Large, diversified and highly-rated reinsurance groups with approved internal capital models will likely have materially lower capital requirements under Solvency II than they already maintain for their ratings. For these reinsurers, rating agencies will remain the final arbiters of capital requirements, while Solvency II will add administrative and regulatory cost and, perversely, encourage a lower standard of solvency. So far, rating agencies have resisted the demand to materially reduce capital requirements, with S&P granting only a limited weight to internal economic capital models in their assessment of risk adjusted capitalization (1).
Reinsurers of all sizes with material non-proportional books of business and/or material catastrophe exposure outside of Europe are essentially forced to apply for internal model approval. This is due to the seemingly high capital charge for this business contained in the standard formula. This increases the compliance costs for those companies that do not already use internal models. QIS 5 results show that progress on internal models has been slow as companies struggle with model construction and validation.
On the other hand, many smaller or unrated reinsurers assessed under the standard model will see capital requirements increase. We may see certain niche reinsurers withdraw from the market or combine with larger companies as a result. While some consolidation will improve the health of the reinsurance market, it may also pressure rates and eliminate some of the risk transfer options available to cedents. Longer-tail lines of business will be particularly prone to rate pressure as it becomes more expensive to match long-term liabilities with long-term assets.
Solvency II May Contribute to More Intense and Volatile Underwriting Cycles
A more precise (or over-calibrated) measure of solvency is naturally more prone to volatility. The widespread use of the Solvency II standard model and internal capital models in conjunction with market consistent accounting of assets and liabilities could contribute to shorter, more volatile underwriting cycles. It could also drive more volatile earnings and balance sheets. Reinsurers, guided by economic capital models based on value-at-risk (VaR), may more actively shed assets and repurchase shares in soft markets, then seek to replace capital in hard markets. While this practice may appear to be sound capital management to investors and some managers, it tends to amplify the market impact of large losses while increasing reinsurers’ cost of capital. It is also based on a potentially spurious measure of risk as it often overly simplifies an organization’s exposure to tail risk as discussed below.
For example, Figure 1 shows the year-on-year rate change and cumulative rate on line index for the global property catastrophe business. The nearly 80 percent year-on-year average increase in pricing seen in 2006, following the shock losses of hurricanes Katrina, Rita and Wilma in 2005, was also preceded by reinsurers returning several billion dollars in capital to shareholders. This happened in response to relatively modest price declines in 2004 and 2005. Following these events, capital flooded into the reinsurance market in response to anticipated rate increases. The establishment of new markets and “side cars” benefited many cedents. However, several reinsurers that had been actively managing capital based on VaR and pricing trends found that they could not replace the capital that they had returned to shareholders only months earlier.
The Solvency Capital Requirement (SCR), the risk-based capital requirement for (re)insurers under Solvency II, is calibrated to a 99.5 percent VaR over a one-year period. Many internal capital models in use today also calibrate to VaR. There are a number of problems with the use of VaR as a measure of risk, many of which were illustrated over the course of the 2007-2009 credit crisis. For example, VaR is the foundation for risk-based capital requirements under Basel II, which not only failed to prevent bank failures, but arguably contributed to the crisis by providing a false sense of security around risky investments.
The three probability density function distributions shown in Figure 2 represent three reinsurance portfolios that all have the same VaR at 99.5 percent probability (the same level of confidence as the SCR under Solvency II ). Yet, the risk profile of each is clearly different. The portfolio represented by the distribution in blue has the highest average expected loss but is actually the least risky, with its short tail. The portfolio represented by the distribution in yellow has the lowest average expected loss but is the riskiest because it has potential for much higher losses in its long tail.
While tail value at risk (TVaR) shares many of the same limitations as VaR and may also contribute to volatility when relied upon as the ole measure of risk, it can be a better measure of underwriting risk. In this example, the VaR at 99.5 percent probability is USD10 million for all three distributions. However, the TVaR at the same level of probability is USD10.7 million for the blue distribution, USD11.4 million for the green and USD13.4 million for the yellow.
The 2007-2009 credit crisis vividly showed that the simplistic use of VaR to manage risk may result in increased concentrations and gross underestimation of exposure to tail events. It can also give a false sense of security that can contribute to the overcorrection in risk appetite following unanticipated events.
It is clear that Solvency II will profoundly impact the reinsurance market - not only within Europe, but globally. Advances in disclosure and overall market strength will come with very real costs to both the industry at large and individual companies. It is imperative that any companies affected by these sweeping changes make preparations now to navigate this changing and increasingly volatile reinsurance market.
Note: 1 Standard and Poor’s, Methodology: Assessing Insurers’ Economic Capital Models, May 15, 2008.