Models, the “skeleton on which you hang your underwriting thinking”: Tom Bolt

Accurate and adequate risk models are a necessity to the operations of the insurance and reinsurance industry, but they are sometimes viewed as a black box which underlines the need for common sense and underwriting judgement to enable a more holistic view of risks.

Executives in the insurance, reinsurance and insurance-linked securities (ILS) space recently discussed the notion that the industry is overly dependent on risk models.

The notion that if you can’t model it you can’t underwrite it, so you simply leave it alone, resulting in heightened competition across easy-to-model, better understood exposures, which in turn creates the challenging pricing environment that the market’s enduring today.

“I look at the model as actually the skeleton on which you hang your underwriting thinking, and that it does a great job of helping you identify exposures, but I don’t believe too much in their predictive power,” explains Tom Bolt, Director, Performance Management, at specialist insurance and reinsurance market Lloyd’s of London.

It’s an interesting way of putting it, and reflects the views of fellow panellists Greg Hendrick and Stephen Catlin, Chief Executive Officer (CEO) of Reinsurance Operations and Deputy Chairman of XL Catlin, respectively, at the 2015 National Insurance Conference of Canada (NICC), that models are needed in the industry but should in no way be solely relied on.

“It’s revolutionised the business. I mean the $69 billion of alternative capital doesn’t come in play if there isn’t a model,” says Greg Hendrick, underlining the reality that for now the flood of third-party capital is focused largely on the easier-to-model business lines.

But also bringing to light the fact that without the models the capacity wouldn’t be in the industry at all, something that will be essential when the market figures out how to better cover emerging, new, large-scale risks and catastrophe events in underinsured and vulnerable regions of the planet.

“I completely agree though,” continues Hendrick, “you cannot use it (models) as a black box; it’s not meant to be that, but it is a necessary tool for the business.”

The reality is that all financial markets suffer from model uncertainties, so the fact that they are wrong sometimes isn’t too surprising. Indeed risk models are seen by many to be directional, an additional source of insight to assist with the underwriting of risks, not a holy grail.

But without them the insurance, reinsurance and insurance-linked securities (ILS) market wouldn’t be as developed and capable as it is today, that’s certain, the development of ever more advanced risk models helps to make insurance and reinsurance capacity more available than perhaps ever before.

However, explains Stephen Catlin, “I think if we go into a situation where we are driven simply by models then we close ourselves down,” referring to the need for common sense and a focus on grass root principles to assist the models.

He continues; “I’m pro people thinking about risk holistically, and thinking about how frequent it might occur, thinking about how you price it and how you limit it as a starting point, and then model it to see whether or not the model is in agreement with what you’re saying. And if it doesn’t agree, then try and work out where the differences are.

“The day that we forget about relationships, the day we forget about grass root principles, the day we forget about common sense, is the day we should go home.”

A viewpoint somewhat echoed by Tom Bolt; “So you need the common sense, you need the underwriting judgement to temper the models, and think about what is it we just haven’t thought of, what is it that can knock us out of the box sideways.”

The message from the panellists is clear; models are essential for the insurance and reinsurance industry but shouldn’t be the only point of decision-making. Industry and underwriting expertise, knowledge and common sense must come into the equation to try and fill the gaps that the models miss or simply cannot see.

“If you compare the modelling to the world I grew up with in property cat, the world’s come a long way in terms of giving a better description of the risks that are exposed. It’s how you use the models and how you think about them in terms of what can happen to you, I think that makes a big difference,” concludes Bolt.