IIG News

Catastrophe Modelling and Closing the Gap for Secondary Perils in South Africa –  Do we have the technology and what are the Commercial Implications?

IIG President Thabo Twalo kicked off this Insights Session with a welcome to all attendees.  Having spent time working for a Re-insurer, Thabo introduced today’s topic by sharing insights into how significantly Cat modelling has progressed in the last few years.

Referencing the recent Floods in Durban and Wildfire events in the Cape reminds us of the potential for Cat events and draws attention to Re/Insurance company liquidity.

Accurate Global Cat modelling is a key issue which Re/Insurers are struggling to come to terms with but is essential to allow all stakeholders to better quantify the risk of Cat events/losses.

At this point Thabo highlighted the sponsor for the session, being Gallagher Re and referenced that they have successfully utilised data to improve Cat Modelling.  Gallagher Re have partnered with Private and State Agencies to obtain data which is then fed into sophisticated cloud-based computer systems and finally analysed to produce models which accurately assist with providing details on risk exposure.

The first speaker was introduced, Martinette Pretorius who holds an Honours Degree in Actuarial Science from The North West University. Her role encompasses risk and data analysis, reinsurance structure analysis and pricing.


Martinette spoke to the Commercial Considerations

  • Secondary perils are becoming loss drivers eg KZN floods 2022; Germany storm 2021; Texas Freeze 2021
  • These secondary perils are challenging to model
  • Climate change will have impact on meteorological perils

Martinette reiterated that Gallagher Re has a robust view of Cat risk across SA and Africa.  By applying benchmarking analysis:

  • clients are able to see how their retention limits compare to the market; 
  • Gallagher Re are able to assist with underwriting of risks 
  • Gallagher Re use data to stress test and execute scenario analysis

In their data analysis, Gallagher Re are able to enhance and develop their own view of risk, taking a dual approach to:

  • Enhance and validate existing models
  • Building and customising new models where no established models exist 

The second speaker was Stephen Tench, who holds a PhD in applied Mathematics and machine learning.  His academic background focused on the application of advanced modelling techniques to capture spatial and temporal patterns and terrorist attacks.

Stephen provided interesting insight into Closing the Gap Using Modelling Techniques through the use of data which is essential in any modelling strategy with Gallagher Re leveraging off the latest sources of innovative datasets which is fed into cloud infrastructure and then utilised in modelling.

Key to being able to provide more accurate modelling is the need for a larger dataset at a more granular level to facilitate modelling in a unique and different way.

Climate change, which is a key consideration in the industry, necessitates the acknowledgement that previously used Cat models may no longer be relevant – models must constantly adapt and change to offer clients the best view at a given point.

Data is therefore an essential element of any modelling strategy.

Gallagher Re datasets include:

  1. Global Human Settlement Layer – provides population concentration and classification.  It gives insight into what’s going on at ground level.  Utilising this data allows better understanding of the geography and how this could affect/result in a loss
  2. Global Earthquake Model – which is one step beyond GHSL – gives characteristics into occupancy, building types, etc to better understand exposure
  3. OpenStreetMap – provides street level details of towns and cities across the world
  4. NASA GPM – data based on satellite imagery which provide information on precipitation pattern (half-hourly) and thereby enables forecasting of extreme weather events and evolving weather phenomenon

Gallagher Re is able to combine all this data to get a more comprehensive view. The data is fed into a cloud-based solution which allows scalability and avoids limitations on hardware. The data is then modelled to the individual client and fed into Power BI for a visual representation of the same.

The end benefit of these Modelling Techniques allows for more accurate capital reserving as well as proving to be beneficial to society as the models indicate the risks and allows for future proofing.

The final speaker was Toby Sansom, Senior Analyst and leads the Sub-Saharan Africa Catastrophe analytics team at Gallagher Re. Toby holds an MSc in Climate Change, focusing on evaluating climate models’ ability to replicate the climate response to volcanic eruptions. 

Toby provided great insight into Successfully Modelling Client exposures to Cat Events

In 2021 Gallagher Re released an innovative hail model which looks at both day and night sets. This is a very challenging model as hail is very localised/narrow and a “small event”.  To produce this model Gallagher Re partnered with KIT and NASA to obtain data from Satellite imagery, validated with historical insurance claims and then modelled the data.

Key challenges in producing this model: 

  • how to disaggregate postal code/cresta level exposure to a high resolution 2km grid
  • use OpenStreetMap to constrain location based on LOB

Further to this Gallagher Re has developed a flood catastrophe model for South Africa – the first of its kind – for producing flood reinsurance loss metrics for the region.

Toby then posed the following questions:

  1. How will extreme rain events change under climate change?
  2. How will climate change impact severity and frequency of hail events?
  3. Can the industry do more to make society more resilient?

Toby highlighted that one can only model with the data you have, and Africa remains largely un-insured.  There is therefore a data gap but working with private and state agencies can provide much needed data source.

Toby ended with the following conclusions:

  • Commercial:
    • 2022 Durban floods shocked the market
    • Need for improved analytics to ensure that premiums reflect the real risk
    • Insurers can then work with policyholders and other stakeholders to mitigate risk
  • Technology:
    • Access to a plethora of data sources
    • Advanced computing power through MS Azure Cloud
    • Combining data and computing resources with innovative modelling
    • Improved understanding of risks being carried which facilitates accurate reserving and risk mitigation.
  • Gallagher Re Solutions:
    • Strong offering of proprietary models for SA, including hail, flood and earthquake
    • Models have been built with in house expertise and academic partnerships
    • Getting a handle on climate risk and collaborating across sectors is key to build resilience

The Insights Sessions was concluded with a brief Q&A session led by Thabo which highlighted the need for the industry to endeavour to work more closely with state/town planning authorities to enable further opportunities for risk mitigation.


Article written by: Linroy Peters