Modelling the next asbestos

The Emerging Risks interview, with Praedicat President and CEO Bob Reville.

“When we first started, which was ten years ago, we called ourselves a casualty catastrophe modelling company. The idea was that we were going to do for the next asbestos what the property cat modellers had done in property insurance. We developed some methods for mining the peer-reviewed scientific literature in order to find chemicals or products or business activities that scientists were concerned might cause bodily injury or damage. The idea was that in litigation, which is what it comes down to, you have to have scientific evidence for these things, so if you can catch the science at the earliest possible stage, when it’s too early to support litigation, you’d be able to have a good early warning vehicle and something comprehensive.

“We developed these initial algorithms and went to market and said: look, we’re building a casualty catastrophe model, and everybody said what is that? I don’t want that, but I really do want your emerging risk information. So we said: ok, we’re a liability emerging risks analytics company!

“Now, ten years in, we’ve kind of come full circle and I think the market now understands what the concept of casualty catastrophe modelling is.” 

What products would you say you are most proud of developing over the past decade? That you regard as innovative?

“Without a doubt, our probabilistic loss model. We are the only liability probabilistic loss model in the market. It was an immense technical achievement for a modelling team, and has come in multiple releases. But just the concept from the very beginning is something I’m incredibly proud of: tracking the science, scoring the science, and then projecting the distribution of potential future outcomes of the science, and using that to create a probabilistic distribution of potential litigation initiation, and then modelling out the size and scope of the resulting litigation. Then, taking that, and translating that into potential losses for 120 000 companies… that was an immense technical achievement and an immense conceptual achievement!

“At the other extreme, I’m also proud of developing our emerging risk framework. At some point we decided that we would characterise our risks into three phases of emergence: emerging interest; emerging damage; and emerging litigation.

“Emerging interest is the point at which scientists are first beginning to write about it, but before we recommend that insurers need to take action. 

“Emerging damage is the phase where the science is mature enough that it could result in litigation initiating within the next seven years, and so therefore we recommend that underwriters take underwriting actions and that insurers manage their total aggregation. 

“Emerging litigation is where the litigation has actually launched, where the science has been matched up with plaintiffs and the plaintiff attorneys have lined up funding and are going after the defendants.

“That three-phase framework is incredibly simple and at the same time, when we introduced it to a lot of a clients that said they loved it and the simple way of breaking it down, and have adopted it for their own internal emerging risk framework.” 

 Why is casualty modelling seen as more difficult than property modelling?

For one thing, whenever there is a major liability event, companies go out of business, products are removed from the market, and insurance forms are changed. The risks that emerged in the past are risks that will never emerge again, so there is this emerging risk identification piece that is required to be able to do this kind of modelling. I think that nothing had happened in the past until the technology associated with trying to mine text and make sense of it at scale was available, and that basically started about ten years ago when we started. We deployed that technology: natural language processing, machine learning and the like. 

“The other key complexity is that, unlike a property cat, which emerges over hours or days, liability cat emerges over decades, it builds up over decades, and it plays out over decades. That is a modelling challenge, and a challenge because you don’t really have the data. You can’t even say what the outcome of asbestos is yet!” 

From your perspective, what are the risks that are most of concern for the casualty market at the moment?

“There are two ways to answer that question. One is from the litigation funding perspective. Litigation finance is something you hear about, but it is truly forming the liability risk environment, and today is a market of $20-30 billion dollars, with the largest part of it investment in mass torts. The return on that investment, according to Swiss Re, is about 26%. It’s high return is uncorrelated with markets, and is expected to continue to grow. That is an enormous amount of capital available to plaintiffs to pursue complex, large-scale litigation.

“From 2011-15, no mass torts were initiated, and people were saying it was a problem of the 1990s and questioned why we were even bothering building models for this? Starting in about 2015, with opioids, talc, pesticides litigation and others, we are in an environment where these new litigations are starting at the pace of two to three a year. That’s quite a stunning change in frequency, and it has real implications, because of the fact that mass torts are involving a longer range of policy years, the latency in any given portfolio in likely to be increasing. What that means is that, unless reserves are adjusted accordingly, there is a risk of widespread reserve inadequacy as these reserves continue to develop, exposed to these mass torts. There is also greater risk of aggregation, so overall that is and should be the number one concern.

“The number one risk we would say at the moment is the PFAS chemicals, that’s per- and polyfluorinated alkyl substances. The originals are PFOA which you would know as Teflon, and PFOS which you would know as Scotchgard. Today, they have come to understand that these chemicals endure in the environment forever, so they are known as the forever chemicals. And not only that, but these forever chemicals, which are used for water-proofing and stain-proofing, among many other things, clearly cause bodily injury and they are in our water supply and in all of our bodies. That is resulting in an expanding amount of litigation: in the past few years there has been litigation against the large producers Dupont and 3M, associated with their contamination of local water in particular places in the process of manufacturing, for example. And more recently there has been the firefighting foam litigation, associated with the run-off of PFAS chemicals in the local water supply. That litigation has expanded into the firefighting gear that is worn, which also has PFAS chemicals in it. 

“Given the incredibly large industrial footprint, and given the potential for serious bodily injury and serious environmental contamination, to us it looks more like the next asbestos than anything we’ve seen.”