Guillermo Franco, Head of Catastrophe Risk Research - EMEA
The quantification of catastrophe risk through the use of numerical models is accepted routine in our industry. Given the large impact of catastrophe risk on (re)insurance portfolios it is no surprise that the methodologies used for such quantifications are becoming the subject of more intense scrutiny by regulators and risk managers. Whether the motivation is internal, external or both, efforts are now in full swing towards acquiring a deeper understanding and a more sophisticated use of catastrophe risk model results - the path to “own your view of risk.”
You “own” your view of risk when you attain sufficient command and knowledge over your risk quantification tools to independently and rationally justify risk management decisions based on these tools. Arguments need to be supported by tested modeling hypotheses - identifying which models or model blends are most suitable to represent your risk.
To assist clients in pursuit of developing their individual view of risk, Guy Carpenter is introducing the Model Suitability Analysis (MSA)SM initiative. The MSA framework, shown in Figure 1, consists of eight components arranged along three groups of critical functions necessary to evaluate model suitability for a given portfolio and peril.
The evaluation of models within MSA is driven by rigorously defined tests. These tests can also be tailored to address particular client requirements, focusing on issues critical to their risk exposures. This differentiates Guy Carpenter’s approach from general model-wide testing campaigns by focusing on a more subtly customized service. The model evaluation protocols encompass sensitivity testing, loss validation using client and industry claims data, and a scientific appraisal of the models that inquires into hazard and vulnerability assumptions.
The MSA scientific appraisal process philosophically relies on comparisons of model characteristics and assumptions with third-party datasets. These datasets are obtained through investigation of research literature as well as through research collaborations with academic partners. To that purpose, Guy Carpenter has established contact with a series of universities and research groups in order to develop the comparative frameworks necessary. An example of one of these initiatives is collaboration with the Department of Civil Engineering and Engineering Mechanics at Columbia University on vulnerability related to wind perils. Another is collaboration with the Istituto Universitario di Studi Superiori in Pavia, Italy on a series of earthquake-related MSA tests involving ground motion collection and damage experience for historical events in Italy.
By studying comparisons with third-party scientific sources, MSA offers clients a view into which scientific hypotheses are more in-line with model assumptions. It also offers a perspective into the uncertainty related to these assumptions, which may be in some instances, quantifiable.
The conclusions from the evaluation tests are collected in the MSA Grid, a table organized by model and testing criterion that constitutes the foundation to risk customization. The MSA Grid provides our clients with a simple way of supporting their decision making process and documenting their model validation efforts. Each grid entry contains the associated conclusion, color-coded according to the performance of a particular model under a particular test. This information yields a clear synthesis of model suitability for the client exposures. The company may then pinpoint model traits that constitute opportunities for model enhancement or risk customization, which ultimately lead to the integration of a well-understood perspective of risk.
MSA also considers the need for a communication strategy, leveraging a modular documentation system and a customized knowledge sharing program. With these two components, clients gain access to all training and material necessary to communicate their view of risk to stakeholders. The documentation system is designed in such a way that each test performed through the MSA process and each conclusion reached is recorded in a standardized fashion. This allows clients to easily extract parts of the MSA documentation and integrate it into their internal model evaluation documentation as well as into their Solvency II reporting documents.
In sum, MSA encompasses all fundamental model evaluation, integration and communication components, establishes a tight link to business needs, and offers a path to successfully own your view of risk. This path is made achievable through this modular and flexible approach, which can be tailored to the company’s needs and resources. A complex and time-intensive effort to understand and use a sophisticated catastrophe modeling environment becomes, through MSA, an endeavor that companies can phase- in over a period of time and tackle component by component. For companies willing to progressively gain command over their cat risk modeling tools and to acquire a position of independence in their cat risk management, MSA offers an appealing approach.