COMMENT: Cat modellers’ credibility issue

COMMENT: Cat modellers’ credibility issue

The US wind model update from Risk Management Solutions (RMS) was on the minds of many in attendance at this year’s annual meeting of the Property Casualty Insurers Association of America (PCI). It seems the US market is still digesting the new RMS model, and insurers are concerned about its impact.

Patrick Hartigan, Beazley’s head of treaty reinsurance, told us at the meeting in New Orleans in October that cedants were concerned by the impact of RMS version 11 ”primarily because the prognosis seems to be that they need to buy more cover”.

Hartigan also believes there has been a lack of leadership from cat modellers themselves. Worsening the problem has been what he describes as a black box approach in which insurers are not able to understand the inner working of the cat models that they are using.

“The problem is I don’t think the cat modelling companies are providing a lead on this and they’re not explaining,” he says.
“I think there is a credibility problem,” argues Hartigan, “and it does make us question over reliance on one model.”

The solution, he therefore suggested, is to use multiple models, rather than a solitary tool, to avoid an insurer or reinsurer becoming too dependent on their latest software. “ are good because they are a prospective gauge, but they’re not meant to be used in isolation,” he said.

Understandably, RMS’s rivals were keen to separate themselves from the industry disquiet. Rather, they backed up the view that the industry should not rely so much on just one model.

“I think the hottest topic for this year end will be how the models are blended, aside perhaps from potential cat losses,” said Paul Little, senior vice-president at cat modeller Eqecat. “Brokers and service providers are busy developing techniques for how to blend models, and how to calibrate them in a manner that they think appropriately affects their portfolio.”

In contrast to RMS, Eqecat makes frequent model updates. It has recently released volume 3.16 of its North American hurricane model. “Overall our loss results for most US regions have gone down slightly – by 10% in some areas. That is based on our market portfolio. It can vary, depending on the concentration level. It’s very much a contrast to RMS,” said Little.

Others reported that the effect of version 11 will not be fully felt by the market until the January 1 2012 renewals. Greg Coda, president of client management in the reinsurance division at Munich Re America, said the RMS model change along with tornado losses earlier this year will put upwards pressure on rates. RMS version 11 was released in March this year, too close to the mid-year renewals to be fully reflected in pricing.

“We saw some rate increases at mid-year. We will continue to see that at the 1/1 renewal with a lot more understanding of the RMS impact than at June/July,” Coda told us.

Roughly two-thirds of insurers are still to integrate the latest catastrophe model changes into their pricing assumptions, according to Hannover Re’s chief underwriting officer for North America, Hans Rohlf.

“Currently we think about 35% of our clients have gone through the process,” Rohlf told Reactions in New Orleans.

Rohlf cited a lack of pressure from rating agencies as the deciding factor, with the exception of insurance specialists AM Best, and says that many will not be sufficiently prepared in time for the renewals period in January.

“We’ve seen some people buying additional layers...but January is not the only renewals period,” he said.

The US market is also still digesting the heavy tornado losses this year, which have forced insurers to look for more reinsurance support. While the tornado losses add up to a large loss, they were too small individually to hit most reinsurance programmes.

“The tornado and hail losses were largely held net so there was not too much impact on reinsurers,” Thomas Luning, managing director for client markets at Swiss Re America, told us. “What we are seeing in response to that is companies taking remedial action – particularly in the Midwest and largely in personal lines.”

In some cases this is taking the form of quota shares. But Luning reports that aggregate deals are also being considered.

“The other thing they are doing is looking to reinsurers for new and innovative products, specifically aggregate products,” says Luning. “It is a great opportunity for the primary insurers and the reinsurers to partner more than they have in the past.”
The issue of how the tornadoes will affect reinsurance demand was also touched on during a reinsurance CEO panel at the PCI meeting.

Tony Kuczinski, CEO of Munich Re America, said on the panel that he agreed that aggregate tornado cover made sense but warned it is not easy to do. “We have offered aggregate cover over the past few years; the problem is it is very expensive,” he said. “The biggest risk is we have is to factor in the one time buyer – where someone comes in, blows out the reinsurance and walks. But it would be a sustainable product in our view.”

Tad Montross, chairman and CEO of General Re, said that the aggregate cover product is a tricky one because it is so expensive that the buyer either feels like he made a great purchase if the policy is triggered or that they wasted a lot of money if it does not.
“You have a 35 rate online or whatever it is so it is a difficult product to deal with from an emotional standpoint. It is a real winner/loser type of product,” he said.

Phil Bowie, executive vice-president at Willis Re, said on the same panel that another fallout from the heavy tornado activity this year may also be development of new models to cover the risk. “We will see development of new tornado models, no doubt,” he said.

Latest Issue

October 2018


In this month's Reactions

  • Monte Carlo roundup
  • National Flood Insurance Program
  • Munich Re roundtable
  • Liberty Mutual roundtable
  • US Cannabis industry
  • Kidnapping & Ransom cover
  • Marsh & McLennan acquiring JLT



Follow Us on Twitter @reactionsnet

Catastrophe Centre

Catastrophe Centre