Europe's natural catastrophe insurance schemes compared

Every system on the table above was built in response to a particular disaster or a particular political moment, and none of them was designed with the others in mind. France legislated in 1982, Norway’s pool opened in 1980, Spain’s Consorcio has roots in 1941, the Swiss private pool goes back to 1936, and Britain’s Flood Re was created by an Act of 2014. Germany and Austria never built a dedicated insurance scheme at all. Comparing them is therefore less a ranking exercise than a way of seeing which choices each country made, and what those choices cost.

The background is shared. EIOPA, the EU’s insurance supervisor, said in April 2026 that only about a quarter of natural catastrophe losses in the EU have been insured over recent decades. The six systems are six attempts to shrink the other three quarters, or at least to decide who carries them.

What follows takes nine questions in turn. After them come three worked comparisons, each following one household or one event through several systems.

1. Is cover compulsory?

In five of the six systems there is some compulsion, but it almost never falls on the homeowner.

Switzerland comes closest to a true obligation, because in the cantons with a monopoly building insurer the building has to be insured with it. Everywhere else the mechanism is the same: catastrophe cover is welded to a policy most people buy anyway. That works well where the base policy is near universal and poorly where it is not.

Germany’s planned opt-out sits between the two camps. Under the framework Justice Minister Hubig presented on 28 September 2026, new building policies would include elemental cover by default and existing customers would be offered an upgrade, but owners could still say no. The best evidence that defaults matter comes from Germany itself: Baden-Württemberg, which had compulsory monopoly building cover until 1993, still had 94% elemental take-up in the GDV’s 2025 figures, far above the national 57%.

2. Which perils are covered?

The lists overlap less than people expect.

• Flood Re: flood only.
• CatNat: floods, earthquakes, overseas cyclones and clay shrink-swell subsidence from drought. Storm, hail and snow are left to ordinary policies.
• Consorcio: earthquake, tsunami, flood, volcanic eruption, atypical cyclonic storm (gusts above 120 km/h) and meteorites, plus terrorism and riot.
• Norway: storm, landslide, flood, storm surge, earthquake and volcanic eruption.
• Switzerland: high water, flooding, storm, hail, avalanche, snow pressure, rockfall, falling stones and landslide. No earthquake.
• Germany: the voluntary extension covers heavy rain, river flooding, backwater and other elemental perils; storm and hail sit in the basic policy.
• Austria: the Katastrophenfonds can support recovery from floods, avalanches, storms and landslides, but as aid, not cover.

Two gaps stand out. Switzerland, with the most complete list of mountain hazards, excludes the one peril that could cause its largest loss: the Swiss Insurance Association estimates a major earthquake could cause CHF 30–40bn of damage, and only 21% of building values are insured against it. And France’s inclusion of drought subsidence has turned out to be its most expensive decision, as the CatNat page explains.

3. How is cover priced?

Here the split is clean. Four systems use a flat rate set centrally; one uses risk-based private prices; one caps the price for a defined group.

The flat-rate systems differ in what they measure. France’s surcharge follows the premium, so it rises with everything else that makes a policy expensive. Spain’s and Norway’s follow the sum insured, so they track the value of the property. Flood Re’s band premium is the only one that is explicitly tied to a proxy for household means.

That British design is also the only one actively being reshaped towards lower-income households. Insurance Business UK reported in July 2026 that from April 2027 the contents-only Flood Re premium for Bands A and B will fall from £52 to £25, while its headline pointed to the rising cost at the top bands. None of the flat-rate systems can make that kind of targeted change, because their rate is the same for everyone by law.

Germany is the outlier. Risk-based pricing gives a clear signal, but the GDV, as reported by Artemis in December 2025, counts more than 400,000 residential buildings facing premiums that are prohibitively expensive.

4. Who carries the tail risk?

The tail is the year nobody wants: the loss far above anything in the records. The six systems handle it very differently.

France and Spain are the only two where a state guarantee stands openly behind the scheme. The UK deliberately has none. Germany and Austria have, in effect, an implicit guarantee for the uninsured that is renegotiated after every flood.

The layers are being thickened almost everywhere. Flood Re’s 2025-26 annual report says that at a £3.2bn loss the scheme would now retain about £347m, against about £130m under the old structure, and that its first catastrophe bond was placed in 2025. The Norwegian Pool reports that its 50/50 quota share made its 2025 result about NOK 500m better than it would otherwise have been, and it is building central capital towards NOK 4bn. The Consorcio’s reserve grew in 2025 even though its extraordinary-risk loss ratio had reached 612% the year before. Each system, in its own way, is buying more distance between a bad year and its guarantor.

5. What does the state promise?

Related, but not the same. A state backstop protects the scheme; ad-hoc aid protects people the scheme does not reach.

The two German figures that best describe ad-hoc aid come from the July 2021 floods. The federal reconstruction fund Aufbauhilfe 2021 was set at up to €30bn, and by 30 June 2026 only €6.2bn of it had been drawn, according to a dpa report. In Austria, aid through the Katastrophenfonds is discretionary, with no legal claim, and the Court of Audit found it covered on average 31% of private damage in Carinthia and 63% in Tyrol between 2015 and 2023.

In the compulsory systems, the state’s promise is narrower and firmer: it stands behind the insurer, and the household’s claim is a contractual right.

6. How many people are actually covered?

Take-up is where the published figures are thinnest. Where a scheme is attached to a policy, the relevant number is how many households hold that policy, and several countries do not publish it in a usable form.

• Spain: 80.8% of homes are insured, according to UNESPA data reported by elDiario.es; in Valencia province the share rose from 78.4% to 84.5% after the 2024 DANA.
• Switzerland: more than 95% of buildings and contents are insured against natural hazards, according to the SVV.
• Germany: 57% of residential buildings had elemental cover in 2024, according to the GDV, up from 41% in 2017.
• Austria: about 5% of the values of households and businesses are insured against flood, according to the Court of Audit.
• UK: Flood Re held about 353,000 policies at 31 March 2026. A national home insurance penetration rate is not given in the sources used here.
• France and Norway: cover is automatic with property and fire policies respectively, but neither a household penetration rate for France nor a coverage rate for Norway could be confirmed from primary sources.

EIOPA’s protection gap dashboard offers a cross-check for the five countries it covers (the UK and Switzerland are outside it). Its 2025 current flood scores, on a 0–4 scale where higher means a larger gap, are 1 for Spain and Norway, 1.5 for France, and 2.5 for Germany and Austria. The historical share of flood losses insured between 1980 and 2024 is 54% for France, 50% for Norway, 26% for Germany, 25% for Spain and 15% for Austria. The protection gap page explains why the two measures can point in different directions.

7. What triggers a claim?

This is the difference a flooded household notices first.

• No declaration needed: Flood Re (the household claims from its own insurer), Spain (the household claims directly from the Consorcio, automatic since 1986), Norway (own insurer, with the Pool deciding whether the damage is a natural peril), Switzerland (insurer or cantonal insurer pays when a covered peril occurs) and Germany (ordinary claim under the policy).
• Declaration needed: France. An inter-ministerial arrêté must recognise a natural catastrophe for the commune, the period and the type of damage before CatNat applies.
• Application for aid: Austria (through the municipality and the Land) and Norway’s state scheme for uninsurable property (through Landbruksdirektoratet).

The French trigger is not a formality. The Senate’s 2024 report found that only about half of communes applying for drought recognition succeed.

Who decides whether an event counts also varies. In Norway it is the Pool, an industry body, that determines whether damage was caused by a natural peril. In Spain it is the Consorcio, applying definitions written into law, such as the 120 km/h gust threshold for an atypical cyclonic storm. In France it is ministers, by decree. The more political the trigger, the more room there is for argument after the event.

8. How fast is the money?

Few schemes publish payment speed in a comparable way, so the evidence comes from rules and from individual events.

The Spanish and Swiss figures are strong for events of that size. The German insured figure is respectable for a much larger event; the public fund is the slowest money in the table.

Speed also depends on how many institutions a claim passes through. In Spain and in the Swiss cantonal monopolies, one body collects the money and pays the claim. In the UK and Norway, the household deals only with its own insurer while the scheme settles with the insurer later, which keeps the process familiar. France adds a public decision before the private claim can be completed, and Austria routes every euro of aid through a municipality, a Land and the Federation.

9. Does anyone get rewarded for reducing risk?

This is the weakest area across the board. Flat tariffs, by design, do not reward a household for moving its boiler upstairs or a commune for refusing to build in a flood corridor.

• UK: Build Back Better funds up to £10,000 of resilience measures during a repair, and Flood Re is piloting Flood Performance Certificates, according to Insurance Business UK in July 2026.
• France: the Senate’s 2024 report proposes lower deductibles for owners who invest in prevention and judges the building rules for new homes on clay soils inadequate.
• Norway: a Norges Bank staff memo of January 2026 notes that the uniform premium weakens prevention incentives.
• Austria: the Katastrophenfonds spent €233.7m on prevention in 2025, its largest single item, and the Court of Audit recommends stricter building bans in red hazard zones.
• Germany: risk-based pricing sends the clearest signal of all, but for the most exposed homes the signal is a price nobody pays.

The honest summary is that no system has solved this. Flat-rate schemes protect the exposed and then struggle to reward the careful; risk-based markets reward the careful and abandon the exposed. The UK’s mix of a capped price, a resilience grant and a fixed end date is the most deliberate attempt to have both, which is why its 2039 transition is worth watching from the rest of Europe.

Worked comparison 1: the same flooded terraced house in Hull, Lyon and Valencia

Take one hypothetical house: a pre-war terrace with a ground-floor kitchen, insured with an ordinary buildings and contents policy, flooded to knee height by a river in the same week in three countries. All figures below are the published rules; the house is invented.

In Hull, the house was built long before 2009 and sits in a low Council Tax band, say Band A. Its insurer has ceded the flood risk to Flood Re and pays the scheme £205 for combined cover in 2026/27. The owner rings the insurer the morning after the flood. No declaration is needed. The insurer pays the claim under the policy terms and recovers it from Flood Re, subject to the £250 excess in Flood Re’s rules. If the insurer offers Build Back Better, the repair can include up to £10,000 of resilience work. The owner may never know Flood Re was involved.

In Lyon, the same house carries a CatNat surcharge equal to 20% of its property premium. The owner reports the damage to the insurer, but the CatNat guarantee opens only when an inter-ministerial arrêté recognising the flood for the commune is published in the Journal officiel. From then, the owner has 30 days to declare the claim if not already done, the insurer must pay a provision within two months and settle within three months of the estimate, and the statutory deductible is €380.

In Valencia, the house carries the Consorcio surcharge of 0.07‰ of its sum insured, which on a hypothetical sum insured of €150,000 comes to €10.50 a year. The owner claims directly from the Consorcio, without any declaration, and pays no deductible on a home. In the 2024 DANA this process handled more than 250,000 claims and had paid over €4,000m by late November 2025.

And if the owner had no policy? In all three cities the answer is the same: no scheme pays. In Spain that applies to the 19.2% of homes without insurance. The schemes protect policyholders, not properties.

Verdict for this case: Valencia is the simplest and cheapest for the household, Hull the most discreet, Lyon the most dependent on a government decision.

Worked comparison 2: a landslide in Norway, Valais and Tyrol

Now a mountain hazard. A slope gives way above a house and its private access road.

In Norway, the house is covered automatically through its fire insurance and the Pool, at 0.08‰ of the fire sum insured. The road is not insurable, so the owner applies to Landbruksdirektoratet under the state natural perils scheme and is paid only once restoration is documented, within three years.

In Valais, one of the seven Swiss cantons without a cantonal building insurer, the house is insured privately against landslide under the FINMA tariff, and 80% of the claim is shared through the Elementarschadenpool. The Blatten rockslide of May 2025 showed this route paying more than 80% of total-loss sums within four months.

In Tyrol, there is no insurance obligation, and private natural hazard cover is rare: the Court of Audit put flood cover at about 5% of values nationally. The owner applies to the municipality for aid from the Land, part-refunded by the Katastrophenfonds. On the Court of Audit’s figures for 2015–2023, Tyrol’s aid has averaged 63% of private damage, the higher of the two state figures the Court cited, but it remains discretionary.

Verdict: Switzerland is the clearest for the house, Norway the most complete once the road is counted, Austria the most uncertain.

Worked comparison 3: a new house on a floodplain

Finally, the case the schemes treat most differently. A house built in 2012 on land that floods.

• UK: Flood Re excludes homes built after 1 January 2009. The owner faces the ordinary risk-priced market.
• France: CatNat cover applies at the same 20% surcharge as for any other house; the age of the building does not matter.
• Spain: the Consorcio surcharge applies at 0.07‰ like any other home.
• Germany: elemental cover is available only at a risk-based price, if an insurer offers it, and the house may be among the 400,000 the GDV considers effectively priced out.

This is the sharpest philosophical divide in the comparison. The UK uses insurance price to discourage building in the wrong place. France and Spain spread the cost of that decision across everyone. Which of the two keeps more new building out of harm’s way is the question the UK’s 2039 deadline will eventually test.

What the comparison does not settle

Each of the six pages goes deeper: Flood Re, CatNat, the Consorcio, the Norwegian pool, the Swiss system and Germany and Austria.

One question cuts across all of them and is not answered by any: what happens when the household without a base policy is the one that floods. Germany’s coming bill is the only live attempt in this group to change the default for that household, and Switzerland’s parliament, on 30 September 2026, has just refused a federal backstop for earthquake.