1.  In preparation for Solvency II, the Committee of European Insurance and Occupational Pensions Supervisors (CEIOPS) conducted its third Quantitative Impact Study (QIS3) in April-July 2007. In this report, we summarise the UK results of that study, and highlight the main issues identified by the UK insurance industry. CEIOPS expects to publish its report on the QIS3 results on 20 November 2007. CEIOPS' report is available on its website: www.ceiops.eu.

OBJECTIVES

  2.  The overall objectives of QIS3 were to test the financial impact on firms and suitability of proposed requirements of Solvency II, such as technical provisions, own funds, the Minimum Capital Requirement (MCR) and Solvency Capital Requirement (SCR). For the first time, QIS3 tested proposed group capital requirements, as well as solo requirements.

  3.  The Solvency II framework has progressed substantially during 2007, with publication of the draft Directive (Level 1) in July, along with the Commission's Impact Assessment [ref, footnote]. QIS3 tested key aspects of the proposed Level 1 text, along with CEIOPS' current thinking on the Level 2 implementing measures that will provide the underpinning detail. CEIOPS is due to present its final advice to the Commission on Level 2 in October 2009.

PARTICIPATION

  4.  UK participation in QIS3 was considerably higher than for QIS2, with 39 life firms and 46 non-life firms participating, comprising 65% and 75% of the market by premium income respectively. Although participation was skewed towards large and medium-sized firms, enough small firms took part to enable us to draw appropriate conclusions. In addition to the quantitative returns, most firms also completed the detailed qualitative questionnaire, including comments on practicability and resourcing implications. Six large and five medium-sized groups submitted group responses.

KEY FINDINGS

  5.  QIS3 showed an overall reduction in firms' solvency ratios compared to Solvency I, but this was expected in view of the significant known deficiencies in the risk sensitivity of the Solvency I requirement. Overall, the industry would have a substantial buffer of capital in excess of the SCR, but the effect varied between firms. Over 80% of UK firms had a surplus of available capital over the standard SCR as proposed in QIS3.

  6.  In addition to submitting calculations as set out in the QIS3 specification, firms were also invited to supply internal model results. Over a half of UK respondents also sent internal model results, based on their ICAS work. This was particularly helpful in supporting our analysis of the suitability of the standard approach SCR.

MCR ISSUES

  7.  A key objective under QIS3 was to evaluate CEIOPS' proposals for a modular MCR. The Commission's proposed Level 1 directive text sets out an MCR calibration range of: 80-90% level of confidence that, over the time-horizon of a year, a firm's assets will remain adequate to meet its liabilities. The Level 1 calibration range indicates that a firm's MCR might be expected to be around 35% of its SCR. In advocating the compact approach for the MCR (under which the MCR would be directly calculated as a proportion of the SCR) the industry has suggested calibration at a similar level: both to allow an adequate operation of the supervisory ladder of intervention, and of SCR internal models.

  8.  QIS3 results for UK firms have revealed a substantial variation in the ratio of the MCR to the SCR, indicating that the modular MCR tested by CEIOPS was insufficiently risk-sensitive. This result was particularly acute for life firms, for which the adjustment for profit sharing was a significant element. Both life and non-life firms voiced concerns that:

(a)  the MCR was not calibrated in line with the draft Directive;

under the modular approach the SCR and MCR would not move consistently from year to year, so providing problems for capital planning.

  9.  CEI0PS is therefore considering several alternatives for the design of the MCR.

SCR ISSUES

  10. The most significant issues with the standard SCR identified by UK firms were:

(a)  the 75% lapse cat component for linked life business; and

(b)  non-life underwriting risk (premium and reserve) calibration.

  11.  In both cases, a comparison of QIS3 results and internal models/ICAS results indicated that the standard SCR overstated the risks of these firms and materially adversely affected their reported solvency ratios. Non-life firms commented on the lack of transparency in CEIOPS' calibration work—on average, the standard non-life underwriting SCR was in excess of 160% of modelled capital requirements. Linked life firms raised similar calibration concerns and questioned potential double-counting of risk addressed in the life underwriting module.

  12.  Comparison with modelled results also identified some areas where the standard approach SCR may understate risk, including:

(a)  KC profit sharing adjustment for life business-firms suggested calculating the Basic Solvency Capital Requirement (BSCR) net of profit-sharing;

(b)  credit risk scope—a number of classes of asset were omitted; and

(c)  operational risk—firms were concerned that the standard approach does not recognise investment by firms in risk management.

  13. Other significant issues firms raised about the suitability of methodology included:

(a)  diversification effects between business written in different countries (diversification in all respects was a major issue for Groups); and

(b)  potential exclusion of free assets for calculating the market risk component.

TECHNICAL PROVISIONS

  14. There were relatively few reported problems in calculating best estimate provisions. However, there were several comments on the calculation of the risk margins under the Cost of Capital approach, and the resulting margin was often considered to be too high to meet the principle of market consistency. Comments on the method related to:

(a)  knock-on effects of over-calibrated aspects of the SCR;

(b)  the inclusion of a component for market and premium risk in year 1;

(c)  the absence of any allowance for potential diversification between lines of business.

  15.  A large number of comments were raised on the 6% cost of capital factor specified in QIS3, including whether a single factor was appropriate for all lines of business, and firms questioned what work had been done to calibrate the factor.

OWN FUNDS

  16.  Some firms said they found the proposed criteria for classifying own funds unclear, and that some of the criteria could conflict with one another, As a result, there was considerable uncertainty about the classification of some financial instruments.

GROUPS

  17.  It was difficult to draw clear conclusions for UK Groups, given the diverse nature of their business and the fact that this was the first QIS for groups. In overall terms, most UK Groups reported a reduction in their overall solvency ratio compared to the current requirements under the IGD. This reduction was largely attributable to differences between QIS3 and Solvency I technical provisions standards and solo capital requirements.

  18. For UK groups, a comparison of the sum of aggregated solo SCRs with the SCR calculated on consolidated group data revealed that the reduction in the group SCR as result of group diversification benefits averaged around 5-10%. A further 10-20% reduction was observed for those groups that provided internal model figures

QIS4 DESIGN AND SPECIFICATION

  19.  The European Commission is seeking input from all interested parties in the design and specification of the next exercise, QIS4. Currently, CEIOPS is drawing up a draft QIS4 specification, informed by the results of QIS3 and is due to pass this to the Commission in December. We expect the QIS4 draft specification to be released for public consultation during January and February 2008.

FEEDBACK ON QIS3

  20.  Since completing QIS3, we have invited participating firms to detailed feedback seminars, and FSA staff have spoken at a variety of industry events. The support of the ABI and other trade associations, as well as the industry more broadly, has been very important to the overall success of the QIS3 exercise.