built on the wrong diagnosis.

India's pool targeted a different gap. The 1.5 billion dollar Bharat Maritime Insurance Pool draws on the shared capacity of domestic insurers behind a 1.4 billion dollar sovereign guarantee, and wrote its first war policies in May. On our reading the pool worked because it addressed an access problem for domestic buyers. The private market was not idle either: in June Lloyd's and Chubb formed a consortium totalling 400 million dollars across hull, liability and cargo, and SCOR confirmed on 30 October that marine, energy and war lines had seen significant rate increases. According to the Congressional Research Service, the facility's 20 billion dollar share can consume 97.6% of the 20.5 billion dollar ceiling allocated to high-income economies.

Climate and cyber: the tail walks to the public desk

In natural catastrophe, the size of the gap has now entered policy documents. Aon put the nine-month protection gap at 55%: of 204 billion dollars in economic damage, 91 billion dollars was insured. On 27 October the European Commission said that historically around 25% of losses have been insured, and announced a Climate Insurance Alliance that will bring forward its first measures by the end of 2027. The text does not propose a publicly backed reinsurance mechanism at EU level; it lists catastrophe bonds only among the options to be assessed. The timetable sits roughly 12 months behind the January renewal.

James Talbot of the Bank of England said on 21 October that the share of uninsured mortgage borrowers could rise from 5% to between 7 and 10% by 2050, and to 16% after a severe flood. That moves the insurance gap into monetary policy and financial stability. Flood Re is doing the same thing on the balance-sheet side: after its first bond of 140 million pounds it plans to add a layer, having found conventional reinsurance limited and expensive. Here the public vehicle is becoming less a carrier of risk than a channel that passes risk to the private market.

In cyber the debate is still at the opening stage. Travelers raised the ceded share in its cyber quota share treaty on 1 October; the effect on net premium is roughly 165 million dollars in the fourth quarter and around 230 million dollars in 2027. Chief executive Alan Schnitzer said he did not know whether a state-backed cyber insurance backstop was needed, but that discussions were continuing. No law or figure currently in force for a cyber pool could be verified.

Türkiye and frontier markets: cede it or carry it

Türkiye stands on the public side of this debate. According to Anadolu Agency data from 16 August, compulsory earthquake insurance covers 11.7 million dwellings, a penetration rate of 58%, rising to 65% in Marmara and falling to 45% on the Black Sea coast. Of the 104 billion dollar loss from the Kahramanmaraş earthquakes, roughly 5.5 billion dollars came from the insurance sector and 1.6 billion dollars from DASK. The fund size DASK disclosed on 27 September is about 30 billion lira; the institution says the fund, its reinsurance and state support where needed can each cover two major earthquakes separately. The amount of state support is not disclosed, so the capacity claim cannot be independently verified.

Abundant capacity gives this structure an opening, on two conditions. Özgür Bülent Koç, acting general manager of Türk Reasürans, said in March that the planned Compulsory Disaster Insurance would extend cover to villages and could raise premiums by 25 to 30%; the average policy costs around 1,700 lira. Whether it has been enacted could not be verified. The first condition is that coverage rises above 58%; the second is whether DASK will use transfer instruments such as catastrophe bonds, on which no statement could be found. Smaller markets point the way: Nepal has signed a mandate with the World Bank for a parametric earthquake bond of 80 to 190 million dollars. Munich Re, with the Africa Finance Corporation, is sharing the risk on trade credits of no more than 12 months within predetermined limits under a 1 billion dollar programme. Private capital will enter frontier markets too, where it sees bounded, measurable risk over a defined period.

What would disprove the thesis

The thesis has three parts and each has its falsifier. First, that abundant capital will not take unmodellable risk: a 500 million dollar cyber catastrophe bond, or 1 billion dollars of total issuance carrying non-catastrophe risk, before 31 March 2027 would disprove it. Second, that public pools do not solve the price problem: if the DFC facility insures voyages without an escort condition in the first quarter of 2027 and brings the Hormuz premium down, this reading breaks. Third, the sorting role of yields: if the US 10-year yield falls below 4.50% while the average multiple stays under 2.4, the collateral yield channel weakens. A multiple above 3 in the Artemis fourth-quarter report would equally disprove the cheap capital thesis.

The near calendar is set. Munich Re publishes its final quarterly report on 12 November. SCOR presents its new strategic plan on 3 December. Reporting dates for the nine-month results of Swiss Re and Hannover Re could not be verified for this piece. No current official data on usage of the DFC facility could be found either.

Probabilities

Scenarios

ScenarioProbabilityTriggerMarket impact
H1The divergence holds55%No major catastrophe before the first quarter of 2027, issuance passes 20 billion dollars, non-catastrophe issuance stays smallProperty catastrophe gets cheaper; war and cyber stay capped, the public debate gathers pace but no new major pool is legislated
H2Private capital moves into the tail25%A new cyber bond or war sidecar is established, and 1 billion dollars of non-catastrophe issuance appears before 31 March 2027Pressure for public pools eases, per-risk limits widen, cyber pricing falls by single digits
H3A shock makes the state unavoidable20%A systemic cyber event, a fresh wave of Gulf losses or a major late-season catastrophe pulls private capacity backCover contracts and states issue guarantees in haste; contingent liabilities land on the balance sheet unplanned

Module A

Constraints Matrix

STRUCTURAL AVG 4.0 · TACTICAL AVG 2.5Structural constraints dominate: the outcome is set more by these limits than by the actors' preferences.

Hard structural constraintspersistent · beyond the actors' will

  • The modelling limit

    5/5

    According to Artemis, none of the 948 million dollars issued in the third quarter carried non-catastrophe risk; Icosa treats the behaviour of cyber in a systemic event as untested.

  • The risk-free yield threshold

    4/5

    The US 10-year yield stood at 5.05% on 28 October and SCOR's reinvestment yield at 5%. Unmodellable risk has to pay a premium above that return.

  • Coverage in Türkiye · Türkiye

    4/5

    Compulsory earthquake insurance reaches 58% of homes and 45% on the Black Sea coast; of the 104 billion dollar Kahramanmaraş loss, roughly 5.5 billion dollars was insured.

  • The DFC statutory ceiling · United States

    3/5

    The Congressional Research Service puts the high-income economy ceiling at roughly 20.5 billion dollars; the facility's 20 billion dollar share can consume 97.6% of it.

Tactical frictiontemporary · eases over time

  • The escort condition weeks

    3/5

    Because DFC cover is tied to a US Navy escort, it goes unused until a convoy system exists; zero voyages had been written as of 3 September.

  • The transparency gap months

    3/5

    Neither the amount of state support behind DASK nor current DFC usage is disclosed; capacity claims cannot be independently measured.

  • Multiple compression weeks

    2/5

    The average multiple has fallen 2.75 points since 2023; if investors demand extra return in the lower layer, transactions may shrink.

  • The EU timetable months

    2/5

    The Climate Insurance Alliance will bring its first measures by the end of 2027; it does not affect the 1 January 2027 renewal and proposes no EU-level mechanism.

Module B

Signal vs Noise

SIGNAL 67% · NOISE 33%

Module C

Asset-Class and Positioning Implications

Asset classExposureTransmission channelH1H2H3ExpectedConvictionHorizonWhat to watch
CreditInsurance-linked securities spreadsHeavy issuance and risks moving down the tower are compressing the excess return−−++−0.40●●●3–12 monthsThe path of the average multiple against 2.4 in the Artemis fourth-quarter report
Freight & insuranceThe cost of Gulf marine war coverPrivate consortium ceilings and a public facility tied to an escort condition+−−+++0.45●●●0–3 monthsThe Strait of Hormuz transit premium and the first policy announced by the DFC facility
Sovereign debtContingent liabilities of states that guarantee disaster and war riskPublic pools and guarantees are accumulating obligations off the balance sheet0−+++0.15●●●12+ monthsDisaster and insurance guarantee lines in budget documents
EquitiesThe European reinsurance sectorThe balance between margin pressure and a 5% reinvestment yield0+−−−0.15●●●3–12 monthsMunich Re's report on 12 November and SCOR's strategic plan on 3 December

How to read: ++ strong structural support · + support · 0 neutral · − pressure · −− strong pressure. “Expected” is the direction weighted by scenario probabilities. H1: The divergence holds · H2: Private capital moves into the tail · H3: A shock makes the state unavoidable.

General, scenario-conditional analysis at asset-class level. It contains no specific security, price target or trade timing and is not personalised investment advice (Turkish Capital Markets Law No. 6362).

Second-order effects

And then what?

Starting point

The average catastrophe bond multiple has fallen from 5.15 in 2023 to 2.4, and no third-quarter issuance carried non-catastrophe risk; abundant capital is concentrating in modellable catastrophe risk.

  1. 1

    Reinsurance capacitywithin weeks

    Cheap capital chases the tower downwards: sponsors open tranches with expected loss of 6.35% and property catastrophe pricing falls on 1 January. As the cost of modellable risk recedes, war and cyber lines stay outside that decline.

    Watch: The final price of the Ursa Re H tranche, and the war and cyber lines in broker reports at the turn of the year

  2. 2

    Public policywithin months

    While private limits on unmodellable risk stay capped, policyholders and regulators press for a public solution; the EU Alliance, the US cyber backstop discussions and Türkiye's compulsory disaster insurance agenda all move the same way.

    Watch: The membership list of the EU Climate Insurance Alliance, the US cyber backstop discussions and the compulsory disaster insurance text in the Turkish parliament

  3. 3

    Budget and sovereign riskwithin months

    As public pools grow, tail risk is booked as a contingent liability in the budget and part of the burden passes into premiums. The product planned in Türkiye could raise policy costs by 25 to 30%, with the remainder entering the sovereign risk account.

    Watch: Disaster liabilities in Türkiye's 2027 budget, the DASK premium tariff and the DFC annual exposure report

What breaks the chain

If a credible model for cyber or war risk gains acceptance and a 500 million dollar catastrophe bond finds investors, the chain breaks at the second step; private capital takes on the tail and the public desk empties.

Triggers

Thresholds to watch

IndicatorThresholdTodayWhat it means
US 10-year yield< 4.505.07If the collateral yield falls, investors must demand total return from the spread instead; unless the multiple turns back up from 2.4, the yield channel of the thesis weakens
US 10-year yield> 5.255.07A risk-free yield above this level raises the excess return demanded for unmodellable risk and makes private capacity in war and cyber lines still more selective

Sources

  1. GTR — Lloyd's unveils US$400mn marine war risk consortium for Hormuz ships (24 June 2026)
  2. Artemis — Q3 2026 Catastrophe Bond & ILS Market Report
  3. Artemis — The Hartford returns with $280m target for Foundation Re IV 2027-1 cat bond (30 October 2026)
  4. Reinsurance News — SCOR delivers Q3'26 net income of €206m and P&C combined ratio of 84.1% (30 October 2026)
  5. Artemis — Yield normalisation, evolving risk profiles demand structural discipline in ILS: Nickerson, Icosa (30 October 2026)
  6. Artemis — There's a lot of hungry and extremely naive capacity available: Chubb CEO Greenberg (21 October 2026)
  7. Insurance Business — Washington's $40 billion Hormuz insurance fix: zero takers (18 May 2026)
  8. Insurance Business — Lloyd's puts a $1.9bn number on the Iran war. Washington's own fix has written nothing (3 September 2026)
  9. US Congressional Research Service — DFC Shipping Reinsurance Facility: Iran Conflict and Strait of Hormuz (6 May 2026)
  10. ReinAsia — India rolls out war risk cover under US$1.5bn Bharat Maritime Insurance Pool (13 May 2026)
  11. Artemis — Global insured natural catastrophe losses reach $91bn in 9M'26, lowest since 2019: Aon (29 October 2026)
  12. Reinsurance News — EU unveils climate resilience framework with re/insurance protection gap in focus (27 October 2026)
  13. Reinsurance News — Bank of England warns climate risks could widen UK insurance gaps (28 October 2026)
  14. Insurance Journal — Flood Re cat bond plans (27 February 2026)
  15. Reinsurance News — Travelers increases ceded percentage to cyber quota share reinsurance: CFO Frey (20 October 2026)
  16. Memurlar.net (Anadolu Agency) — Half of all homes have no compulsory earthquake insurance (16 August 2026)
  17. Capital — DASK has paid out on some 702,000 claims in 26 years (27 September 2026)
  18. Bloomberg HT — Türk Reasürans/Koç: compulsory disaster insurance could raise premiums by 25-30% (4 March 2026)
  19. Reinsurance News — Nepal and World Bank Group sign mandate to develop cat bond for earthquake (16 October 2026)
  20. Reinsurance News — Munich Re and Africa Finance Corporation launch $1bn managed debt mobilisation programme (27 October 2026)
  21. Daba Finance — AFC and Munich Re launch a $1 billion risk-sharing program for African trade (27 October 2026)
Share

Sourcing and verification rules: methodology · Report an error: contact

Related reports