The devastation wracked by Hurricane Katrina has set a new precedent Nigel Allen looks at the potential impact on this year's balance sheets and what it means for the future of catastrophe modelling.

Is it wrong to be awed by the devastation brought by Katrina to New Orleans? The world has unfortunately grown accustomed to the images of the utter carnage left in the wake of hurricanes, but little could prepare people for the images of a virtually submerged New Orleans. The only comparison that can be drawn is with that of the almost unearthly destruction wrought by the Indian Ocean Tsunami.

Over a month after Hurricane Katrina hit and we have only now received confirmation of lives lost as a result of this tragedy. Despite initial reports, putting the death toll at 10,000, the final number is 1,209. But it is no surprise that there is such a massive gap between the inital figure and the final toll as officials were severely hampered by the fact that they were unable to gain access to these regions devastated by the storms.

Any attempts to estimate overall economic losses from the event and in turn the insured losses are similarly scuppered by the inability of loss adjusters to gain access to the region. However, despite numerous insured loss estimates having been issued, most of the larger reinsurers seem to be basing preliminary losses on a total insured loss of approximately $45bn, with most expecting Rita to only add marginally to this figure.

Down but not out

There was a degree of inevitability about the decision by Standard & Poor's to revise the outlook on the reinsurance industry to negative. The outlook revision said Simon Marshall, a credit analyst at Standard & Poor's, reflects a number of concerns which the rating agency has about the overall financial strength of the reinsurance sector and also the degree of uncertainty which surrounds the preliminary company insured loss estimates from Katrina, which is further compounded by the losses from Rita.The revised outlook also factored in concerns over overall operating performance in 2005, the degree of financial flexibility of those seeking to access the capital markets, the availability and affordability of retrocession, the potential for further catastrophe losses coupled with the increased frequency of such events and the difficulties in attempting to model them.

With expectations that rating downgrades will be "modest", despite the negative outlook, the axe has already started to swing with three companies, Olympus Re, PXRE and OIL all being struck with downgrades in the aftermath of Katrina and Rita. Olympus Re was the first to fall, seeing its financial strength rating cut from "A-" to "B+" by AM Best. Since then both PXRE and OIL have undergone a similar rating slash. OIL saw its Standard & Poor's financial strength rating tumble two notches to "A-" from "A+" following the announcement that the company planned to up its reserves by $1bn at the end of the third quarter to tackle expected Katrina claims, while PXRE was downgraded one notch by both AM Best and Standard & Poor's to "A-". While both agencies highlighted PXRE's $450m capital raising initiative, neither believed this would satisfactorily dampen the capital volatility in its monoline property catastrophe, with the company putting its preliminary estimate of Katrina losses at between $235m-$300m with a net loss from Rita expected to range from $30m-$40m. Jeffrey Radke, president and CEO of PXRE, commenting on the downgrades, said that while he was disappointed by the decisions, he believed that with the successful completion of the company's capital raising plan, "we expect to enter the key January 1, 2006 renewal period with our strongest balance sheet ever."

Simon Marshall added that the number of downgrades might rise if loss estimates for Katrina proved inadequate or the industry is hit by another major catastrophe. However, he added that the outlook may return to stable "if the near-term strains on financial strength are addressed and if the uncertainties surrounding Katrina are resolved," and highlighted a strong January renewals season as a potential trigger for a stable outlook.

Capital rush

While for the majority of companies affected the financial impact of these two storms will be captured in the earnings safety net, for some insured losses will burst through this barrier and burrow into the capital below. The rush to the capital markets to plug these predicted balance sheet holes has already begun in earnest, and the signs are already promising that there is sufficient capital set to enter the market, evidenced by the speed at which Montpelier Re was able to raise $600m, a fact which resulted in Standard & Poor's removing the company from CreditWatch. "This is significant for other companies," explains Thomas Upton, a senior credit analyst at Standard & Poor's, "in that there are ample sources of new capital willing to come into the industry." But how great an appetite can investors be expected to have in a sector that has had to withstand seven of the largest catastrophe events ever experienced in the last two years?

Ready capital could well stave off downgrades for a number of reinsurers, but in order to ensure that this inflow of capital does not dry up there are a number of factors which must come into play. The first factor is one the market is powerless to control, that there be no further catastrophic losses this year. With weeks still remaining before the "official" end of the hurricane season and sea temperatures in the Atlantic basin still above normal levels, the market can do little but cross its fingers. Second, capital inflow is dependent upon a dramatic hardening in rates, and not just on those lines directly affected. Marshall believes that, on a loss of this scale, "there will undoubtedly be rate hikes in many lines of business in the US and perhaps elsewhere also."

The combination of Katrina and Rita will provide clear opportunities for reinsurers to avail of the inevitable hardening in market conditions. However, those companies that see their balance sheets riddled by storm losses will almost certainly struggle to exploit these post-Katrina market conditions if funds raised are funnelled only into refloating their balance sheets, while the more financially buoyant reinsurers can exploit new capital to increase their capacity to write more business.

Expectations are that new capital entering the market will flow predominantly into the established market players, rather than into new start-ups. However Simon Marshall adds "We believe that there will be some new start-ups this time, but fewer than after 9/11." However, for those wishing to exploit the low barriers to entry offered by Bermuda as has been the case in the past, this may prove difficult as there is virtually no space left on the island to actually establish a new operation.

Model behaviour

Katrina and Rita have exposed the fallibility of the catastrophe models. Initial insured loss estimates from the three risk modellers, EQECAT, RMS and AIR varied dramatically, with the individual ranges of such forecasts showing levels of uncertainty that many practitioners assumed such models were designed to remove. Hemant Shah, president and CEO of RMS, admits that the severity of the flood damage from Katrina was unexpected. "It was not expected that there would be multiple failures of the levee system that resulted in the whole city filling with water," he confirms. "When the storm hit, we and many others, were tracking it as a terrible storm event with some storm surge along the Mississippi coast, but by Tuesday it became clear that there was another story unfolding and that was the flooding of the city itself."

At the time of the interview, some four weeks after the event, RMS was maintaining its loss estimate of $40bn-$60bn, the $20bn gap a reflection, said Shah, of the extraordinary nature of the Great New Orleans flood. "In many ways this flood is unprecedented and difficult to quantify in that most flood losses are not 'ponding losses', where water rises up like in a bath and sits there. The kind of damage that this causes is fairly uncertain."

Damian Magarelli, an analyst at Standard & Poor's believes that Katrina has served as a reminder to the reinsurance industry of how inexperienced the market is in dealing with such large scale losses and a clear indicator of the limitations inherent in the catastrophe models. Magarelli is convinced that the degree to which catastrophe reinsurers rely upon their models to evaluate rate levels will be a key differentiator in adequate pricing. "Those reinsurers which have placed a heavy reliance upon catastrophe models are likely to have underpriced catastrophe risks, while those who have used models as one output of many and have added sufficient margin and underwriting fundamentals are less likely to have done so."

Shah takes the view that while there are certain things that will emerge from Katrina, which will needed to be reflected in the models, people need to acknowledge that the catastrophe model is only a single weapon in the reinsurer's arsenal. "They are powerful tools for risk assessment, but they are models and you need to be able to understand the models and contextualise them and make decisions appropriately." What will come from this event he believes is a greater degree of sophistication in the way that reinsurers use their models.

Simon Marshall's take on how the market might use its catastrophe models going forward does not appear to reflect this hope for greater sophistication. "We may see that reinsurers take the output of these models and then add a surcharge to reflect the uncertainty of an event like this ... Reinsurers may take the upper limit of the model range and potentially add more in order to achieve adequate pricing levels for the uncertainty."

But is there not a danger that if reinsurers cannot remove sufficient levels of uncertainty from the pricing environment for property catastrophe risks in regions such as the Mexican Gulf Coast that such cover will become either prohibitively expensive or the risk itself will no longer be deemed an acceptable one? Addressing the Houston Marine Insurance Seminar, Andrew Beazley, chief executive officer of Beazley, said, "It may be the case that, this year and last, a combination of high sea temperatures in the tropical North Atlantic and faster than normal trade winds has made hurricanes more likely. And it is unquestionably the case that property concentrations and values in areas exposed to these hurricanes have risen and continue to rise as more and more people seek to live in coastal regions. But neither of these developments makes the risk impossible to quantify and therefore insure. They just make it more costly to insure and the size of exposures difficult for the domestic and world markets to digest."

However, upping the cost of cover is not the only method that can be employed to mitigate the potential impact of natural catastrophes on a company's balance sheet in these catastrophe prone regions. After the Florida hurricanes in 2004 the area became a "special case" due to the increased level of hurricane activity in the region, with some insurers establishing separate subsidiaries for the region in order to ring-fence such risks from the parent's broader risk base. This option is now being considered by a number of companies for all of the states along the Gulf Coast. Polina Chernyal, an analyst at Standard & Poor's, believes that this is a worrying trend. She believes these subsidiaries may only receive limited financial backing from their parent in order to reduce their overall capital exposure but by doing so they are bringing about lower rated capacity in areas prone to catastrophes.

The industry has clearly come a long way since the terrible events of Hurricane Andrew in 1992 and 9/11 in terms of its ability to assess risk. Increased sophistication in the ability to estimate overall exposures has created an industry which, as is evidenced by the apparent ability of the reinsurance industry to bear the brunt of the largest loss in the market's history on their earnings, is much more resilient than it has been in recent years. However, Katrina and Rita have clearly shown that the reinsurance market still has much to learn. "The important lessons to be learned," says Hemant Shah, "are not the reactive ones after the event occurs, but the proactive ones - to think about how risk needs to be managed going forward. More thoughtful discussions will take place over the next year as we work with hundreds of insurers and reinsurers to bring the lessons learned from Katrina into the modelling process and the risk management process."

Nigel Allen is editor of Global Reinsurance.