Eye of the Storm

Can good truly come from bad?

Consider the good that came from an event as horrific as Hurricane Katrina, destruction so severe and draining – physically, financially and emotionally – that now, as then, it has come to represent more than a single event.

But Katrina, racking up insured losses of in the range of US$45 billion, is no orphan; it has a handful of sisters and brothers over the course of the 2004 and 2005 Atlantic Hurricane seasons who, both individually and together, wreaked devastation and helped to set in motion or cement a number of important changes.

Alone, Katrina, Rita, Wilma, Charley and Frances each ranks among the 10 most costly hurricanes in the history of the United States; together, they form a lesson that disturbingly, graphically and publicly came to light.

How does one describe the legacy of devastation? Fortunately, that may be where the good comes in.

The United States and its Atlantic hurricane seasons may continue to be a main measuring consideration for insured catastrophe losses, but something has shifted. And that shift seems to reflects the global nature of the insurance and reinsurance industries, which have grown more open to applying lessons learned to other geographies, other perils and other situations.

Indeed, the story of Katrina (not yet ended) may be an optimistic tale of recognition of weakness, willingness to take action, advances in catastrophe modelling, improved efforts around collaboration and the ever-present talk of resiliency that now moves beyond specific cities, regions and countries.

But the most lasting and significant lesson to wash ashore as the damage has receded, the biggest benefit of things gone so bad, is an understanding that good enough, covered enough or safe enough should not be part of the vocabulary of losses, insured or otherwise. There is no place for complacency; only the need to adjust and adapt.

THE AFTERMATH

“In the broadest terms, Katrina taught the industry to be prepared for the worst,” suggests David Huebel, chief executive officer, XL Catlin – Toronto. When the winds died down, it was anticipated wind damage would not be as severe as expected, Huebel says. “No one anticipated the levees breaking and the flooding that ensured,” he adds.

“Arguably, Katrina was the first ‘supercat’ event of the 21st century, meaning that Katrina’s winds and storm surge also triggered a second, man-made disaster that dramatically increased insured losses from the event, in this case the overtopped and breached levees that caused the New Orleans flood,” says Mark Bove, senior research meteorologist at Munich Reinsurance America.

“I think about it in my mind and I replay what I saw on TV, and you see things like people standing on their roofs trying to get rescued. You see families with doors floating their children and their belongings that they had left to rescue. And you recognize resilient cities are so important, as well as being able to respond to those types of events,” says Monica Ningen, Swiss Re’s head of property underwriting for the U.S. and Canada.

Following Katrina were several more significant hurricanes in the North Atlantic, including Rita and Wilma, says James McCarney, senior vice president of Willis Re Canada. “In total, the insured damage from these three hurricanes was US$65 billion, including damage in the Caribbean,” McCarney reports.

“Katrina forced insurance companies to look at the ways they respond to large-scale disasters and make improvements,” notes Patrick Lundy, president and chief executive officer of Zurich Canada. Citing his company specifically, he notes the global resource model used in Katrina “is now well-established and has been used in other natural disasters since then, such as the 2010 earthquake in Chile and flooding in the United Kingdom in 2013-2014,” Lundy says.

“Seeing the level of destruction that happened in Katrina certainly changed many perspectives, our perception of risk. It’s one thing to think about it in mathematical terms versus it’s a very different thing when you start seeing it in front of your eyes,” Jayanta Guin, executive vice president of AIR Worldwide, comments.

Rohan Dixon, executive vice president and chief broking officer at Aon Risk Solutions in Toronto, notes that Katrina, without a doubt, had a material impact on the insurance industry.

There was impact “around Cat modelling, on specific risks rather than on a portfolio basis, on the preparedness of clients and understanding where their risks are, on underwriters understanding of the policy wordings,” Dixon says.

Katrina’s key influences on the insurance industry “were a greater focus on monitoring and controlling exposure accumulations in surge-prone areas, the importance of building codes and building code enforcement to reduce damage, the inclusion of different views of hurricane risk with catastrophe models, and highlighting the large loss potentials of supply chain risk and contingent business interruption,” adds Bove.

“The debate within the industry of the proper treatment of storm surge within policy coverage and perils began with Hurricane Katrina,” comments Michael J. Hudson, a managing director for Marsh in Los Angeles.

Rising sea levels are driving up expected economic and insurance losses from hurricane-driven storm surge in coastal cities across the U.S., states an analysis released in August by catastrophe risk management firm RMS. “Hurricane Katrina was the first time in decades that more than 50% of losses from a hurricane resulted from storm surge,” Robert Muir-Wood, RMS’s chief research officer, says in a statement. “Katrina can be seen as a milestone in the long-term shift from ‘wind’ to ‘water’ as the main driver of hurricane loss,” Muir-Wood says.

Wind continues to be an issue, of course, but one that has seen advancements since Katrina. Research by the Insurance Institute for Business & Home Safety (IBHS), Katrina 10 Years Later: Improving the Resilience of Roofing in the Gulf States, “shows that stronger building codes and standards, along with more stringent requirements for inspections, building permits and contractor licensing, all have contributed to safer, stronger roofs.”

Margareta Wahlström, head of the United Nations’ Office for Disaster Risk Reduction, notes in a statement marking Katrina’s 10th anniversary that the true legacy of the U.S.’s costliest hurricane disaster was to raise the bar for disaster risk management worldwide. “Hurricane Katrina exposed weaknesses in disaster risk management which are common to many hazard-prone locations around the world,” Wahlström argues.

There also has been was plenty of focus around the unified management of event response, Ningen says. “And I say unified management because that really translates to whatever sort of structure you’re dealing with in a specific country or in a specific county,” she suggests.

Katrina further “spurred the creation of more robust tools to manage natural catastrophe risk,” Bove says, including multiple views of risk within Cat models and increased demand for spatial analytics of a portfolio of risks.

Guin would likely agree. “Katrina was a big landmark event in our industry, in the insurance industry, broadly speaking, but more so from a catastrophe modelling point of view.”

MODEL APPROACH

“This was the advent of today’s rigor around catastrophe modelling and modelling’s universal acceptance and use,” Hudson comments.

“Katrina revealed some weaknesses in the state of practice in the insurance industry,” Guin says, including the quality of exposure data. “Companies were using models, but oftentimes, the data that was going into the models was not the quality that it expected. Companies would find themselves surprised that they had underwritten risks which were, oftentimes, very vulnerable,” he says.

“I think Katrina further consolidated the recognition that extreme risk cannot be managed or quantified without the application of science. Past historical data is simply not enough to deal with such risk,” Guin argues. “From a catastrophe modelling point of view, I think it gave further impetus to us, as modellers, to improve our science,” he says.

“Essentially, Katrina and the other hurricanes of 2004 and 2005 forced modellers to allow for different views of risk within their models, allowing insurance companies to conduct more robust sensitivity testing of models and adjust model output to their own view of risk,” Bove explains. “This ‘democratization’ of Cat modelling is one of the biggest impacts on the Cat management community after Katrina,” he contends.

“Insurers used to use Cat modelling more for managing aggregates to purchase reinsurance. Today, Cat modelling is much more of an everyday underwriting tool to help better understand individual risks,” Huebel suggests.

Katrina fuelled the two most significant innovations since the models were first developed – open loss modelling platforms and the new characteristic event (CE) method for monitoring exposure accumulations – reports Karen Clark, chief executive officer of Karen Clark & Company. With an open model, Clark says, model assumptions are visible and accessible to the model users, allowing insurers to “truly ‘own’ the risk.”

But another important lesson learned from Katrina was that numbers derived from the EP (exceedance probability) curves “are not the right metrics for monitoring exposure accumulations,” she maintains. Why? The EP curves do not answer the following questions:

• How much business can I write in specific geographical areas?

• Where can I have surprise losses?

• Where can I have outsized losses relative to my competitors?

“These are questions most frequently asked by the CEOs and Boards of Directors, and the traditional models cannot answer them effectively,” Clark argues. “A new approach is needed to identify and manage exposure concentrations,” she points out.

Improvements have been made for other perils as well. Catastrophic losses drove the push for greater granularity in flood models, which are now getting down to 10 by 10 metres squared, Dixon points out. “The level of granularity that comes out of that is phenomenal, because now you can [see] this location is different from the building next door, and here’s the reasons why.”

Guin notes that, until very recently, flood was deemed to be an uninsurable risk by many thinkers in the insurance industry. However, with “the availability of data, the advances in science, the availability of higher computing power, I think that has all changed,” he says.

“We are regularly modelling flood, the tools are available, the data is getting better and better every day,” he adds.

But is it possible to rely too much on models? “Catastrophe models are always wrong. Yet, they are one of the best tools available to us to assess the loss potentials of rare, extreme loss events,” Bove says. “When used correctly and with knowledge of the model’s strengths and weaknesses, they can be invaluable in the underwriting process. But models should only inform, not replace, the underwriting process.”

Models are tools, an attempt to incorporate science, data and experience all into a solution, says Guin. “They are not going to solve every problem. The use for models is increasing, and that’s for a good reason. But at the same time, model users need to be aware of what the model is trying to do, what are its strengths, what are its weaknesses.”

One negative impact around modelling may be insurer perception that models – which some considered to be “broken” before Katrina – are now “fixed,” Clark says. “One event does not fundamentally change the scientific understanding of hurricanes,” she says.

“Because more data is being collected, there can be a temptation to put too much weight on the new data, thereby over-calibrating a model to the most recent event,” she says. “Over-calibrating a model to the most recent event makes the model a backward-looking tool,” she adds.

“If you put junk data into a model, you are going to get a junk number back out,” Ningen says plainly. “At the end of the day, if you have a strong underwriting entity, they know how to use the models appropriately. I think where companies go wrong is if they don’t have a strong underwriting culture and they’re driven by the numbers that come out of the model without remembering that there’s so many assumptions that go into that,” she adds.

“Once again, storms like these illustrated the fallibility of the models,” says McCarney. “Of particular note was the susceptibility and vulnerability of commercial risks to damage from these storms. The models had really under-estimated this damageability,” he adds.

“The oil refining, forestry, agriculture and tourism industries were severely impacted for many months following Hurricane Katrina and, in some cases, more than two years,” he reports.

Most agree risk management has benefited greatly by changes in the aftermath of Katrina. One potential threat to that progress, however, is if things start becoming a transaction, Dixon suggests. “That takes away from risk management around Cat events being at the forefront of a corporation’s mind around how they protect their assets to falling within treasury or procurement, and it being more of a transaction,” he says.

OTHER INFLUENCES

Catastrophe models were hardly the only areas influenced by Katrina and its destructive cohort. “People weren’t talking much about Big Data in 2005 like they are today. Katrina exposed a problem in what data was available: inconsistent quality or incompleteness,” Huebel says.

“Today, everybody is aware we live in an era of Big Data, and analytics can help inform our decision-making,” he says.

Clark agrees. “Katrina had a positive influence on exposure data quality in the U.S. Katrina revealed several issues about the data insurers were putting into the models – namely, erroneous coding by construction and occupancy, and undervalued building replacement costs.”

The insurance industry also witnessed changes around new capital. “Katrina brought with it the ‘Class of 2005’, providing an inflow of new capital,” Huebel says. Many insurers set up sidecars, and “Cat bonds and other insurance-linked securities also gained traction and continue to build in popularity today.”

Hudson points out that “the hard market following Katrina spawned the first significant use of sidecars, giving non-traditional investors a means of quick entry into, and exit from, the world of property insurance.”

McCarney reports that “new and well-documented reinsurance capital and capacity flowed into the market after the 2005 losses, similar to what happened following Hurricane Andrew and the attacks on September 11, 2001.” The use of Cat bonds to augment traditional catastrophe reinsurance capacity increased significantly since 2005, he points out.

“In terms of the Canadian market, buyers of catastrophe reinsurance experienced a gentle firming in rates, but no significant tightening,” McCarney says. “The U.S. and Caribbean Cat rates went up significantly, as did the retentions for most cedants, even ones not impacted by the hurricanes.”

Dixon points out there is now more capacity in the insurance market and in the reinsurance market than ever before, and it grows every year.

“So what would cause it to change?” he asks. Answering his own question, he says it has been posited it would need to be an event of $200 billion or more.

What would do that? “If someone attacked a major cloud or a major hosting site, which the banks use, large corporations use, etc., and there are estimates that that’s in the hundreds of billions of dollars of economic losses very quickly. That would be potentially something that would change the market,” he says.

“I don’t see the market at any time soon, in the next five years, or potentially ever, turning to these hard and soft cycles,” Dixon says. “I think what you’ll find is the more that we use data and the more that we try and innovate, you will find it becomes very granular on where we see hardening within the market.”

Underwriting organizations today are trying to drive underwriting results at a product and industry level, “because they don’t have as much buffer with their investment returns as they used to,” Dixon says.

“Now, the underwriting organizations drill it right down, by portfolio, by location, and have a much more granular view of their performance, and adjust accordingly. I think that’s what’s changing in the cycle.”

BORDER CROSSING

Bove says that Katrina clearly showed the need for better understanding of supply chain risk. “Katrina’s lessons can be applied to other regions, whether domestically or internationally, to help other countries improve the resilience of their built environment, as well as to address important life safety concerns.”

For Canada, though, Dixon suggests Katrina’s influence took some time to travel northward. “It wasn’t really until 2013, with the Calgary floods, did we really start to see insurance companies, clients and underwriters and brokers start to look at Superstorm Sandy, Katrina and the Calgary floods and start to pull all those together,” he suggests.

Now, though, what large clients are “thinking about as they go and either build or acquire assets, is they do look at the Cat impact,” he says. “They look at their aggregation, they model that to specific sites,” he reports.

“People will take lessons and try to extrapolate,” suggests Guin. “I think the world is now connected enough that it causes people to think about other risks they are taking and what lessons can we learn from events that are happening in other parts of the world.”

Learning however one can is key when trying to tackle the persistent issue of the gap between insured and non-insured losses. “Many business owners did not purchase the right type and amount of coverage, including business interruption (business income) insurance. Many were never able to open their businesses again,” Huebel says of Katrina.

The protection gap is an important topic in the U.S., Ningen agrees, “but I think that resonates to a lot of other geographic areas as well. The question is often asked: ‘Can we afford to take steps to mitigate against these types of events?’ And after Katrina, the question really became, ‘Can we afford not to?'”

Huebel notes that like most major catastrophe events, Katrina caused many companies to examine their risk aggregates. “Policy language at the time generally did not contain sub-limits on high-hazard flood areas. Policies do now, though, and there is much less ambiguity on the subject of wind versus water damage,” he suggests.

That said, “there is still a massive gap between the economic losses and insured losses. And I think there’s a real opportunity for insurance companies to step into that space,” Dixon maintains.

INNOVATION NEEDED

A number of sources say ensuring that things are covered – and covered more successfully than in past – demands innovation and recognizing opportunities.

“Hopefully, innovation on the modelling front will encourage innovation on the product front,” Clark says. “Storm surge risk can certainly be quantified by traditional and capital market participants who are willing to invest in understanding the risk utilizing advanced tools versus waiting for a black box model to give an ‘answer’,” she contends.

“I haven’t seen a real change in innovation within the insurance industry. I think that’s a shame,” Dixon comments. “The insurance industry is very, very slow to respond and there are clients that are pushing the industry to change. It should be the other way around. We have more capacity and more capital in the marketplace than we have ever seen before,” he says, adding insurers must think outside of the traditional box.

“Currently, alternative capital is providing a lot of additional protection for the peak peril regions, such as U.S. hurricane and earthquake,” Clark notes. “There are fewer offerings for second- and third-tier peril regions, but this is where a lot of future growth potential lies,” she suggests.

“We need to start thinking about innovation in how do we help our clients achieve that so the insurers get that information and can understand the risk in an appropriate manner. The changes, the challenges, that have come out of Katrina and other Cat events is this sudden requirement for it,” Dixon says.

STRENGTH IN RESILIENCE

Innovation will only help to enhance industry strength that has been demonstrated by Katrina and other events. “For many insurers, Katrina was more of an earnings event than a capital event,” Huebel says, “and that showed the strength of the industry as a whole.”

Guin suggests the insurance industry did navigate through Katrina in a very robust way, pointing out very few companies were in financial trouble, unlike after Hurricane Andrew, when several companies basically went insolvent. “That, in our view, spoke to the resilience in the insurance industry in terms of managing these extreme risks.”

That sort of strength and resilience will likely be key should another such disaster occur. But what if it did?

Noting that the insured losses from Katrina were approximately US$45 billion – excluding some losses related to offshore casinos – “if the storm were to happen again today, we estimate, excluding the platforms, it will actually still cost US$45 billion of loss in that scenario where the city doesn’t get flooded. If the city gets flooded the way it happened in 2005, then that insured loss would be closer to US$60 billion,” Guin reports.

However, if “the levee systems failed in addition to all of the damage it caused elsewhere, and flooded the entire City of New Orleans, that particular subcomponent of the overall aftermath would probably be mitigated with the raised levee system,” he says.

“Even after all we’ve learned from Katrina, losses to another major catastrophe could still be great,” Huebel says. “I’d attribute that mainly to continued growth along our coasts. We continue to populate our coasts,” he says.

Although catastrophe models have improved, challenges remain, Dominic Casserley, chief executive officer at Willis Group Holdings, says in a recent press release from A.M. Best.

“The risks keep rising because of the constant building along the coastline of the United States,” Casserley said in an episode of A.M.BestTV. “Recognizing the potential fragility of our infrastructure, there is over US$10.5 trillion of exposures along the coastline, so the risks are still there.”

Cities and regions, however, would do well to match the resiliency of the insurance industry. “The need to build resilient communities will continue to be critically important over the next decade, given the ongoing potential for significant losses,” Lundy emphasizes.

“More people are moving to high-risk coastal areas, so the value and concentration of assets at risk is growing. Recent estimates of insured property values for coastal exposures in the U.S. alone have been calculated at US$16 trillion,” he says.

“All over the world, I think there is a tremendous amount of focus on improving resilience,” Guin offers. “The insurance industry is taking note of it and there is a lot of discussion out there of how resilience has to be improved, and from a financial point of view, how much of the capital that is out there in the industry today could be deployed to actually strengthen the resilience of major cities, of major countries.”

The recently released IBHS roofing research notes that “while hurricanes will continue to threaten coastal areas year after year, there is no reason to allow homes and businesses to remain vulnerable to damage from even relatively low-level storms.”

The report recommends breaking the “cycle of destruction,” by using the scientific knowledge currently available to “build and retrofit safer, stronger homes that are more hurricane-resistant.”

A new white paper released to mark Katrina’s 10th anniversary – issued by Zurich North America and the Wharton School of the University of Pennsylvania’s Wharton Risk Center – also suggests more needs to be done to create resilient communities to guard against risk in the face of extreme weather.

“As the past decade has shown, we have entered a new era of catastrophes,” Erwann Michel-Kerjan, executive director of Wharton’s Risk Management and Decision Processes Center, says in a statement. “Either we pretend they will not happen to us, or we join forces to make resilience a national priority.”

Regardless of locale or peril, that may be good advice for all.