Jeff Mohrenweiser looks at the increasingly sophisticated capital requirement models employed by the reinsurance industry.
The insurance industry is hurtling towards a stochastic world where capital and risk management decisions are increasingly guided by the results of thousands and thousands of scenarios. Coupling improved computer tools with regulatory directives, management teams are anxious to review modelling results of simulated natural catastrophe events or economic scenarios and to develop countering action strategies. Unfortunately, there is no recipe for creating these “rocket science” models. General principles and guidelines are promulgated but no simple solutions are available. As such, insurers are spending millions of dollars on solutions and the ante is rising.
The good news for investors, rating agencies and others is that these unique models focus on the particular risk drivers of each insurer. The bad news is the parameters and output of these models can vary quite substantially.Fitch reviewed several in-house capital models and noted areas of model divergence. An example occurred with two similar property casualty companies. The first focused considerable energy around asset modelling at the expense of catastrophic analysis. The second spent large amounts of time on catastrophic analysis but used factors to monitor credit risk. Incorporating other model decisions creates endless possibilities.
Fitch developed Prism, a global capital model, as a consistent platform, or risk yardstick, in analysing insurers around the world. It uses local market parameterisation, robust simulation engines employing leading actuarial methods and modern financial theory. In short, a motor insurer in France will have capital charges based on French motorist experience derived from thousands of possible scenarios.
In contrast, competing factor-based models are limited in capturing the interaction between assets and liabilities and can shoehorn parameters derived from one region into others. Extreme tail events or diversification benefits may be simplified or ignored. This “one-size-fits-all” model cannot incorporate the differences that exist between products in various countries or allow for common enterprise risk management techniques such as hedging, product redesign or reinsurance.