By GRAHAM BUCK who covers European risk management issues.
MONTE CARLO, Monaco -- Delegates at this September's annual Reinsurance Rendez-Vous in Monte Carlo had a particularly costly year to look back on. The first half of 2011 was the costliest ever experienced for the reinsurance sector, according to Swiss Re's sigma report.
Earthquakes and aftershocks in New Zealand, Japan's quake and tsunami, as well as extensive floods in both the United States and Australia left the industry with a total bill estimated at around $70 billion.
Insured earthquake losses to date this year, at $30 billion, are already the highest ever. New Zealand property damage and business interruption alone account for as much as $12 billion out of the $30 billion.
However, that New Zealand figure appears to be equaled or exceeded by the severe weather that afflicted the Midwest and southern United States during April and May. Inevitably, this year's main debate at the Rendez-Vous focused on the consequences for the industry.
Titled "Japan, Australia, New Zeal--Implications for the Insurance Market. Time to Change the Way we Look at Risks?" much of the discussion focused on the role of risk modeling.
Despite the technological advances that have produced increasingly sophisticated models, risk models ability to evaluate the losses that extend beyond the direct and most evident damage remains limited, reinsurance executives said.
The catastrophe that has highlighted these limitations has been the Japanese quake and tsunami of March 11 and the resulting disruption suffered by various industry sectors. The shortage of brake pads for Formula One motor racing is one example. Supply chains providing components from key Japanese manufacturers have been fractured, resulting in heavy contingent business interruption losses.
Reinsurers appear to be waking up to the massive changes that have transformed businesses in recent years, as specialization and outsourcing has led to just-in-time delivery of components across several continents.
Torsten Jeworreck, a member of the management board of Munich Re, said reinsurers hadn't fully realized that over the past 10 to 20 years the original contingent business interruption product has changed due to globalization.
The industry "needs to change its product," he said, but customers also need to explain just how contingent business interruption has changed, and how supply change management has evolved in response. "We have to understand these new structures in our customers' organizations," he said.
However, in a subsequent interview the chief executive of Lloyd's of London begged to differ. Richard Ward said that he disagreed with suggestions that reinsurers had not kept abreast of developments in supply chains.
He cited a conversation three years ago with a German manufacturer, who identified around 1,000 single-source suppliers for his company.
Ward invited the manufacturer to outline these dependencies to underwriters, and said that businesses are now looking more where potential single points of failure exist and whether they want their insurance to cover key suppliers.
Ward added that although risk models have their limitations, they had proved an accurate guide for estimates such as losses from Japanese earthquake. "Events never match the models and we are not modeling reality, so they need to be continually revised,'' he said.
"What happens in real life is usually very different to the models' scenarios, but they are nonetheless useful in gauging what our potential exposures are," he said.
While risk models may be constantly improving in predicting direct damage, their ability to assess contingent business interruption exposures is more restricted.
Another panelist, Aon Benfield Chairman Grahame Chilton, said that the exposure is particularly hard to calculate, and that the unexpectedly large business interruption losses since March demonstrated that ultimately "a model is still only a model."
To emphasize that fact, the real world continues to produce surprises that challenge many prevailing assumptions. Another example of the unexpected turned up in the flooding in Australia last December and January, which extended so far across the state of Queensland that it exceeded the combined area of France and Germany.
Hemant Shah, chief executive of Risk Management Solutions (RMS) said risk modelers had been "working furiously" over the past four years to assimilate the lessons of the hurricane losses of 2004-05, and to update their models accordingly.
So-called "surprise" events regularly challenged the assumptions regarding risk, and exposed the gaps, he said. The string of worldwide catastrophes over the past 18 months offered new lessons, and the challenge lies in deciding which of them should be incorporated when revising the models.
The debate also welcomed guest speaker Gerry Brownlee, the New Zealand government's minister for economic development, who more recently has taken on the specific post of minister for earthquake recovery.
He recently joined a government delegation whose itinerary in September included visiting Lloyd's underwriters in London, and delivering a presentation in Monte Carlo.
Brownlee, who met with Munich Re and Swiss Re, estimated that losses from two major earthquakes and aftershocks in and around the city of Christchurch since September 2010 amount to nearly $24.5 billion, of which around $9.9 billion is covered by the state-backed Earthquake Commission.
"The impact has been catastrophic in certain areas; for example, we've lost around 1,400 commercial buildings," he said.
To put the figure in perspective, U.S. hurricane losses from Katrina were equal to around 1 percent of national gross domestic product, while Japan's earthquake and tsunami in March cost around 4 percent of the country's GDP.
Economic loss from the Christchurch earthquake equates to around 8 percent of New Zealand's GDP.
He said the government wanted the reinsurance sector to continue to regard New Zealand as a favorable long-term prospect. "We could ask the question 'Why don't we take over all of the risk ourselves?' but we remain confident of the market's ability to deliver."
Asked whether the industry had responded well to the disaster, Brownlee described the response from insurers as "extremely good" although the speed of payments from reinsurers to cedants had been less impressive overall.
Although ministers are keen for the industry to take over a greater degree of participation from the state, he admitted that many New Zealanders "want things to be fixed immediately" and believe that earthquake cover should be government-led. "Even the right-wingers are looking to a socialist solution," he said.
Brownlee also directed criticism at risk-modeling agencies already designing new models for New Zealand quake. "Some of them evidently haven't seen first-hand the impact that the quake had on Christchurch, where the government is paying for information to be collated."
Nor had they partnered with the government-owned Crown Research Institutes that compile data on natural disasters.
This year's Rendez-Vous offered a combination of brickbats and bouquets for risk models and their accuracy. While their limitations are recognized, the industry also recognized their value, due in part to the developing science.
Offerings from risk-modeling agencies Risk Management Solutions, EQECAT and AIR Worldwide face more competition from broking groups such as Aon Benfield, which was promoting its own products such as ImpactOnDemand.
RMS's recently-released version 11.0 of its U.S. hurricane model has already proved its value by predicting the path and extent of Hurricane Irene, said the group's chief research officer, Robert Muir-Wood.
The revised estimates for losses have presented many cedants with the options of purchasing more coverage, boosting their capital reserves or cutting back on their exposure.
Jean-Philippe Thierry, a former board member of German insurer Allianz who now serves as president of the Rendez-Vous, warned the industry that catastrophe risk models should be handled with care. "Recent crises have highlighted the fragility and dangers of models when they are used blindly," he said.
His view was supported by Dan Fortuit, head of commercial nonlife business at Allianz IARD and secretary general for the Rendez-Vous, who said that while the models had grown in sophistication they are no more than an approximation of reality.
"Often after big catastrophes you realize that the model is not correct," he said. "The problem is that some users think the models are the truth. But they are a tool, not the truth."
October 15, 2011
Copyright 2011© LRP Publications