28 June 2010
28 June 2010
Joint CRO Forum and CFO Forum detailed feedback on the draft IMs discussed at the Expert
Group meeting on 14 June 2010
This note was written as a joint effort by the CRO Forum and the CFO Forum, describing our view and key
issues regarding the draft implementing measures.
We would like to thank you for the invitation to comment at this critical point in time in the development of
the Solvency II Directive. This letter is a follow up to our previous high level comments sent to you on 11
June 2010 when we provided you with our most immediate concerns ahead of the Expert Group meeting
on 14 June 2010.
As expressed in our initial feedback sent to EC on 14th June, Main concerning IMs are IM13 - technical
provisions (definition and scope of boundaries of contracts, diversification in the risk margin between
entities), and IM34 - spread risk (calibration & design of bonds, credit derivatives, structured products,
covered bonds, sovereign).
For your convenience, we have also prepared a revised mark-up on these draft IMs that present our
suggestions and concerns.
We look forward to discussing this further with you.
IM13 – Technical Provisions
Article TP2: Boundaries of insurance and reinsurance contracts
We are disappointed that the cover note to the Solvency Expert Group does not adequately represent the
industry perspective on an issue that is fundamental to the measurement model underlying Solvency II. In
our view the issue at stake is a wider point of principle and not just associated with the interpretation of
"unlimited ability". Notwithstanding our discussions on this issue we are of the view that the respective
positions are not yet fully understood.
Consistently with Alberto Corinti‟s note circulated on 23 rd June, we believe that the proposals for QIS5 and
the draft Implementing Measures1 differ from the Industry position in that they exclude all cash flows after
the point at which the insurer can change the premiums charged, even if the insurer cannot re-underwrite
at individual policyholder level (and so treat the policy as a new contract). This means that it will be
excluding the value/ cost of the contract within the technical provisions, as well as excluding the risks
associated with this contract within the SCR. This is not in line with the current methodology used by
insurers in their Internal Models. Insurers have an ongoing obligation to the policyholders of these contracts
and it is not at all considered appropriate to ignore the risks associated with these.
We should note that the Industry position would be expected in many cases to result in a more prudent
treatment that the current proposals drafted by the Commission. Future premiums likely to be affected and
excluded from the technical provisions computations (and related SCR calculations) are group insurance
contracts where the contractual premium to be paid by the insured cannot be re-priced at the individual risk
profile of the policyholder (e.g. many mortgage insurance policies) and a significant portion of universal life
Both QI5S and IM13 seem to differ from the industry position and also appear to be inconsistent with
28 June 2010
As a result, we believe that at least TP2 3. should be deleted as not in line with the principle supported by
the industry that premiums that cannot be repriced at the individual risk profile of the policyholder should be
included in the future cash flows to be projected out in the technical provisions. TP2. 2. should be deleted
as adding confusion. In addition, a recital should be added in order to clarify the intention of TP2.
"In determining the boundary of a contract it is important to reflect the obligations of the insurer/reinsurer.
An insurer/reinsurer is considered to have a contractual obligation to a policyholder up until the earlier of
the point at which the insurer/reinsurer either is no longer required to provide coverage; or has the right to
reassess the risk of the individual policyholder and, as a result, can set a price or amend the benefits to
fully reflect that risk. The boundary of a contract is the point at which the insurer/reinsurer's obligations
Article GTP: Risk margin
We would like to reiterate our concern that the risk margin does not recognise group level diversification in
the current draft. Only by its inclusion can the risk margin fully reflect the underlying economic principles of
market consistent valuation.
This is a critical issue. An economic value must reflect what is being transferred. As the CEA report clearly
demonstrates, in nearly all cases companies or group of companies rather than individual business lines
are transferred into another group structure and not into an empty shell, where group diversification benefit
will continue to prevail. In addition, from the conceptual perspective on transfer value, the potential buyer
would only be compensated for non-diversifiable risks, ie after recognition of all diversification effects. This
demonstrates that diversification within the group should be fully taken into account within the risk margin
at group level as the transfer value is the value of the company. In addition, we believe that diversification
benefits between life and non-life companies are appropriate and therefore would like TP18 (b) to be
deleted. This full group diversification in the risk margin shall be explicit on the fact that the company or
single entity would not be sold or transferred to an empty shell, but into another group structure which is
Article TP18: Residual market risk
With regard to calculating the risk margin, we would like to restate our support for the assumption that
assets have been selected so the SCR for market risk has been eliminated (rather than simply minimised).
The residual market risk may be minimal and difficult to estimate, so assuming its exclusion is desirable.
When difficult to assess or expected to be minimal, in such cases, we suggest undertakings would
have to document in the context of Pillar II processes the absence of residual market risks being
taken into account.
Under Article TP 18 paragraphs (g) and (h) the unavoidable market risk is still included in the Risk margin.
We recognize that some discussion may still be required on how to practically treat the question of
unhedgeable risks in the risk margin in relation to the question of the definition of the discounting curve and
its extrapolated part. This can be done within QIS5. However it is imperative that any solution needs to be
practical and guarantee no double-counting of risks with the SCR. In particular, the interest rate shocks in
IM27 have also not been adjusted to exclude the non-hedgable interest rate part. If this is not addressed
then the unavoidable market risk will be accounted for three times, once in the risk margin, once in the
extrapolation and once in the interest rate SCR. We have already suggested amendments to IM27. For the
consistency with the extrapolation we suggest to add text that the Risk margin for unavoidable market risk
should be set in a consistent manner with the risk incorporated in the extrapolation.
Tax: TP18.1.(j) “the is no loss-absorbing capacity of deferred taxes as referred to in Article 108 of Directive
for the reference undertaking” This should be aligned with the other items in Art 108. The loss-absorbing
capacity of deferred taxes should be the same as in the original undertaking. This is more in line with
economic reality and would align with (i). The concrete drafting suggestion would be to delete this
Article TP26: Lines of business
The definition of lines of business proposes segmentation by Member State. We believe this requirement is
excessively granular and will be onerous for undertakings to implement. Further, it is not clear why
segmentation in addition to that set out in Annex I is required as this is not a requirement of the Directive.
IM20 –Technical Provisions
The text (in IR2 (3)) states that government rates shall be used when swap rates no longer available/not
liquid as set out in TP21.
28 June 2010
A government bond curve should only be considered if it is reliably available at longer tenors than swaps.
Further, there should be a smooth transition from swaps to government bonds if this applies to avoid
The illiquidity premium should apply to the forward curve rather than the spot curve. Please see our QIS 5
technical specification on the risk free rate dated 1 April.
Article IR6: Illiquidity premium observed in the financial markets
We reiterate our response that the calculation of the illiquidity premium observed in the financial markets
should be based on a clear methodology that builds on the experiences and data in that market and avoids
uncertainty around how the calculation is carried out based on those observations. Consequently, we
strongly believe that the formula based on a transformation of the observed credit spread, as developed by
the CEIOPS led taskforce (reported 1 March) and presented in our QIS 5 risk free rate technical
specification (1 April), should be included in the Level 2 text.
We further recommend that the calibration of the illiquidity premium formula as per IR6 is considered as
part of the review mechanism outlined in IR9. We note also that the calibration of the formula mitigates
undue pro-cyclical effects with the illiquidity premium being significant in periods of stress and otherwise
close to zero. It would be prudent to review the calibration after a period of time to ensure this structure is
working in different market conditions.
Critically, this would also negate the need for anybody (be it EIOPA or anyone else) to be tasked with
determining whether a market is or is not in “stress” which causes uncertainty and runs the risk of
exacerbating the volatility in the market. Introducing such uncertainty cannot be a sensible basis for the
regulatory framework as it requires capital to be assessed on an on-going basis. We do not believe
anybody (including EIOPA) should make this subjective judgement. The formula (subject to periodic
review) should be allowed to operate automatically to determine the level of illiquidity premium relative to
The lack of an explicit formula and this text puts a lot of uncertainty to the industry of when illiquidity
premium can be applied. So we object this text as it does not recognise that liquidity is in fact a continuum.
Article IR7: Portion of the observed illiquidity premium corresponding to insurance or reinsurance
We welcome the developments to date over the application of the illiquidity premium to liabilities and
recognise the increased scope of liabilities that will attract an illiquidity premium. With reference to our
paper to the EC dated 20 May, we believe that a 75% portion of the illiquidity premium should be applied to
other products with discretionary benefits, including, US style fixed annuities.
As noted in previous responses, we welcome the recognition that in an economic based approach the
liabilities would be re-stated in stress conditions and would therefore reflect the changing liquidity premium
in the case of the stress of spreads. We are therefore concerned to see that this is not recognised in the
most recent relevant draft IM on spread risk. The CFOF would be happy to work with the Commission to
ensure that this reflects the recognition of the LQP in the base case and in particular the "predictability
buckets” and the interaction with the transitional provisions to ensure a consistent treatment.
Article IR8: Transitional Provisions
CFO Forum and CRO Forum will come back to the Commission at a later stage on general principles to
be applied to transitional provisions.
Transitional provisions are welcome and should be re-assessed once all implementation measures
IM34 - Concentration Spread Risk
This is a very concerning IM. We strongly recommend a decrease in the spread risk factors (bonds, credit
derivatives, structured products, covered bonds, sovereign).
As presented in our CRO Forum study on market risk calibrations (March 2010), there are 2 options for
calibrating the corporate bond spread stress tests:
Option 1 is a recalibration of the same stress test framework proposed by CEIOPS, based on
CDS. It involves an alternative calibration to the Credit Default Swap (CDS) market removing some
of the biases in the original CEIOPS calibration, based on more representative indices.
Option 2 is a more theoretically sound approach but more complicated for implementation. This
method involves calibrating to Corporate Bonds but allows for a liquidity spread widening offset on
28 June 2010
the liabilities. The degree of liquidity allowed for in the liabilities depends on the predictability of
their cash-flows. It follows the work of the CEIOPS Liquidity Premium Task Force.
Since the illiquidity premium has been omitted in the spread module, the spread shocks should be
calibrated net of the liquidity premium. Therefore we propose that the draft IM34 uses the CRO Forum
proposed calibration for option 1. (See marked up IM attached)
We also propose reductions in Maximum duration (to values in CEIOPS‟ final advice), with the Max
duration on a single exposure without credit assessment being 6 years.
• CDS (not qualifying as hedges) subject to 600% spread widening or 75% spread narrowing. We
strongly disagree with this arbitrary approach. In the calibration paper, CEIOPS focused on the CDS
sold by institutions that defaulted during the crisis which created inconsistency with the calibration and
the treatment of bonds. The issue of the default of the counterparties in CDS transaction is more
relevant in the calibration of the counterparty sub-module.
• Credit Default Swaps should be treated in the same way as corporate bonds (i.e. based on the rating
of the underlying name). For the calibration of CDS in the spread module, we recommend using the
same calibration than for bonds – what we call Option1 (ie. CDS data and not real bonds index). The
credit quality should be set equal to the rating of the underlying bonds of the credit derivative while the
duration should be set equal to the maturity of the credit derivatives.
• We propose that maximum durations should equal the values in CEIOPS‟ final advice. (See marked up
• A different CDS shock would give the wrong risk management incentives: companies will optimize their
SCR by including hedges either in the bonds module or in the CDS module (if there is a large part of
CDS where protection is sold, it will be cheaper to treat hedges in the CDS module since they will lead
to a netting).
• Although in IM34 the method for Structured assets has been simplified (no minimum 10% and no
recovery rates), the structure is still very complex and, more importantly, incorrectly calibrated. The
current Risk Weights are calibrated based on AAA level which corresponds to a 99.98 confidence level
which is inconsistent with other modules being calibrated to a 99.5%.
• The CRO Forum would be happy to provide the evidence that the proposed calibration based on the
underlying assets is incorrect. We strongly propose that the methodology referring to the rating of the
asset pool should be removed. We therefore propose a deletion of the method based on the rating of
the underlying assets, and including methodology based on the rating the actual structured credit using
CRO Forum proposal. (See marked up IM attached)
• As expressed in previous contribution, we believe that shocks of the structured credit should be based
on the actual rating of these instruments. The CRO Forum proposes that the stress tests from our
Option1 based on the rating of the structured products are re-calibrated to be equivalent to the
corporate bond stress tests multiplied by a scaling factor of 1 for investment grades (ie. applying
directly the corporate bond stress) and a scaling factor of 1.5 for other ratings (ie. applying 1.5 x
corporate bond stress for BB and ratings below).
• We propose that maximum durations should equal the values in CEIOPS‟ final advice. (See marked up
• We continue to believe that the requirements for the insurance company to prove that the originator
retains on a continuous basis a 5% economic interest in structured finance instruments (Art. SR7.2) is
very difficult to apply in practice. Consequences with non-compliance are punitive.
• The text proposes to have different treatments of rating of covered bonds under spread risk and
concentration. This is inconsistent and contradicts the rating in CRD (no distinction between AAA and
AA). We find that both 1A and 1B covered bond ratings should have a reduced spread shock of 0.6%.
(See marked up IM attached)
• We think that the very different design of covered bonds (safeguard mechanisms such as collateral
pools) in different countries renders a more individual treatment more adequately. Thus, the shock
levels for country-specific covered bonds (Pfandbriefe, UK covered bonds, foncière, cedulas, etc.)
should be calibrated based on an analysis similar to the one carried out for the corporate bond.
Similarly, the different design of hybrids has also impact on its classification into tiers such that there
should be an analogous, more sophisticated and differentiated treatment of covered bonds regarding
the spread risk.
28 June 2010
• We do not understand the logic of excluding the mortgage module. The fact is that many insurance
companies have significant holdings of residential mortgage loans and that should be recognised in the
Govies/ Treatment of Sovereign credit risk:
• For the treatment of spread for govies, we note inconsistency between the cover memo (where it‟s
written that EC has decided to remove that OECD govies should not be subject to spread risk, but ask
to refer to the treatment in CRD – govies issued by a country with a rating A or lower should have a
shock) and draft IM (that do not refer explicitly to CRD).
• We propose that:
• EEA govies (even if the rating is A or lower) should not be shocked
• Sovereign bonds from non-EEA governments that invested to back liabilities in the same currency
are exempted form shocks.
• Sovereign bonds from countries considered as competent authorities (even with a rating A or
lower), should not be shocked [eg. Korea is considered as competent authorities under CRD]
• Sovereign bonds from non-EEA governments with a rating A or lower, should have spread shocks
set equal of those of corporate bonds with a rating one notch higher.
Concentration risk for bank deposits:
It is unclear why insurers are required to consider concentration risk against bank deposits, the most liquid
form of assets. Moreover, the criteria defined in the CO6(5), such as government guarantee schemes are
usually reserved for consumer protection and hardly ever extended to industry. Therefore, the structure of
the proposal gives very little chance, if any, to insurers to ever meet the criteria.
In addition companies in general, including insurers, tend to build up a reserve of cash in times of stresses
and the monies wait on the side lines waiting for the market to settle. This was clearly evident in the recent
market upset. Therefore for an insurer to have to require further capital in times of stresses could result in
elements of procyclicality being introduced into the SII framework.
Moreover, as a “customer” of the bank, insurers should be sufficiently protected under the revised Basel III
framework. Therefore to hold capital against a very liquid asset held in a bank governed by Basel III makes
very little sense.
Therefore we propose that bank deposits should not attract a concentration risk charge.
IM35 - Health UR
• In general, we recognise that this has been a really complicated process, due to the complexity and
variety of the products over the various countries, and due to the limited availability of data, and we
advise to continue the process of calibration in order to improve robustness of the calibration factors.
• Health Risk Equalisation Systems (HRES) are now recognised, which is important. A positive aspect is
also that the scope of the HRES is not limited to one Line of Business within Health, but may be
• However the applicability is limited to the area of premium risk. In our opinion, this should also be
extended to reserve risk.
• We have seen no adjustment on the use of Undertaking Specific Parameters (USP‟s) in the standard
formula. The industry is of the opinion that more credibility could be attributed to the use of USP‟s in
the health segment.
• The issue regarding segmentation of Health business is now included in IM 13, and is generally in line
with what was asked by the industry. Segmentation remains a complex issue for health insurance
across Europe, as there is a large variety of insurance forms that are covered under health.
• On the calibration we should reiterate that because of the complexity and variety of this segment, there
is an ongoing need to continue the process in order to improve robustness of the calibration factors.
IM36 - SCR Simplifications
Proposals for SLT health shouldn‟t be simply copied from life, since special risk measures i.e. CaR (Capital
at Risk) or surrender value don‟t exist for SLT health. We suggest reworking the proposal and will come up
with an alternative proposal to the EC.
IM37 - Procedure For Updating Correlation Parameters
It would be very burdensome for companies to report all quantitative data relevant to update all correlation
parameters within the annual report to the supervisor. The CRO Forum suggests establishing - prior to a
review by EIOPA of the correlation parameters in the standard formula - a joint industry and supervisors
working group to ensure that appropriate quantitative information are provided to the supervisory
28 June 2010
IM38 - MCR
There is a general issue regarding the calculation of the MCR. On the one hand, it has to be simple by
following a linear approach and on the other hand, it has to be risk-sensitive in order to ensure consistence
with the calculation of the SCR in order to enable an appropriate ladder of supervisory intervention. These
objectives cannot be achieved at the same time sufficiently which is the reason why a cap and a floor is
introduced in the calculation of the MCR. We observed that the current determination of the MCR lacks
consistency with the calculation of the SCR even where consistency would have been possible to be
ensured and all these issues will lead to an unjustified level of prudence leading to a more regular
application of the 45% SCR cap in the calculation of the MCR. The main issues are as follows:
The calculation of the MCR is based on a linear approach where presently diversification benefits are
not accounted for. This will lead to an unjustified level of prudence leading to a more regular application
of the 45% SCR cap. The solution of this problem would be to introduce a factor on every level of
aggregation in the calculation of the MCR that represents the diversification benefits. This factor should
be derived from the calculation of the SCR (standard approach). This approach would preserve the
linear approach prescribed in the Directive and it would allow for diversification benefits in a prudent
way as it is reasonable to assume that there are less diversification benefits at the 99.5%-quantile than
at the 85%-quantile (MCR calibration target) due to non-linear dependence structures including tail-
The proposed method does not take into account loss-absorbing capacity of deferred taxes. Although
the Directive allows for not using all variables mentioned in Art. 129(2), among them deferred taxes,
omitting deferred taxes would introduce an inconsistency with the calculation of the SCR and thus
introduce an unjustified level of prudence leading to a more regular application of the 45% SCR cap.
The cover note of the Solvency Experts Group from 14 June 2010 does not provide any justification for
omitting deferred taxes in the calculation of the MCR. We strongly propose to take the loss-absorbing
capacity of deferred taxes into account.
Currently, it seems that internal models are allowed to calculate the MCR. A reliable ladder of
supervisory intervention can only be established when the MCR is calculated consistently with the
SCR. Otherwise, the only consistency with the calculation of the SCR, when the latter is determined
according to an internal model, would be the cap and the floor applied to the MCRlinear. Although the
Directive only refers to capital add-ons, replacements of a subset of parameters or the use of an
internal model in case where the “risk profile of the insurance or reinsurance undertaking concerned
deviates significantly from the assumptions underlying the standard formula calculation” (cf. e.g. Art.
110) with exclusive reference to the SCR and not to the MCR, the calibration target of 85%-VaR
referred to in Art. 129(1)(c) cannot be achieved by a standard formula MCR when it is already clear the
standard formula SCR is not appropriate for the calculation of the undertaking-specific SCR. By
insisting on a standard formula MCR in this case, the resulting MCR would be meaningless, the
calculation would be necessary but irrelevant as generally the 45% of the (internal model) SCR cap
would apply which then even would have to be explained by the undertaking by Art. 129(4).
Inconsistent treatment of composite insurance undertakings
The calculation of the MCR according to the standard formula is done by first adding all linear Minimum
Capital Requirements and then applying the cap and the floor to arrive at the combined MCR whereas
for composite insurance undertakings the cap and the floor are applied before aggregating the life and
the non-life part. This treatment is inconsistent. As the formulas for calculating MCRNL and MCRL are
less sophisticated than the formulas for calculating SCRnl and SCRlife, we propose to take over the
approach currently proposed in MCR6 to the standard formula for the MCR in order to obtain some
basic alignment between the intermediate MCR- and SCR-results. This means there should be a
combined MCRNL and MCRL with caps and floors taken into account in Articles MCR3 and MCR4.
IM39 - Equivalence
The provisions have adopted wording in the criteria to be assessed that contain greater prescription
than set out in the CEIOPS advice. We encourage the commission to maintain a principle based
approach in the IM and use level 3 (subject to transparent consultation procedures corresponding to
the standards of the European Commission) to identify the indicators proposed by CEIOPS, and
subject to consutation, that will help assess whether the substance of a third country regime is
equivalent. We have therefore proposed some rewording and removal of the third requirement under
„System of Governance and Public Disclosure‟.
28 June 2010
It is important that it is clear that the objective of equivalence is to encourage convergence rather than
achieve convergence to enable equivalence assessments to be dynamic while recognising the
differences that may exist between different markets.
The industry in its response to CP78 set out the need for a transitional to address concerns that key
jurisdiction may not meet the criteria to be determined as equivalent and the failure to be equivalent
could have material commercial implications. We encourage the Commission to set out clearly its
proposal to address this concern and are prepared to work with the Commission to establish a
workable solution that promotes convergence.