Real-time Catastrophe Exposure Modelling

Insurance companies are in business to assume the financial risk of various hazards their clients face. In so doing, insurers have been practicing risk management in one form or another by their very nature. However, regulators and rating agencies are increasingly looking for more than just tactical risk management. Companies are being evaluated based on their enterprise risk management (ERM), including new elements such as a strong culture of risk management, consistency across the company in the identification and measurement of risks and an enterprise- wide view toward the management of risk.

Risk arising from the accumulation of insured exposures to catastrophes — either natural or man-made — typically represents one of the biggest risk factors facing a property and casualty insurer. In a January 2008 white paper outlining its rating methodology, Risk Management and the Rating Process for Insurance Companies, A. M. Best states: “[We] consider catastrophic loss, both natural and man-made, to be the Number 1 threat to the financial strength and policyholder security of property and casualty insurers because of the significant, rapid and unexpected impact that can occur.”

Clearly, prudent risk management as applied to catastrophe risk should be an integral part of a company’s overall ERM program. So then what should a company do in the way of catastrophe portfolio management as applied in an enterprise-wide framework? A variety of tools and processes currently define a set of industry best practices:

• a high level of data quality;

• use of advanced catastrophe models;

• identification of key risk factors;

• explicit articulation of corporate risk tolerances;

• constant monitoring of aggregate and individual exposures against risk tolerances; • risk underwriting discipline and compliance;

• real-time catastrophe monitoring and operational response;

• strategic portfolio optimization;

• use of internal capital models to integrate cat risk with other company exposures;

• an integrated reinsurance program; and

• financial strength and flexibility.

The remainder of this article will touch upon several key elements of the above list.

Models describing the frequency, severity, and consequences of natural catastrophes such as earthquakes, hurricanes, severe storms, winter storms and floods have become increasingly sophisticated over the past 20 years. Although far from perfect, they are a powerful class of tools in portfolio management. That said, cat models are no exception to the rule of “garbage-in, garbage-out.” Data quality is of the utmost importance in the application of cat models; this is increasingly the focus of scrutiny by industry observers. Following the 2004 hurricane season, in which four hurricanes struck Florida, cat model vendor RMS conducted a post-season review of actual losses versus predicted losses. Their research found that data quality issues caused models to understate the Florida hurricane loss estimates by 20% to 30%. The point is this: there is little sense establishing a state-of-the-art portfolio management or ERM practice if your company does not have the input data to support it.

EXPOSURE MONITORING

If good data is a first step, then simple exposure monitoring is a necessary second step. It is difficult to know whether or not your company has “too much” exposure if you don’t know what you have at all.

Computer systems have provided a modern-day alternative to an old stick-pin map of exposures. (Please see Figure 1.) In this view, the insured portfolio is shown by location, with an overview of clusters by total insured value (TIV).

Alternatively, exposures can be aggregated to a particular geographical region. (Please see Figure 2.) Colour coding quickly identifies hot spots of concentrations. Queries can be run against portfolio attributes such as construction type, protection class or insured value.

Exposures can be accumulated around key locations — for example, a potential terror target (Please see Figure 3) or an earthquake epicenter — to evaluate exposures to specific events. Catastrophe models can be run against these specific locations to help identify the largest risks facing the company.

Exposure management systems such as these are useful and go beyond traditional monitoring of existing exposures. New risks can be mapped and evaluated using the same system; in this way, new exposures can be judged not just on their own merits, but also based on the marginal impact to the company’s existing portfolio.

REAL-TIME EVENT MONITORING

The functionality of a best-practice portfolio management system is not restricted to aggregate portfolio monitoring or risk underwriting. Real-time event monitoring allows a company to trace the development of an actual occurrence — a hurricane, for example, or an earthquake, wild fire, thunderstorm, etc. — and superimpose it on the insured portfolio. (Please

Such a tool enables the company to respond to catastrophes quickly. Teams of claims adjusters can be mobilized, for example. Reinsurers can be notified and informed. Financial departments can make provisions for appropriate liquidity or even capital depending on the expected impact.

PORTFOLIO OPTIMIZATION

Perhaps the ultimate use of a portfolio management system is the optimization of the portfolio’s expected performance subject to certain constraints. For example, an insurer may wish to maximize profit or minimize losses in a layer, subject to volume constraints in any one postal code. Best-practice portfolio optimization tools can identify the exposures that should be eliminated from the current portfolio, as well as the sorts of exposures that should be added to the exposure. Optimizing or even just incrementally improving the expected performance of a portfolio is not simply an issue of writing some new business or non-renewing some existing business. A full portfolio tune-up should include all the tools at management’s disposal, such as pricing, terms (deductibles, limits), agency management, and reinsurance to name a few.

CAPITAL MODELLING

Inclusion of cat model output into an internal capital model is probably the clearest departure from the traditional risk management of catastrophe exposure. Inclusion of catastrophe risk distributions — integrated with all other meaningful company risk distributions — allows insurers to better judge the level of required capital for the company or, alternatively, judge the adequacy of the capital on hand.

In this enterprise-wide context, risk tolerances can be best expressed. It is often the case that a significant risk to one business unit is not, in fact, significant to the company as a whole. Or the risk is hedged or diversified away by the firm’s other activities.

Re i n s u r a n c e strategies are best devised at this company- wide, integrated level because the true risks to the firm are understood. Furthermore, since reinsurance reduces the volatility of the company’s results, it is a tool in the company’s overall capital management strategy.

The integration of portfolio management tools into the operations of the company (including underwriting, claims, reinsurance or finance) is an important element to outside observers. Rating agencies and regulators look for evidence that the tools and processes defined in the company’s ERM program are actually put to use. Use is self-evident when a tool is engrained into the fabric of the company. In the end, the use of these best-practice tools and processes is not really an issue of compliance, but of good management.