" class="no-js "lang="en-US"> EXCLUSIVE: "Calm before the Storms" - Jamie Rodney, Reask in 'The Fintech Magazine'
Friday, June 14, 2024

EXCLUSIVE: “Calm before the Storms” – Jamie Rodney, Reask in ‘The Fintech Magazine’

Traditional models underestimate the role of climate as a driver of past events. Jamie Rodney, CEO of Reask, explains why a new approach, using climate-aware AI, is needed

Tropical cyclones are one of the most destructive natural disasters. According to Aon, over the last 10 years, they have caused in excess of $1trillion in economic damages globally, with less than 50 per cent of those damages insured. This presents a major issue – and opportunity – for the financial services industry and insurance markets, and the clients they serve.

The ability to accurately forecast the number, impact, and location of hurricanes, cyclones, and typhoons has always been very challenging, especially as they have grown in intensity due to climate change. As a result, the financial services sector and re-insurance markets have developed a much greater need to understand climate risks. These businesses are reliant on a clear view of how weather systems are changing in an increasingly volatile climate, so they can prepare for the extent to which their physical assets, infrastructure, business models, and customers are exposed.

Not only does this help to inform insurers’ strategies around climate variability, but it improves capital management and relative volatility of global exposures, and helps to quantify and hedge risk in response to shorter-term extreme weather expectations. Fortunately, through advances in technology, research, and the use of AI and machine learning, new solutions, such as those from Reask, are emerging to provide decision-makers with the data they need to confidently change their assessment of potential catastrophe risks over multiple time horizons.

The weather does not see borders, which is one reason why the limitations of traditional natural catastrophe (Nat Cat) models, built region by region, are being challenged. Fresh thinking, improved AI and machine learning, and increased computing power are now enabling experts to understand climate variability better than ever, including the potential impact of the frequency and intensity of natural disasters, worldwide. Today’s climate models, with a more automated and scalable process, enable experts to get a more interconnected understanding of global climate risk.

At Reask we have built robust algorithms to transform this climate model output into risk data that dynamically captures tropical cyclone wind variability down to a neighbourhood scale, over multiple time horizons. Nevertheless, the ultimate aim is to have a consistent view, born out of one methodology, irrespective of where the exposure happens, from the South Pacific near Australia, to the north Indian Ocean, the Gulf of Mexico, or the Atlantic.

THE LEARNING CURVE

Traditionally, historical data and statistical modelling techniques have been used in models over the last 30 years to approximate the likely risk of future tropical cyclones.

Creating these models involves many PhD scientists building one model for one country at a time, then switching to another, or maybe region by region, with each of these models manually taking months or even years to build.

Further, while these models still have some value, they leave many questions unanswered, such as what happens when risks we observe hit new levels? What if they behave in ways that are different from what we have observed in the historical data? And how can we make a model aware that the underlying climate may be different to the historical data?

To successfully and accurately create one global model requires climate pattern extraction with a globally connected framework and machine learning approach. This involves embedding the latest climate knowledge into algorithms that allow the model to know the physics that underpin tropical cyclone risk and build risk distributions accordingly.

By looking into the atmosphere and, to a certain extent, the hydrosphere, we can now understand how the climate impacts the extreme perils that are phenomena of that particular environment. Furthermore, we need to look at so-called pattern recognition often referred to as ‘unsupervised learning’. These are algorithms that can go through vast volumes of data and figure out what is worth using. Afterwards, human experts need to assess the data to make

sure we understand the physics being picked up, but it is close to a fully automated process. A lot of what’s been done in the past has been by experts looking at data and picking up global patterns. The typical example would be the ENSO patterns, which cover things like water temperature.

“Beyond insurance, the climate-aware nature of our model is being used across operational risk and resilience to assess the impact of climate change on companies and the investments they makeToday”

They are usually, however, very simplistic. Unsupervised learning allows us to use machine learning algorithms to do that job instead, faster and more accurately. What we have found at Reask is that we gain much in terms of predictive skills, and we can trust the machine to pick up the right signals to speed up the process considerably.How often this data is updated depends on what one is trying to do with the data and what part of the model is being looked at.

The global climate products available from either the US agencies or the European centres are excellent; their data is available every month and provides the ability to look at hurricane activity in the coming season. However, for tropical cyclone wind structure and what happens to the winds when they travel overland, new datasets are needed. At Reask, we use a weather forecasting model and run it at very high resolution on hundreds of historical cases.

This has been the training data we use for machine learning.Many firms will have invested heavily in Nat Cat models over the years, thus it is important to understand that their existing models still have value and can be integrated with the models and approach we are discussing.

Organisations can have a collaborative view of risk, and augment their existing models with this new, next-gen approach. For example, where firms don’t have coverage in certain regions, these models can provide a consistent view across all territories. One way to think of the operational side is that newer models, like Reask’s, answer questions that other models don’t because of the global approach. A client can look at relativities in the long-term view of risk we have, versus a particular season in the Caribbean or US, for instance.

Today, Reask’s model is being utilised across the climate risk and insurance value chain. The resolution and accuracy of the model is being used to underwrite insurance policies exposed to tropical cyclones more efficiently.The global consistent nature and granularity of the risk data is being used to innovate the parametric insurance space, by allowing risk carriers to offer globally consistent tropical cyclone insurance coverage.

This is a huge area of opportunity, and it is already plugging protection gaps, worldwide.Beyond insurance, the climate-aware nature of our model is being used across operational risk and resilience to assess the impact of climate change on economic and operational resilience of companies and the investments they make.We have also successfully provided operational forecasts for the ILS industry for four years in a row. Now including where we see more risk, including for hurricanes hitting landfall.What we are discussing is new, breakthrough science.

The market determines, in the end, what is an appropriate approach to take about assessing this risk, and that encompasses all of their internal risk appetite and approach to risk management. Communication is vital and we are in regular exchange with decision-makers across the market to constantly inform others and evolve our methodology. It is a very transparent process. We also write publications that are peer-reviewed and we make that information available to them.Managing tomorrow’s climate risk today remains a challenge; one that we at Reask are addressing.

However, we are very fortunate that all our clients are sophisticated users of risk assessment, catastrophe modelling, and hazard modelling tools. From ILS firms to reinsurance intermediaries and global insurers, the way they incorporate and utilise our information is varied and unique. The one consistent is how our clients are turning a climate challenge into an opportunity.


 

This article was published in The Fintech Magazine Issue 28, Page 62-63

People In This Post

Companies In This Post

  1. EBANX Obtains ISO/IEC 27701:2019 Certification and Reinforces Its Commitment to Data Privacy Read more
  2. LTX, A Broadridge Company, Releases GenAI-Powered List Trading Functionality Read more
  3. Laura Bayley, Head Clearing Services, SIX, Appointed SWIFT Board Member Read more
  4. Paytiko Expands to Dubai, UAE Read more
  5. Lumin Digital Announces Partnership With Pinwheel to Offer Banks and Credit Unions Advanced Digital Deposit Switching Solution, Pinwheel Prime Read more