Skip to main content
Article

Model tackles rising subsidence risk across France

Reinsurance
Climate Risk and Resilience|Insurer Solutions

By Hani Ali | November 15, 2021

A recently released Willis Re model will help insurers of buildings in France address the rising threat posed to claims levels from subsidence.

Subsidence is characterised by vertical land movement that can cause damage to building structures. This stems mainly from the presence of clay in the soil that swells and shrinks in wet and dry weather respectively.

Whilst most insurers are reinsured against subsidence risk in France, the retention rate remains high.”

Tina Thomson
Head of Catastrophe Analytics, EMEA West and South, Willis Re

“Whilst most insurers are reinsured against subsidence risk in France, the retention rate remains high. It is therefore important for insurers to develop their own view of risk to this growing hazard. Willis Re analytics solutions helps our clients to estimate their exposures to subsidence.“ Tina Thomson, Head of Catastrophe Analytics, EMEA West and South.

France has experienced an increase in the frequency and intensity of subsidence events since the beginning of this century, with extreme events in 2003 and 2011. The last three years, 2018 to 2020, have also been extremely severe years for subsidence risk, with more than 1billion euros in average losses per year. Consequently, Willis Re has built a new subsidence model to enable our clients to understand their portfolio’s exposure to this risk.

This graph shows the number of communes in each year - description below
from 1988 to 2020, that are recognised for the first time (yellow), recognised (purple), or not recognised (green) to have declared to the government for subsidence risk nat cat cover. A commune can be classed as ‘not recognised’ but still have subsidence claims. The number of communes in all categories has increased significantly over recent years. details the information on the infographic to meet accessibility requirements.
This image shows an aerial view of France. There are three coloured subsidence risk categories, from pink (low) to red (high) depicting the level of subsidence risk in areas across France, based on our model.
Figure 1: Communes Affected by Subsidence

Top: The number of communes in each year, from 1988 to 2020, that are recognised for the first time (yellow), recognised (purple), or not recognised (green) to have declared to the government for subsidence risk nat cat cover. A commune can be classed as ‘not recognised’ but still have subsidence claims. The number of communes in all categories has increased significantly over recent years.
Bottom: The spatial distribution of the communes in each year from 1988 to 2020, that are recognised for the first time (yellow), recognised (purple), or not recognised (green) to have declared to the government for subsidence risk nat cat cover.
Source: Willis Re based on CCR data

Subsidence represents a large proportion of the natural catastrophe losses for insurers in France and was integrated into the French Natural Catastrophe regime starting from 1989. The NAT CAT regime uses geotechnical and meteorological criteria to accept or refuse natural catastrophe declarations (see figure 1).

Under this regime subsidence claims for properties are usually covered. However, the change in criteria in 2018 means it is more difficult to build a predictive model for the upcoming years based on previous claims data.

This graph is showing model estimates of cost and count of subsidence claims (split by CATNAT and non CATNAT) from the Willis Re subsidence model versus CCR as if numbers, for the last 5 years. The model estimates are close to the CCR as if numbers.
Figure 2: Market Estimation for Last 5 Years in France

This graph is showing model estimates of cost and count of subsidence claims (split by CATNAT and non CATNAT) from the Willis Re subsidence model versus CCR as if numbers, for the last 5 years. The model estimates are close to the CCR as if numbers.
Source: Willis Re

Willis Re’s approach has been to provide clients with an alternative view of risk by using multiple open access explanatory variables to calibrate the subsidence model: Standardized Soil Water Index (SSWI), Standardized Soil Temperature Index (SSTI) and the Standardized Precipitation Index (SPI) along with the soil type layer. The French Natural Catastrophe regime classification has been used, which determines whether the event has been declared a natural catastrophe. The model has been calibrated using market historical claims and exposure information at commune level, which aims to predict a first estimate of the cost of subsidence risk for the French market as earlier as September for the current year (see figure 2). This estimate is then fine-tuned with subsequent updated data.

If temperatures continue to rise, as predicted by different Intergovernmental Panel on Climate Change (IPCC) reports, there will be an increase in the amount of property exposed to subsidence risk in France.”

Thomas Kroely | Executive Director, Willis Re

“If temperatures continue to rise, as predicted by different Intergovernmental Panel on Climate Change (IPCC) reports, there will be an increase in the amount of property exposed to subsidence risk in France.” Thomas Kroely, Executive Director, Willis Re.

This image shows an aerial view of France. There are three coloured subsidence risk categories, from pink (low) to red (high) depicting the level of subsidence risk in areas across France, based on our model.
Figure 3: 2019 Risk Map

This image shows an aerial view of France. This map of France shows three coloured subsidence risk categories, from pink (low) to red (high) depicting the level of subsidence risk in areas across France, based on our model.
Source: Willis Re

Informing (re)insurance decision-making through risk quantification

By building our own View of Risk for subsidence, Willis Re can provide support varying from reserving and underwriting, to reinsurance, helping our clients to protect earnings volatility. A first estimation for the number and cost of claims at commune level is produced in early September of each year (see Figure 3).

Willis Re can provide multiple risk underwriting layers based on our subsidence model for clients, as either a general risk map and/or risk maps for a particular year (see Figure 3) to demonstrate how the risk has been evolving. The areas with the highest risk predicted by the model are consistent with the historical Nat Cat declarations. By quantifying subsidence risk, insurers can adapt their underwriting strategy and their reinsurance covers to take into consideration the strong evolution of this risk in terms of physical phenomenon and the regulatory context.

Going forward Willis Re plans to update the model each year with the new claims data obtained from the market and using the latest physical indicators. Additionally, two climate change scenarios have been tested, which have both been created using an increase in the soil temperature based on projected climate scenarios. The first results show an increase in property damage from subsidence risk in France, which is in line with previous reports published by the Fédération Française de l'Assurance (FFA) on the same subject.

For more information regarding our model please contact us.

With acknowledgements to Tina Thomson, Rachel Gillespie, Nikita Aksenov for their support and to all our modelling and actuarial team members in particular Sarah Lochhead, Alec Hong and Lucas Grandperrin who have contributed to the development of our model. The contribution of Molly James and the technical support of Arthur Charpentier are also gratefully acknowledged.

Author

Head of Cat Analytics for France and Belgium

Hani Joined Willis Re in 2019 as the head of the Cat analytics team in Paris. Hani has a PhD in Applied Mathematics and as an extensive experience in CAT modeling especially flood modeling. Prior to joining Willis, Hani was working at AXA as an internal model developer for floods for European and Asian countries (e.g France, Belgium, Germany, Italy, Malaysia, Indonesia). Hani has published scientific papers in peer reviewed journals.


Related content tags, list of links Article Reinsurance Insurance Climate Change Insurer Solutions France

Related Capabilities

Contact Us