INTERVIEW: DAN DICK, AON’S REINSURANCE SOLUTIONS
Cedants must form their own view of risk
Aon has recognised that comprehensive risk analysis does not result from taking today’s event sets and adjusting them, but rather building event sets using a future climate model.
As we move into a tougher renewals market, it’s vital that cedants develop their own view of risk, according to Dan Dick, global head of property analytics at Aon’s Reinsurance Solutions.
“It used to be that you just pushed everything through the models, everyone would look at the model results, and the reinsurance transaction would occur,” said Dick, speaking to Monte Carlo Today.
He added: “What’s important now is that companies and cedants understand how the models perform on their book, what their expectations are of performance, and what adjustments they make for things that may or may not be in the model to clearly differentiate themselves. This enables them to make better business decisions.”
Dick said that having a unique view of risk for each client should lead to a better predictability outcome on the models when an unforeseen event occurs.
“It’s essential when we go into a marketplace where capacity is constrained from the reinsurance side, that we are able to tell that story on behalf of clients. We enable our clients to develop a view of risk, own that view of risk, and be able and willing to talk to others about why it’s the best view for their unique business,” he said.
The climate change journey
Dick noted that modellers are taking the past several years of experiences and building those into the models for many perils, so climate change works itself into the models.
However, on the underwriting side, the issue for the individual perils is whether the models are “adept enough to truly predict or look at underwriting individual risks in an individual location due to a high frequency severity peril”.
“AI is going to let us process a massive amount of data much faster.”
Dan Dick, Aon’s Reinsurance Solutions
Dick said: “Think about the tornado that went through Kentucky and southern Illinois late last year. Could that event have been predicted and helped you underwrite the warehouse or candle factory that were destroyed? That’s where some of the uncertainty comes in, but the models are improving. It’s been an interesting journey.”
In a bid to enhance its view of climate change, Aon has announced a series of academic partnerships, the output from which will inform its modelling suite in order to help clients navigate climate volatility.
“We quickly realised when trying to think about how to build a view of climate risk into the models, that many of the people who were doing it were taking an event set representing the risk today and adjusting the frequency or toying with that event set to try and replicate the future,” said Dick.
He added: “It became clear to us that the climate of tomorrow is going to generate for some perils—but not necessarily all perils—events that look different from what we experience today.”
Aon has recognised that comprehensive risk analysis does not result from taking today’s event sets and adjusting them, but rather building event sets using a future climate model.
The “immediate answer” to this problem was to go to the experts who are building those climate models and have them work in tandem with Aon to build future event sets, said Dick.
“That became the foundational springboard to get us to that next step of having our partners build those event sets. It’s very differentiating in the industry and gives us a very insightful basis for conversation with clients,” Dick added.
“We’re running tens of millions of events on a global basis.”
Aon is addressing a lack of modelling in certain geographies across the world. Citing South African hail and South American flood, Dick said that while these areas may not have large insured losses on a relative scale to other parts of the world it is still very large for that region.
He added: “One of the things we’ve worked hard on is making sure our own Impact Forecasting platform is open and transparent for bringing models in so that it makes it easier to work with local governments or research universities to build out some of these perils.”
Finally, Dick considered the developments of artificial intelligence (AI) and quantum computing. He explained that the industry has a lot of data and information in paper form that could be more efficiently translated for cat models and risk assessment, while also holding information that is not necessarily captured on a policy form but can be important for underwriting attributes.
“AI is going to let us process a massive amount of data much faster, but it’s also going to let us identify the key underwriting attributes or risk decision attributes that aren’t captured and use other means to go out, capture and bring them forward,” he said.
Quantum computing is set to benefit the industry greatly.
Dick concluded: “When you start thinking about global perils, global regions and very robust simulations to predict what could occur, we’re running tens of millions of events on a global basis.
“Anything we can do to speed up that process and turn over those results is going to help the industry to react when events do occur.”
Dan Dick is global head of property analytics at Aon’s Reinsurance Solutions. He can be contacted at: dan.dick@aon.com
Main image: Shutterstock / Max4e Photo
