The Impact of Extreme Weather Events on Agricultural Insurance in Europe

Abstract

In Europe, climate change has a big impact on agriculture, due to an increase in the frequency and severity of extreme weather events. Many and prolonged droughts, heatwaves, floods, and hailstorms cause major economic losses that affect crop quality and generate instability in supply chains. In this study, we analyse the evolution of extreme weather events across Europe starting from the 1980s. The economic losses caused by extreme events were divided into three categories: heatwaves, frost, and fires; floods; and storms. In order to identify the trend and any shifts of the trend of the extreme weather events, we calculated moving averages over different periods: 5, 10, 20, and 30 years. The moving average analysis shows how climate change has altered from causing isolated and temporary economic losses to generate a consistent upward trend in losses, with an increasingly significant impact in the short, medium, and long term. In the second part of this study, we conducted a correlation analysis between the economic losses caused by extreme weather events and variations in property insurance premiums (fire and other property damage—which includes crop insurance premiums) and we calculated correlation coefficients directly, with a one-year lag, and with a two-year lag. Thus, we analysed whether insurance markets respond immediately to incurred losses or whether, depending on climate trends, there are delays in premium adjustments.

1. Introduction

Agriculture is very exposed to weather variations and is therefore also vulnerable to extreme climate events. The main weather events are temperature and precipitation. These events can damage agricultural production by causing physical harm to both crops and livestock. The impact of past extreme events is usually measured by production losses, which are translated into economic losses using datasets such as crop yields and/or disaster monitoring data. It is important to note that, in most studies forecasting future yields, the effects of extreme weather conditions are often oversimplified [1]. For example, extreme high temperatures and/or precipitation events occurring during the anthesis phase of crops can significantly reduce grain/fruit formation and hinder flowering. The accuracy of impact projections also depends on the quality of projected climate data and how well future extremes are captured. This means that many studies may, in fact, underestimate the future impacts of climate change on crop productivity, particularly in Europe, where the impact of extreme events is expected to be quite substantial [2].

Research tools are based on crop-climate statistics or agroecological zone indicators derived from climate and soil information, combined with simple estimates of soil water budgets. These approaches are useful for assessing large-scale trends in crop-climate relationships but cannot capture nonlinear responses and tipping points. Furthermore, the representation of agricultural management practices is limited in these approaches [3,4]. Process-based approaches represent detailed biophysical processes and are better suited to capture the effect of various agricultural management practices. These require a substantial amount of data for parameterization and calibration. Advances in agricultural datasets have enabled these approaches to produce large-scale projections of climate change impacts on crop productivity. However, some factors remain overlooked and require further research and development, such as the effect of elevated CO₂ on crop quality and the interaction between extreme weather conditions and elevated CO₂ on crop development.

To estimate both the economic damages caused by climate change and the costs of responding to it researchers use integrated models that bring together economic data and the effects of climate change on agriculture. However, these approaches can oversimplify agricultural processes, and uncertainties propagate through the climate–crop–economic modelling modules [5]. Interestingly, some recent developments in impact assessment are capable of considering multiple components of food systems to include mixed farming systems representing interactions between livestock, crops, and trees [6].

Researchers, through statistical methods and modelling, have explored how and to what extent each type of extreme climate affects agricultural production. Compared to wet conditions, hot and dry conditions are found to be the most damaging to crops [7,8]. Some researchers argue that heat stress alone does not affect agricultural productivity more than average climate changes [9]. Base on the literature drought contributes the most to the combined impact of drought and heat [10]. Moreover, the study reflects that storms do not have a significant impact—unlike floods, drought, frost, and heatwaves. For example, refs. [11,12] forecast that, in The Netherlands, the strongest impacts will come from extreme temperatures rather than drought, which aligns with global findings [13].

Table 1. Mechanisms by which extreme climate events affect agriculture.

Extreme Event	How It Affects Agriculture	Cumulative Impact
Hail	Physical damage to crops, especially harmful to perennial plants (orchards and vineyards).	Low probability
Flooding/heavy rainfall	Damages cultivated soils and roots, may delay planting, and can indirectly reduce crop productivity. Extreme wet events have reduced the yield of tuber crops (potatoes, sugar beet, onions) in The Netherlands in 1998 and 2018.	“Excessive soil wetness is less economically relevant” [14]
Frost	- Perennial plants (e.g., vineyards, fruit trees such as almonds, pistachios, cherries, apricots, apples) are sensitive to both reduced winter chilling (vernalization requirements for exiting dormancy may not be met) and late spring frosts, which can damage buds and reduce yields [15]. - Most annual crops (soybeans, corn, wheat, potatoes) benefit from fewer frost episodes	Significant economic losses due to high value per hectare for perennial plants
Heatwaves	- Heat stress affects crop growth. If extreme temperatures occur during flowering, it can lead to sterilization (no grain formation) and yield losses. - For livestock, heatwaves combined with humidity affect reproduction, milk production, and can lead to excess mortality. Heat stress during flowering is a minor factor currently but becomes major with climate change [16].	High probability
Droughts (agricultural, meteorological, hydrological)	- Agricultural drought, defined by soil moisture during the crop growing season, harms plant growth and can lead to yield reductions and crop loss. - Hydrological drought reduces water resources for irrigation, affecting the adaptability of irrigation systems. - Drought stress around flowering is already a key factor and will intensify with climate change	High probability
Cold spells	Can indirectly disrupt agriculture by affecting food transportation.	Low probability
Combined events
Drought combined with extreme temperatures	- The combination of multiple stressors can intensify impacts on crop growth and create new risks. For example, drought combined with extreme temperatures can wilt crops and create fire hazards, endangering farmland through vegetation or crop fires. - Statistical regressions show significant yield losses for winter wheat in Germany, wheat and barley in Spain.	High probability
Moisture combined with extreme temperatures	The combination of heavy rainfall and strong winds creates storm-like conditions that can destroy farmland, damage soils, and trigger landslides.	Low probability

Source: [17].

from below presents the mechanisms through which various extreme climate events influence the agricultural sector. Each type of phenomenon affects crops and agricultural systems differently, resulting in variable losses depending on the intensity, duration, and affected region. This classification helps to better understand how climate change puts pressure on European agriculture and highlights the need for implementing adaptation measures and risk management strategies.

Based on the literature review, we identified that extreme climate events have a significant impact on European agriculture. Droughts and heatwaves are the most dangerous and have a high probability of occurrence and severe consequences for crop yields, water resources, and livestock production. In contrast, events such as hail, frosts, and cold spells have a lower probability but can still cause major economic damage in certain sectors. The most destructive effects occur when extreme events combine (example: prolonged drought accompanied by extreme temperatures or excessive humidity combined with strong winds). These conditions promote the formation of storms and landslides which worsen agricultural losses and reduce the farmers’ adaptive capacity. The analysis of impact mechanisms highlights the need for effective adaptation and climate risk management strategies.

Over the last two decades, the scientific literature has increasingly addressed the economic impacts of climate change. The references were mainly to the rising frequency and severity of extreme weather events. Many studies have documented the growing vulnerability of both developed and emerging countries to climate risks [18,19]. Most existing analyses focus on short-term assessments, typically associated with isolated events and their immediate economic consequences [20,21]. However, more recent contributions emphasized the need for time series-based approaches that allow researchers to identify the underlying structural trends in economic losses [22,23]. These studies argue for moving beyond year-to-year volatility to uncover persistent trajectories in loss patterns that may signal deeper climatic transformations. Nonetheless, these works primarily concentrated on absolute losses or normalized data (e.g., by GDP or population) and have rarely adopted territorial (e.g., losses per km²) or per capita indicators to capture the uneven spatial and social distribution of climate risks across Europe. In this study, we introduced information about all these indicators and compared the result addressing this important methodological gap in this way.

Another underexplored area in the literature concerns the use of moving average methodologies to assess climate-related economic losses. Although well established in economics and technical analysis, moving averages remain rarely applied in climate impact research. They are particularly effective for identifying inflection points, smoothing out short-term fluctuations, and confirming the presence of long-term trends [24,25]. Moreover, intersections between moving averages (e.g., between 5 year and 10 year windows) can serve as analytical signals of structural change in loss dynamics. This technique is known as a crossover analysis in time series research which has yet to be systematically applied in the climate economics domain.

Additionally, the correlation between economic losses and the insurance sector’s financial response remains insufficiently addressed. While insurance is widely recognized as a key tool for managing climate risks [26,27] few studies have empirically examined whether changes in insurance premiums correlate with observed climate-related losses—either in the same year, or with a 1-year or 2-year lag. This study fills that gap by analysing these correlations, also testing the statistical significance of the relationships. This helps assess whether the insurance market reacts proactively or reactively, and how it contributes to building long-term financial resilience.

Given these gaps in existing research, this study aims to improve our understanding by combining an analysis of long-term trends, a regional perspective on how losses vary across areas, and an assessment of how quickly and effectively the insurance sector reacts over time. The goal is to offer a clearer picture of how climate change is affecting the economy in Europe and to help create better policies that reduce climate risks and improve financial protection.

2. Material and Methods

The main objective of this paper is to analyse the relationship between economic losses caused by extreme weather events and the evolution of the insurance premiums in the “Fire and Other Property Damage” (which includes crop insurance premiums) category in Europe. Using this approach, we aimed to determine how the insurance market reacts to climate change. It is important to analyse the trends in economic losses caused by extreme meteorological and climate events in Europe over a long period of time (1980–2023). Another important aspect of this study is to examine the geographical distribution of economic losses and the level of insurance coverage in different European countries. By comparing total economic losses and the percentage of insured damages, the analysis highlights discrepancies between developed countries and emerging ones, where financial protection against climate risks is significantly lower.

To analyse the impact of extreme climate events on economic losses in Europe and the insurance market’s response to these events, we used a methodology based on the statistical analysis of historical data, the correlation of economic trends, and the modelling of the relationship between losses and financial market adjustments. This approach combines quantitative and comparative analysis techniques to highlight trends and the dynamics of the insurance sector in the context of climate change.

The data analysis was structured into several stages:

Classification of climate-related events (stage 1)—in order to analyse the total annual economic losses in Europe during the period 1980–2023, we grouped extreme climate events into three categories that contributed to these losses: heatwaves, frost, and wildfires; floods; and storms.

Examination of economic loss trends (stage 2)—the evolution of annual total losses was analysed through a time series based on yearly values by type of extreme event. To identify the general direction of trends, we used several moving averages of the respective total losses;

Evaluation of regional and sectoral impact (stage 3)—we analysed the differences between EU member states regarding economic losses and insurance coverage levels, identifying disparities between regions with developed markets and those with limited financial protection;

Correlation analysis between economic losses and insurance premiums (stage 4)—we calculated correlation coefficients to assess the relationship between losses caused by extreme events and variations in insurance premiums for fire and other property damage, both in the same year and with one- and two-year lags. We also calculated the p-value to identify the statistical significance of each result.

Stage 1—The total annual economic losses were calculated by summing the losses from the three categories as follows:

$$L_t={\sum }_{i=1}^n{L}_{i,t}$$

(1)

where the following definitions apply:

i—extreme event category (heatwaves, frost, wildfires; or floods; or storms);

t—year under consideration;

n—number of extreme event categories;

L_t—total economic loss in year t;

L_i,_t—economic loss caused by extreme event i in year t.

Stage 2—In order to identify the trend of losses caused by extreme weather events, we used moving averages of economic losses over different periods (5, 10, 20, and 30 years). The intersection of two moving averages is a concept specific to technical analysis and can indicate either a trend reversal or confirmation of an existing trend. The formula used is:

$$M_t={\displaystyle \frac{1}{N}}\times {\sum }_{j=t-(N-1)}^t{L}_j$$

(2)

where the following definitions apply:

t—the year for which the moving average is calculated;

j—each year within the interval [t − (N − 1), t];

N—the length of the moving window (5, 10, 20, or 30 years);

$L_j$—the annual economic loss in year j;

$M_t$—the moving average of economic losses in year t;

The trend change analysis was performed by comparing the 5-year moving average of losses with the 10-year moving average. In addition, we compared the 10-year with the 20-year moving average. The intersection between the 5-year and 10-year moving averages indicates the start of a new trend while the intersection between the 10-year and 20-year moving averages serves as a confirmation of that trend. The formula used is derived from the difference between moving averages:

$$D_t^{\prime }={\displaystyle \frac{d}{dt}}\times \left(M_{5\left(t\right)}-M_{10\left(t\right)}\right) \mathrm{and} D_t^{\prime }={\displaystyle \frac{d}{dt}}\times \left(M_{10\left(t\right)}-M_{20\left(t\right)}\right)$$

(3)

where the following definitions apply:

D_t′—the derivative of the difference between moving averages;

M₅(t), M₁₀(t), M₂₀(t)—the 5-, 10-, and 20-year moving averages at time t.

Interpretation:

• If D_t′ is positive before the intersection and negative after the intersection than we have a downward trend reversal. For example, when the 5-year moving average intersects the 10-year moving average from above, it signals the beginning of a downward trend.
• If D_t′ is negative before the intersection and positive after the intersection, then we have an upward trend reversal. For example, when the 5-year moving average intersects the 10-year moving average from below it signals the start of an upward trend.

Stage 3—The comparative analysis between regions was used to examine the differences between European countries in terms of exposure to climate risks and the level of insurance coverage. Countries were classified according to the total value of economic losses and the percentage of insured damages. Thus, groups of countries with distinct patterns of exposure and financial protection were identified, highlighting the discrepancies between developed and emerging countries. Developed countries benefit from better insurance coverage, while emerging countries remain vulnerable due to low levels of insurance. Based on the data, we created a ranking of countries with the highest cumulative losses during the period of 1980–2023, as well as a ranking of countries in terms of the level of insurance coverage for damages.

Stage 4—To evaluate the relationship between economic losses and insurance premiums, the correlation coefficient was calculated between the annual variation in c and the variation in premiums for fire and other property damage insurance. The countries that we took into account are Austria, Belgium, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, The Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Iceland, Liechtenstein, Norway, Switzerland, and Turkey. By comparing the direct correlation, the one-year lag, and the two-year lag, we were able to observe whether adjustments in the insurance market are immediate or delayed. We thus found a weak correlation in the same year, but a stronger one with a one-year lag, especially in the case of heatwaves, droughts, and wildfires. This indicates a delayed reaction of the insurance market. The formulas used included the annual variation of insurance premiums in the “Fire and Other Property Damage” category, the annual variation of total extreme events, as well as for the three categories of extreme events.

$$V_t=\left({\displaystyle \frac{C_t-C_{t-1}}{C_{t-1}}}\right)\times 100$$

(4)

where the following definitions apply:

t—calendar year under consideration;

V_t—annual variation of the categories in year t;

C_t—value of the category (insurance premiums, extreme events) in year t;

C_t−1—value of the category in the previous year t − 1.

Based on the results of annual variations, we performed a correlation analysis using the Pearson correlation coefficient, which measures the degree of linear association between two variables—in our case, between the variations in insurance premiums and the economic losses caused by extreme events for the same year:

$$\rho \left(X,Y\right)={\displaystyle \frac{\sum (X_i-\overline{X}\left)\right(Y_i-\overline{Y})}{\sqrt{\sum {(X_i-\overline{X})}^2\sum {(Y_i-\overline{Y})}^2}}}$$

(5)

where the following definitions apply:

$\rho$—Pearson correlation coefficient;

X_i—variation in fire and other property damage insurance premiums in year i;

Y_i—variation in economic losses caused by extreme events in year i;

$\overline{X}$– arithmetic mean of the insurance premium variations;

$\overline{Y}$– mean of the variations in economic losses.

$$\rho \left(X_i,Y_{i-1}\right)={\displaystyle \frac{\sum (X_i-\overline{X}\left)\right(Y_{i-1}-\overline{Y})}{\sqrt{\sum {(X_i-\overline{X})}^2\sum {(Y_{i-1}-\overline{Y})}^2}}} \mathrm{and} \rho \left(X_i,Y_{i-2}\right)={\displaystyle \frac{\sum (X_i-\overline{X}\left)\right(Y_{i-2}-\overline{Y})}{\sqrt{\sum {(X_i-\overline{X})}^2\sum {(Y_{i-2}-\overline{Y})}^2}}}$$

(6)

where the following definitions apply:

$\rho$(X_i, Y_i−1)—correlation coefficient with a 1-year lag;

Y_i−1—economic losses caused by extreme events one year prior to the variation in premiums;

$\rho$(X_i, Y_i−2)—correlation coefficient with a 2-year lag;

Y_i−2—economic losses caused by extreme events two years prior to the variation in premiums.

To evaluate the statistical significance of the correlation coefficients between climate-related economic losses and property insurance premiums, we calculated the p-value associated with each Pearson correlation coefficient. This was done to test the null hypothesis that there is no linear association between the two variables. The p-value was derived using the t-distribution, based on the following transformation of the correlation coefficient r:

$$t={\displaystyle \frac{\mathsf{\rho }\sqrt{n-2}}{\sqrt{1-{\mathsf{\rho }}^2}}}$$

(7)

where the following definitions apply:

t—follows Student’s t-distribution with n − 2 degrees of freedom;

n—number of paired observations (years);

$\rho$—Pearson correlation coefficient.

Once the t-statistic was computed, the two-tailed p-value was obtained by referencing the cumulative distribution function (CDF) of the t-distribution. The two-tailed test was selected to detect any significant linear relationship, regardless of direction (positive or negative).

$$p=2(1-T_{cdf}\left(\left|t\right|\right)) \mathrm{or} \mathrm{Excel} \mathrm{function}: \mathrm{T.DIST.2T(ABS(t)}, n - 1$$

(8)

where the following definitions apply:

p—p-value;

$T_{cdf}\left(\left|t\right|\right)$—cumulative distribution function (CDF) of the t-distribution;

n—number of paired observations (years).

The threshold for statistical significance was set at α = 0.05. Correlations with p-values below this threshold were considered statistically significant at the 5% level. Values slightly above this level (e.g., 0.05 < p < 0.1) were interpreted as marginally significant.

It is observed that the applied analytical model combines quantitative and comparative methods to provide a detailed picture of how climate change affects the European economy and the insurance market. The results reflect a significant increase in economic losses, a delayed reaction of the insurance market, and a high level of vulnerability in certain countries. This analysis is essential for the development of effective adaptation policies and for expanding financial protection against growing climate risks.

3. Results

3.1. Extreme Weather Losses Trend in Europe

For this study, we used data from relevant official industry sources, including historical data on economic losses caused by extreme events (1980–2023), published by the European Environment Agency [28] 2023, and the analysis of the European insurance market based on reports from the European Insurance and Occupational Pensions Authority [29]. The dataset used was analysed from two perspectives: a temporal one, by examining a long-term period (1980–2023) to gain a clearer view of extreme weather trends, and a geographical one, by analysing data at the country level across Europe.

Between 1980 and 2023, in Europe, the extreme meteorological and climate events caused estimated economic losses of USD 738 billion, out of which over USD 162 billion (22%) were recorded between 2021 and 2023. Analysing trends in economic losses is challenging, mainly because of the high annual variability. Statistical analyses have shown that economic losses have increased over time, and in the last three years, we have the highest annual economic losses. The main risks that influenced the losses were extreme temperatures, heavy precipitation, and drought.

The EU strategy aims to strengthen resilience and ensure that it is well prepared to manage these risks. All these strategies are focused on better mitigating the climate change impact. The ultimate goal of the EU is to reduce the total financial losses caused by meteorological and climate events [30,31,32].

As shown in Figure 1 below, we can very clearly see the financial impact of extreme weather events between 1980 and 2023. The economic losses are presented in billions of EUR at 2023 prices in order to eliminate the inflation effect and divided in three categories: storms; floods; heatwaves, frost, droughts, and wildfires. In addition, we included a 30-year moving average, which shows a long-term upward trend in economic losses, as climate change intensifies the frequency and severity of these events.

In connection to stage 1, we can see in the analysis from Figure 1, that in the 1980s, the annual economic losses were generally below EUR 20 billion, except for years like 1990s, when they exceeded EUR 30 billion. Most of these losses were caused by storms and floods. After the year 2000, there is a significant increase in losses with a major peak in 2002, reaching approximately EUR 45 billion, due to devastating floods. During the recent period of 2021–2023, the losses exceeded EUR 55 billion annually, showing the increasing frequency of floods and heatwaves.

In the years 1990 and 1999, storms had the highest impact in the total losses, reaching over EUR 15 and EUR 25 billion during years with extreme events. Although still present, the weight of storm-related losses has decreased in recent decades compared to other categories.

In recent years, the weight of heatwaves, frost, droughts, and wildfires has increased contributing to a substantial rise in economic losses. Thus, during the period 2022–2023, heatwaves and drought were responsible for major losses, reaching an estimated average of over EUR 20 billion per year. By analysing the 30-year moving average we can see an increase from below EUR 10 billion annually in the 1980s to over EUR 40 billion annually in the 2020s. These results reflect the intensification and increasing frequency of extreme weather events.

We can conclude that there is an important increase in economic losses caused by extreme weather in Europe. The level reached in recent years is over 60 billion annually. Floods and heatwaves have become the main causes of these losses. The upward trend in the moving average line shows the need for proactive climate risk adaptation and mitigation policies.

In stage 2, we focus on identifying the trend of climate related losses using moving averages. Moving averages and their intersections (crossovers) are widely used in trend analysis, especially in financial time series [33]. Their application in climatology can support the identification of turning points and long-term shifts in the dynamics of extreme weather losses [34].

To illustrate the short-, medium-, and long-term trends of economic losses caused by extreme weather events in Europe, we created Figure 2 below, presenting 5, 10, and 20-year moving averages of extreme events in the EU during the period 1985–2023. All values are in billions of EUR adjusted at 2023 prices. By using moving averages, this graph highlights annual fluctuations and the accelerated growth trends in economic losses, emphasizing the increasing impact of climate change on the European economy. This analysis offers a clear perspective on how extreme events have evolved and become a structural challenge for the EU.

As a result of the moving average analysis, we observed three distinct periods:

• The period 1985–2000 is considered a stable phase. Economic losses caused by extreme events were constant with the 5-year moving average. The moving average fluctuated between EUR 5 and EUR 10 billion. In this period, the extreme weather events were less severe, and the impact of climate change was not so important.
• The period 2000–2010 is characterized by moderate increases. We observe the first signs of intensifying climate phenomena. The 5-year moving average begins to rise after 2000 and reached an average of EUR 17.5 billion. The main reason of this result was due to extreme events (as example the floods from 2002 in Central Europe had a big impact in the economic losses).
• The more recent period 2010–2023 shows accelerated growth. All moving averages indicate a sharp increase during this period. The section with the heatwaves, droughts, wildfires, and floods have become more frequent and more severe leading to big economic losses. The 5-year moving average has tripled between 2015 and 2023, from EUR 12 billion to nearly EUR 40 billion. The 10- and 20-year moving averages confirm this medium- and long-term trend.

The intersection in 2013 represents a turning point marking the beginning of an upward trend in extreme events, which suggests a period dominated by increasingly higher economic losses. It signals the growing impact of climate change on European economies and highlights the need for faster and more effective adaptation measures to manage these risks.

The intersection from 2019 between the 10-year and 20-year moving averages represents a second inflection point in the trend of economic losses caused by extreme events in Europe. This intersection marks a confirmation and consolidation of the long-term increase in economic losses initiated by the accelerated trends observed after 2013. Unlike the 2013 intersection which reflected a rapid increase in recent losses, the 2019 intersection indicates that this growth has become a consistent trend, now influencing long-term dynamics as well.

Based on the above data, we can conclude that the period 1980–2000 is considered a stable period with only some significant spikes, such as 1983 and 1990, which were followed by immediate declines, indicating reactive losses driven by specific events rather than a persistent trend. Even between 2000 and 2010, the pattern remained inconsistent, with high-loss years such as 2002 quickly offset by years with considerably lower losses. This pattern illustrates a phase of occasional and irregular variability. In contrast, the quantitative progression of the 5-year moving average: from EUR 13 billion in 2013 to EUR 21 billion in 2019 (+60%), and EUR 40 billion in 2023 (+94% from 2019), confirming a sustained and accelerating upward trend. This sustained growth supports the conclusion that climate-related losses are now part of a structural and ongoing trend, rather than isolated disruptions. The 2013 intersection served as a warning signal about the intensification of recent extreme climate events, while the 2019 intersection confirms that these phenomena have become a lasting trend that systematically affects the European economy.

3.2. Extreme Weather Losses and Insured Events in Each Country from Europe

In stage 3 we consider that it is very important to also identify the most affected countries in Europe. For this reason, in Figure 3 from below, we analysed losses by country for the entire period (1980–2023). This figure presents the total economic losses (in millions of euros at 2023 prices) caused by extreme weather events for each country in Europe, alongside the percentage of damages covered by insurance. In this way, we managed to highlight the significant differences between countries in terms of the value of losses and the level of insurance coverage. This approach provided a clear perspective on the different levels of exposure to climate risks across European states and their capacity to manage these losses.

From the analysis in Figure 3, we can observe that the main countries affected by economic losses, particularly Germany, which has the highest total economic losses (approximately EUR 180,000 million). The losses are mainly caused by major floods from 2002 and 2021. Italy, France, and Spain follow and have significant losses between EUR 80,000 and EUR 135,000 million. These countries are frequently affected by the following risks: heatwaves, floods, and droughts. Countries with medium-level losses such as Poland, Switzerland, and Romania show values around EUR 20,000 million, reflecting moderate exposure to extreme events. The countries with the lowest losses are Malta, Iceland, and Liechtenstein. These have insignificant total losses mainly due to their small territorial sizes and lower exposure to climate risks.

From the perspective of the percentage of insured damages, we observe that the countries with the highest insurance coverage are Norway (70%), Denmark (62%), and Luxembourg (50%). In these countries, the insurance market is very developed and there is a high level of risk awareness. Countries with moderate insurance levels include Belgium and The Netherlands (39%), France (35%), Switzerland (37%), and Germany (30%). Countries with low insurance coverage are Romania (1%), Slovenia (2%), Croatia (2%), and Bulgaria (2%). In these countries, there is a low awareness of the need for insurance as well as limited access to insurance products.

We can conclude, based on Figure 3, that developed countries such as Germany, Italy, France, and Spain have the highest economic losses caused by extreme events but have moderate insurance coverage rates. Northern countries such as Denmark and Norway have a high percentage of insured damages. Emerging countries such as Romania and Bulgaria experience lower losses but have an extremely low level of insurance protection, making them vulnerable to future risks. This analysis highlights the need for adaptation policies and for expanding the insurance market, especially in regions with limited coverage.

Next, Figure 4 provides a comparative representation of the economic losses caused by extreme climate events, expressed per unit of surface area (EUR/km²) across European countries, and indicates the share of losses that were covered through insurance mechanisms. This metric approach allows for a more equitable comparison between countries.

The analysis offers a territorial perspective that is essential for understanding the intensity and concentration of climate risks across Europe. Figure 4 shows that countries with smaller territories but high economic and agricultural density—such as Slovenia, Belgium, Germany, and Luxembourg record the highest losses per square kilometre indicating significant geographical exposure to climate-related events. The discrepancies between these losses and the level of insurance coverage reveal differing degrees of adaptation to climate risks. For instance, in Slovenia, losses exceed EUR 850,000/km², while insurance coverage stands at only 2%, highlighting a structural vulnerability. In contrast, Austria and The Netherlands manage to offset moderate levels of losses (approximately EUR 175,000–300,000/km²) with substantial insurance coverage (19% and 39%, respectively).

Countries in Central and Eastern Europe—such as Romania, Bulgaria, Slovakia, and Poland—show relatively low losses per square kilometre (approximately EUR 40,000–82,000/km²), but also very low levels of insurance coverage (1–7%). This reflects not only a potentially lower exposure but also a lack of development in the financial infrastructure needed for effective risk management.

This approach highlights that the territorial dimension of exposure is a critical factor in designing effective climate adaptation policies, particularly in regions with high agricultural density or concentrated economic infrastructure. Figure 4 confirms that, in the absence of a direct correlation between the intensity of losses per area and the level of insurance coverage, there is an increased risk of systemic vulnerability in regions that are heavily affected yet insufficiently protected. This finding supports the need for territorially tailored policies that promote the development of insurance markets aligned with the actual exposure to climate risks. Next, Figure 5 provides a comparative representation of the economic losses caused by extreme climate events reported per capita (EUR/capita) for each EU member state. This standardized approach allows for a more equitable assessment of citizens’ exposure to climate risks and contributes to understanding the social dimension of economic vulnerability. The figure also presents the percentage of losses covered by insurance highlighting the level of development and the effectiveness of national risk transfer markets.

The results show significant variation between countries. Slovenia has the highest per capita losses exceeding EUR 8693 per person, while insurance coverage remains extremely low (2%). This discrepancy indicates a high level of vulnerability among the population driven by the absence of effective financial protection instruments. Similar situations are observed in Italy (EUR 2311/person, 4% coverage) and Spain (EUR 2258/person, 5% coverage) where the losses are considerable but the insurance systems remain underdeveloped relative to the level of exposure. In contrast, countries such as Luxembourg (EUR 2694/person, 50% coverage) and Switzerland (EUR 2685/person, 37% coverage) demonstrate a high degree of insurance market maturity and strong capacity to mitigate climate risks. These countries combine significant economic exposure with efficient compensation systems suggesting the existence of well-established public policies and support mechanisms. In Central and Eastern European countries (e.g., Romania, Poland, Bulgaria, Slovakia), per capita losses are lower (below EUR 1000/person), but insurance coverage is also very limited (1–7%). This reflects not only the potential under reporting of losses but also a lack of institutional infrastructure for effective risk management.

The analysis of Figure 3, Figure 4 and Figure 5 highlights the complexity and heterogeneity of the economic impact of climate change across European countries. Each of these perspectives offers a distinct yet complementary angle on countries’ exposure and adaptive capacity in the face of extreme weather events.

Figure 3 was essential for understanding the overall economic scale of the losses and identifying the countries with the highest absolute values while also revealing disparities in the development of agricultural insurance markets. Such raw figures may lead to misleading conclusions in the absence of adjustments for structural differences between states. Accordingly, Figure 4 introduced a territorial intensity indicator relating losses to the national surface area which enabled the identification of regions with a higher spatial concentration of damages. This approach is particularly relevant for spatial planning and risk management at the regional level. Complementing the previous perspectives, Figure 5 added a socio-economic dimension by evaluating the losses’ impact on the average citizen through the per capita indicator. This allowed for a more accurate assessment of individual vulnerability and of the effectiveness of financial protection provided by national insurance systems.

Taking into account all three perspectives, we have an integrated approach to better understand climate risks. The recent evolution of climate risks are generating a severe economic impact on European agriculture. This multi-perspective analysis reveals an uneven distribution of vulnerability among member states shaped by geographic demographic and institutional factors. Western and Northern European countries generally benefit from more developed insurance systems, which mitigate financial losses and contribute to the stability of the agricultural sector. In contrast, Central and Eastern European countries exhibit significant levels of under insurance and untreated exposure, rendering them more vulnerable to future climate crises.

We can conclude that there is a need for an integrated and differentiated approach that accounts for the regional specificity of climate risks as well as the economic and institutional capacity of each country to respond. Strengthening agricultural insurance markets alongside proactive public policies and territory-adapted preventive measures will be essential in building a more resilient, equitable, and sustainable European agricultural system in the face of ongoing climate change.

3.3. Correlation Between the Evolution of Extreme Events and the Development of Insurance at the European Level

As already concluded, climate change has a significant impact on the increase in the frequency and the severity of extreme weather events. In this context, it is essential to understand the extent to which insurance markets are able to respond to these risks by adjusting insurance premiums. In order to see how the insurance industry responds to extreme weather events, we analysed the correlation between economic losses caused by extreme weather events and the variation of property insurance premiums (fire and other property damage).

In stage 4, the main objective is to determine whether there is a significant relationship between the economic losses caused by extreme events and the evolution of premiums. Also, we want to identify the types of events that have the greatest impact on agricultural sector. This is important because climate change affects different regions and industries in distinct ways and financial markets must adapt in order to maintain economic stability and provide effective protection solutions against these risks. By comparing the correlation coefficients calculated directly, with a one-year lag, and with a two-year lag, we examine whether insurance markets react immediately to incurred losses or whether there are delays in premium adjustments, depending on climate trends. A weak correlation could indicate that the insurance market is not sufficiently adapted to climate risks or that the level of asset insurance is inadequate—especially in countries with low insurance coverage for extreme weather events.

Through this analysis, we aim to answer the question of whether the segment including heatwaves, cold spells, droughts, and wildfires has the greatest impact on agriculture and whether the insurance market successfully integrates this risk into premium adjustments. If the correlation between economic losses from this segment and premium variation is significant, we can conclude that the market recognizes these risks and responds accordingly. On the other hand, a weak correlation may indicate the need for adjustments in insurance policies and risk management strategies in order to cope with the increasing impact of climate change.

This analysis is essential for understanding the gaps and effectiveness of adaptation mechanisms within the insurance sector, as well as for proposing solutions to improve economic protection against growing climate risks.

Next, we developed

Table 2. Annual percentage variations in property insurance premiums and economic losses by extreme weather events in Europe (2005–2023).

Year	Variation in Property Insurance Premiums	Total Variation in Losses from Extreme Weather Events	Variation in Losses from Heatwaves, Cold Spells, Droughts, and Wildfires	Variation in Flood-Related Losses	Variation in Storm-Related Losses
2005	4%	337%	312%	1281%	127%
2006	1%	−80%	−66%	−88%	−84%
2007	0%	164%	−28%	−23%	876%
2008	−2%	−48%	−84%	131%	−56%
2009	4%	36%	48%	−10%	57%
2010	8%	189%	641%	737%	16%
2011	2%	−70%	−71%	−64%	−79%
2012	4%	−37%	183%	−76%	−44%
2013	1%	426%	−100%	1055%	1104%
2014	0%	−41%	1500%	−52%	−27%
2015	5%	−10%	2523%	−19%	−50%
2016	9%	−16%	−79%	52%	−50%
2017	2%	205%	2337%	−67%	466%
2018	4%	−24%	−58%	−26%	37%
2019	3%	10%	16%	374%	−42%
2020	3%	−42%	−80%	−83%	46%
2021	−9%	310%	332%	2980%	−42%
2022	9%	−11%	457%	−92%	24%
2023	8%	−22%	−80%	585%	12%

, in which we present the annual percentage variations of property insurance premiums and economic losses caused by extreme weather events in Europe, categorized into heatwaves, cold spells, droughts, wildfires, floods, and storms, for the period 2005–2023.

This allows us to highlight the volatility of economic losses and the dynamics of the insurance market in relation to the increasing intensity and frequency of extreme climate events, offering a perspective on the impact of climate change on the insurance sector and the economy.

From the analysis of Table 2 above, we observe that the value of insurance premiums has increased steadily, with moderate annual variations—except for 2021, when it decreased by 9%. The highest increases in premiums were recorded in 2016 (9%), 2022 (9%), and 2023 (8%), reflecting a possible heightened awareness of the risks associated with extreme climate events. Regarding the variation in total damages caused by extreme events, we note significant fluctuations with very high values in the years marked by severe events and lower values when frequency and intensity were reduced:

• 2010 (+189%)—a year with multiple major events, including floods and heatwaves.
• 2013 (+426%)—the overall increase was driven by massive floods and extreme storms.
• 2021 (+310%)—losses increased dramatically, especially due to catastrophic flooding in Europe (e.g., Germany).
• 2006 (−80%) and 2020 (−42%)—years with lower frequency and severity of events.

From the category level analysis of extreme weather events, we observe that heatwaves, cold spells, droughts, and wildfires show extremely high variation rates, with surges in 2015 (+2523%) and 2017 (+2337%) due to devastating wildfires and prolonged droughts, and significant decreases in 2016 (−79%) and 2020 (−80%), indicating years with less severe events.

Floods are responsible for the largest positive variations. For example, 2013 (+1055%) and 2021 (+2980%) are the years with major floods that caused massive economic losses. Years with substantial decreases include 2006 (−88%) and 2022 (−92%), reflecting lower intensity and frequency of flood events. Storms have shown more moderate variations compared to other categories. For instance, 2007 (+876%) and 2013 (+1104%) reflect years with particularly violent storms, while years such as 2006 (−84%) with significant negative variations indicate a temporary decline in storm-related damages.

To determine the level of correlation between extreme events and the evolution of premiums for fire and other property damage insurance, we present

Table 3. Correlation between extreme events and the evolution of fire and other property damage insurance premiums.

Correlation Coefficient	Total Extreme Events	p-Value (Total)	Heatwaves, Cold Spells, Droughts, and Wildfires	p-Value (Heatwaves…)	Floods	p-Value (Floods)	Storms	p-Value (Storms)
Correlation (no lag)	−0.2994	0.2130	−0.0109	0.9646	−0.4899	0.0332	−0.1710	0.4838
Correlation (1 year lag)	0.0039	0.9873	0.4201	0.0733	0.2182	0.3694	−0.2982	0.2149
Correlation (2 year lag)	0.1186	0.6286	0.1080	0.6598	0.1894	0.4372	0.1105	0.6522

below, which includes the Pearson correlation coefficient. This was calculated as a direct correlation in the same year, with a one-year lag and a two-year lag. Corresponding p-values (two-tailed, based on t-distribution) are provided next to each coefficient. The goal is to assess how and when the insurance market reacts to the losses generated by these events.

From the analysis of the results, we observe that, in the same year, there is no direct correlation between the variations in property insurance premiums and the losses caused by extreme weather events. The negative result in the no-lag scenario (i.e., comparing results within the same year) indicates that premium variations do not immediately respond to economic losses caused by extreme events. All results are not statistically significant, as in almost all cases, the p-value is higher than 0.5. There is only one exception in case of flood where we see an indirect correlation with a statistical importance as the p-value is 0.0332. This indicates that, in case we have lower losses, but maybe with a higher frequency, the insurance market increases the premiums in advance of an high flood loss.

When introducing a one-year lag, the correlation becomes positive and more significant for phenomena such as heatwaves, droughts, and wildfires, suggesting a delayed reaction to these types of losses. For this category, the one-year lag correlation is positive and the highest overall, indicating a stronger relationship between major losses in these segments and the adjustment of premiums one year later. From a statistical perspective, even though the p-value is showing that the correlation is not significant at 5% level, we can consider that we have a marginal significance as the p-value is very close to 0.05. This shows that the insurance market responds with premium adjustments in the year following such losses.

With a two-year lag, the correlations remain positive but are lower than with a one-year lag. This suggests that the market reacts primarily in the year following the event, with more limited adjustments occurring two years later. Anyway, from a statistical point of view, the correlations are considered very weak, as the p-value is very high (p-value > 0.4).

4. Discussion

The findings related to climate risk trends demonstrate that climate change has evolved from generating sporadic high-impact events to producing a systematic and sustained escalation of economic losses. Starting in 2019, the trend underlines the urgent need for anticipatory and adaptive strategies in climate risk governance. European institutions, policymakers, and insurance systems must prioritize long-term resilience planning and accelerate investment in mitigation and adaptation to limit the deepening socioeconomic consequences of climate-related disasters.

Based on the correlation analysis, we can say that the adjustment of premiums for fire and other property damage insurance is weakly correlated with the economic losses caused by extreme events. This weak correlation may be significantly influenced by the low insurance penetration in many European countries, the lack of coverage for certain risks, and the delays between losses and insurance market reactions. Improving insurance coverage, diversifying insurance products, and adapting the market to emerging climate risks could increase the correlation and better reflect real losses in premium variations. However, another key conclusion we reached is that the segment comprising heatwaves, cold spells, droughts, and wildfires has a significant impact on agriculture, as supported by the correlation results. The one-year lag correlation coefficient for this segment (0.4201) is the highest among all the categories analysed, suggesting that these extreme phenomena have the greatest influence on variations in fire and property insurance premiums. This stronger correlation can be explained by the fact that heatwaves, droughts, and frosts directly affect agricultural production, through crop losses, soil degradation, and reduced water resources. Wildfires can destroy farmland and rural infrastructure, having a devastating effect on the agricultural economy.

Taking into account our main of objective regarding the correlation of climate events with insurance market, we can say that agriculture is a sector that is highly sensitive to climate change, and agricultural insurance, especially when linked to crop loss, is better adapted to address these risks. For this reason, premiums may increase as a reaction to the losses incurred in the following year.

5. Conclusions

The security and stability of European agriculture are increasingly threatened by climate change, which is leading to more frequent and intense extreme weather events. Drought, heatwaves, and floods have become recurring problems putting pressure on farmers, agricultural production, and the overall economy. This study showed that these events have a direct impact on the farmer yield. Extreme events also have an impact on product quality, soil health, and water resources. For example, droughts and heatwaves can lead to smaller, deformed, or nutrient-deficient crops, and floods or heavy rain can have impacts such as soil erosion, nutrient leaches, and the contamination or depletion of water resources. An important aspect is that climate change affects these regions differently. If southern and central Europe are more exposed to severe droughts, the northern and western regions are more affected by floods and storms. These differences highlight the need for tailored strategies for each region and agricultural sector.

The insurance market plays a crucial role in managing climate risks, but the analysis shows that it responds with delays and it is not uniformly developed across European countries. While western and northern states have high levels of agricultural insurance coverage, eastern countries remain vulnerable with low percentages of compensated damages. Furthermore, the correlation between economic losses and rising insurance premiums indicates a delayed financial market reaction, which may leave farmers exposed to future events. To face these challenges, the implementation of long-term solutions is essential. Developing more climate resilient agricultural practices (modern irrigation technologies and cultivation of varieties adapted to extreme conditions) can help reduce losses.

As already mentioned, the findings of this study highlight not only the structural rise in climate-related economic losses, but also the delayed and uneven responsiveness of insurance markets. The insurance premiums only adjust with a lag of one or two years, and are often more reactive than preventive. To address this, there is a clear need for expanding and strengthening agricultural insurance coverage across the European Union, with a particular focus on accessibility, affordability, and the integration of climate risk in premium-setting models.

In conclusion, climate change is no longer a distant threat but a reality that directly impacts European agriculture. Due to the increase in the frequency and severity of extreme events, we need to have more effective measures to protect farmers and ensure food security. Adapting to new climate conditions must become a priority for the agricultural sector, governments, and insurance companies in order to reduce risks and ensure the long-term sustainability of agricultural production.

Future research should focus on developing climate risk models at the regional level within countries to better understand the impacts across European agriculture. This paper presents the impact of climate change, but it is also very important to identify which is the best approach in order to mitigate the climate risk at the farm level. Future research should also be on how to integrate digital technology in order to have better predictive models.