The global reinsurance broker Guy Carpenter, a partner to Nasdaq and user of the Nasdaq Risk Modelling for Catastrophes service, continuously develops new catastrophe risk models across perils and regions of the world. We sat down with Mark Weatherhead, head of Model Development at Guy Carpenter, to talk about their approach to risk modelling, as well as the trends and challenges they see in the future.
Why have you chosen to partner with Nasdaq Market Technology?
The Nasdaq Risk Modelling for Catastrophes is powered by the OASIS Loss Modelling Framework, and it’s designed to be an open-source platform for members of the insurance community, like Guy Carpenter, to explore and analyse natural peril risk. There’s a lot of complexity that goes into catastrophe risk modelling. The scaling possibilities and the robust structure that the Nasdaq platform provides have helped our operations and enabled us to run multiple risk models across perils and countries. So the platform offers a possibility to easily scale the modelling, which in theory means that we can run thousands of CAT risk analyses per year.
If we dive deeper into the Guy Carpenter business, can you share some background to your mission and how you serve the re/insurance industry?
We serve the insurance industry. Our main goal at Guy Carpenter is to help our clients, insurance companies, manage risk. In this case, that’s natural peril risk such as earthquakes and floods. In terms of managing risk, we offer advisory and solutions, and we have invested heavily into filling the gap in the mainstream catastrophe risk models and analytics. We have focused on creating tools and models for our clients to help them better understand their risk exposure. With the Nasdaq Risk Modelling platform, we can run multiple models to analyse natural peril risk without being tied down on the technology development or software side. The transparency that comes with the open-source set-up has essentially helped democratize the CAT risk space, allowing us to focus on the analytics rather than building and maintaining the software.
If we look at your model development more specifically, do you focus on any specific regions or perils? And how does your model development enable efficient management of risk and climate change challenges?
Guy Carpenter’s model development has historically focused on addressing the gaps in knowledge; in other words, make sure there are models where models are needed, which for the most part has been for flooding. Our goal is always to help our clients manage their risk exposure, meaning that we focus on perils and regions where our clients have exposure. Our clients increasingly want to have a view on their exposure and risk on a global basis, not just for the traditional peak areas. In the past year or so, we have released flood models for parts of Asia such as Malaysia, Indonesia, Vietnam and Thailand. In addition to developing and providing models across regions and territories to support our insurance clients, we also increasingly work with governments, NGOs and other organizations such as World Bank to address natural disasters and climate risk. In many countries, when a natural disaster strikes, the responsibility mostly falls on the government. That in itself can cause a strain on the government and take resources from other more pressing issues. Disaster risk financing, through mechanisms such as reinsurance or cat bonds, can go a long way to improve the resilience of communities to natural disasters and to significantly decrease the recovery time. That’s where our modelling comes into place, as you need to have a view on the frequency and severity of natural perils in order to secure investment. We can use a framework from Nasdaq that isn’t tied to a specific model vendor, which is often more attractive to public-sector clients and creates a common basis for understanding risk.
How has partnering with Nasdaq Risk Modelling for Catastrophes enabled your business to service re/insurers worldwide?
Partnering with Nasdaq has simplified the process for us to run models on OASIS at a production-level scale. As we can use one platform to run both our own models and those of other vendors, it means we only need to train our analysts and modellers in how to use NRMC while still being able to access the models that our clients require. That allows them to focus on what they are there to do – analyse natural peril risk – instead of retraining to learn different platforms.
Since the beginning of the partnership, we have collaborated closely with the Nasdaq team on the software development side. They’ve taken our advice and viewpoints into consideration when developing the new SaaS-based platform. The data our analysts work with looks different in every country. So the platform needs to be flexible enough to be able to handle the data, regardless of country, region or peril.
Additionally, the models are used and run differently depending on the end-user and the questions we are asking. An analysis to inform on placement of reinsurance can look very different from that used for internal portfolio management purposes. Likewise, a governmental client may be primarily interested in the total damage cost rather than a post-re/insurance perspective. The end goal is still risk management. And having access to a unified platform that our analysts and modellers can use helps us service clients, no matter the region or industry or peril.
Nasdaq’s CAT risk modelling platform is powered by the OASIS Loss Modelling Framework, an open-source framework that enables easy access and industry standardization for catastrophe risk models. How has the OASIS framework been relevant for your business and model development?
The open-source perspective with OASIS has certainly reduced some barriers of entry for smaller model developers, users and academic groups. As I mentioned, the OASIS initiative democratizes knowledge in risk and CAT risk modelling. I also have to emphasize the ecosystem aspect of the framework and the platform because the community around this has grown, and it’s become a collective effort that feeds back into the framework for continuous evolution.
The democratization aspect is one point. But beyond that, there’s the standardization aspect too. When there’s a standard in building, deploying and running models that we can all gravitate towards, there are efficiencies in both time-spent and learning. There’s value in having a single platform that the insurance community can all unify around and develop together.
What are you looking forward to in the rest of 2021 for Guy Carpenter and for the industry?
From the OASIS perspective, I’m looking forward to continued improvements to the platform and increased ability to run at scale, and we look forward to launching a new suite of models that are in the pipeline.
Improving knowledge around climate change and understanding the impact of different climate change scenarios is important for us and the rest of the industry. There is no doubt that climate change is going to affect the industry as a whole, and it is a good opportunity for us to work together going forward to share data and insights for the benefit of the communities that we all support.
CAT modelling is very different today than it was 30 years ago when it all began, and even in the last 15 years since I first started, we’ve seen an amazing increase in the amount of data and computing power available. Platforms like OASIS and NRMC have the potential to change the industry, and I look forward to seeing what’s next in terms of analysing large data sets at scale and at high performance. It’ll be interesting to see how this is going to grow and if we can get to the stage of running truly global models that can allow us to understand everything from regional patterns right down to the impact on individual buildings.