Unless the insurance industry finds a way to close the gap between the prediction and the pricing of weather-related risks, South Africans might see increased insurance costs or even risk being uninsured for climate risk.
According to Ronald Richman, chief actuary at Old Mutual Insure, reinsurers are not particularly happy with South Africa, after mounting industry losses in the past few years. This, together with the increased cost of footing the bill for catastrophic events, is presenting a significant challenge to the ability of the insurance industry to provide the relatively cheap insurance products consumers are used to.
According to Old Mutual Insure data, the frequency of severe weather events has increased from six to 36 per decade since 2012.
“Across the world, we are seeing the trend of insurers leaving areas that are presenting too many risks, and the price has become too expensive – for example, in California and Florida,” says Richman.
“We are only a few catastrophic events away from insurers declaring some areas in South Africa uninsurable. This is why the need for data projects with innovative solutions is critically important.”
Challenges of physical risk modelling
Richman explains that the rapid change in the risk environment around climate change and catastrophic events means that insurers have been struggling with several challenges when it comes to physical risk modelling around flood, wildfire, hail, and rising sea levels, among others.
This has been particularly difficult in South Africa, which has traditionally been a low or “no” catastrophe zone, leading to an underdevelopment of models and analytics for the local market. This has been compounded by changes in the environment because of climate change.
“Traditionally, these risks have been handled with catastrophe modelling, where macro simulation models are run to find out the impact of these events on insurance portfolios. However, there is no set of standardised models for some of the most important weather-related risks in South Africa, presenting challenges,” says Richman.
He adds that another complexity is how to reconcile the outlook on climate change, which is expected to manifest over 10 or 15 years or more, with the traditional one-year period that insurers use when pricing risk or calculating regulatory capital needs.
“There are also many differences when it comes to regions, so changes are not happening in a homogenous way. This makes it even more complex.”
Bringing multi-disciplines together
Last month, Old Mutual Insure announced a partnership with JBA Risk Management. JBA, the UK-based flood science specialist, will provide Old Mutual Insure with detailed flood maps of the country, enabling it to underwrite flood risks better in personal and commercial lines.
Read: Old Mutual Insure teams up with UK flood specialist to tackle rising climate threats
The alliance forms part of a pioneering approach that integrates climate data with claims data to enhance risk forecasting and pricing accuracy in response to mounting climate change insurance challenges.
Old Mutual Insure says the new project is the result of multi-disciplinary work with meteorologists, actuaries, and data scientists.
“The work we are doing allows us to join massive climate datasets with claims data in the South African context. This means we can make better forecasts of insurance experiences, which will ultimately benefit the policyholder as we will be able to charge for weather-related risks at the right premium, as well as help policyholders better understand and manage their risks,” says Richman.
Moving beyond traditional modelling solutions
At the Actuarial Society of South Africa’s 2023 annual convention, actuaries from Old Mutual Insure presented a talk on a micro approach to managing catastrophe risks. The micro approach to managing risk is concerned with a much more granular view of how weather-related risk factors impact insurance experiences, which is the primary concern of the project Old Mutual Insure is spearheading.
Richman says that the project aims to link climate data directly to the insurer’s pricing, to quantify the effect of climate change.
“It does this by incorporating highly granular precipitation data, curated by meteorologists, into traditional short-term insurance pricing datasets. It then fits statistical and machine learning models to observe the predictive value of this addition and then quantifies the potential impact of using future predicted precipitation levels in rating processes,” he explains.
The project aims to quantify the impact of increased precipitation (driven by climate change and the La Niña weather phenomenon) on insurance risk. The company has partnered with researchers from the University of the Witwatersrand, University of Pretoria, and ETH Zürich on this groundbreaking piece of work.
The project shows that new micro-level datasets can enhance the accuracy of actual predictive modelling and that partnerships are essential in tackling these complex challenges in an innovative way.
“By more accurately being able to price risks arising from weather events, insurers can establish actuarially fair rates that can ensure the stability and longer-term health of insurance markets,” says Richman.
