From an underwriting perspective 2004 and 2005 have been a strong test for our underwriting philosophy with hurricane events testing the Group's ability to withstand both severe and frequent catastrophe events. The benefits of withstanding these events well are not just relatively good calendar year financial performance. Insurers that have suffered higher than anticipated losses have been placed under greater scrutiny from clients, rating agencies, equity analysts and commentators.
We believe that we have come through these events well and, with Amlin Bermuda now operating, we have increased the options available to the Group for the future.
The tenets of the philosophy that were set out in last year's annual report remain the same and are
the foundation for our recent good performance. We believe in profit focused underwriting across a diverse range of short tail classes of business with the business expanding into improving market conditions and contracting with deterioration in conditions. We aim to have highly experienced individuals underwriting our classes who share our overall philosophy. Control of risk, both at individual policy and portfolio level, remains a core feature that needs continuous focus and improvement. Underwriting risks and associated controls are covered further within the risk disclosure section of
the notes to the accounts.
Risk profile for 2005
2005 classes in profit/loss
Amlin's market share of Lloyds
Senior underwriter tenure
Managing through 2005
Given the nature of Amlin's business we were not immune to the hurricanes, with our ultimate gross losses from these events estimated to be US$860.9 million, reducing to US$236.7 million net of reinsurance recoveries. However, crucially, we still expect to deliver a 2005 net underwriting year profit. This differentiates the business significantly from many of its peers in the London and Bermudian markets.
We believe that this outperformance is due to three fundamental reasons, namely the diversity of our business, our net risk appetite which was commensurate with our balance sheet and strong risk management over our assumed exposures.
Hurricanes |
|
Total cost US$m |
Property insurance
US$m |
Property reinsurance
US$m |
Energy
US$m |
Other
US$m |
|
Katrina |
527.4 |
217.0 |
267.4 |
32 |
11.0 |
|
Rita |
160.7 |
17.8 |
89 |
51 |
2.9 |
|
Wilma |
172.8 |
16.5 |
155.9 |
- |
0.4 |
|
Gross loss |
33 |
252.5 |
512.3 |
83 |
13.1 |
|
Reinsurance recovery |
-619.5 |
-214.2 |
-319.6 |
-73.2 |
-12.5 |
|
Reinstatement premiums |
241.4 |
38.3 |
192.7 |
9.8 |
0.6 |
|
Net loss |
-4.7 |
11.9 |
-28.8 |
10.1 |
2.1 |
|
Sub total |
236.7 |
50.2 |
163.9 |
19.9 |
2.7 |
|
|
Enlarged Amlin Group pro forma
Geographic source of income
Diversity and aggregation of risk
Diversity was crucial to our financial performance in 2005. In excess of one third of our 2005 premium income was not exposed to natural catastrophe events and equally our exposure to these events is well spread worldwide and also within the United States.
We are also well aware of the possibility of multi-line catastrophe events and we are careful to control property insurance and reinsurance, energy and other exposures which may clash.
It is our policy to obtain as much transparency as possible over the catastrophe coverage being given. We limit our underwriting of worldwide retrocessional exposure, nationwide or worldwide coverage and give catastrophe coverage for direct accounts sparingly where it clashes with our reinsurance account.
Line size
The size of the Group's capital base allows us to offer attractive lead lines to brokers and insureds whilst maintaining an acceptable relationship between the maximum indemnity cost per risk and overall gross premium income. Amlin's largest individual risk line in 2005 was approximately 10% of our gross premium income and the average lines written were considerably lower as shown in the table below. We believe that this risk profile makes us less prone to single material losses relative to our capital compared to some of our peers.
Maintaining relatively modest gross lines as a percentage of premium income and capital also makes us less reliant on reinsurance for most of our direct classes of business and makes it easier to adjust exposures if reinsurance becomes unavailable or unacceptably expensive.
Gross line limits in 2005 |
Normal Maximum line
size £million |
Average Line size £million |
|
Catastrophe per programme |
44 |
3.8 |
|
Risk XL |
18 |
1.9 |
|
Property large risk |
18 |
4.8 |
|
Cargo |
15 |
2.3 |
|
Energy |
21 |
3.1 |
|
War/Terrorism |
15 |
5.5 |
|
Specie |
35 |
7.9 |
|
Airline |
72 |
24.8 |
|
Airports liability |
49 |
22.3 |
|
Space |
46 |
10.5 |
|
|
Catastrophe risk management
The correct identification of accumulating exposures to a possible event and the modelling of potential damage factors are crucial components of catastrophe risk management. We control catastrophe coverage both by way of policy limits and territorial or zonal limits.
Historically, we believe that we have taken a prudent approach to potential accumulating exposures. For example, in estimating the possible loss from an earthquake, we have assumed significant damage to aircraft on the ground which is likely to exceed exposures at any given time.
Modelling software is used to assess the losses that may arise from a certain specified event. Having upgraded our software in 2004, we commenced a review of our systems for capturing aggregating risk in 2005 with a view to significantly increasing the granularity of the assessment of catastrophe exposures. This has included the use of modern mapping technology with an ability to show, in very detailed form, the exposure written in our direct accounts.
This is particularly beneficial for our large property and terrorism portfolios where the information can be used to test the potential size of losses by applying different damage factors. For example we can vary a wind path or strength or, in the case of terrorism risk, vary the size of a potential bomb blast.
The actual damage caused by wind alone has been reasonably well modelled by systems available to insurers. However, damage caused by storm surge or flood, which is often included and sub-limited in commercial policies, is more difficult to model and has been less reliably estimated. Furthermore, the collateral insurance losses that arise from coverage that is given for extra living expenses, business interruption cover and contingent business interruption needs to be identified and added to modelled losses.
|