EQECAT’s senior vice-president discusses risk modelling in a region exposed to every natural catastrophe going

Q: Asia is a huge region – what are some of the hazards it’s exposed to?

A: It is exposed to pretty much all the natural catastrophes we have in the world: earthquakes, tsunamis and on the weather side, tropical cyclones with some winter storm and tornado hazard. Overall, flood is a very significant peril for the region.

Q: Tell us about EQECAT’s new typhoon model and what it encompasses.

A: This July, we released a brand new typhoon model covering the entire East Asian region. It’s covering eight territories that are subject to Western North Pacific typhoons, including Japan, South Korea, China, Taiwan, Hong Kong, the Philippines, Thailand and Malaysia. For that entire region, the model has also been built to accommodate expansion to additional areas, such as Vietnam and Guam.

What’s new about the model is that it has all eight models on one stochastic event set that has 150,000 typhoons, representing the full range of possible typhoons in Asia. That’s important for companies that are concerned with risk across the region. There are a lot of examples where there are significant overlaps of risk – such as Taiwan and China, or the Philippines and China, or Korea and Japan – and that is captured in the event set of a commercially available model for the first time.

We have also spent a lot of time developing some new flood components. Traditionally, models have focused heavily on the wind side so with this model, in addition to the wind, we have two flood components. This includes storm surge – where the typhoon wind field drives very high surge heights at the coast and can cause damage in low-lying areas – and then there is rainfall-induced flooding. With this robust new model, EQECAT helps clients set rational expectations about risk.

Q: China has a long historical record of events such as earthquakes – is this helpful from a cat modelling perspective?

A: Such information is very helpful, particularly on the hazard side. We have this information in China and Japan; in these countries, it probably goes further back than many other countries in the world, as there have been societies recording such things there for many centuries. But as a caveat, the built environment changes very rapidly, so if you look at a place like China, what was there 20 or 30 years ago is already very different and that means damageability is very different.

Q: Could China one day become a major peak zone?

A: I think it’s only a matter of time.

The economic environment is pretty much already there, so it’s just a question of how quickly the insurance market develops. But there’s no question it will be a peak zone because, from the typhoon, flooding and earthquake side, there are enormous levels of risk in China.

Q: Catastrophe risk evaluator CRESTA recently increased its resolution for China – how significant is that?

A: We accommodate the input of exposure data and reporting of results at all these levels in the model, so it’s just a question of how fast the market evolves to the point where they’re capturing significant amounts of data at these levels. I think we’ll see that, but there is still a lot of data coming in at a province level. It’s also a question of how well defined the categories of data are.

Q: How can we assess some of the major loss scenarios for Asia?

A: We have the capacity to run events in our models on a modelled historical basis, which for older events is often very instructive. In Japan, the 1959 typhoon is probably the most significant event that’s occurred there in the last few decades. Obviously, the built environment is very different now than it was in 1959, so running it through a cat model is instructive in a lot of ways to assess what would be the impact if it was to occur today.

Q: How is climate change likely to affect exposures in the future?

A: We’re looking at numerical modelling to get a handle on event frequencies. It’s not so much that we’re going to see severities that are way outside the range we’ve seen in the past. It’s more a question of how cyclicality in the climate and long-term climate change is likely to affect frequencies of events and where they occur.

We’ve made some inroads in some of our weather peril models already and, while we haven’t yet directly applied that in Asia, there’s certainly the capacity to do so. One thing that’s very clear with respect to Asia typhoons is cyclicality in levels of activity – high and low periods that last a decade or two is certainly something we’ve seen historically. GR