Captives
and Solvency ii
Captive
Insurance and Captive Reinsurance Companies after the Solvency ii
Directive
from the Solvency ii
Association, the largest Association of Solvency ii Professionals
in the world
PROPORTIONALITY - LEVEL 2
MEASURES
CEIOPS-CP-45/09, 2 July 2009,
Consultation Paper No. 45
Draft CEIOPS’
Advice for Level 2 Implementing Measures on Solvency II: Technical
Provisions – Article 85 h, Simplified
methods and techniques to calculate technical
provisions
3.2 Simplified
methods
3.107 This sub-section
considers the use of simplified methods for the
valuation of technical provisions in the context of a proportionality assessment as described in the
previous sub-section.
3.108 At first, it
generally considers the role of simplified
methods in the valuation framework, in particular with regard
to whether it would be appropriate to include a
specification of simplified methods on Level 2.
It then analysis the
availability and applicability of simplified methods for two
specific components of the valuation where the use of simplified
methods is especially relevant.
3.2.1
Specification of simplified methods on Level 2
3.109 Article
85(h) of the Level 1 text states that:
“The Commission shall
adopt implementing measures laying down […] where necessary,
simplified methods and techniques to calculate technical provisions,
in order to ensure the actuarial and statistical methodologies
referred to in point (a) and (d) are proportionate to the nature, scale and complexity of
the risks supported by insurance and reinsurance undertakings
including captive insurance and reinsurance undertakings.…
”
3.110
It is therefore necessary to
consider:
• The
circumstances under which simplified methods would need to be
specified under Level 2; and
• For which
specific valuation components these circumstances would
apply.
Circumstances
which would necessitate specification of simplified
methods
3.111 As has been
previously mentioned, Solvency II envisages a principles based approach to the valuation of
technical provisions.
This means that the regulatory
requirements relating to the valuation process would generally not prescribe any specific approaches to
carrying out the valuation.
3.112 Instead, in carrying
out the valuation according to the principles-based requirements set
out in the Level 1 text the (re)insurance undertaking should be flexible to reflect the specificities
of its business and risk profile, taking into account the principle of
proportionality.
3.113 To support this,
supervisory guidelines on Level 3 could be
developed.
It is also expected that national and international associations of the
actuarial profession will issue guidelines on technical issues,
including the valuation of technical
provisions.
3.114 In light of this, it
may be argued that it would be excessive to include any detail on
specific simplified methodologies for the valuation of technical
provisions on Level 2.
There are a number of further
aspects which would support this view:
• It would generally
not
be possible to define any “default” or reference method for the valuation of
technical provisions; instead, a continuum of
methods would typically be available to the undertaking, differing
in their degree of complexity and
sophistication.
Hence any
choice of simplified methods would necessarily be incomplete
and to some extent arbitrary;
• Actuarial
methodologies and techniques for the valuation of technical
provisions are subject of continuous scientific research and development, so that it
is likely that a description of specific simplified techniques and
their application criteria would need to be regularly reviewed and
updated.
The legal
framework provided by Level 2 may not be sufficiently
flexible to achieve this aim;
• Inclusion of simplified
valuation techniques in Level 2 may be contradictory to the aim of incentivising
(re)insurance undertakings towards using more sophisticated
techniques, and may lead to an over-reliance on simple
techniques which may not be fully appropriate in all
circumstances;
• The
principle of proportionality provides an adequate framework
to ensure that undertaking s apply appropriate methodologies for the
valuation of technical provisions, including the use of simplified
techniques.
The expectations and
requirements on the undertaking’s proportionality assessment are
principles-based rather than rules based and would not need to be
supplemented by providing simplified methods on Level
2.
• CEIOPS supports the
establishment of European technical standards to be to be applied by
the actuarial function in exercising its
tasks.
As part of these tasks,
the actuarial function has to coordinate the
calculation of technical provisions and to ensure the
appropriateness of the methodologies and underlying models used
therein.
Therefore, such European
technical standards are expected to provide comprehensive support to
undertakings for calculating their technical provisions, and a high
level of convergence in the guidelines to be
used.
3.115 On the other hand,
not including to some extent a specification of simplified valuation
techniques on Level 2 may erode the intention of the Level 1 text
to achieve an increased harmonisation of
quantitative and qualitative supervisory methods, including
technical provisions.
3.116 In this regard,
recital 30 of the Level 1 text states that
(30) […] The
principles and actuarial and statistical methodologies underlying
the calculation of those technical provisions should be harmonised
throughout the Community in order to achieve better comparability
and transparency.
3.117 In the same vein,
recitals 11 and 33 stipulate that
(11)
[…] Harmonisation should be increased by providing specific rules
for the valuation of assets and liabilities, including technical
provisions.
(33) […] The
use of effective and harmonised actuarial methodologies should be
required.
3.118 A need to achieve harmonisation would be particularly
relevant in areas where the use of simplified methods in the
valuation is widespread, and where a common understanding of an
actuarial “best practice” is still evolving.
3.119 Not including any
detail on specific simplified valuation methods in Level 2 may also
raise concerns with respect to the needs of small and medium sized
(re)insurance undertakings.
Whereas it could be argued
that these needs are sufficiently addressed in a
proper application of the principle of
proportionality, an
inclusion of specific simplified valuation methods in Level 2 may be
helpful for small and medium-sized undertakings in
• Making
available valuation techniques which are tailored to the
specificities of their business; and
•
Providing legal certainty on the
appropriateness of such techniques under the Solvency II
framework.
3.120 This consideration
is also reflected in recital 34 of the Level 1 text which states
that
“In
order to reflect the specific situation of small and medium sized
undertakings, simplified approaches to the calculation of technical
provisions should be provided for.”
3.121 In light of the
considerations above, an inclusion of specific simplified valuation
methodologies in Level 2 should only be foreseen for components of
the valuation where:
• The
use of simplified methods is expected to be widespread, and a common
understanding of an actuarial “best practice” is still evolving;
or
• There is a
particular need for small and medium-sized undertaking s for such an
inclusion.
3.122 Where such
simplified methods would be specified on Level 2, they should be
available for all undertakings and be subject to appropriate
application criteria.
Such criteria
should
•
Specify the circumstances and conditions under which they are
intended to be used (in terms of the risk profile of the portfolio
to be valued); and
• Have
due regard to the model error inherent in an application of the
method.
Components of the valuation where such
circumstances would apply
3.123 Considering the
conditions mentioned in para. 3.121, we note that in the QIS4
exercise a number of simplifications and proxies were included in
the specifications, which covered the valuation of the best estimate
technical provisions, including the valuation of reinsurance
recoverables, and also the calculation of the risk
margin.
3.124 In QIS4, a widespread use of simplified methods for
the valuation of technical provisions was observed for the valuation
of reinsurance recoverables and the risk
margin.
3.125 It is expected that
in these areas there is also a need for small and medium-sized
undertakings to have simplified methods available.
Hence these valuation
components are analysed in more detail,
below.
3.2.2 Thresholds
determining the allowance of simplified
methods
3.126 This sub-section
discusses the extent to which it would be
appropriate to introduce external thresholds guiding the use of
simplified methods for the valuation of technical provisions.
The idea of such thresholds
would be to provide a cut-off point below which it would be regarded
as justifiable to use specific (simplified) valuation techniques (or
a class of such techniques).
3.127 Such external
thresholds may be specified:
• In
implementing measures on Level 2;
• As part of
supervisory guidance on Level 3; or
• As part of
supplementary technical guidance or
standards.
3.128 We note that
where external thresholds are introduced in Level 2,
these would be legally binding, which would not be the case for the other
levels.
This may lead to different
conclusions on the feasibility and appropriateness of an
introduction of thresholds on different levels.
In the following, the analysis
is focused on a potential inclusion of thresholds in Level 2
implementing measures.
Types of
thresholds to be considered
3.129 Where external
thresholds are considered in the context of a valuation of technical
provisions, these would typically apply to either
• the scale of
the underlying risks; or
• the degree
of model error inherent in valuation
methods.
3.130 Usually, they would
be defined as materiality thresholds, i.e. where they are not exceeded it would be considered
that the scale of the risk (or, respectively, the degree of model
error in the calculation) is immaterial, so that an application of
certain simplified valuation techniques would seem
appropriate.
3.131 It is also useful to
distinguish between
•
Thresholds which are proposed to apply to individual valuation
techniques; and
• Thresholds
which apply more broadly to all methods or to a specific class of
methods.
3.132 The following figure
illustrates these different types of
thresholds:
3.133 Most often a threshold
would be expressed quantitatively, either
in relative or in absolute terms.
However, it would also be
possible to specify a threshold in qualitative
terms.
3.134 An example of a quantitative “Type 1” threshold
(expressed in relative as well as absolute terms) is given by
the (indicative) materiality threshold specified by CEIOPS for the
use of simplified methods for the valuation of technical provisions
in QIS4.
The intention of this
threshold was to indicate when the liability that is valued would
not be material in absolute terms or relative to the overall size of
the total best estimate.
It was
• To
be applied broadly to the set of all simplified methods;
and
•
Based on simple volume measures (size of the best estimate of
technical provisions) related to the scale of the underlying
risks.
3.135 An example of a
(qualitative) “Type 2” threshold is given
by Step 2 of the proportionality assessment process outlined in
section 3.1.
Here, it was set out that a
valuation technique (simplified or not) would be considered
proportionate if it could be expected that the degree of model error
inherent in an application of the method would not be material.
In this context, “materiality”
was expressed in qualitative terms, considering the degree to which
the decision-making or judgment of the intended user of the
information could be influenced.
This establishes a general
materiality threshold which
•
Applies to all valuation methods which the (re)insurance undertaking
may consider for calculating its technical provisions;
and
• Is directly
related to the degree of model error inherent in the application of
the method.
3.136 Whereas “Type 1” and
“Type 2” thresholds would apply broadly to all methods or to a
specific class of methods, “Type 3” and “Type 4”
thresholds would be specific for individual simplified methods.
This means that, where
specific simplified methods would be introduced on Level 2, this
could be supplemented by including specific “Type 3” or “Type 4”
thresholds in their application criteria with the intention to limit
or restrict the use of the method depending on the scale of the risk
(in case of “Type 3” thresholds) or on the degree of model error
expected from an
application of the method (in case of “Type 4”
methods).
3.137 To illustrate this
by way of an example, consider the “Discounting
Proxy” technique tested in QIS4, which provided a means to discount
technical provisions by applying a pre-specified discounting
factor.
Consider further that such a
technique was specified on Level 2, with some general application
criteria setting out certain minimum conditions on the risk profile
of the portfolio to which the method could be applied.
In this context, it may be
decided to introduce a threshold specific to the “Discounting Proxy”
technique which would e.g. specify that the technique may only be
used to value up to 50% of the best estimate of the portfolio, where
the percentage of “50%” would have been derived by assessing the
degree of model error expected from an application of the
techniques.
3.138 In the following,
the appropriateness of introducing thresholds – in relation to the
different types as described above – is considered further.
Thresholds
relating to the scale of the risks
3.139 As was mentioned
above, it may be contemplated to implement external thresholds on
basis of an assessment of the scale of risks, so that an
(re)insurance undertaking would be allowed to use simplified methods in case the threshold is not
exceeded.
However, such an approach
could lead to a number of
problems:
• Relying on a threshold
based on the scale of risks may not be sufficient. It is important
to also consider the nature and complexity of the risks to which an
undertaking is exposed;
• Ultimately, it is not the
scale of risk which is the deciding factor in a proportionality
assessment, but whether the chosen method is
proportionate to the risks and whether the degree of model error in
the calculation is material.
This aspect may not be
sufficiently addressed in this type of
threshold.
3.140 Moreover, where
thresholds based on the scale of the risk are introduced, they would
often rely on simple volume measures (such as
the amount of premiums or technical provisions) related to the size
of the undertaking.
This may be problematic since:
• Size in itself may not be an adequate approximation to the risk to
which an undertaking is exposed. In general, neither the premiums
nor the technical provisions can be considered as a sufficient
benchmark to specify a threshold below which the undertaking would
no longer be vulnerable to the risk.
• Undertakings within the
scope of the Solvency II Level 1 text should not be classified
differently on the basis of size.
Indeed, policyholders should
not expect a lower degree of protection simply because their cover
is provided by a smaller undertaking.
3.141 Therefore, it would
not seem appropriate to introduce thresholds based on the scale of
the risks (e.g. with respect to the size of the undertaking or the
size of the risks) to determine the allowance for a simplified
approach for the calculation of technical provisions within
implementing measures on Level 2.
Thresholds
relating to the degree of model error
3.142 When considering
model error in context of Step 2 of the proportionality assessment
process, it was observed that in practice an assessment of model
error may be rather demanding on undertakings, leading to additional implementation
costs.
3.143 In view of this, it
could be contemplated to quantify the model
error of simplified valuation methods centrally (by CEIOPS) before
Solvency II is introduced.
Following such an approach,
specific thresholds (externally specified on Level 2 or 3) for the
use of individual simplified valuation methods could be set which
would reflect the assessed degree of model
error.
As long as these thresholds
would not be exceeded, it would be considered that the degree of
model error resulting from an application of the method would not be
material, and hence it would not be necessary for the undertaking to
calculate or quantify model errors.
3.144 However, in view of the ultimate aim of Solvency II to improve
risk assessment and risk management processes across (re)insurance
undertakings, it is believed that a holistic approach – which
integrates an assessment of model error into the valuation process
as part of actuarial best practice - would be more suitable than an
approach which stresses a need to avoid an assessment and potential
quantification of model error.
3.145 Moreover, it seems
likely that an approach to introduce thresholds as described would
be difficult to implement in practice:
• The
degree of model error incurred by an application of a method does
not only depend on the method, but rather is determined by the
degree to which the method is able to capture the undertaking’s
individual risk profile.
However, since the same
threshold would need to be specified for all undertakings, the
assessment of the model error
of the valuation method to which
the threshold is attached would need to make some generalising
assumptions on the characteristics of the risk profiles of the
undertakings which would use the threshold.
It seems likely that this
would make the calibration of the calculation of
such thresholds very demanding.
It may also lead to a
situation where for some undertakings the (central) assessment of
the model error implicit in the determination of the threshold would
not appropriately reflect the actual model risk which the
undertaking incurs in applying the method.;
• It would seem difficult to integrate the calculation of such
thresholds into the actuarial reserving process in a reasonable way;
and
• Under this approach,
thresholds would be established for a selection of simplified
methods which would be externally specified (on Level 2 or Level 3).
However, these methods would
only represent part of the spectrum of (simplified) methods which
would be available for the
undertaking. For these other (possibly
similar) methods, the thresholds would not apply, and the
undertaking would assess their
appropriateness on basis of the
proportionality assessment process outlined in section 3.1.
This may
create inconsistencies, where for similar methods this
process would lead to a different assessment of the degree of the
model error than is indicted by an application of the
threshold.
3.146 Considering this,
where external thresholds applicable to specific (simplified)
valuation methods and relating to the degree of model error are
introduced, care should be taken to ensure
that:
• This is consistent with
the principles-based proportionality assessment process outlined in
section 3.1 of this paper;
• In implementing the
threshold it can be ensured that the assessment of model error
implicit in the calibration of the threshold adequately reflects the
actual degree of model error incurred when the method is applied by
individual undertakings;
• This
should not lead to the impression that it would no longer be
necessary for the undertaking to undertake an own assessment of the
appropriateness of the method, including an assessment of the degree
of model error.
3.147 Overall, the above
discussion has shown that the usual interpretation of thresholds as
providing a cut-off point below which it would be regarded as
justifiable to use specific (simplified) valuation techniques (or a
class of such techniques) may be problematic.
Rather than as a
criterion for the allowance for specific simplified techniques it
may therefore be more appropriate to use thresholds as a criterion
for their rejection.
This means that, where the
threshold is exceeded it would be considered that the degree of
model error in the calculation is material, so that an use of the
simplified method would not seem appropriate.
However, where the threshold
is not exceeded the undertaking would need to conduct further
analysis and assessment before it can be decided whether an
application of the simplified method would be proportionate in
regard to its risk profile.
PROPORTIONALITY - LEVEL 2
MEASURES
CEIOPS-CP-45/09, 2 July 2009,
Consultation Paper No. 45
1.
Introduction
2.
Advice - Proportionality
3.
Proportionality Assessment – A three step process
4.
Simplified Methods
5.
Reinsurance Recoverables
6.
Annex A: Gross-to-net Techniques
New:
Solvency ii and Captives, CEIOPS Level 2 measures
|
