The Intelligent Quarterly from the publishers of The Insurance Insider

Spring 2012

Search archive

Advanced search

Winston Churchill didn't quite speak these words, but perhaps he would have penned them and delivered this speech if he was a contemporary international property catastrophe underwriter.

For 2011 will not only be remembered as the year of international catastrophes, but as the year when the post-Katrina diversification paradigm was challenged.

Perhaps there is no other better demonstration of this than New Zealand, home to three significant earthquake events within the past 12 months that have caused excruciatingly expensive reinsurance bills for international underwriters (not to mention immense human tragedy).

The two events in September 2010 and February 2011 alone amassed total insured losses of more than $20mn, and total reinsured losses of at least $13.7bn, according to Aon Benfield estimates. The June aftershocks produced an insured loss estimate from Eqecat of $3bn-$5bn, although this was subsequently disputed by the New Zealand government.

Click to enlarge Against these losses, Aon Benfield points out that total premium from New Zealand and Australia combined was $680mn in 2010, while insured losses for New Zealand as a percentage of GDP amounted to a whopping 17.9 percent for the Darfield/Lyttelton quake.

Furthermore, rates-on-line for this premium averaged just 4 percent in the two countries - 2.5 percent below the global average and far below the 10 percent average paid by US insureds.

Learning the wrong lesson
It is perhaps ironic that some of those worst hit by Q1 losses from the New Zealand, Australia and Japan Tohoku events, were those who embraced the lessons of the new prudence - the disciples of international diversification following the US storm clustering events in 2005 such as Flagstone Re, PartnerRe, Platinum and RenaissanceRe.

In 2005, the Gulf Coast of the US was battered by three devastating windstorms (Katrina, Wilma and Rita) that ratcheted up insured losses never seen before and which have not even been matched this year, notwithstanding the pending loss notifications from Thailand (see graph page 7).

In the aftermath of these losses, a new consensus emerged that heavy exposure to the peak zone US risks was bad, while geographical diversification was the silver-bullet answer.

And here is the real irony. Perhaps the lesson of 2005 was not that the concentration of risks in the US was the real problem, but that events can cluster in ways not thought of before.

Andrew Carrier, president of Argo Re, says the surprise aspect of 2011 was the earthquakes clustering in an unprecedented manner that challenged the norms of the reinsurance market.

"If 2004/05 taught us about the potential clustering of windstorms, I personally didn't anticipate a cluster of earthquakes in the same spot in New Zealand in such a short space of time," he said at The Insurance Insider roundtable in Monte Carlo this September.

"And when you look at the cat reinsurance product, which has this automatic reinstatement, I just wonder to what extent the industry is addressing the nature of the product when the second event potentially becomes more likely with the first event having occurred."

One example of this flawed approach is thought to be the reinsurance for Civic Assurance, a local government insurer. Sources suggest the insurer received payouts from its reinsurers for all three quakes having paid minimal reinstatement costs. The programme, understood to be Aon Benfield-brokered and dominated by Swiss Re, Munich Re and Axis, has subsequently been axed. The insurer has withdrawn coverage for earthquakes risks in the area, having been unable to obtain coverage.

Unsurprisingly, New Zealand's earthquake minister Gerry Brownlee was quickly on the charm offensive both in the London market and in Monte Carlo in September.

Click to enlarge Speaking to The Insurance Insider at the time, Brownlee urged against a kneejerk reaction to the sequence of costly quakes. "Everyone is being cautious because we are just starting to go into what we hope will be a more seismically settled period, so I wouldn't rush to any particular statements about the appropriateness of [reinsurance] pricing or anything else," he said.

However, a recent scientific paper on giant earthquake clusters by Paul Thenhaus, Dr Kenneth Campbell and Dr Mahmoud Khater for Eqecat seemed to support Carrier's assertion that cover for earthquakes might need to be rethought. The paper said there was "mounting empirical evidence" that giant earthquakes (near or above 9 magnitude) occurred in clusters.

"The statistics of great and giant earthquake occurrences indicate that the historical temporal clustering of these earthquakes on a global scale cannot be attributed to chance," it said.

The paper highlighted a trend of clustering on available data around 15-year timeframes, with the last sequence of mega-quakes occurring between 1950 and 1965.

Since the end of that period, the first quake to break the 9 magnitude barrier was the December 2004 Sumatra quake, which triggered the tragic Christmas day tsunami. This has been swiftly followed by the 8.8 magnitude quake in Chile in 2010 and the 9 magnitude quake in Japan on 11 March this year.

"If the current cycle follows the one that occurred in the 1950-1965 timeframe, we may be only about half way through the cycle, and the largest earthquake in the current cluster may not have yet occurred," the report warned.

Market reacts
In response to the heavy loss burden, cat reinsurance underwriters are looking at international policies not just for their diversification credits, but also for their potential aggregate loss exposures.

Click to enlarge Speaking to The Insurance Insider in July Kevin Allchorne, head of reinsurance at Amlin, explained the logic for pulling back from some international property cat excess of loss cover with a small premium base that does not justify the severity of the potential loss.

"A fair portion of our losses in 2010/11 were from pretty low rate-on-line layers, in some cases less than 1 or 2 percent rate-on-line where significant events had not happened for many years and modelling suggested a remote risk," he said. And, despite the increase in premiums heralded by some players in the space in subsequent renewals, Allchorne noted: "A 50 percent [increase] on not a lot, is still not a lot."

Speaking on the Axis Q2 earnings call, CEO John Charman said that underwriters had "fooled themselves" into thinking that rate increases alone solved the problem of international exposures. "It does not address the issue at all. There are some structural issues within the programmes, especially in Australia and New Zealand, relating to retentions of the individual cedants, relating to free reinstatements, which have not been properly priced into the original contracts," he said.

"Knowing what we now know, we believe rate change on international catastrophe-exposed business in the second quarter, and at the 1 July renewals, was still not at a level to make us feel fully comfortable utilising our risk appetite for those zones," he said.

All eyes on IAG
Another significant feature of the losses from New Zealand has been the constant deterioration of reserving announced by (re)insurers, as is typically the case with earthquake losses.

Part of the problem has been generated by the difficulty in establishing which event triggered losses when buildings were damaged by some events and then collapsed during others.

Out of the mega reinsurance programmes to have been triggered in the region, Suncorp, QBE and the giant New Zealand Earthquake Commission have all seen significant loss creep. Meanwhile, hosts of smaller (re)insurers have also raised their loss estimates (see graphics).

Yet the focus has now turned to IAG, as its A$3.8bn excess of A$250mn catastrophe reinsurance programme renews on 1 January. IAG has continued to stand by the ground-up A$1.9bn loss estimate it made in May for the second New Zealand earthquake in February, although sources have suggested the market considers IAG's current estimate to be unrealistic. It remains to be seen who will be proved right when the claims finally filter through.