Catastrophe Management

Rating agencies are unanimous in their assessments that catastrophe exposure is the most significant risk facing most insurance companies today. To manage this exposure, Holborn helps its clients to:

• improve the quality of their policy/location data,


• identify accumulations of exposure,


• identify the sources and quantify the magnitude of the risk,


• quantify the impact of changes in underwriting or policy terms on the exposure, and,


• examine various reinsurance and other risk financing alternatives, to pick the most cost effective way of managing the risk.

Holborn uses all of the commercially available models (e.g., RMS, EQE, AIR), as well as proprietary models, to estimate our client's risk from natural and man-made perils. Our relationships with the modeling firms ensure that we are current with all model updates. Holborn uses catastrophe models to analyze hurricane, earthquake, tornado/hail, workers compensation and terrorism exposures. Holborn has been using these models since 1994.

Holborn has also developed a proprietary Dynamic Reinsurance Analysis ("DRA") program that simulates thousands of years of company results and compares the effectiveness of various risk management programs (including reinsurance and capital market options). DRA can quickly evaluate even the most complex reinsurance structures against success criteria established by Holborn and its clients (e.g., minimum BCAR, volatility in net income or policyholder surplus).

Key considerations in catastrophe analyses are:

• all models are a simplification of reality and the results must be viewed in that context,


• use of multiple models is key to understanding the range of expert opinions on these challenging questions,


• an in-depth understanding of the models is necessary to reach meaningful conclusions,


• there needs to be independent research on frequency, demand surge, secondary uncertainty, etc., and


• all analyses should be backstopped with deterministic studies a a review of actual exposure accumulations.

Holborn has a full range of capabilities to help its clients manage it catastrophe exposures, including:

PML drivers

A fundamental goal of catastrophe modeling is to gain an understanding of the magnitude of the risk. This is typically described as a company's Probable Maximum Loss ("PML"), which is the gross loss to the company at a predetermined probability, e.g., the 1-in-100 year or 1-in-250 year event.

Equally important is an understanding of what drives the company's losses, particularly the PML. Holborn's analyses provide insight into what part of a client's book presents the most catastrophe risk, e.g., coverages, locations, construction, etc.

Portfolio Optimization

Optimization is the constant goal of every business manager: "How do we get the most performance for my company within the operating environment we face?" The logical answer to this question needs three clear points of reference:

• Goals: How does the manager know one result is better or the best? Profit, revenue, market share, costs or some other measure?


• Options: What are the ranges of possible actions? What products? What prices? What distribution?


• Constraints: How does the environment limit the firm's possible performance? Regulation, competition, capitalization, market inertia?

These references can be expressed in a limited number of very simple mathematical equations, which can be used to scientifically test possible strategies against the goals and constraints. Managers can use these results to identify better strategies, and sometimes even the best single strategy.

With respect to insurance portfolios, optimization techniques face a number of challenges, including:

• a constantly changing environment, with many elements beyond the control of the insurer,


• the size of the uncertainty introduced by property catastrophes,


• insurers can't quickly change their exposures to achieve specific goals, and


• in most cases, insurers cannot neatly define a limited set of strategies that respond to their environment.

Holborn has derived a proven mathematical approach to guide companies in business planning for their catastrophe exposed portfolios that:

• uses the company's available information to measure potential changes in exposure, e.g., policy or risk count, premiums, values, etc.;


• divides the portfolio into sub-portfolios (e.g., county, zip or even an individual risk) such that all of the current business is mapped into one and only one sub-portfolio. The sub-portfolios are defined so that no individual sub-portfolio constitutes the company's PML event by itself;


• ranks the sub-portfolios based on the change in modeled PML loss resulting from an increase of the amount of business in each sub-portfolio. "PML loss" can be defined as any desired level of loss, such as the 1-in-100 or 250 year return time, or that would cause concern on RBC or BCAR scores.

By combining the output of Holborn's analysis with information on the profitability or attractiveness of each sub-portfolio, this analysis shows companies where they can grow while adding the least to their PML losses. This analysis can also identify which sub-portfolios in the current book of business create unacceptable PML loss exposures.

Mapping exposures/accumulations

Sometimes a picture is worth a thousand words, so Holborn creates customized maps summarizing exposures, average annual loss or impact on PML by geographical region (e.g., zip codes, county, proximity to terror target, etc.). These summaries may be for the entire portfolio or a sub-portfolio (e.g., homeowners separately from commercial property).

Data quality auditing

To improve the quality of the data, Holborn's cat modeling team evaluates it for completeness and reasonability against prior year data, industry standards, other company reports, etc. The team will also check the output of the models for consistency with the underlying data. If any discrepancy arises, we will work with our clients to figure what caused the discrepancy and make changes/corrects where necessary. The goal is to avoid being penalized for poor data, through higher reinsurance costs, inadequate coverage, etc.

Holborn helps its clients understand the cost/benefit of improving the quality and completeness of their data for modeled loss estimates, rating agency views and reinsurance pricing and capacity.

Deterministic storms

Probabilistic analyses provided by catastrophe modeling can be useful in understanding the long-term average risk of a portfolio; however, they may not adequately capture the risk of peak accumulations. To analyze such peak exposures, Holborn has developed models that run user defined storms (hurricane and tornado/hail) through these peak exposures to ensure that clients are aware of their "worse-case" events. For example, the user would define a hurricane by "drawing" the path of the eye on the map and then selecting the size (Rmax), forward speed and maximum wind speed.

Stochastic modeling

Holborn has developed models to simulate losses from all perils and apply various risk mitigation strategies (including reinsurance, cat bonds, contingent capital, etc.) to the losses to get an annual net EP curve that accurately reflects the following by contract:

• coverage,


• retentions,


• co-participations,


• limits (per occurrence and aggregate), and


• reinstatement premiums.

The impact on annual gross and net losses from changes in such factors as storm frequency, demand surge, loss adjustment expenses can be quickly analyzed by simply changing simulation parameters.

In addition, losses can be ceded to a specific reinsurer or group of reinsurers, if so desired. Based on the reinsurers' ratings, the probability and impact of reinsurer default can be determined.

Post-event modeling

Holborn produces detailed industry-wide analyses of major catastrophes (e.g., Katrina and 9/11). It can help individual clients estimate their losses from such events by running their property portfolios through the various catastrophe models. While recent events have shown that the modeled loss estimates from actual storms can widely miss the mark, they still provide useful information to clients in planning their post-event responses.

Dynamic Financial Analysis/Dynamic Reinsurance Analysis

The effectiveness of the client's catastrophe management strategies can be analyzed through Holborn's proprietary DFA/DRA model, which simulates a client's catastrophe losses (based on the output of the models or as determined by the client), as well as large losses from other lines, e.g., WC, property per-risk, other casualty), other loss experience, expenses, premium growth, etc. Each simulated underwriting result is run through a simplified balance sheet, income statement and cash flow statement to analyze the volatility of the client's gross and net experience. Different catastrophe management strategies can be examined and evaluated against client defined metrics, e.g., likelihood of a ratings downgrade or not meeting plan.