Continued from the Minimum_Capital_Requirement_2.htm

1/ Modular approach
19. QIS3 participants were asked to calculate the Minimum Capital Requirement (MCR) as the aggregation of capital charges covering market risk and underwriting risk, adjusted by deducting a reduction for profit sharing (RPS).

The two modular approaches tested in the QIS3 shared a common base for the life and non-life underwriting components, both factor-based, with only a top-level adjustment to capture the risk mitigating properties, similar to the approach tested for the SCR in QIS2.

20. For market risk, two alternatives were tested under QIS3:

• Alternative 1: a simple factor-based approach based on asset-side volume measures;

• Alternative 2: a more sophisticated factor-based approach, taking also into account liabilities and durations.

Assessment according to the criteria

21. Simplicity and auditability:

The calculation is simple and the required input data is relatively simple, more so in the first alternative than in the second.

The first alternative is auditable (the calculation is based only on data available in the audited annual accounts, see also paragraph 53), while the second alternative is not.

22. Safety net:

It is a stand-alone calculation that provides capital in the Minimum Capital Requirement (MCR) to cover market and underwriting risks.

For these risks it provides a safety net for the SCR calculation, which can be subject to model errors.

23. Calibration and interplay:

The modular approach was calibrated to the defined confidence level. However, calculating the reduction for profit-sharing in a different way in the Minimum Capital Requirement (MCR) and the SCR (KC factor) seems to have been a source of poor interplay for life firms.

24. The modular Minimum Capital Requirement (MCR) is an independent calculation from the SCR and therefore there may be instances of the MCR being higher than the SCR.

QIS3 feedback – quantitative and qualitative

25. Quantitative results can be summarised as follows:

o For non-life firms, the results for both Minimum Capital Requirement (MCR) alternatives were broadly consistent with the calibration target and indicated an adequate interplay with the SCR.

In the first alternative, the Minimum Capital Requirement (MCR) nowhere exceeded 70% of the SCR.

Somewhat larger deviations and more outliers were observed under the second alternative, including some cases where the Minimum Capital Requirement (MCR) was higher than the SCR.

o For life firms, both Minimum Capital Requirement (MCR) alternatives showed poor interplay with the SCR in some countries, and wide differences between countries and individual insurers.

In those countries where reductions for profit sharing were used, negative MCR results were common (i.e. before applying the AMCR floor).

In the case of life firms, the results under the second alternative were more stable than under the first alternative, but even so they did not show a proper interplay with the SCR in all countries. The different approaches to market risk in the SCR and Minimum Capital Requirement (MCR) were also singled out as a cause for poor interplay.

In some members’ view, these problems are effectively mitigated if the second alternative of the modular Minimum Capital Requirement (MCR) is applied.

In life, these findings are largely attributable to the fact that the design of the MCR (factor-based) is very different from that of the SCR (scenario based) and that the calculation of the reduction for profit sharing (RPS in MCR and KC-adjustment in SCR) is one of the key factors distorting the interplay between the MCR and the SCR.

26. Qualitative feedback – regarding simplicity and auditability:

The first alternative relied on a small number of basic volume measures (like market value of equity, property, technical provisions, capital at risk etc.), so it was easy to calculate in QIS3. It could be considered as auditable in non-life.

Some respondents commented that the modular Minimum Capital Requirement (MCR) calculation is unduly complex; in particular, the calculation of market risk in the second alternative was difficult for some smaller firms.

There was also some uncertainty about how to interpret the adjustment for profit sharing.

27. Regarding the safety net function, some respondents noted that, for life firms, the negative (or, after applying the absolute floor, very low) results observed in life due to the deduction of the profit sharing adjustment do not provide an adequate safety net.

Most respondents and members regard these very low results as a flaw of the methodology, since the resulting MCR is unable to provide a meaningful protection for policyholders.

o However, one member commented that a negative Minimum Capital Requirement (MCR) may be a correct outcome from a risk perspective if future profit sharing is able to completely absorb losses in the context of the ultimate intervention6, and that that the absolute floor will override negative MCR results

28. Regarding calibration and interplay with the SCR: Respondents noted the lack of a proper interplay specifically in life, and expressed concern about the resulting poor supervisory ladder, the inconsistency between theMinimum Capital Requirement (MCR) and the SCR, and the difficulties this would entail in their capital planning and management.

A number of respondents also criticised the modular approach for inhibiting incentives to good risk management.

29. However, some members note that the Minimum Capital Requirement (MCR) is a requirement that is not constructed for risk management purposes. In practice, insurers should hold capital far in excess of the MCR, so that inconsistencies between MCR and SCR do not pose a risk management problem.

2/ A fixed percentage of the SCR

30. This approach, so-called compact approach, could be calculated in two ways:

Alternative 1: in its proposal for QIS3, the CEA suggests that the MCR should be calculated as a fixed percentage of the SCR of an insurer calculated by either the standard formula or an internal model.

Alternative 2: MCR calculated as a fixed percentage of the standard formula SCR, even when the firm uses in fact an internal model.

Assessment according to the criteria

31. Simplicity and auditability: The calculation is simple and requires minimal additional effort. There is no stand-alone MCR calculation required in alternative 1.

In alternative 2, while this is also true for the firms using the standard formula, firms using an internal model have a major administrative burden because they need to calculate the standard formula in addition to applying their internal model permanently.

32. The clear burden for firms using an internal model in the second alternative could be a disincentive to build an internal model.

33. The calculation is simple in appearance, but it relies on the very complex SCR calculation in fact, involving both a very high number of input data and complex calculations.

34. Interim calculation would be very difficult due to the complexity of the SCR calculation. To avoid this difficulty, the CEA suggested re-expressing the SCR as a percentage of technical provisions for interim calculations.

The CEA has argued that such reference to technical provisions would also ensure auditability.

In those cases however when (due to a significant change in the risk profile) the Directive
requires a full interim SCR update, the MCR would be automatically updated as well.

o Some members however comment that this shortcut adds another layer of complexity, making the MCR dependent on both the SCR and the technical provisions at the previous year’s end, creating uncertainty if the amounts are very different.

There is also a possibility that the SCR and the technical provisions would not move in parallel during the year.

35. The result of the SCR calculation cannot be audited.

o It could be argued that this would affect the legal certainty of the calculation.

o A few members envisage that the SCR could be made auditable in the future.

36. Safety net: This approach puts into question a fundamental design choice of Solvency II which is to have two capital requirements: an Minimum Capital Requirement (MCR) equal to a percentage of the SCR is equivalent to having only one capital requirement.

37. The Minimum Capital Requirement (MCR) will always be lower than the SCR, even when the latter is very low.

Therefore this approach would not provide a safety net against a very low SCR.

However, under this approach, the Minimum Capital Requirement (MCR) will provide capital to cover all the risks the company is exposed to and ensures that the MCR cannot be exceptionally low compared to the SCR.

38. Any potential flaws in the design of the SCR standard formula would be duplicated in the MCR, with the risk that 'ultimate supervisory action' is taken too late.

39. Alternative 2 provides more comfort for those supervisors who are uneasy about basing a safety net MCR on an internal model.

40. Calibration and interplay with the SCR: In the first alternative (and in the second alternative only for firms using the standard formula for their SCR calculation), proper interplay with the SCR is automatically ensured, allowing for a proportionate and escalating ladder approach to intervention (unless the absolute floor has to be used for the MCR).

41. The Minimum Capital Requirement (MCR) adapts instantly to any changes in the SCR design. Given that it is a simple percentage, the MCR is very flexible: if through experience the percentage is deemed too low or too high, then it can easily be revised.

It allows for risk mitigation to be taken into account to the same extent as in the SCR calculation.

42. Unless all distributions follow the same law, there is no linear relation between a x% VaR and a y% VaR. Therefore the Minimum Capital Requirement (MCR) defined as a percentage of the SCR, i.e. a percentage of the 99,5% VaR, will NOT reflect a uniform level of confidence,
unless all insurers share the same risk profile.

43. Link with the group support regime in the Directive Proposal: Supervisors, who are solo supervisors of firms belonging to a group under the group support regime, would lose the Minimum Capital Requirement (MCR) as control level if it is linked to the SCR, which is a
source of major concern.

Under the group support regime laid down by the Directive Proposal, the Minimum Capital Requirement (MCR) serves at the same time as the intervention level for calling on group support, and (in case this failed) as the trigger level for reverting to solo supervision and for the ultimate intervention.

Therefore the Minimum Capital Requirement (MCR) is a key control level for the solo supervisor.

Under group supervision, internal models are approved at the group level (Article 238).

Furthermore, according to Article 245, under the group support regime, other key decisions affecting the SCR – while they may be proposed by the solo supervisor – are ultimately delegated to the group level: these include the imposition of capital add-on on an internal model or the standard formula, and the decision to require a subsidiary to revert to the standard formula.

These decisions will have a direct impact on the Minimum Capital Requirement (MCR). The lack of a decision – either because the group and the solo supervisors disagree, or simply because of the delay resulting from involving the group supervisor in the loop – will also impact the enforceability and the timeliness of Minimum Capital Requirement (MCR) level intervention.

44. Link with capital add-ons: In the compact approach, it is an open question whether SCR add-ons should also apply to the MCR. There are two options, both
of which raise further issues:

o The first option is that (a fixed percentage of) a capital add-on would also apply to the Minimum Capital Requirement (MCR). A capital add-on, imposed by the supervisor, may bechallenged by an insurer before the court, thus complicating the enforcement of the MCR as an ultimate trigger.

o Alternatively, the Minimum Capital Requirement (MCR) could be exempted from the capital add-on. Given that an add-on is imposed when the SCR is an inadequate measure of the insurer’s risk profile, or when the insurer’s risk management is deficient, not applying the add-on to the MCR would arguably erode the level of protection provided by the MCR.

45. Small firm effects: For small firms, if the Minimum Capital Requirement (MCR) is linked to SCR internal models, concerns have been expressed that this could leave small market participants – who are less likely to benefit from diversification or internal model effects – at a
competitive disadvantage.

QIS3 feedback – qualitative
46. Most of the supervisors reported that QIS3 participants who commented on the CEA compact approach expressed support. In several countries, most respondents preferred this approach over the modular approach because it would provide a proper supervisory ladder.

47. In one country, an explicit comment was made on the preference of smaller firms, who preferred the modular approach in non-life to the compact approach.

Return to Index