When it comes to risk, Chaucer works with a “three lines of defence model,” says James Wright, the company’s Chief Risk Officer. “You have policy underwriters and they manage their own risks and balance risk and reward. They’re the first line of defence. I’m responsible for the second line of defence, which is reviewing the risks being taken in the business – we’re a bit of a control function.” The third line of defence is the functions that provide independent assurance such as internal audit.
Wright’s work is split into two main areas. “One is quantitative modelling and the other is qualitative risk analysis. So, I have two teams. The quantitative people are actuaries – and I’m an actuary by trade. What they’re interested in building is quite complicated models of risks.”
Perhaps surprisingly, in our information-overloaded era, there is often less data involved in these models than you might expect. “If you’re working with, say, motor insurance, you have large amounts of data,” says Wright. However, “If you take the big claims, which are for over $1 million, you might have only had one in the last ten years.” For this reason, he says, data-based modelling can be difficult. “We can build the best model in the world but if there are only four data points to build it from, the model’s always going to be rubbish.”
These big, rare claims tend to come from disasters like the Deepwater Horizon oil spill. But it varies by class of business. “It could be an aeroplane falling out of the sky or it could be e-cigarettes and a battery exploding in someone’s pocket. These things happen and they can cost us a lot of money.” Regulators, he explains, require all insurers to create these models in order to assess how much capital they need. “From an economics point of view, this capital is really just a buffer. It’s like saving for the rainy day. So, our teams will look at the risks and ask, ‘How much do we need?’ ”.
This is the quantitative side – and every syndicate in Lloyd’s does this. “What we do with quantitative modelling, which makes us a bit different, is that we make much greater use of it in terms of helping us run the company. If we were to go into a new line of business, someone would come and ask the capital team: ‘What’s the impact of your model on the capital we need? How does it affect our likelihood of making a return or making a loss?’.”
Overall, Wright says, these models are a very useful tool. “We use them all the time when we’re looking at purchasing reinsurance. I often think of our work the way a bookmaker would. We’re laying a bet. These models will tell us if the bet represents fair value for us.”
The other side of Wright’s work is the qualitative team. This is rather different and nowhere near as common in the sector. “We don’t need actuaries for this,” he explains. “It’s very subjective and we need people with good common sense and a bit of an imagination.”
If you take climate change as an example here, Chaucer has a qualitative exposure management team, which will look at risks related to this. However, “The qualitative team will look at climate change much more widely to try and say how it affects other parts of our risk profile.” This will include the company’s investment portfolio – and might take in risks such as having money in coal-fired power stations (whose use is likely to decline in the future) and the offices in Miami (which are likely to be at greater risk of hurricanes caused by warmer weather and flooding caused by rising sea levels).
Much of the qualitative team’s work is identifying emerging risks; in a way it’s quite similar to the fashion industry identifying new trends. “When we say risks, they’re not good or bad things. Rather we’re trying to understand how new risks will affect our business in the future.” One such risk is the automation of vehicles. This will impact motor reinsurance and could be an opportunity, particularly if premiums stay at similar levels. “I think 90 per cent of the work is the identification of risks. Once we’ve done that, we can start to think about how we can manage those risks.”
Wright’s teams also look at how Chaucer’s risk strategy and risk profile compares to other companies in the sector. “It’s that old adage isn’t it?” he says. “Two people are in the jungle being chased by a tiger – the first person turns to the other and says, ‘You’re quicker than this. Why aren’t you running any faster?’ The other says, ‘I don’t need to run faster than the tiger. I just need to run faster than you’ ”.
To this end, Wright’s teams do a lot to understand what Chaucer’s exposure is to events relative to its peers. “We pick disasters around the world and try to understand our exposure and how much we could lose.” Some of these models are quite exotic and even outlandish, “We have an aviation scenario that tries to look at the extreme worst case where aircraft of two airlines we insure collide mid-air above New York City, coming down on the Museum of Modern Art, which we also insure.”
Of course, you can only do so much with models and you need to recognise that they can go very wrong (as they did in the financial crisis) and have limits. “We’re very, very clear about where models are good, where they’re bad – and we’re big believers in that quote from George Box, the famous statistician. He said ‘All models are wrong but some are useful’.”
These useful models, says Wright, help make Chaucer a well-run company. “You look at some people out there and they’re underwriting some really wild risks. We don’t want to be one of those reckless risk takers. We want be a sensible insurer who will be there when you need us.”