EXPOSURE: Assigning a Return Period to 2017

This is a taster of an article published in the latest edition of EXPOSURE magazine. For the full article click here or visit the EXPOSURE website.

Hurricanes Harvey, Irma and Maria (HIM) tore through the Caribbean and U.S. in 2017, resulting in insured losses over US$80 billion. Twelve years after Hurricanes Katrina, Rita and Wilma (KRW), EXPOSURE asks if the (re)insurance industry was better prepared for its next “terrible trio” and what lessons can be learned.

According to Aon Benfield, HIM caused total losses over US$220 billion and insured losses over US$80 billion — huge sums in the context of global catastrophe losses for the year of US$344 billion and insured losses of US$134 billion. Overall, weather-related catastrophe losses exceeded 0.4 percent of global GDP in 2017 (based on data from Aon Benfield, Munich Re and the World Bank), the second highest figure since 1990. In that period, only 2005 saw a higher relative catastrophe loss at around 0.5 percent of GDP.

But, it seems, (re)insurers were much better prepared to absorb major losses this time around. Much has changed in the 12 years since Hurricane Katrina breached the levees in New Orleans. Catastrophe modeling as a profession has evolved into exposure management, models and underlying data have improved and there is a much greater appreciation of model uncertainty and assumptions, explains Alan Godfrey, head of exposure management at Asta.

“Even post-2005 people would still see an event occurring, go to the models and pull out a single event ID … then tell all and sundry this is what we’re going to lose. And that’s an enormous misinterpretation of how the models are supposed to be used. In 2017, people demonstrated a much greater maturity and used the models to advise their own loss estimates, and not the other way around.”

It also helped that the industry was extremely well-capitalized moving into 2017. According to Aon Benfield, global reinsurance capacity grew by 80 percent between 1990 and 2017 (to US$605 billion), against global GDP growth of around 24 percent. The influx of capacity from the capital markets into U.S. property catastrophe reinsurance has also brought about change and innovation, offering new instruments such as catastrophe bonds for transferring extreme risks.

Is 2017 the New Normal?

One question being asked in the aftermath of the 2017 hurricane season is: What is the return period for a loss year like 2017? RMS estimates that, in terms of U.S. and Caribbean industry insured wind, storm surge and flood losses, the 2017 hurricane season corresponds to a return period between 15 and 30 years.

However, losses on the scale of 2017 occur more frequently when considering global perils. Adjusted for inflation, it is seven years since the industry paid out a similar level of catastrophe claims — US$110 billion on the Tohoku earthquake and tsunami, Thai floods and New Zealand earthquake in 2011. Six years prior to that, KRW cost the industry in excess of US$75 billion (well over US$100 billion in today’s money).

So, does this mean that a US$100 billion-plus (or equivalent in inflation-adjusted terms) loss year like 2017 is now a one-in-six-year event? As wealth and insurance penetration grows in developing parts of the world, will we begin to see more loss years like 2011, where catastrophe claims are not necessarily driven by the U.S. or Japan peak zones?

“Increased insurance penetration does mean that on the whole losses will increase, but hopefully this is proportional to the premiums and capital that we are getting in,” says Asta’s Godfray. “The important thing is understanding correlations and how diversification actually works and making sure that is applied within business models.

“In the past, people were able to get away with focusing on the world in a relatively binary fashion,” he continues. “The more people move toward diversified books of business, which is excellent for efficient use of capital, the more important it becomes to understand the correlations between different regions.”

And while a warming climate is expected to have significant implications for the level of losses arising from storms and other severe weather events, it is not yet clear exactly how this will manifest, according to Tom Sabbatelli, senior product manager at RMS. “We know the waters have risen several centimeters in the last couple of decades and we can use catastrophe models to quantify what sort of impact that has on coastal flooding, but it’s also unclear what that necessarily means for tropical cyclone strength.

“The oceans may be warming, but there’s still an ongoing debate about how that translates into cyclone intensity, and that’s been going on for a long time,” he continues. “The reason for that is we just don’t know until we have the benefit of hindsight. We haven’t had a number of major hurricanes in the last few years, so does that mean that the current climate is quiet in the Atlantic? Is 2017 an anomaly or are we going back to more regular severe activity? It’s not until you’re ten or 20 years down the line and you look back that you know for sure.”

For the full article click here or visit the EXPOSURE website for free access to over 40 articles.