Contribution of CCR models to measure the effectiveness of prevention measures on insured losses
(1) Jérémy Desarthe - Public Funds and Prevention Department, CCR
(2) David Moncoulon - Modelling, R&D Department, CCR
(3) Roxane Marchal - Modelling, R&D Department, CCR
INTRODUCTION
The partnership with the DGPR
In the fall of 2021, a multi-year agreement was signed between CCR and the Direction générale de la prévention des risques (DGPR) of the French Ministry of Ecological Transition. This agreement entrusts CCR with missions of general interest relating to knowledge of natural risks and prevention measures. By using CCR data and models, the work specifically aims to improve diagnoses of the exposure of territories to natural hazards, particularly up to 2050; to analyse the implementation of public prevention measures, to evaluate their benefits and to provide support for the structuring of public prevention policies and assistance to the State’s decentralised services in their work and their thought process.
Modelling and projections to 2050 in summary
Through its role in the natural catastrophe compensation scheme, CCR has been developing a platform to model hazards and insured losses for over fifteen years. The goal of this modelling work is, on the one hand, to provide an estimate of the damage resulting from extreme events a few days after their occurrence, and, on the other hand, to provide the necessary elements for the sustainability of the Nat Cat scheme over the long run. With this in mind, several studies have been carried out with Météo-France to assess the impact of climate change on the losses from natural catastrophes by 2050. An original methodology was implemented, with the ARPEGE-Climat scenarios with constant climate: i.e. the simulation of 400 years of ‘2000’ type climate and the simulation of 400 years of ‘2050’ type climate according to different greenhouse gas concentration scenarios resulting from IPCC work. In 2018, the study focused on the consequences of the RCP 8.5 scenario, which is the most severe one and corresponds to a probable temperature increase of +4°C in 2100. The results showed that climate change would result in an average 35% increase in average annual insured losses due to an increase in the frequency and severity of natural catastrophes. By including the impact related to the concentration of populations, as forecast by INSEE, in risk-prone areas, this increase would be +50%.
The most significant impact is that on coastal flooding. Indeed, the damage would increase by 82%, mainly due to a 23cm rise in sea level. This increase, which may seem small, becomes an important factor in increasing damage in the event of coastal flooding due to a coastal storm.
Studies related to the prevention and reduction of damage
Impact of the Elan Act
Geotechnical drought linked to clay shrinking and swelling (CSS) is now a significant risk. Over the 2016 - 2021 period, it amounted to almost 49% of the compensation paid out under the Nat Cat scheme. Geotechnical drought is thus ahead of floods, which have accounted for 53% of insured losses since 1982.
In 2018, the Elan Act attempted to provide an answer to this issue by first proposing a mapping of the territory’s exposure to the CSS phenomenon. This mapping included recommendations on the studies to be carried out for new building projects. The goal is to allow housing to be built while not increasing the exposure of issues to the phenomenon of clay shrinking and swelling.
Thus, modelling the evolution of damage resulting from the CSS drought phenomenon was carried out by ignoring the damage to new builds in the areas at risk, which is considered to be covered by prevention. To this end, the identification of these new builds in relation to their location with respect to BRGM zoning was carried out and a new portfolio of insured property for the market was built.
The simulations of 400 years of current climate and 400 years of future climate were repeated by altering the population parameters in these high-risk areas. The results show that damage will increase slower by 2050, highlighting the effectiveness of the Elan Act in limiting the consequences of climate change by fo - cusing on the issue of new builds. The strict application of this new regulation should thus make it possible to reduce the average annual loss by €100 million by 2050.
Adding up these benefits over the 2022 - 2050 period, we obtain a €2.8 billion ‘saving’ for the natural catastrophe compensation scheme.
Figure 1 - Changes in average annual losses by hazard type under the impact of climate change by 2050.
Investments in hydraulic structures
Intended to reduce the frequency and intensity of floods, the creation or rein - forcement of hydraulic structures is one of the main levers of the prevention policy. Over the last decade, they have represented 38% of the gross delegations of the Fonds de prévention des risques naturels majeurs (FPRNM). Over the 2016 - 2020 period, this represents an average annual investment of €96 million by the FPRNM. This sum only represents a part of the investments in the hydraulic structures as local authorities also participate in financing them. A total of €215 million is invested in hydraulic structures every year.
# modelling
# prevention measures
# insured damage
# cost-benefit analysis
To assess the impacts of these works, CCR sought to determine the benefits of investments on claims experience up to 2050. The methodology used is based on that established by the General Com - mission for Sustainable Development (CGDD) and is currently being imple - mented through the Flood Prevention Action Programmes (PAPI). To determine the benefits, i.e. the damage avoided, over the long run, it is necessary to have the B/C indicator, which determines, for an investment cost (C), the benefits ex - pected from the measure (B). The analy - sis of the reports of the PAPIs submitted for labelling by the Joint Flood Commis - sion for the 2018 - 2020 period has made it possible to define an average B/C ratio of 3.5, i.e. for every €1 invested, the expected benefits are €3.5 over a 50-year period.
If the annual investments observed over the 2016-2020 period were maintained, the loss experience due to flooding and coastal flooding events by 2050 would be reduced by €130 million. In the end, the application of the Elan Act and the pursuit of investments in flood prevention through the financing of hydraulic structures should make it possible to partially curb the expected increase in the loss experience, which would amount to 32% (compared to 50% without these prevention measures).
Figure 2 - Modelling the effect of great lakes (GLs) on the damage caused by the January 1910 flood at municipal level.
Impact of the Seine great lakes
Following the drought of 1920 and the floods of 1924, several actions were implemented to reduce the vulnerability of the Seine basin to hydrological risks. A vast programme for the creation of reservoir lakes to fight against both flooding and low water levels in the Seine was launched. The first to be built was the Pannecières reservoir on the Yonne River in 1949, followed by the Seine River in 1966, the Marne River in 1974 and the Aube River in 1991. In total, these four reservoirs have a capacity of 807 million m3 and are responsible for regulating the upper Seine basin. The La Bassée project, for which construction of the pilot site was launched in 2021, is supposed to eventually increase this capacity to 862 million m3 .
Following the 2018 floods, CCR in collaboration with EPTB Seine Grand Lac sought to assess the effectiveness of these great lakes on insured losses. The aim was to determine the benefits, i.e. the damage avoided, that the presence of these reservoirs had allowed.
First, the effectiveness of the Seine great lakes was measured during the winter flood of 2018, during which the great lakes were able to play their role to the full as flood control structures. Of the damage estimated at the time, their effectiveness was to reduce the height of water by 65cm on the Seine and 75cm on the downstream Marne, thus reducing the insured damage by nearly €100m, i.e. 30% of the damage estimated at the time of the event.
To assess the effectiveness of these protection measures, the flood of January 1910 was simulated, as if it occurred nowadays, with and without the presence of the Great Lakes. It was shown that, without the Great Lakes, the insured damage for this event could reach €15-33bn. On the other hand, with appropriate flood management by the Great Lakes, this damage would be reduced by up to €12-28bn, i.e. a saving, on average, of €3.5bn (including €2.5bn in damage to professionals). The annual frequency of occurrence of such an event is estimated to be between 1/100th and 1/200th.
Table 1 - Modelling the effect of great lakes (GLs) on the damage caused by the January 1910 flood at commune level.
CONCLUSION
In conclusion, the models developed by CCR over many years are being used to conduct prospective studies to measure the effectiveness of prevention measures. These models have been calibrated based on historical damage observed by all stakeholders in the French insurance market and are therefore the most robust tools to assess damage in this area. By altering the vulnerability of assets in these models, it becomes possible to carry out fine-scale studies of the impact of prevention measures. Then, using stochastic event catalogues, to generalise these impacts to the entire country.
This approach is currently being implemented as part of the agreement between the DGPR and CCR to estimate the avoided damage caused by the presence of hydraulic structures during the floods that affected the Aude in 2018 and the Var in 2020./
CITATION
Desarthe et al, Implementation of CCR models to measure the efficiency of prevention measures on insured losses. In Rapport Scientifique CCR 2022 ; CCR, Paris, France, 2022, pp. 44 - 47