Participation in Quantitative Impact Study 4 (QIS4) exceeded European Commission expectations for small, medium, and large companies. The results, published by the Committee of European Insurance and Occupational Pensions Supervisors (CEIOPS) in November, suggest that 98.8 percent of the carriers participating will meet the Minimum Capital Requirement (MCR) and 89 percent satisfied the Solvency Capital Requirement (SCR), though the ongoing financial catastrophe could cause some changes to this result. Quantitative Impact Study 5 (QIS5), originally planned for early this year, has been deferred because of the potential impact of the financial crisis.
The turmoil caused by the ongoing financial crisis has been a pretext for hard bargaining between the European Parliament and the EU’s Economic and Financial Affairs Council (ECOFIN) around the concept of “group support” and contracyclical measures (like the equity dampener). Finally, a common draft Solvency II framework directive was adopted in early December, without the group support concept. The European Parliament will formally adopt this position this Spring and will thus avoid further jeopardizing the targeted implementation of Solvency II by 2012.
CEIOPS will take advantage of the delay caused by deferring QIS 5 to review some issues raised by QIS 4.
Participation in QIS4 exceeded the ambitions goals set by the European Commission. Thirty-four percent of European carriers participated, compared to an objective of at least 25 percent — with 65 percent of cross-border groups participating compared to an objective of 60 percent. Both the number of insurers and the number of participating countries increased relative to QIS3. All 30 EEA member countries were represented – including Romania and Liechtenstein, which had not been previously represented. France had the highest number of participants, and the United Kingdom had the largest number of life carriers.
The number of small and medium-sized carriers participating increased, as well. The former grew by 58 percent (to 667 companies) from QIS 3 to QIS4, and the latter was up 25 percent (to 522 companies). Additionally, 220 large carriers submitted their data (up 18 percent compared to QIS3). The total number of respondents (life and non-life) was 1,412, an increase of 37 percent over QIS3. Of these, 351 came from the life sector.
Under Solvency II, a company’s eligible capital (or “own funds”) represents the financial resources available to serve as a buffer against risks and absorb losses when necessary. QIS4 reported that, at the time of the study, eligible capital was 27 percent higher than the currently reported own funds, mainly because of valuation adjustments following the move to market consistent valuation, reclassification of equalization provisions and the inclusion in full of hybrid capital instruments, subordinated liabilities, and ancillary own funds.
Solvency ratios have risen for life carriers relative to Solvency I. In QIS4 the median life insurance company recorded a solvency ratio of 230 percent, which is an increase of 30 percentage points. For carriers as a whole (life and non-life), QIS4 showed that the vast majority of companies (98.8 percent) will meet the MCR criteria; however, 7 percent of captives will not. Eleven percent of all participants did not meet the SCR under QIS 4, the most affected being captives (28 percent), large carriers (13 percent) and non-life carriers (11 percent). The calculation date for the exercise was December 31, 2007. Given that financial markets have declined significantly since then, these results are likely to have been adversely affected.
“Not meeting the SCR does not necessarily imply having to raise capital upon the introduction of Solvency II for a number of reasons. In particular, undertakings can anticipate the introduction of Solvency II or, for example in the case of entities forming part of a group, they can reallocate own funds between entities. In absolute amounts the aggregated capital surplus of participating undertakings remains fairly stable, with a reported aggregate decrease of 3 percent.
For the European insurance industry as a whole, no additional capital is needed. However, the redistribution process of capital between risks and undertakings is confirmed again by QIS4 results.”1
Reliability of Results
Some regulators have questioned the reliability of the results – in particular the treatment of deferred taxes, the inclusion of the future premiums, and inconsistencies for solo entities for which IFRS (International Financial Reporting Standards) are not in force.
Cost of Implementation
Participating carriers spent only 3.2 person months on average to complete the exercise and were invited to take part on a best efforts basis, which may indicate that many applied the simplified calculations as opposed to the full approach.
Some carriers found the modeling requirements for QIS4 to be onerous. Calculation of the best estimate reserve and risk margin involves building models to project future cash flows, taking into account all products, options embedded in the products and all risks involved. Building such a model requires a lot of time and resources. In the life insurance sector, the complexity also relates to the policy-by-policy basis of the calculations, particularly for the lapse and catastrophe risk modules. Smaller carriers lacked sufficient flexibility in their models, for example, to incorporate the use of a term structure for interest rates rather than a single discount rate. Larger carriers commented that run times were very long, especially for stochastic policy by policy calculations.
Some participants indicated that, the more complex the model is, the more the results will not be readily understandable — and the higher the likelihood of difficulty in identifying the real source of risk of the business.
Participating carriers usually considered their data input for the QIS4 as sufficiently reliable. It was mentioned that the data was mostly derived from IFRS figures and was also used for other purposes. Carriers in some countries, however, pointed out that their data for life business might be of lower quality than data for non-life business.
Many carriers indicated that the segmentation of policy contracts as used in QIS4 was difficult or not clear, including the segmentation into proportional and non-proportional reinsurance treaties.
The approach adopted by life carriers for calculating technical provisions was generally more consistent across countries than in QIS3. For most lines of business, most carriers calculated technical provisions using a deterministic projection of best estimate future cash flows. Calculations were done on a policy-by-policy basis, and best estimate assumptions were derived based on analysis of past experience. Best estimate cash flows were generally discounted using the interest rate curves supplied by CEIOPS.
For lines of business that include embedded options and guarantees, stochastic techniques tended to be used with a “model point” approach.2 Carriers in some countries, though, used closed form solutions (such as the Black-Scholes model). In general, carriers in many countries supported market consistent valuation of options and guarantees. Supervisors recognized that many small carriers would probably not have the resources required to implement a stochastic model for the valuation of options and guarantees at the current time. Some simplifications, therefore, may be forthcoming in this area.
For unit-linked business, technical provisions were set equal to the unit fund (i.e., applying a surrender value floor) or the unit fund less the present value of future profits emerging from unit-linked business.
Solvency Capital Ratio
Market risk is on average the largest component of the Basic Solvency Capital Requirement (BSCR) for life insurance companies, representing about two thirds after diversification effects are reflected. The largest component of this risk module is the interest rate risk sub-module, which forms more than half the market risk capital charge – followed by equity risk, which comprises about 44 percent.
For the equity risk sub-module, many carriers and supervisors stated that the 32 percent calibration of the equity stress was too low for a 99.5 percent calibration and suggested that a figure of approximately 40 percent might be more appropriate.
For the equity risk sub-module of the SCR, a “dampener” formula was tested for liabilities with a duration of more than three years as an alternative to the standard approach. The underlying rationale for this feature is that (a) the capital needed to cover a decrease in equity values for carriers with long duration liabilities is smaller than for carriers with short duration liabilities (based on the assumption of mean reversion in equity markets over time) and (b) the probability of an increase in the value of equity indices is small when the index is high (and vice versa). The dampener approach resulted in a reduction of around 13 percent in equity risk capital for life undertakings.
The duration aspect of the dampener approach was opposed explicitly by many supervisors and carriers. The main reasons given for this objection were a lack of theoretical and empirical justification, inconsistency with the 99.5 percent one-year VaR level and inappropriate incentives for risk management. There was a suggestion that procyclicality3 should be taken into account in supervisory intervention (i.e., Pillar 2) rather than in the capital charge. However, the French regulator, supported by the industry in that country, was in favor of having an equity capital charge determined in relation to the duration of the insurance liabilities (or to the holding period of assets) and to the current point in the financial cycle.
For life underwriting risk (which combines various subrisks such as mortality, longevity, sickness, lapses and expense inflation through a correlation structure), some concerns have arisen – especially on lapse risk calibration which was considered to be too high. It was also suggested that a scenario approach would be more appropriate for lapse risk, allowing impacts on differing lines of business to be captured: the stress would be taken as the worst of (up, down) scenarios for the whole undertaking.
Several carriers argued for an age- and duration-dependent treatment of longevity, reinforcing more general comments that a one-off shock is not the most appropriate form of stress for biometric risks. An improvement in mortality rates by a regulator-defined percentage (over base mortality) was suggested as an alternative.
The life catastrophe risk sub-module uses a factor-based approach to capture risks from extreme or irregular events (e.g., a pandemic). Some carriers suggested distinguishing by health status (e.g., smoker/non-smoker) and allowing for regional diversification. The possibility of applying entity-specific parameters for catastrophe risk was suggested by some carriers, particularly in the context of annually renewable contracts. The application of this module in the case of reinsurance was questioned, particularly where the coverage is based on specified restrictions (e.g., time period, number of injured, or type of claim).
It was observed that the adjustment for loss absorbency through profit sharing appears to be one of the key elements in the calculation of the SCR for life and health insurers. There is an adjustment in about half the countries, and its impact is material (i.e., a reduction of more than 5 percent of the BSCR) in about a third of the countries, with a wide range of values (reduction of 5 percent to up to 75 percent).
An area of particular importance in QIS4 has been the collection of information on internal models. To this end, CEIOPS asked participants to provide information on the current and future potential use of internal models. Approximately 50 percent of solo participants provided some information on internal models. This finding is an increase from the 13 percent of participants submitting information on internal models under QIS3. Only 10 percent of carriers provided quantitative results from their internal model calculations in QIS4.
Some carriers found that the standard formula works reasonably well and therefore would not consider developing an internal model at this stage (13 percent of the respondents, around 90 carriers). Full and partial internal models are a possible route for many carriers (63 percent would consider using a partial or full internal model under Solvency II, around 450 carriers), though companies are in different stages of development. The key drivers for the development of an internal model are better risk management and governance.
Carriers were also asked to benchmark their internal models to the SCR standard formula in order to compare internal model outcomes with the QIS4 results. For all companies (i.e. life and non-life), the internal model results for the total solvency capital requirements were lower, with half of the carriers expecting a 20 percent decrease in their total capital requirement. Some individual risks seem to generate on average a lower capital requirement than under the standard formula (e.g., interest rate risk, longevity risk, lapse risk or premium, and reserve risk). Other risks would require higher capital charges when calculated using an internal model (e.g., operational risk, equity risk or property risk). For life companies, the life underwriting risk sub-module of the SCR gave a 30 percent lower risk capital charge using internal models than the standard formula on average.
With regard to group internal models, only seven groups provided complete data on a group SCR calculated with internal models. Therefore, no general pattern can be drawn. For those groups, on average there was only a very small difference between the internal model SCR and the standard formula, but there was a wide range of results. A number of groups reported a higher group SCR when applying their internal models. General comments on the use of partial or full internal models — and on the reasons for developing an internal model — are the same for group and solo entities.
There are several important considerations when evaluating the findings regarding internal models:
- Internal models vary among companies.
- Carriers may structure their risks differently than what is foreseen in the standard formula.
- Internal models may apply different correlations among their risks than the ones prescribed in the standard formula.
- Internal models may include risks that are not considered in the standard formula, hence raising capital requirements. The opposite also applies, where companies did not provide information on specific risks as they are not included in their internal models.
- None of these internal models would currently be considered as fully “Solvency II compliant.”
Under the QIS4 approach, recognition of risk mitigation techniques in reducing capital charges is supported. A set of principles on financial risk mitigation tools has been laid out. However, it is not yet clear what risk mitigation techniques are allowed and how the reduction would apply in practice. In particular, the implementation in QIS4 for non-proportional risk mitigation techniques, such as non-proportional reinsurance and securitization, is not accurate. Such an objective cannot be reached in the standard formula framework proposed by Solvency II and the QIS grids, whereas it can be reached in the frequency/severity approach supported by the Swiss regulator in the Swiss Solvency Test (SST).
At present, a (partial) internal model is the only approach to fully recognize the effect of non-proportional risk mitigation techniques. This approach is increasingly supported by the different stakeholders to Solvency II, but it will be necessary to convince local regulators of the adequacy of risk transfer and the industry’s ability to measure it.
The QIS4 exercise was based on 2007 balance sheet figures and therefore does not reflect the current balance sheet distortions caused by the financial crisis. A QIS5 exercise was originally planned for early 2009, but in view of the current turmoil in the financial market, its consequences on the parameterization of the standard formula, and to free up the necessary time to deeply review the results of QIS4, QIS5 has been postponed until early 2010.
Status of the Political Process and Group Support
The most recent draft Solvency II framework directive, written by the EU’s Economic and Financial Affairs Council (ECOFIN), abandons the notion of “group support,” originally a key concept of the directive. Group support acknowledges the benefits of geographic diversification for groups that have subsidiaries and risks in various countries by recognizing that, for a group, less risk capital is necessary than the sum of risk regulatory capital requirements across the local entities. The original proposal allowed local entities to hold less than the SCR (target capital) than they would have needed on a standalone basis, provided that this lower capital ratio was supplemented by group support. It required that capital adequacy be supervised primarily group-wide under the responsibility of a lead supervisor located in the country in which the head of the group is domiciled. According to an informal agreement on the compromise Solvency II framework directive reached at the end of March 2009, group support will be scoped out of the initial launch of Solvency II but will be picked up again three years after the launch. Formal adoption of the Solvency II framework Directive by the European Parliament and ECOFIN is likely to take place beginning of May 2009.
A Parliamentary committee approved a version of Solvency II in October that included group support. The European Commission also views it as a key element of the proposed new capital regime for insurers. Negotiations will now begin between the ECOFIN and the European Parliament to reach a compromise that can be put to a vote of the full Parliament in February. The European Commission has stated that it is still confident that a joint text can be agreed and the Framework Directive adopted before the European Parliament elections in June 2009.
In 2012, Solvency II will introduce a new (and, for many, fundamentally different) approach to the calculation of capital adequacy for European insurers. But Solvency II is much more than a simple compliance exercise. The adoption of the Solvency II risk management philosophy and the development of sophisticated internal models will allow companies to analyze and manage the risks their businesses face more effectively. Profitable risks can be managed through internal processes, reinsurance and capital market activity. Non-performing risks can be eliminated. Each Euro of capital must earn its return.
Guy Carpenter is in an excellent position to help our clients succeed in this new and challenging environment with our long standing, in-depth knowledge of risk management solutions and access to advanced actuarial modelling software such as MetaRisk®.
- CEIOPS report on Quantitative Impact Study 4 (QIS4) for Solvency II, page 6.
- A model point is a single policy which is used to represent a group of similar policies within a larger portfolio. Instead of modeling each policy in the group individually the model point approach involves modeling a single representative policy and scaling up the result in line with total sum assured.
- Procyclicality here refers to the tendency of insurers’ actions in the capital markets to amplify existing movements. For example, a fall in the equity market may reduce the value of an insurer’s asset portfolio, reducing its excess capital. To recover its capital position, the insurer is forced to reduce its investment risk by selling equities and buying bonds: its actions cause the equity market to fall further and a vicious cycle ensues.
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.