There has been much discussion regarding the state of the marine and energy markets following last year's hurricanes. Nick Bonnar explains how many previously held assumptions have changed.
While the devastation wrought by the 2005 hurricane season has been the driver for the upward movement in pricing and for the restructuring of the marine and energy markets, there is still a feeling that the sector has let one opportunity slip though its fingers.
Hurricane Ivan was the only loss of the 2004 season which affected the offshore energy and marine market, but it highlighted many of the issues that were to impact the market a year later. At the time there was an assumption that there was an automatic loss correlation differential between the onshore property and offshore marine and energy markets. Ivan marginally changed that assumption but the events of last year have simply reinforced the fact that the two exposure sets can no longer be viewed separately by underwriters, management, regulators and rating agencies.
Post Ivan understanding
Ivan did not have a significant impact on premium levels but it did raise questions over the underwriters' approach to business interruption (BI) and loss of production income (LoPI) covers. Whereas prior to Ivan the indemnity basis for loss of production was the dollar cost of a barrel of oil and contingent business interruption (CBI) triggering events that could occur anywhere in the supply chain, post-Ivan the long march to coverage changes began and became a sprint after Katrina and Rita. LoPI indemnities reduced to reflect better the cost of loss of production and CBI coverage required named assets in the supply chain to be specifically identified and priced for as triggering loss assets. Additional changes included aggregate sub limits for named Gulf of Mexico windstorms, higher deductibles and greatly reduced capacity.
Rates for Gulf of Mexico risks have seen significant premium increases in the region of 200% to 400% and for some loss-affected risks the increases have been in the region of 600%. For non-Gulf of Mexico risks, rates have been rising by around 20% to 25%. Additionally, with a desire to avoid renewal during the North American hurricane season many clients have brought forward their inception dates. They have sought to renew earlier amid fears of the difficulty of renewing cover if there is another active hurricane season.
For reinsurers, they have sought to differentiate between the Gulf of Mexico and the rest of the sector with the result that Gulf of Mexico risks are now generally shown in a separate pillar of a cedant's programme. The resulting transparency in pricing can be seen as a positive step for the market.
It is, however, how the market has approached the use of aggregate accumulation control where I believe it has missed a clear opportunity. There are various off-the-shelf software packages that can be used to monitor offshore energy exposures but none currently combine the accumulation part with a stochastic model to give credible results, principally due to a lack of tested historic data for what are new frontier assets. As a result, return period talk is premature so the widely-used deterministic approach is the Lloyd's market Gulf of Mexico realistic disaster scenario (RDS), which prescribes a single windstorm path, $10bn of loss, affected licence blocks and probable maximum loss (PML) percentages. This PML had never been tested and then along came Katrina and Rita, which on current estimates created market losses in the region of $9bn for Katrina and $5bn for Rita. Almost without exception these two losses each blew-out the RDS calculated amounts. The fact that Katrina took a different path to that prescribed in the Lloyd's RDS and delivered losses far in excess of those from the RDS should raise questions over the use of this single path estimated PML by underwriters and management.
The losses should have given the market a clear opportunity to address the issue of natural peril accumulation control and modelling and prompted a wider debate on the way models are created and utilised. A shift to a whole Gulf of Mexico PML assessment, and not a single path event PML, should have followed.
The underwriter, whether insurer or reinsurer, who uses the RDS as an assumption for a peak loss or to drive a pricing curve is forewarned of the failings in the use of this matrix, as it is not a realistic benchmark for what it is being used for. There have been specific examples where people have used this single path knowledge to their advantage as a way to circumnavigate the RDS' restrictions - a fool's paradise.
If this year's season ends with no major storm losses impacting the market then the year will be used to start to rebuild the losses of 2005. However, insurers and reinsurers have to take a more holistic approach to how they are managing their exposures within the Gulf of Mexico and other natural catastrophe prone areas.
- Nick Bonnar is head of speciality at Aspen Re.