“It’s surprising anyone can get any sleep,” says Bill Keogh, president of Eqecat, a global catastrophe risk modeling provider and consultancy. “There’s a lot out there that we don’t know about.”
Keogh is referring to the March 11 Tohoku earthquake and tsunami that devastated Japan. Aside from its magnitude, the catastrophe was notable for its “gray swan”-like quality. Scientists knew a quake in that area was likely, explains Miaki Ishii, associate professor of earth and planetary science at Harvard University’s Harvard Seismology Group. But they didn’t think the entire Tohoku fault, which was believed to be broken into pieces, “could move all at once.” Yet it did.
Despite the gap between what was expected in Japan and what actually happened, the availability of previously inconceivable amounts of computing horsepower is enabling insurers to develop catastrophe models that can narrow the divide between what humans can imagine and what nature can do.
“Our [insurance] clients use multiple servers and run billions of calculations, depending on the size of their portfolio,” says Keogh. “Imagine you have a million records in a portfolio and you simulate 10,000 possible events. That’s 10 trillion possibilities. And for each policy and location you have to estimate the loss to the deductible and/or the loss to reinsurance. That’s a lot of number-crunching.”
Before insurers start their computers crunching, they license a catastrophe model from a provider such as Eqecat, Risk Management Solutions (RMS), or AIR Worldwide (the Big Three of global catastrophe modelers). The insurers put into the model each and every property insured in a specific geography, along with details pertaining to potential loss of life, business loss, and so on. The model then runs simulations of what could possibly happen to all those policies in thousands of disaster scenarios.
Constructing these simulations requires an enormous amount of computational juice. For example, AIR currently is modeling 10,000 years of weather to build a probabilistic model for flooding in Germany. “At best, we have 100 years of historical data,” says Jayanta Guin, AIR’s senior vice president of research and modeling. “That’s not enough.” So instead of simply using the data from those 100 years to make a model, AIR ran 10,000 possible permutations of that historic data to provide what Guin calls “the full universe of possible outcomes.” To do that, AIR ran its program nonstop for six months on its own computing grid. It would take 18 months of run time to do a similar simulation for the United States, estimates Guin. (Not surprisingly, in view of those resource-intensive run times, the company is exploring the possibility of leveraging the computing power of independent clouds.)
The catastrophe modeling industry is about 25 years old, says Guin, and back then “a gigabyte [of data storage] was unthinkable. Now we’re talking terabytes. This allows us to analyze risk in much more detail. Back in the ’90s, we analyzed risk at a county level. Now every single house is quoted individually.”
“Based on satellite data and GPS systems, we know the building stock at each locale,” says RMS
co-founder and chief risk officer Robert Muir-Wood. Using data collected from on-the-ground motion sensors and transmitted through a wireless network, models estimate damage based on the intensity and character of ground movement and translate that into cost.
But even with the increasing comprehensiveness of catastrophe models, the ability to quantify loss and risk is still constrained by lack of data and the fundamental unpredictability of nature.
The most difficult losses to model, AIR’s Guin says, are from what the insurance industry calls contingent business interruption: the impact of a catastrophe on a company’s downstream suppliers and partners. “That can be modeled,” he says, “but it requires input data that doesn’t exist. The physical losses due to ground shaking and the tsunami are easy to estimate, but those $20 billion to $30 billion [insured] loss estimates [of the Tohoku disaster] do not include business interruption.”
The reason the data needed to model business-interruption losses doesn’t exist is simple, says Guin: “More data means more costs.” It’s expensive for companies to collect data about all their suppliers and partners — the location of their factories, the nature of their building stock, and so on — and about those suppliers’ and partners’ factories and supply chains, especially in less-developed corners of the world. Many companies are just not willing to make that investment.
But RMS’s Muir-Wood believes it’s critical for companies to map their suppliers’ risks as well as their own. “Then,” he says, “I’d like to understand if I’m too dependent upon a supplier with exposure to catastrophe. Then I’d diversify. If I couldn’t, I’d specify insurance to cover damage at the supplier.” Does a company in Michigan really need earthquake insurance? If it depends on a supplier in an earthquake zone, it might. Especially as the uncertainty of where a catastrophe might occur is magnified by the uncertainty of when.
“Catastrophes,” says Eqecat president Keogh, “don’t happen in an orderly fashion. It would be nice if we could say that every six months we’ll have this or that catastrophe, but it doesn’t work that way.
“What is happening is that the world economy has grown, and some regions have grown a lot,” says Keogh. “The hazards aren’t changing, but there’s more stuff in harm’s way, particularly in Asia. Therefore, there’s increased risk and loss to the insurance industry.”
According to a recent study conducted by global reinsurer SwissRe, earthquakes, storms, floods, and explosions (remember Deepwater Horizon?) resulted in economic losses of about $218 billion in 2010 and cost insurers $43 billion, an increase in insured loss of more than 60% over 2009. Some of that rise can be attributed to economic development in affected areas.
And if you think 2010 was bad, get ready for 2011.
Already in the first quarter there have been severe winter storms in the United States; a magnitude 6.3 earthquake in New Zealand; cyclone Yasi in Australia (with an estimated insured loss of $5 billion); and, of course, the earthquake and tsunami in Japan, where the insured loss has been variously estimated so far to be between $12 billion and $34 billion. And, as Guin points out, “The U.S. hurricane season is still ahead of us.”
Meanwhile, off the northwest coast of the United States stretching from Vancouver Island to Northern California, the Cascadian subduction zone (where one tectonic plate is plunging beneath another, deforming the earth’s crust and storing scary amounts of potential energy) looks a lot like Tohoku. Scientists, says Harvard’s Ishii, believe the inevitable release of energy — that is, the earthquake — could well be as large as 9.0 in magnitude.
That quake could happen tonight; it might not happen for hundreds of years. But it will happen. In the meantime, Keogh suggests that his insurance clients “stress-test their portfolios to account for things not in their models. For example, one of the unanticipated consequences of the Tohoku event was a nuclear plant with its electrical backup at a relatively low elevation knocked out by a larger-than-anticipated tsunami. That’s not something you’re going to pick up in a catastrophe risk model.
“Chief risk officers, to paraphrase Donald Rumsfeld, need to think about the known knowns (what happened before), the known unknowns (possible events that could occur), and unknown unknowns (everything else),” says Keogh.
And that’s why global CROs, CFOs, and insurance company executives are not getting much sleep these days.