Review

Risk management

Phil Ellis
Risk Committee Group Chief Actuary

Modelling and exposure management

The Group is exposed to underwriting risk from an increase in severity (including catastrophe losses), or frequency (attritional loss) beyond the expectation of the business plan. These risks are quite different and require different methods of control.

Catastrophe risk modelling has developed significantly in recent years as reinsurers and modelling agencies have reacted to a series of events, particularly US hurricanes. However, the impact of Hurricane Ike is another lesson of the issue that each major event has characteristics which differentiates it from losses that have occurred before, thus defying the projection of the models. Ike’s uniqueness was a combination of the sheer size of the storm and the distance that the weather system penetrated the US mainland. Therefore to rely on models as the key risk management tool is folly and the approach to writing catastrophe exposed business is the key ingredient of our risk control.

Property insurance and reinsurance are major business lines at Amlin, and they have strong potential for correlating catastrophe losses. In response, we have a tried and tested formula for the way in which these classes are written which insulates us from significant individual catastrophe loss model error as described above. First, we write the exposure predominantly through catastrophe excess of loss reinsurance which means that, for much of the book of business, the potential loss caps out at the levels to which insurers buy protection. If they do not buy enough cover, the additional loss falls to them and not the reinsurance market. Second, the reinsurance account we write is made up of regional writers with original portfolios of small commercial, or homeowners, business which is much easier to model. Third, the reinsurance book is written with contained maximum exposures for any one zone. The property book is written with care taken over aggregates and cover given for business interruption, which is very hard to model. This set of underwriting policies allows easier management of the clash of portfolio exposure. In the methodology used to estimate the impact of events, we do not simply rely on a modelled output, but actively consider the potential loss for each of our classes of business independently.

Progression of Risk Appetite
Click here for full sized image

The accounts written and the care taken to avoid exposures which are difficult to model, as well as our approach to reserving, has borne fruit in the relative stability of our loss estimate for major events. As an example of this Hurricanes Gustav and Ike have loss reserves of US$302 million at year end, compared to US$285 million at the end of October.

In 2008 we have further developed our central catastrophe modelling capability within the DFA model to analyse further the annual aggregate expected catastrophe loss rather than just the single event potential. To this end we have produced levels of probable outcomes for annual aggregate catastrophe loss cost for 1 in 50, 1 in 100 and 1 in 250 years.

Pricing

Attritional losses can materialise due to inadequate pricing of risk or due to an increase in the quantum of claims through sheer number or through inflationary pressures on settlements. The Group uses a number of controls to mitigate this risk.

The pricing of risk is critical because adequate pricing for non-catastrophe exposed risks should cover a mean level of loss experience and still leave a margin for profit.

The greater this margin, the greater the cushion for unexpected loss development. Underwriters are required to record renewal rate movement and the technical benchmark price.

Increased focus on credit risk

Reinsurance protection is an important element of our underwriting risk. However, the use of reinsurance in itself gives rise to further risk of counterparty failure or dispute. Therefore, the Group has a number of controls in place to manage this risk. The list of acceptable reinsurance security is determined at the Reinsurance Security Committee chaired by the Chief Risk Officer and attended by the Group Finance Director, an analyst and representatives of the underwriting divisions. Whilst rating agency input from Standard & Poor’s and AM Best is used, the Committee grades each reinsurer according to established criteria. There are limits of usage applicable in total for each reinsurer dependent on the grade of security.

During 2008, the impact of investment market volatility, particularly in the equity and debt security markets, has had an impact on the financial strength of a number of Amlin’s reinsurance counterparties. This impact has been reviewed by our reinsurance security analysts and appropriate action taken. The security risk from certain counterparties has been enhanced by collateralisation of limits and this has helped contain the potential for bad debt.

The potential for disputed reinsurance is also considered and appropriate provisions made. A history of reinsurer payment performance is important so that the Reinsurance Security Committee can make rating decisions based on willingness to pay, as well as pure financial suitability.

Positive action to manage market risk

Since the appointment of the Head of Investment Risk in early 2008 there have been a number of enhancements to our analysis of market risk. First, all asset managers have been visited and have provided feedback via a questionnaire to improve our understanding of their own risk management policies and procedures.

Second, we have developed an asset risk system which allows the investment portfolio to be analysed using a more developed range of risk measures including value at risk, drawdown and return/volatility or Sharpe ratio. This system also enables loss simulation and stress scenarios to be run over the portfolio.

Finally, we have scoped a project which will assist in portfolio optimisation against a prescribed risk budgeting framework. This will be completed by the middle of 2009.

These developments will add further robustness to the already strong framework operated by the Chief Investment Officer.

Liquidity risk management

Our approach to liquidity risk management is explained in the financial management section of this review.

Operational risk monitoring enhanced

The adoption of our improved risk management framework will be a key step in the analysis of internal operational risk because the benefit of controls (and therefore the potential cost of nonperformance) is specifically evaluated in the risk assessment process. Amlin has a zero appetite for some operational losses, including those which could involve legal or regulatory risk, and human safety. However, it is recognised that within the entire risk framework, there has to be an acceptable balance between the cost of control and the size or probability of the residual risk. Few risks can be fully avoided, transferred or prevented and, as in most organisations, the benefit of agreed values, culture and training are enormous. Nevertheless, better risk evaluation and reporting will lead to fewer surprises and more accountability for risk management personnel.

The key external operational risks, such as building risks or external fraud risks, are managed by a mixture of controls, including insurance or postmitigation provisions such as business continuity plans and back up capability.

We also intend to capture and evaluate operational risk within the DFA model and therefore improve upon the capital modelling of this most intangible risk category.

Proactive strategic risk management

The threats posed to Amlin from the appropriateness of our business model, strategy and location in the face of potential changes to customer behaviour, the commercial and economic environment or political and legislative changes are reviewed by the Chief Executive and the Boards of Directors. Areas of emerging risk are discussed at Risk Committee and long-term strategic plans formulated to deal with such threats or to develop strategic diversification so that the company is not reliant on few or single sources of revenue.