The Insurance Protection Gap for Natural Catastrophe Losses in Europe

<iframe src="https://www.enhanceproject.eu/embed/protection-gap/" width="640" height="560" style="border:0" title="Insured share of flood losses by country (EIOPA 2025)" loading="lazy"></iframe>

Over the 45 years from 1980 to 2024, only around a quarter of the losses from natural disasters in Europe were insured. That is the headline finding of the 2025 update of the European Insurance and Occupational Pensions Authority’s dashboard, and it describes the insurance protection gap in a single sentence: three euros in every four had to be found somewhere else, from savings, from public budgets, from EU funds, or not at all.

The rest of the story is in the detail. The gap is not spread evenly. It is concentrated in a handful of countries and perils, and it is closely tied to how each country has organised its catastrophe cover.

What the insurance protection gap measures

The protection gap is the difference between the economic cost of a disaster and the part of that cost that insurance pays. If insurers pay a fifth of what a flood destroys in homes, roads and business stock, the gap for that event is the other four-fifths.

Two things make the measure harder than it looks. The first is that “economic loss” mixes very different assets. Public infrastructure such as bridges, river banks and schools is often uninsured by design, because governments carry that risk themselves. A gap that consists mostly of damaged motorways says something different from a gap made of uninsured houses.

The second is time. A single year is noisy. A quiet year in a high-risk country can show almost no uninsured losses, then one earthquake or one river flood can wipe out decades of savings. That is why supervisors look at both a modelled view of current risk and a long historical record. EIOPA does both, and it warns that the current and historical scores are not fully comparable with each other.

How EIOPA scores the insurance protection gap, peril by peril

EIOPA’s dashboard covers 30 countries: the 27 EU Member States plus Iceland, Liechtenstein and Norway. It looks at property insurance and five perils:

• windstorm;
• wildfire;
• flood, meaning river and pluvial (surface water) flooding together;
• coastal flood;
• earthquake.

For each peril and each country, the dashboard gives a score from 0 to 4: 0 means no protection gap, 1 low, 2 medium, 3 high and 4 very high. According to EIOPA’s technical description (October 2025), the current score comes from a matrix. One axis is a modelled risk score from 0 to 4, built from hazard, exposure and vulnerability. The other axis is insurance penetration, sorted into four bands: 0–25%, 25–50%, 50–75% and 75–100%. High risk combined with low penetration produces a high score. High risk with near-universal cover produces a low one.

The five peril scores are then added together. The maximum total is therefore 20.

Reading a score without misreading it

A few rules of thumb help:

1. Look at the thresholds, not only the ranking. EIOPA treats a peril score of 3 or more as a relevant gap and 2.5 as a level that “should be monitored”. A country can rank high on the total while having no single peril above 3.
2. Break the total down. A total of 12 can come from one very high score and several medium ones. Greece’s 12 includes a 4 for earthquake and a 3 for wildfire; its flood score is 1.5.
3. Remember what penetration means. A low score does not mean a country is safe. Spain scores 1 for every peril in 2025 because cover is attached to ordinary policies, not because Spain does not flood.
4. Compare with the historical column. The chart above sets each country’s flood score next to the share of historical flood losses that was insured. Where the two disagree, the reason is usually in how cover is sold.

The chart can be embedded on other sites; the snippet sits directly under it.

The 2025 totals at a glance

Source: EIOPA dashboard, 2025 update (dashboard page last updated 5 December 2025).

Where the natural catastrophe gap is widest

Four countries stand out in the 2025 scores: Greece, Italy, Romania and Portugal. Croatia (11) and Bulgaria (10) are close behind.

Italy is the case EIOPA itself singles out. Italian earthquake risk, Italian river and pluvial flood, and German flood together account for about 43% of uninsured losses in Europe, according to the dashboard’s summary. The uninsured share of their historical losses is 98% for Italian earthquake, 97% for Italian flood and 74% for German flood. Only 3% of Italy’s historical flood losses were insured, according to the CATDAT figures in EIOPA’s data file. Italy has reacted: Law 213/2023 made catastrophe cover compulsory for businesses (earthquake, flood, landslide, inundation and overflow), with SACE reinsuring up to 50% of claims. The start date has slipped several times, from 31 December 2024 to a staggered timetable by company size through 2025, with some sectors given until 31 December 2026, according to Italy’s Ministry of Enterprises (MIMIT, updated 3 June 2026). The obligation applies to businesses.

Greece shares the top total with Italy, but its profile is different: the very high earthquake score and a wildfire gap of 3. Only 9% of historical flood losses were insured.

Romania is one of only two countries, with Croatia, where EIOPA finds a relevant flood gap; it scores 3 for flood and 3.5 for earthquake. Its historical flood insured share is 2%, and the European Environment Agency lists Romania among the countries where less than 3% of climate-related losses were insured.

Portugal is the only one of the four whose total rose in 2025, from 10 to 10.5. Its wildfire score of 3 is a relevant gap on EIOPA’s definition, and only 2% of historical flood losses were insured.

What these countries have in common is not the hazard. Austria and Czechia also flood. What they share is that natural catastrophe cover is optional, sold separately, or bought by few households.

Why Spain and France score lower

Spain (5) and France (6.5) sit at or near the bottom of the table, and neither is short of floods. The 2024 Valencia floods were the worst event in the history of Spain’s scheme. The reason is structural: both countries attach catastrophe cover to the ordinary insurance people already buy.

In Spain, the Consorcio de Compensación de Seguros covers flood, earthquake, tsunami, volcanic eruption and certain extreme storms for anyone with a home, motor, business or life policy in the prescribed lines. A compulsory surcharge, 0.07‰ of the sum insured for homes, funds it, and claims are paid automatically with no disaster declaration needed. Penetration therefore follows ordinary home insurance: 80.8% of Spanish homes are insured, according to UNESPA data reported by elDiario.es. The remaining 19.2% have no policy and therefore no cover. The weakness shows in the CCS’s own accounts: its loss ratio was 612.0% in 2024, the year of the Valencia floods, against 87.2% in 2025 (CCS Informe Anual 2025).

In France, the CatNat regime adds natural catastrophe cover to every property policy by law, for homes and businesses alike. The surcharge rose from 12% to 20% of the property premium on 1 January 2025, a change decided in December 2023 and expected to add EUR 1.2bn of capacity a year, according to the Ministry of the Economy. The state-owned reinsurer CCR stands behind insurers with an unlimited State guarantee. France’s historical flood insured share in EIOPA’s data is 54%, the highest among the countries discussed here.

The comparison carries a lesson about the scores. Spain’s historical flood insured share is 25%, well below France’s, yet Spain’s current score is lower. The current score reflects today’s penetration; the historical share reflects decades of losses, including years before cover was as widespread. EIOPA’s caution about comparing the two views applies here directly.

The UK sits outside the EIOPA dashboard

The UK is not in EIOPA’s dashboard. EIOPA covers only EU and EEA countries, so there is no UK score to compare with the figures above, and no official European source sets UK flood cover on the same 0-to-4 scale.

The UK model is different again. Flood cover is part of standard private home insurance, and Flood Re reinsures the flood element of eligible policies, funded by a levy on all home insurers. Eligibility is limited to homes built before 1 January 2009; businesses and newer homes fall outside it. The UK side, including how flood risk is mapped and checked in England, is covered in the flood risk guides.

Losses from 1980 to 2024: the EEA record for Europe

The European Environment Agency publishes a separate indicator on economic losses from weather- and climate-related extremes. Its October 2025 edition, based on RiskLayer’s CATDAT database, gives these figures for the EU-27:

• Total: EUR 822bn of losses over 1980–2024.
• Recent years: more than EUR 208bn of that, or 25%, came in just four years, 2021–2024.
• Annual average: EUR 8.6bn a year in the 1980s, rising to EUR 44.9bn a year in 2020–2024.
• Worst year: 2021, at EUR 65.2bn, the year of the July floods in Germany and Belgium.
• By hazard: floods 47%, storms 27%, heatwaves 18% and other hazards 8%.

On the insured side, the EEA says less than 20% of total losses were privately insured. The share varies by type of event: more than 35% for meteorological events such as storms, about 15% for hydrological events such as floods, and slightly over 10% for climatological events such as heatwaves and droughts. Storm damage is widely covered by ordinary home policies; flood often is not.

By country, the EEA puts the insured share below 3% in Bulgaria, Croatia, Cyprus, Lithuania, Malta, Romania and Iceland, and above 35% in Belgium, Denmark, France, Luxembourg, the Netherlands, Liechtenstein, Norway and Switzerland.

Why does the EEA say “less than 20%” while EIOPA says “around a quarter”? The two do not measure the same thing. The EEA indicator covers weather- and climate-related events only, so earthquakes are left out; EIOPA’s historical view includes earthquake and covers a different set of countries. Neither figure is wrong, but they cannot be swapped for one another in an argument.

Reinsurer figures and their scope caveats

The two largest reinsurers publish annual loss tallies that are often quoted in the press. Both are useful, and both need their small print read.

Swiss Re. The Swiss Re Institute’s sigma No 1/2026 (19 March 2026) puts Europe’s 2025 losses at USD 10bn economic and USD 6bn insured, with 2,514 victims. Those Europe figures cover natural catastrophes and man-made disasters together, so they cannot be read as a natural catastrophe figure alone. Globally, sigma gives USD 235bn of economic losses in 2025 and a natural catastrophe protection gap of USD 112bn, or 51% of economic losses.

Munich Re. Munich Re’s figures for 2025 (13 January 2026) give Europe about USD 11bn of overall losses, of which “around half” was insured. That was a light year by recent standards: Munich Re’s ten-year average for Europe is USD 35bn overall and USD 12bn insured. In 2024 the figures were USD 31bn overall and USD 14bn insured losses, a share of about 45%.

Three cautions apply. Reinsurers report in US dollars, so year-on-year comparisons partly reflect exchange rates. Their definitions of “Europe” may differ from the EU or the EEA. And early estimates move: Munich Re’s January 2025 figure for the Valencia floods (USD 11bn overall, USD 4.2bn insured) was overtaken within months: the Consorcio’s payments alone passed EUR 4bn on 25 November 2025, according to the Spanish government. The shares quoted here are therefore always calculated from a single source’s own pair of figures, never by mixing one source’s total with another’s insured losses.

Worked example: Emilia-Romagna, May 2023

The floods in Emilia-Romagna between 2 and 22 May 2023 show what a high score means in practice. PERILS, the industry loss-index provider, put the economic loss at EUR 9bn, of which EUR 495m was insured property loss (PERILS, 22 May 2024).

The calculation is short:

• insured share: EUR 495m ÷ EUR 9,000m = 5.5%;
• uninsured part: EUR 9,000m − EUR 495m = EUR 8,505m, or 94.5% of the loss.

About half of the losses hit public infrastructure, according to PERILS, so part of that uninsured sum would never have been insured anywhere. Even so, the result fits Italy’s flood score of 2.5 and its 3% historical insured share closely.

Who paid the rest? Some of it came from the EU Solidarity Fund. Italy declared EUR 8,533m of direct damage to the Commission, a figure assessed differently from the PERILS total and not interchangeable with it. The Fund granted EUR 378.8m for Emilia-Romagna and EUR 67.8m for Tuscany, according to the Commission’s proposal COM(2024) 325. Set against either damage figure, EU money covered a small fraction. The rest fell on the Italian state, regions, businesses and households. The full event is set out in the Emilia-Romagna case study.

For comparison, Munich Re’s figures for the July 2021 floods in Germany and neighbouring countries give EUR 46bn overall and EUR 11bn insured, an insured share of 23.9%. The gap was still large, but proportionally far smaller than in Emilia-Romagna two years later.

What closes an insurance protection gap, and what each fix costs

Four approaches to narrowing the insurance protection gap recur in European policy debates. None is free, and each moves the cost somewhere else.

Compulsory or bundled cover

Attaching natural catastrophe cover to fire or home insurance, as France, Spain, Norway and Switzerland do in different ways, raises penetration almost automatically. Switzerland insures more than 95% of buildings and contents against natural perils.

The trade-off is the price signal. A flat surcharge, the same for a riverside house and a hilltop flat, does not tell anyone that their home is at risk. Critics of Spain’s tariff and of Norway’s uniform premium make this point. In France, the criticisms gathered in the Sénat’s May 2024 report include a depleted CCR reserve and claims costs projected to rise 40–60% by 2050. Compulsion also raises legal questions: Germany’s 2025 coalition agreement points towards elemental cover in every new policy with an opt-out, a design shaped by constitutional and EU-law doubts about a true mandate. Germany’s take-up of elemental cover stood at 57% of homes in 2024, up from 41% in 2017 (GDV).

Public-private schemes and state backstops

A second approach lets private insurers sell the cover while a public or industry body carries the extreme layer. Flood Re, CCR’s State-guaranteed reinsurance and the German insurers’ “Elementar Re” proposal, which would add a state stop-loss only above EUR 30bn (Artemis, December 2025), all work this way. The cost is a contingent liability for the taxpayer and the risk that a backstop becomes permanent. Flood Re, which has no taxpayer guarantee, is meant to close in 2039, and its own annual report warns of a cliff edge for the households it supports if it left the market.

Where no scheme exists, the state tends to pay anyway, after the event. Austria’s Katastrophenfonds pays discretionary aid through the Länder, and only about 5% of household and business values there are insured against flood, according to the Austrian Court of Audit (October 2025). Germany’s federal reconstruction fund after the 2021 floods was set at up to EUR 30bn. Ad-hoc aid of this kind is the most common criticism of the German system: it weakens the incentive to insure.

Parametric cover

Parametric insurance pays a fixed sum when a measured value, such as water depth or wind speed, passes a threshold, without assessing the damage. It pays quickly and can cover assets that conventional policies leave out. Its weakness is basis risk: the payout may not match the loss. The BIS Financial Stability Institute and IAIS asked directly, in December 2024, whether it can help bridge natural catastrophe protection gaps, and they flag the mismatch between trigger and loss as a threat to trust, cost and effectiveness.

Risk reduction

Insurance moves money around after a loss; prevention shrinks the loss. Flood Re’s Build Back Better scheme pays up to £10,000 for flood resilience measures during a repair. Austria’s Katastrophenfonds spent more on prevention in 2025 than on any other item. The Sénat has proposed lower deductibles for French policyholders who take prevention measures. Risk reduction is slow to show in scores, though, because EIOPA’s matrix weighs modelled risk, and models change only when the physical exposure does.

What the scores leave out

The dashboard measures the gap at the moment of loss. It says little about how fast money arrives. In Germany, only EUR 6.2bn, about 20%, of the up to EUR 30bn reconstruction fund had been drawn by 30 June 2026, nearly five years after the Ahr valley floods, according to dpa. A country can have generous public money on paper and still leave people waiting, a part of the insurance protection gap that no 0-to-4 score captures.