Scor’s Brohmi: Non-peak perils are driving global insurance losses

Maria Brohmi, head of P&C for Western Europe, Middle East and Africa at Scor, answers questions on how modelling of non-peak perils needs to evolve.

The floods that struck central Europe across May and June, triggering estimated insured losses of $2.1bn-$3.2bn according to Moody’s RMS, marked the latest impact of secondary or ‘non-peak’ perils on P&C carriers.

In a Q&A, Maria Brohmi, head of P&C Western Europe, Middle East and Africa at Scor, explained the challenges of analysing these perils as well as social unrest risks.

How has your outlook on secondary perils changed?

Secondary perils are no longer ‘secondary’ – these non-peak perils are driving global insurance losses. Every country, regardless of geographic location, is now a potential catastrophe zone which demands proactive risk management.

A comprehensive approach that includes improved forecasting and risk modelling is required to address this challenge. Risk prevention and collaborative efforts between governments, insurers and reinsurers are also essential to develop reinsurance frameworks and to enhance the resilience of industries. There are discussions in several markets to address this and we are part of those.

Are there particular challenges to non-peak peril modelling?

There are specificities, these risks are difficult to model for multiple reasons. One is that they are a localised phenomena; for example, hail or tornadoes have relatively small footprints of high damage. These perils require not just detailed modelling to capture the different parameters affecting the risk, but also high-resolution exposure data.

Another factor is what we call secondary modifiers of the risks protected. That means the level of the defence systems as well as the characteristics of the risks. These can have a high impact on the potential loss and it’s challenging to capture and account for this appropriately in the modelling.

Then there’s the lack of historical damage information and events data on non-peak perils as well as an unknown effect of climate change. We need to capture the current and future climate when developing risk models.

The cat modelling community has historically dedicated efforts and resources to developing risk quantification models for perils like earthquakes and hurricanes, now the community is shifting to better capture non-peak perils but it’s running a couple of years behind.

Three years in a row – 2021, 2022 and 2023 – we had exceptional events in Europe. When each year is exceptional, we cannot call it an exception anymore.

Social unrest has been another prominent topic this year, and the New Caledonia riots reflected increasing political volatility globally. How has the risk outlook for the strikes, riots and civil commotion (SRCC) class changed for Scor in 2024?

I would say that it hasn’t changed specifically this year, we’ve been observing this trend for several years. Rising inequalities, socio-economic complexity and an increasing political polarisation leading to more frequent and intense protests that can quickly escalate, have all contributed to this trend.

In addition, social media and digital communication can rapidly mobilise large groups of people, which can amplify the scale and speed of civil commotion.

As political instability has spread worldwide, will SRCC be assessed differently?

There is more work to do to properly assess the risk, to make sure it is well defined, relying on robust underwriting guidelines and clear wordings. It is important to define the boundaries of the risk borne and to ensure that there is a probability of having no losses. Otherwise, the question of insurability arises. We need to engage with clients to assess their SRCC exposures and tailor the coverage accordingly. For this, we should consider past SRCC loss experience, but also on-going socio-political developments.

As the risk is increasing and is more present, it requires an increased price, but pricing is also linked to the way the reinsurance programmes are shaped, in terms of structure and wording.