European insurers and reinsurers will face requirements for full compliance with the new Solvency II capital regime requirements in just over two years. Even if this introduction is phased in — as the European Commission has reportedly indicated it could be — these requirements will have a wide-ranging and profound impact on the insurance industry throughout Europe.
From a quantitative perspective (”Pillar 1″ of the Solvency II regime), insurance companies may opt for an internal model or for the Standard Formula in establishing their own solvency capital requirements (SCR). The Standard Formula is still under development and has been tested in detail in the most recent CEIOPS quantitative impact study (QIS5). A new approach for defining catastrophe scenarios under natural perils has been proposed.
We have undertaken an analysis to measure the industry-wide impact of the natural catastrophe scenarios provided in QIS5 and determine whether they are in line with current industry assumptions. Our findings indicate that, at least at the aggregate level, the QIS5 scenarios are roughly within the range of current industry models — but nevertheless, most insurers will likely need to make adjustments to optimize their positioning and operations under the new capital regime.
The QIS5 Scenarios
The process of defining the QIS5 catastrophe scenarios was supported by an industry-wide task force. This group, consisting of major industry players (including Guy Carpenter) as well as regulatory authorities, was charged with developing standardized European scenarios for all major perils that would be suitable for both regional and global insurers (and reinsurers).
The scenarios that were developed include windstorm, flood, earthquake, hail and subsidence. The task force proposed parameters aimed to define each catastrophe scenario at a 99.5 percent VaR level - which corresponds to a 1-in-200 year loss. The proposed factors are applied to Catastrophe Risk Evaluating and Standardizing Target Accumulations (CRESTA) zone data for each country and each peril. These calculations are supported by different correlation matrices (between CRESTA, between countries and between perils) to arrive at the final catastrophe solvency capital requirement (SCRCat).
Windstorm Loss Analysis
PERILS AG, located in Zurich, Switzerland, has recently published QIS5 windstorm scenario property loss estimates based on its 2010 Industry Exposure Database. As shown in the chart below, the PERILS gross occurrence-based loss estimate for the nine major European markets, including geographical diversification benefit, amounts to almost EUR37 billion.
QIS5/PERILS vs. Industry Loss Models
The first step in our analysis was to compare the loss figures computed with PERILS data under the QIS5 scenarios to the 1-in-200 year loss figures in the premier vendor models - AIR, EQECAT and RMS. All figures were computed based on ground-up figures without deductibles and limits, on an occurrence basis.
As the charts below illustrate, the loss projections computed with PERILS data under the QIS5 scenarios fall within the range of the three vendor models, with a tendency toward conservatism. The QIS5 results are higher than some vendor model figures and lower than others; they appear to have been computed in most cases more prudently in the three largest European markets: France, Germany and the UK. The diversification effect, while substantial in all four approaches, is distinctly lower in the QIS5 model.
Using the PERILS market exposure data, 200-year losses under QIS5 standard scenarios by country are shown next to the corresponding losses from the major three vendor models.
Benchmarking QIS5 Scenario Return Period Against Vendor Models
We have established that losses in the QIS5 windstorm scenarios are generally within the range of the major vendor models. But how do return frequencies compare? The QIS5 standard scenarios are tailored to represent a 200 year return period in each territory. Looking at the QIS5 loss estimate on the modeled exceedance probability curves reveals the corresponding modeled return period. Plotting the implied vendor model return frequencies against the QIS5 scenarios’ 1-in-200 year level yields an interesting (albeit similar) spread, as shown in the chart below.
Illustrated above are return period levels (logarithmic scale from the center) of QIS5 standard scenario losses and vendor models’ return periods of the same loss amount. Both are based on PERILS market exposure by country. EU represents the scenario for all countries combined.
As seen with Model A in most countries, the lower exceedance probability curve in this vendor model leads to lower frequencies – i.e., higher return periods for a given loss level.
Reinsurance Considerations in QIS5 Windstorm Scenarios
In order to consider the reinsurance implications of our findings, we must first convert our per-occurrence calculations into an annualized capital scenario. We do this with the following formula provided in QIS5:
- Case A: Event 1 = 0.8 * CATWind; Event 2 = 0.4 * CATWind
- Case B: Event 1 = 1 * CATWind; Event 2 = 0.2 * CATWind
- Net capital requirement: max (Case A, Case B)
However, a simple global formula cannot realistically capture the nuances of real-world situations, and this fails to incorporate such important considerations as aggregate protections, multi-peril structures and stop-loss features that insurers may have purchased to protect their portfolio. Nor does it adequately reflect the diversification benefits of pan-European reinsurance covers. Such implementation issues are best addressed on a case-by-case basis with the assistance of a professional reinsurance advisor.
Impact of reinsurance
Based on our market benchmarks for reinsurance structures in each of the nine countries covered by PERILS AG’s study, we can compare the overall amount of reinsurance purchased to the levels implied by these scenarios. In most of these countries, the current market reinsurance limit (including cat bonds and comparable structures) is below the amounts suggested by these scenarios. In other words, on average, companies buy reinsurance below the 1-in-200 return period in most countries. On an individual company basis, however, the level of purchase may sometimes reach and even exceed the amount defined by the QIS5 formula for windstorms. There is no specific return period-capital requirement linkage in the current solvency regime; Solvency II will create a direct link between reinsurance purchase levels and the capital needed to cover a 1-in-200 year loss scenario.
For multinational companies, the benefit of diversification significantly impacts the gross scenario, as illustrated in the PERILS study. As a consequence, it also impacts the level of reinsurance purchase and the net scenario.
In the following chart, we illustrate the impact of a synthetic catastrophe excess-of-loss (XOL) reinsurance program on a QIS5 wind scenario.
The total is composed of:
- The retention amount from the first event
- The reinstatement premium payable following the first event
- The retention amount from the second event (which could be lower than that of the first event due to the impact of annual aggregate deductibles)
- The reinstatement premium to be paid following the second event (generally limited to a second reinstatement premium on some low layers)
To obtain the final SCRCat, these net exposures should be added to that from other perils, with the proper recognition of the diversification benefit between perils as provided by the QIS5 correlation matrices.
How Guy Carpenter & Company can assist
While the analysis above produces useful QIS5-scenario loss metrics at the aggregate level, it is not enough to successfully position a company for the new Solvency II regulatory requirements. This must be done on an individual basis, taking into account each firm’s unique attributes, situation and strategy. Guy Carpenter is a leader in reinsurance advisory services, and can help companies prepare for Solvency II by:
- Assisting in understanding and managing the new regulatory framework
- Benchmarking a company’s individual QIS5 result against vendor model results using a (partial) internal model approach
- Properly assessing the potential capital impact
- Advising on how to adequately reflect impact of reinsurance, including under the most complex structures
- Providing tools to optimize reinsurance programs, with a particular focus on regulatory capital implications
- Assessing the impact of group protections for individual local entities on their own SCRs
Statements concerning, tax, accounting, legal or regulatory matters should be understood to be general observations based solely on our experience as reinsurance brokers and risk consultants, and may not be relied upon as tax, accounting, legal or regulatory advice which we are not authorized to provide. All such matters should be reviewed with your own qualified advisors in these areas.