Reinsurers’ Almanac: Your 2006 Forecast

In mid-August, I presented our firm’s preliminary 2006 budget numbers at an operating committee meeting. Barring the unexpected, we would see moderate rate reductions of between 5% and 7% in our property treaty portfolio January 1, 2006. Auto casualty treaty business, priced at or above technically adequate levels, would likely be renewed at expiring rates. All indicators suggested an uninspiring renewal season, much like what we had seen the previous year – a calm and relatively issue-free process.

Three days later, on August 19, the industry incurred the second largest cat loss – currently at $500 million – in Canadian history. Ten days later, Katrina pummeled the gulf coast, generating insured losses that will inevitably rise above $50 billion. When you factor in $10 billion to $15 billion or more for Rita and Wilma, it becomes clear the 2006 renewal season in Canada or elsewhere will not be calm. We have witnessed an industry changing event.

If I look back to the days following Sept. 11, 2001, I recall a local reluctance to accept that the destruction of the World Trade Center would influence Canadian treaty rates. Many argued that since the loss did not take place in Canada – even though Canadian insureds incurred considerable losses – there should be no impact on Canadian rates.

We know of reinsurance brokers who told their clients that Canadian treaties would be immune and that they could confidently secure “as is” renewals for 2002. No doubt, these same brokers will do so again this year and I imagine there are buyers in this market who will listen.

But Katrina is a market changing event and reinsurance is a global business with no Canadian-owned players. Canadians will undoubtedly feel the impact of this loss. We only need to look at some of the history of this business and at the financial fundamentals.

Hurricane Andrew hit Florida in September of 1992 and produced an insured loss of $22 billion (in today’s dollars). Global cat rates more than doubled on Jan. 1, 1993. In the local market – which at that time was operating with underwriting autonomy that has long since vanished — Canadian cat rates moved up 50% despite the fact that Andrew produced no Canadian losses.

After the World Trade Center attacks, Canadian catastrophe rates increased by more than 40% despite brave predictions and assertions to the contrary from brokers out to make a quick sale. To suggest at the time that a loss of $34 billion, which instantly extinguished 12% of the reinsurance market’s capital and a corresponding amount in industry share value, would not affect local prices was quite simply absurd.

Like all large reinsurance losses, Katrina has its unique properties. The industry loss may indeed be twice that of the 9/11 attacks, but it’s not yet clear how much of the payout will come from reinsurers. As Katrina was forming, global reinsurance capital was at an all time high of about $400 billion – almost 60% greater than its value on Sept. 11, 2001. Furthermore, the reinsurance industry has been enjoying favourable market conditions, with combined ratios 10-15 points better than those in the years leading up to 2001. So the market is clearly in a better position to absorb Katrina.

Working against the buyer is the sheer and shocking recognition that a city the size of New Orleans could actually be shut down. Sure, we’ve seen all the models and the worst-case scenarios, but these have all seemed so theoretical. The graphic evidence of so many media photographs brings home the reality of urban destruction in a way that not even the most sophisticated modeling demonstration could. The virtual eradication of a major urban centre ceases to be hypothetical. And that inevitably weighs heavily in the psyche of the treaty underwriter.

All of this leads us to a discussion about models and the modeling firms. How much stock should an insurer take in the output of the models? With or without the complication of flood, the science clearly displayed its limitations during Katrina (and also during Ivan in 2004, when storm surge had not been adequately modeled).

On Sunday, Aug. 28, while CNN’s Anderson Cooper did his theatrical utmost to convince us of the impending doom, the three major modeling firms were at opposite ends of the spectrum as to how much damage Katrina would wrought in the ensuing hours. Eight hours before Katrina made landfall, the models ranged from a low of $3 billion to a high of $46 billion. The range in and of itself speaks to the complexity of what can only be viewed as an inexact science.

The range of estimates as a storm approaches is but one poor measurement of the efficacy of the models. Indeed, there are so many variables as a hurricane approaches that a broad range of estimates is hardly surprising. Of greater concern is the variability of output from the wind and quake models that are generated annually. There is virtually no consensus among models. Based on our experience (our firm is the largest user of catastrophe models in the world), it is the rule rather than the exception to see 250-year or 500-year Canada quake estimates that vary by multiples of two or more.

So, today’s buyer of reinsurance is confronted with a difficult decision in terms of limit purchased when a 250-year event is projected at, say, $150 million by one model and $325 million by another. This is not to say that the models lack value, or that the professionals who are developing these models are culpable. It is only to say that this is a new and extremely complex science; it becomes more and more credible with each catastrophic experience. But we are far away from pointing to an exact number with confidence.

Katrina emphasizes another reality – one that gives some credence to the argument that Canada and other less cat-exposed regions should be at least somewhat differentiated from the United States. If we look at a history of global cat losses over $5 billion (below), a clear trend emerges.

LARGEST INSURED CATASTROPHE LOSSES

Eighty-six per cent, or about $140 billion, of the value of these global cat losses is derived from losses that occurred in the United States. With its wealth and increasing population densities (particularly in coastal locations), America is a dramatically exposed region.

So while catastrophe underwriters are contemplating an overall change in the manner in which catastrophe reinsurance is priced, one would expect that a re-weighting of certain geographies will be uppermost in their minds. If Katrina is a market changing event – and it is – then the market should probably change in varying degrees according to the sources of loss. Canada will inevitably see an increase in cat rates, but any adjustment should be relative to much larger increases in what are clearly recognized as the catastrophe-prone regions of the United States. In short, regional cross subsidies, as they relate to catastrophe business, cannot be continuously unilateral.

As we approach the 2006 renewal season, there are other burning issues in the reinsurance market. Specifically, I’d like to discuss reinsurance security and the spectre of uncollectibles, and also the state of casualty treaty pricing.

Reinsurance security is, of course, related to Katrina. Massive catastrophic losses generally trigger the demise of the least credit worthy reinsurers. The losses in 2001 resulted, in one form or another, in the departure of such notables as Centre Re, CNA Re, Gerling Re, Hart Re and others. Downgrades after the loss were frequent. We are already seeing the wounds of Katrina and I would suspect that some reinsurers will not be able to raise sufficient new capital to participate in the next round of renewals. Unfortunately, some run-off reinsurers have been slow to discharge their claim responsibilities. Vigilance, therefore, in choosing the best security going forward, will continue to be required.

Beyond this – and this is something so self-evident that it is a wonder that some buyers miss it – is the fact that re
insurance brokers are handling millions upon millions of premium and claims dollars without facing the same scrutiny as the reinsurers whose money they hold. Reinsurance intermediaries in Canada are not regulated. The big houses like Aon (Aon Re) and MMC (Guy Carpenter) with large, publicly traded parents with billions of dollars in shareholder equity are fully disclosed. Many of the others are not.

Error and Omission exposures alone could be sufficient to threaten the finances of a number of the Tier 2 players and the unregulated and privately held brokers. Frozen broker assets at the time of a large catastrophe could be enough to endanger an insurer’s solvency.

I’ll close with a very brief discussion of casualty pricing and reserving. I don’t think anyone would debate that large loss case reserves at the insurer level are at their highest level in history. Accident year loss ratios are such that there is more than sufficient financial latitude to build case reserves to appropriate levels. Reinsurance actuaries, however, appear slow to recognize this reality; their pricing tends to reflect dated loss development patterns. For example, it is not valid to assume that a large 2004 loss reserve will develop at the same rate as a reserve that was posted by the same insurance company for a loss in 1998. Yet reinsurers regularly apply these historical loss development factors to current case reserves. The resultant treaty pricing is moving retentions away from loss activity, such that reinsurers will ultimately regret that their portfolio of casualty business has been diminished to a volatile collection of unbalanced clash covers.