An insurer's Board and Senior Management have the responsibility to ensure that the insurer has adequate and appropriate capital to support the risks it undertakes. Capital serves to reduce the likelihood of failure due to significantly adverse losses incurred by the insurer over a defined period, including decreases in the value of the assets and/or increases in the obligations of the insurer, and to reduce the magnitude of losses to policyholders in the event that the insurer fails.

From a regulatory perspective, the purpose of capital is to ensure that, in adversity, an insurer’s obligations to policyholders will continue to be met as they fall due. Regulators should establish regulatory capital requirements at the level necessary to support this objective.

In the context of its own risk and solvency assessment (ORSA), the insurer would generally be expected to consider its financial position from a going concern perspective (that is, assuming that it will carry on its business as a going concern and continue to take on new business) but may also need to consider a run-off and/or winding-up perspective (eg where the insurer is in financial difficulty). The determination of regulatory capital requirements may also have aspects of both a going concern and a run-off^[10] or winding-up perspective. In establishing regulatory capital requirements, therefore, supervisors should consider the financial position of insurers under different scenarios of operation.
 

^[10] In this context, “run-off” refers to insurers that are still solvent but have closed to new business and are expected to remain closed to new business.

From a macro-economic perspective, requiring insurers to maintain adequate and appropriate capital enhances the safety and soundness of the insurance sector and the financial system as a whole, while not increasing the cost of insurance to a level that is beyond its economic value to policyholders or unduly inhibiting an insurer’s ability to compete in the marketplace. There is a balance to be struck between the level of risk that policyholder obligations will not be paid with the cost to policyholders of increased premiums to cover the costs of servicing additional capital.

The level of capital resources that insurers need to maintain for regulatory purposes is determined by the regulatory capital requirements specified by the supervisor. A deficit of capital resources relative to capital requirements determines the additional amount of capital that is required for regulatory purposes.

• reduce the probability of insolvency by absorbing losses on a going concern basis or in run-off; and/or
• reduce the loss to policyholders in the event of insolvency or winding-up.

The extent to which elements of capital achieve the above outcomes will vary depending on their characteristics or “quality”. For example, ordinary share capital may be viewed as achieving both of the above, whereas subordinated debt may be viewed largely as only protecting policyholders in insolvency. Capital which achieves both of the above is sometimes termed “going concern capital” and capital which only reduces the loss to policyholders in insolvency is sometimes termed “wind-up capital” or “gone concern” capital. It would be expected that the former (ie going concern capital instruments) should form the substantial part of capital resources.

For an insurer, the management and allocation of capital resources is a fundamental part of its business planning and strategies. In this context, capital resources typically serve a broader range of objectives than those in Guidance 17.2.6. For example, an insurer may use capital resources over and above the regulatory capital requirements to support future growth or to achieve a targeted credit rating.

It is noted that an insurer’s capital management (in relation to regulatory requirements and own capital needs) should be supported and underpinned by establishing and maintaining a sound enterprise risk management framework, including appropriate risk and capital management policies, practices and procedures which are applied consistently across its organisation and are embedded in its processes. Maintaining sufficient capital resources alone is not sufficient protection for policyholders in the absence of disciplined and effective risk management policies and processes (see ICP 16 Enterprise Risk Management for Solvency Purposes).

The supervisor should require insurance groups to maintain capital resources to meet regulatory capital requirements. These requirements should take into account the non-insurance activities of the insurance group. For supervisors that undertake group-wide capital adequacy assessments with a group level focus this means maintaining insurance group capital resources to meet insurance group capital requirements for the group as a whole. For supervisors that undertake group-wide capital adequacy assessments with a legal entity focus this means maintaining capital resources in each insurance legal entity based on a set of connected regulatory capital requirements for the group’s insurance legal entities which fully take the relationships and interactions between these legal entities and other entities in the insurance group into account.

It is not the purpose of group-wide capital adequacy assessment to replace assessment of the capital adequacy of the individual insurance legal entities in an insurance group. Its purpose is to require that group risks are appropriately allowed for and the capital adequacy of individual insurers is not overstated, eg as a result of multiple gearing and leverage of the quality of capital or as a result of risks emanating from the wider group, and that the overall impact of intra-group transactions is appropriately assessed.

Group-wide capital adequacy assessment considers whether the amount and quality of capital resources relative to required capital is adequate and appropriate in the context of the balance of risks and opportunities that group membership brings to the group as a whole and to insurance legal entities which are members of the group. The assessment should satisfy requirements relating to the structure of group-wide regulatory capital requirements and eligible capital resources and should supplement the individual capital adequacy assessments of insurance legal entities in the group. It should indicate whether there are sufficient capital resources available in the group so that, in adversity, obligations to policyholders will continue to be met as they fall due. If the assessment concludes that capital resources are inadequate or inappropriate then corrective action may be triggered either at a group (eg authorised holding or parent company level) or an insurance legal entity level.

The quantitative assessment of group-wide capital adequacy is one of a number of tools available to supervisors for group-wide supervision. If the overall financial position of a group weakens it may create stress for its members either directly through financial contagion and/or organisational effects or indirectly through reputational effects. Group-wide capital adequacy assessment should be used together with other supervisory tools, including in particular the capital adequacy assessment of insurance legal entities in the group. A distinction should be drawn between regulated entities (insurance and other sector) and non-regulated entities. It is necessary to understand the financial positions of both types of entities and their implications for the capital adequacy of the insurance group but this does not necessarily imply setting regulatory capital requirements for non-regulated entities. In addition, supervisors should have regard to the complexity of intra-group relationships (between both regulated and non-regulated entities), contingent assets and liabilities and the overall quality of risk management in assessing whether the overall level of safety required by the supervisor is being achieved.

For insurance legal entities that are members of groups and for insurance sub-groups that are part of a wider insurance or other sector group, capital requirements and capital resources should take into account all additional reasonably foreseeable and relevant material risks arising from being a part of any of the groups.

The supervisor should establish control levels that trigger intervention by the supervisor in an insurer’s affairs when capital resources fall below these control levels. The control level may be supported by a specific framework or by a more general framework providing the supervisor latitude of action. A supervisor’s goal in establishing control levels is to safeguard policyholders from loss due to an insurer’s inability to meet its obligations when due.

The solvency control levels provide triggers for action by the insurer and supervisor. Hence they should be set at a level that allows intervention at a sufficiently early stage in an insurer’s difficulties so that there would be a realistic prospect for the situation to be rectified in a timely manner with an appropriate degree of urgency. At the same time, the reasonableness of the control levels should be examined in relation to the nature of the corrective measures. The risk tolerance of the supervisor will influence both the level at which the solvency control levels are set and the intervention actions that are triggered.

When establishing solvency control levels it is recognised that views about the level that is acceptable may differ from jurisdiction to jurisdiction and by types of business written and will reflect, amongst other things, the extent to which the pre-conditions for effective supervision exist within the jurisdiction and the risk tolerance of the particular supervisor. The IAIS recognises that jurisdictions will acknowledge that a certain level of insolvencies may be unavoidable and that establishing an acceptable threshold may facilitate a competitive marketplace for insurers and avoid inappropriate barriers to market entry.

The criteria used by the supervisor to establish solvency control levels should be transparent. This is particularly important where legal action may be taken in response to an insurer violating a control level. In this case, control levels should generally be simple and readily explainable to a court when seeking enforcement of supervisory action.

Supervisors may need to consider different solvency control levels for different modes of operation of the insurer - such as an insurer in run-off or an insurer operating as a going concern. These different scenarios and considerations are discussed in more detail in Guidance 17.6.3 - 17.6.5.

In addition, the supervisor should consider the allowance for management discretion and future action in response to changing circumstances or particular events. In allowing for management discretion, supervisors should only recognise actions which are practical and realistic in the circumstances being considered ^[11].
 

^[11] The supervisor should carefully consider the appropriateness of allowing for such management discretion in the particular case of the MCR as defined in Standard 17.4.

• the way in which the quality of capital resources is addressed by the supervisor
• the coverage of risks in the determination of technical provisions and regulatory capital requirements and the extent  of the sensitivity or stress analysis underpinning those requirements;
• the relation between different levels (for example the extent to which a minimum is set at a conservative level);
• the powers of the supervisor to set and adjust solvency control levels within the regulatory framework;
• the accounting and actuarial framework that applies in the jurisdiction (in terms of the valuation basis and assumptions that may be used and their impact on the values of assets and liabilities that underpin the determination of regulatory capital requirements);
• the comprehensiveness and transparency of disclosure frameworks in the jurisdiction and the ability for markets to exercise sufficient scrutiny and impose market discipline;
• policyholder priority and status under the legal framework relative to other creditors in the jurisdiction;
• overall level of capitalisation in the insurance sector in the jurisdiction;
• overall quality of risk management and governance frameworks in the insurance sector in the jurisdiction;
• the development of capital markets in the jurisdiction and its impact on the ability of insurers to raise capital; and
• the balance to be struck between protecting policyholders and the impact on the effective operation of the insurance sector and considerations around unduly onerous levels and costs of regulatory capital requirements.

While the general considerations in Guidance 17.3.1 to 17.3.7 above on the establishment of solvency control levels apply in a group-wide context as well as a legal entity context, the supervisory actions triggered at group level will be likely to differ from those at legal entity level. As a group is not a legal entity the scope for direct supervisory action in relation to the group as a whole is more limited and action may need to be taken through co-ordinated action at insurance legal entity level.

Nevertheless, group solvency control levels are a useful tool for identifying a weakening of the financial position of a group as a whole or of particular parts of a group, which may, for example, increase contagion risk or impact reputation which may not otherwise be readily identified or assessed by supervisors of individual group entities. The resulting timely identification and mitigation of a weakening of the financial position of a group may thus address a threat to the stability of the group or its component insurance legal entities.

Group-wide solvency control levels may trigger a process of coordination and cooperation between different supervisors of group entities which will facilitate mitigation and resolution of the impact of group-wide stresses on insurance legal entities within a group. Group-wide control levels may also provide a trigger for supervisory dialogue with the group’s management.

A range of different intervention actions should be taken by a supervisor depending on the event or concern that triggers the intervention. Some of these triggers will be linked to the level of an insurer’s capital resources relative to the level at which regulatory capital requirements are set.

In broad terms, the highest regulatory capital requirement, the Prescribed Capital Requirement (PCR), will be set at the level at which the supervisor would not require action to increase the capital resources held or reduce the risks undertaken by the insurer^[12]. However if the insurer’s capital resources were to fall below the level at which the PCR is set, the supervisor would require some action by the insurer to either restore capital resources to at least the PCR level or reduce the level of risk undertaken (and hence the required capital level).
 

^[12] Note that this does not preclude the supervisor from intervention or requiring action by the insurer for other reasons, such as weaknesses in the risk management or governance of the insurer. Nor does it preclude the supervisor from intervention when the insurer’s capital resources are currently above the PCR but are expected to fall below that level in the short term. To illustrate, the supervisor may establish a trend test (a time series analysis). A sufficiently adverse trend would require some supervisory action. The trend test would support the objective of early regulatory intervention by considering the speed at which capital deterioration is developing. 

The regulatory objective to require that, in adversity, an insurer’s obligations to policyholders will continue to be met as they fall due will be achieved without intervention if technical provisions and other liabilities^[13] are expected to remain covered by assets over a defined period, to a specified level of safety. As such, the PCR should be determined at a level such that the insurer is able to absorb the losses from adverse events that may occur over that defined period and the technical provisions remain covered at the end of the period.
 

^[13] To the extent these liabilities are not treated as capital resources.

The Minimum Capital Requirement (MCR) represents the supervisory intervention point at which the supervisor would invoke its strongest actions, if further capital is not made available^[14]. Therefore, the main aim of the MCR is to provide the ultimate safety net for the protection of the interests of policyholders.
 

^[14] Note that this does not preclude such actions being taken by the supervisor for other reasons, and even if the MCR is met or exceeded.

These actions could include stopping the activities of the insurer, withdrawal of the insurer’s licence, requiring the insurer to close to new business and run-off the portfolio, transfer its portfolio to another insurer, arrange additional reinsurance, or other specified actions. This position is different from the accounting concept of insolvency as the MCR would be set at a level in excess of that at which the assets of the insurer were still expected to be sufficient to meet the insurer’s obligations to existing policyholders as they fall due. The PCR cannot be less than the MCR, and therefore the MCR may also provide the basis of a lower bound for the PCR, which may be especially appropriate in cases where the PCR is determined on the basis of an insurer’s internal model^[15] approved for use in determining regulatory capital requirements by the supervisor.

^[15] The term “internal model” refers to “a risk measurement system developed by an insurer to analyse its overall risk position, to quantify risks and to determine the economic capital required to meet those risks”. Internal models may also include partial models which capture a subset of the risks borne by the insurer using an internally developed measurement system which is used in determining the insurer's economic capital. The IAIS is aware that insurers use a variety of terms to describe their risk and capital assessment processes, such as “economic capital model”, “risk-based capital model”, or “business model”. The IAIS considers that such terms could be used interchangeably to describe the processes adopted by insurers in the management of risk and capital within their business on an economic basis. For the purposes of consistency, the term “internal model” is used throughout.

In establishing a minimum bound on the MCR below which no insurer is regarded to be viable to operate effectively, the supervisor may, for example, apply a market-wide nominal floor^[16] to the regulatory capital requirements, based on the need for an insurer to operate with a certain minimal critical mass and consideration of what may be required to meet minimum standards of governance and risk management. Such a nominal floor might vary between lines of business or type of insurer and is particularly relevant in the context of a new insurer or line of business.

^[16] In this context, a market-wide nominal floor may, for example, be an absolute monetary minimum amount of capital required to be held by an insurer in a jurisdiction.

Regulatory capital requirements may include additional solvency control levels between the level at which the supervisor takes no intervention action from a capital perspective and the strongest intervention point (that is, between the PCR and MCR levels). These control levels may be set at levels that correspond to a range of different intervention actions that may be taken by the supervisor itself or actions which the supervisor would require of the insurer according to the severity or level of concern regarding adequacy of the capital held by the insurer. These additional control levels may be formally established by the supervisor with explicit intervention actions linked to particular control levels. Alternatively, these additional control levels may be structured less formally, with a range of possible intervention actions available to the supervisor depending on the particular circumstances. In either case the possible triggers and range of intervention actions should be appropriately disclosed by the supervisor.

• measures that are intended to enable the supervisor to better assess and/or control the situation, either formally or informally, such as increased supervision activity or reporting, or requiring auditors or actuaries to undertake an independent review or extend the scope of their examinations;
• measures to address capital levels such as requesting capital and business plans for restoration of capital resources to required levels, limitations on redemption or repurchase of equity or other instruments and/or dividend payments;
• measures intended to protect policyholders pending strengthening of the insurer’s capital position, such as restrictions on licences, premium volumes, investments, types of business, acquisitions, reinsurance arrangements;
• measures that strengthen or replace the insurer’s management and/or risk management framework and overall governance processes;
• measures that reduce or mitigate risks (and hence required capital) such as requesting reinsurance, hedging and other mechanisms; and/or
• refusing, or imposing conditions on, applications submitted for regulatory approval such as acquisitions or growth in business.

Transparency as to the regulatory capital requirements that apply is required to facilitate effective solvency assessment and supports its enhancement, comparability and convergence internationally.

The supervisor may develop separate approaches for the determination of different regulatory capital requirements, in particular for the determination of the MCR and the PCR. For example, the PCR and MCR may be determined by two separate methods, or the same methods and approaches may be used but with two different levels of safety specified. In the latter case, for example, the MCR may be defined as a simple proportion of the PCR, or the MCR may be determined on different specified target criteria to those specified for the PCR.

The PCR would generally be determined on a going concern basis, ie in the context of the insurer continuing its operations. On a going concern basis, an insurer would be expected to continue to take on new risks during the established time horizon. Therefore, in establishing the regulatory capital level to provide an acceptable level of solvency, the potential growth in an insurer’s portfolio should be considered.

Capital should also be capable of protecting policyholders if the insurer were to close to new business. Generally, the determination of capital on a going concern basis would not be expected to be less than would be required if it is assumed that the insurer were to close to new business. However, this may not be true in all cases, since some assets may lose some or all of their value in the event of a winding-up or run-off, for example, because of a forced sale. Similarly, some liabilities may actually have an increased value if the business does not continue (eg claims handling expenses).

Usually the MCR would be constructed taking into consideration the possibility of closure to new business. It is, however, relevant to also consider the going concern scenario in the context of establishing the level of the MCR, as an insurer may continue to take on new risks up until the point at which MCR intervention is ultimately triggered. The supervisor should consider the appropriate relationship between the PCR and MCR, establishing a sufficient buffer between these two levels (including consideration of the basis on which the MCR is generated) within an appropriate continuum of solvency control levels, having regard for the different situations of business operation and other relevant considerations.

It should be emphasised that meeting the regulatory capital requirements should not be taken to imply that further financial injections will not be necessary under any circumstances in future.

Regulatory capital requirements may be determined using a range of approaches, such as standard formulae, or other approaches, more tailored to the individual insurer (such as partial or full internal models), which are subject to approval by the relevant supervisors.^[17] Regardless of the approach used, the principles and concepts that underpin the objectives for regulatory capital requirements described in this ICP apply and should be applied consistently by the supervisor to the various approaches. The approach adopted for determining regulatory capital requirements should take account of the nature and materiality of the risks insurers face generally and, to the extent practicable, should also reflect the nature, scale and complexity of the risks of the particular insurer.
 

^[17] A more tailored approach which is not an internal model might include, for example, approved variations in factors contained in a standard formula or prescribed scenario tests which are appropriate for a particular insurer or group of insurers.

Standardised approaches, in particular, should be designed to deliver capital requirements which reasonably reflect the overall risk to which insurers are exposed, while not being unduly complex. Standardised approaches may differ in level of complexity depending on the risks covered and the extent to which they are mitigated or may differ in application based on classes of business (eg life and non-life). Standardised approaches should be appropriate to the nature, scale and complexity of the risks that insurers face and should include approaches that are feasible in practice for insurers of all types including small and medium sized insurers and captives taking into account the technical capacity that insurers need to manage their businesses effectively.

By its very nature a standardised approach may not be able to fully and appropriately reflect the risk profile of each individual insurer. Therefore, where appropriate, a supervisor should allow the use of more tailored approaches subject to approval. In particular, where an insurer has an internal model (or partial internal model) that appropriately reflects its risks and is integrated into its risk management and reporting, the supervisor should allow the use of such a model to determine more tailored regulatory capital requirements, where appropriate^[18]. The use of the internal model for this purpose would be subject to prior approval by the supervisor based on a transparent set of criteria and would need to be evaluated at regular intervals. In particular, the supervisor would need to be satisfied that the insurer’s internal model is, and remains, appropriately calibrated relative to the target criteria established by the supervisor (see Guidance 17.12.1 to 17.12.18).

^[18] It is noted that the capacity for a supervisor to allow the use of internal models will need to take account of the sufficiency of resources available to the supervisor.

The supervisor should also be clear on whether an internal model may be used for the determination of the MCR. In this regard, the supervisor should take into account the main objective of the MCR (ie to provide the ultimate safety net for the protection of policyholders) and the ability of the MCR to be defined in a sufficiently objective and appropriate manner to be enforceable (refer to Guidance 17.3.4).

The supervisor should address all relevant and material categories of risk - including at least underwriting risk, credit risk, market risk, operational risk and liquidity risk. This should include any significant risk concentrations, for example, to economic risk factors, market sectors or individual counterparties, taking into account both direct and indirect exposures and the potential for exposures in related areas to become more correlated under stressed circumstances.

The assessment of the overall risk that an insurer is exposed to should address the dependencies and interrelationships between risk categories (for example, between underwriting risk and market risk) as well as within a risk category (for example, between equity risk and interest rate risk). This should include an assessment of potential reinforcing effects between different risk types as well as potential “second order effects”, ie indirect effects to an insurer’s exposure caused by an adverse event or a change in economic or financial market conditions.^[19] It should also consider that dependencies between different risks may vary as general market conditions change and may significantly increase during periods of stress or when extreme events occur. “Wrong way risk”, which is defined as the risk that occurs when exposure to counterparties, such as financial guarantors, is adversely correlated to the credit quality of those counterparties, should also be considered as a potential source of significant loss eg in connection with derivative transactions. Where the determination of an overall capital requirement takes into account diversification effects between different risk types, the insurer should be able to explain the allowance for these effects and ensure that it considers how dependencies may increase under stressed circumstances.
 

^[19] For example, a change in the market level of interest rates could trigger an increase of lapse rates on insurance policies.

Any allowance for reinsurance in determining regulatory capital requirements should consider the possibility of breakdown in the effectiveness of the risk transfer and the security of the reinsurance counterparty and any measures used to reduce the reinsurance counterparty exposure. Similar considerations would also apply for other risk mitigants, for example derivatives.

The supervisor should be explicit as to where risks are addressed, whether solely in technical provisions, solely in regulatory capital requirements or if addressed in both, as to the extent to which the risks are addressed in each. The solvency requirements should also clearly articulate how risks are reflected in regulatory capital requirements, specifying and publishing the level of safety to be applied in determining regulatory capital requirements, including the established target criteria (refer to Standard 17.8).

The IAIS recognises that some risks, such as strategic risk, reputational risk, liquidity risk and operational risk, are less readily quantifiable than the other main categories of risks. Operational risk, for example, is diverse in its composition and depends on the quality of systems and controls in place. The measurement of operational risk, in particular, may suffer from a lack of sufficiently uniform and robust data and well developed valuation methods. Jurisdictions may choose to base regulatory capital requirements for these less readily quantifiable risks on some simple proxies for risk exposure and/or stress and scenario testing. For particular risks (such as liquidity risk), holding additional capital may not be the most appropriate risk mitigant and it may be more appropriate for the supervisor to require the insurer to control these risks via exposure limits and/or qualitative requirements such as additional systems and controls.

However, the IAIS envisages that the ability to quantify some risks (such as operational risk) will improve over time as more data become available or improved valuation methods and modelling approaches are developed. Further, although it may be difficult to quantify risks, it is important that an insurer nevertheless addresses all material risks in its own risk and solvency assessment.

As has already been noted, a standardised approach, by its very nature, may not be able to fully and appropriately reflect the risk profile of each individual insurer. In cases where the standardised approach established for determining regulatory capital requirements is materially inappropriate for the risk profile of the insurer, the supervisor should have the flexibility to increase the regulatory capital requirement calculated by the standard approach. For example, some insurers using the standard formula may warrant a higher PCR and/or group-wide regulatory capital requirement if they are undertaking higher risks, such as new products where credible experience is not available to establish technical provisions, or if they are undertaking significant risks that are not specifically covered by the regulatory capital requirements.

Similarly, in some circumstances when an approved more tailored approach is used for regulatory capital purposes, it may be appropriate for the supervisor to have some flexibility to increase the capital requirement calculated using that approach. In particular, where an internal model or partial internal model is used for regulatory capital purposes, the supervisor may increase the capital requirement where it considers the internal model does not adequately capture certain risks, until the identified weaknesses have been addressed. This may arise, for example, even though the model has been approved where there has been a change in the business of the insurer and there has been insufficient time to fully reflect this change in the model and for a new model to be approved by the supervisor.

In addition, supervisory requirements may be designed to allow the supervisor to decrease the regulatory capital requirement for an individual insurer where the standardised requirement materially overestimates the capital required according to the target criteria. However, such an approach may require a more intensive use of supervisory resources due to requests from insurers for consideration of a decrease in their regulatory capital requirement. Therefore, the IAIS appreciates that not all jurisdictions may wish to include such an option for their supervisor. Further, this reinforces the need for such variations in regulatory capital requirements to only be expected to be made in limited circumstances.

Any variations made by the supervisor to the regulatory capital requirement calculated by the insurer should be made in a transparent framework and be appropriate to the nature, scale and complexity in terms of the target criteria. The supervisor may, for example, develop criteria to be applied in determining such variations and appropriate discussions between the supervisor and the insurer may occur. Variations in regulatory capital requirements following supervisory review from those calculated using standardised approaches or approved more tailored approaches should be expected to be made only in limited circumstances.

In undertaking its ORSA, the insurer considers the extent to which the regulatory capital requirements (in particular, any standardised formula) adequately reflect its particular risk profile. In this regard, the ORSA undertaken by an insurer can be a useful source of information to the supervisor in reviewing the adequacy of the regulatory capital requirements of the insurer and in assessing the need for variation in those requirements.

The following outlines a number of approaches a supervisor could use for the determination of capital resources in line with this requirement. The determination of capital resources would generally require the following steps:

• the amount of capital resources potentially available for solvency purposes is identified (see Guidance 17.10.3 - 17.10.21);
• an assessment of the quality and suitability of the capital instruments comprising the total amount of capital resources identified is then carried out (see Guidance 17.11.1 - 17.11.29); and
• on the basis of this assessment, the final capital resources eligible to meet regulatory capital requirements and their value are determined (see Guidance 17.11.30 - 17.11.44).

In addition, the insurer is required to carry out its own assessment of its capital resources to meet regulatory capital requirements and any additional capital needs (see Standard 16.14).

The IAIS supports the use of a total balance sheet approach in the assessment of solvency to recognise the interdependence between assets, liabilities, regulatory capital requirements and capital resources so that risks are appropriately recognised.

Such an approach requires that the determination of available and required capital is based on consistent assumptions for the recognition and valuation of assets and liabilities for solvency purposes.

From a regulatory perspective, the purpose of regulatory capital requirements is to require that, in adversity, an insurer’s obligations to policyholders will continue to be met as they fall due. This aim will be achieved if technical provisions and other liabilities are expected to remain covered by assets over a defined period, to a specified level of safety^[22].

^[22] Refer to Guidance 17.3.1 - 17.9.5.

• the extent to which certain liabilities other than technical provisions may be treated as capital for solvency purposes (Guidance 17.10.8 - 17.10.10);
• whether contingent assets could be included (Guidance 17.10.11);
• the treatment of assets which may not be fully realisable in the normal course of business or under a wind-up scenario (Guidance 17.10.12 - 17.10.19); and
• reconciliation of such a “top down” approach to determining capital resources with a “bottom up” approach which sums up individual items of capital to derive the overall amount of capital resources (Guidance 17.10.20).

Liabilities include technical provisions and other liabilities. Certain items such as other liabilities in the balance sheet may be treated as capital resources for solvency purposes.

For example, perpetual subordinated debt, although usually classified as a liability under the relevant accounting standards, could be classified as a capital resource for solvency purposes.^[23] This is because of its availability to act as a buffer to reduce the loss to policyholders and senior creditors through subordination in the event of insolvency. More generally, subordinated debt instruments (whether perpetual or not) may be treated as capital resources for solvency purposes if they satisfy the criteria established by the supervisor. Other liabilities that are not subordinated would not be considered as part of the capital resources; examples include liabilities such as deferred tax liabilities and pension liabilities.

^[23] However, adequate recognition should be given to contractual features of the debt such as embedded options which may change its loss absorbency.

It may, therefore, be appropriate to exclude some elements of funding from liabilities and so include them in capital to the extent appropriate. This would be appropriate if these elements have characteristics which protect policyholders by meeting one or both of the objectives set out in Guidance 17.2.6 above.

It may be appropriate to include contingent elements which are not considered as assets under the relevant accounting standards, where the likelihood of payment if needed is sufficiently high according to criteria specified by the supervisor. Such contingent capital may include, for example, letters of credit, members’ calls by a mutual insurer or the unpaid element of partly paid capital and may be subject to prior approval by the supervisor.

Supervisors should consider that, for certain assets in the balance sheet, the realisable value under a wind-up scenario may become significantly lower than the economic value which is attributable under going concern conditions. Similarly, even under normal business conditions, some assets may not be realisable at full economic value, or at any value, at the time they are needed. This may render such assets unsuitable for inclusion at their full economic value for the purpose of meeting required capital.^[24]

^[24] In particular, supervisors should consider the value of contingent assets for solvency purposes taking into account the criteria set out in Guidance 17.11.21.

• own shares directly held by the insurer: the insurer has bought and is holding its own shares thereby reducing the amount of capital available to absorb losses under going concern or in a wind-up scenario;
• intangible assets: their realisable value may be uncertain even during normal business conditions and may have no significant marketable value in run-off or winding-up; Goodwill is a common example;
• future income tax credits: such credits may only be realisable if there are future taxable profits, which is improbable in the event of insolvency or winding-up;
• implicit accounting assets: under some accounting models, certain items regarding future income are included, implicitly or explicitly, as asset values. In the event of run-off or winding-up, such future income may be reduced;
• investments^[25] in other insurers or financial institutions: such investments may have uncertain realisable value because of contagion risk between entities; also there is the risk of “double gearing” where such investments lead to a recognition of the same amount of available capital resources in several financial entities; and
• company-related assets: certain assets carried in the accounting statements of the insurer could lose some of their value in the event of run-off or winding-up, for example physical assets used by the insurer in conducting its business which may reduce in value if there is a need for the forced sale of such assets. Also, certain assets may not be fully accessible to the insurer eg surplus in a corporate pension arrangement.

• directly, by not admitting a portion of the economic value of the asset for solvency purposes (deduction approach); or
• indirectly, through an addition to regulatory capital requirements (capital charge approach).

As outlined above, an application of the deduction approach would lead to a reduction in the amount of available capital resources, whereas an application of the capital charge approach would result in an increase in regulatory capital requirements. Provided the two approaches are based on a consistent economic assessment of the risk associated with the relevant assets, they would be expected to produce broadly similar results regarding the overall assessment of the solvency position of the insurer.

For some asset classes, it may be difficult to determine a sufficiently reliable economic value or to assess the associated risks. Such difficulties may also arise where there is a high concentration of exposure to a particular asset or type of assets or to a particular counterparty or group of counterparties.

A supervisor should choose the approach which is best suited to the organisation and sophistication of the insurance sector and the nature of the asset class and asset exposure considered. It may also combine different approaches for different classes of assets. Whatever approach is chosen, it should be transparent and consistently applied. It is also important that any material double counting or omission of risks under the calculations for determining the amounts of required and available regulatory capital is avoided.

The approach to determining available capital resources as broadly the amount of assets over liabilities (with the potential adjustments as discussed above) may be described as a “top-down” approach - ie starting with the high level capital as reported in the balance sheet and adjusting it in the context of the relevant solvency control level. An alternative approach which is also applied in practice is to sum up the amounts of particular items of capital which are specified as being acceptable. Such a “bottom-up” approach should be reconcilable to the “top-down” approach on the basis that the allowable capital items under the “bottom-up approach” should ordinarily include all items which contribute to the excess of assets over liabilities in the balance sheet, with the addition or exclusion of items as per the discussion in Guidance 17.10.8 - 17.10.19.

• the way in which the quality of capital resources is addressed by the supervisor, including whether or not quantitative requirements are applied to the composition of capital resources and/or whether or not a categorisation or continuum- based approach is used;
• the coverage of risks in the determination of technical provisions and regulatory capital requirements;
• the assumptions in the valuation of assets and liabilities (including technical provisions) and the determination of regulatory capital requirements, eg going concern basis or wind-up basis, before tax or after tax, etc;
• policyholder priority and status under the legal framework relative to other creditors in the jurisdiction;
• overall quality of risk management and governance frameworks in the insurance sector in the jurisdiction;
• the comprehensiveness and transparency of disclosure frameworks in the jurisdiction and the ability for markets to exercise sufficient scrutiny and impose market discipline;
• the development of the capital market in the jurisdiction and its impact on the ability of insurers to raise capital;
• the balance to be struck between protecting policyholders and the impact on the effective operation of the insurance sector and considerations around unduly onerous levels and costs of regulatory capital requirements;
• the relationship between risks faced by insurers and those faced by other financial services entities, including banks.

The considerations set out in Guidance 17.10.3 - 17.10.21 above apply equally to insurance legal entity and group-wide supervision. The practical application of these considerations will differ according to whether a legal entity focus or a group level focus is taken to group-wide supervision. Whichever approach is taken, key group-wide factors to be addressed in the determination of group-wide capital resources include multiple gearing, intra-group creation of capital and reciprocal financing, leverage of the quality of capital and fungibility of capital and free transferability of assets across group entities. There may be particular concerns where such factors involve less transparent transactions eg because they involve both regulated and non-regulated entities or where there is a continuous sequence of internal financing within the group, or closed loops in the financing of the group.

• a method by which an insurer determines its own economic capital^[32]needs; and
• a means to determine an insurer's regulatory capital resources and requirements, where appropriate.

^[32] Economic capital refers to the capital which results from an economic assessment of the insurer's risks given the insurer’s risk tolerance and business plans.

One of the main purposes of an internal model is to better integrate the processes of risk and capital management within the insurer. Among other uses, internal models can be used to determine the economic capital needed by the insurer and, if an insurer has supervisory approval, to determine the amount of the insurer’s regulatory capital requirements. As a basic principle, an internal model that is to be used for regulatory capital purposes should already be in established use for determining economic capital. The methodologies and assumptions used for the two purposes should be consistent, any differences being explainable in terms of the difference in purposes.

In particular, where the risk profile of an insurer which is using a standardised approach for calculating its regulatory capital requirements is such that the assumptions underlying this approach are inappropriate, the supervisor may use its powers to increase the insurer's capital requirement, or to require the insurer to reduce the risks it bears. However, in such circumstances the supervisor should also consider encouraging the insurer to develop a full or partial internal model which might enable its risk profile to be better reflected in its regulatory capital requirements.

Where the supervisor is aware that an insurer has an existing internal model but has not sought approval to use it to calculate the regulatory capital requirement, the supervisor should discuss this decision with the insurer.

• produce regulatory capital requirements that are more risk sensitive and better reflect the supervisor’s target criteria; and
• assist the integration of the internal model fully into the insurer's strategic, operational and governance processes, systems and controls.

Where a supervisor allows the use of internal models to determine regulatory capital requirements, the supervisor should determine modelling criteria, based upon the level of safety required by the supervisor, to be used by an insurer adopting an internal model for that purpose. These criteria should require broad consistency between all insurers within the jurisdiction being based on the same broad level of safety requirements applied to the overall design and calibration of the standardised approach to determining regulatory capital requirements. Discussions with the insurance industry in a jurisdiction may also assist in achieving consistency. The supervisor should set out for which of the different levels of regulatory capital requirements the use of internal models is allowed and determine the modelling criteria for each level.

In particular, when considering whether an internal model may be used in determining the MCR, the supervisor should take into account the main objective of the MCR (ie to provide the ultimate safety net for the protection of policyholders) and the ability of the MCR to be defined in a sufficiently objective and appropriate manner to be enforceable. If internal models are allowed for determining the MCR, particular care should be taken so that the strongest supervisory action that may be necessary if the MCR is breached can be enforced, for example if the internal model is challenged in a court of law.

If an internal model is used for regulatory capital purposes, the insurer should ensure that its regulatory capital requirements determined by the model are calculated in a way that is consistent with the objectives, principles and criteria used by the supervisor. For example, the insurer may be able to apply the confidence level specified in the supervisors’ modelling criteria directly to the probability distribution forecasts used in its internal model. Alternatively, depending on the insurer’s own modelling criteria for its economic capital, an insurer may have to recalibrate its internal model to the modelling criteria required by the supervisor in order to use it for regulatory capital purposes. This will allow internal models to have a degree of comparability to enable supervisors to make a meaningful assessment of an insurer's capital adequacy, without sacrificing the flexibility needed to make it a useful internal capital model in the operation of the insurer's business. Further elaboration is provided in Guidance 17.15.1 - 17.15.2.

It is noted that, due to the insurer-specific nature of each internal model, internal models can be very different from each other. Supervisors, in allowing the use of an internal model for regulatory capital purposes, should preserve broad consistency of capital requirements between insurers with broadly similar risks.

The IAIS supports the use of partial internal models for regulatory capital purposes, where appropriate. A partial internal model typically involves the use of internal modelling to substitute parts of a standardised approach for the determination of regulatory capital requirements. For example, an insurer could decide to categorise its insurance contracts along business lines for modelling purposes. If the regulatory capital requirements for some of these categories are determined by modelling techniques, while the capital requirements for other categories are determined using a standardised approach, then this would constitute the insurer using a partial internal model to calculate regulatory capital.

Partial internal models are often used to smooth an insurer's transition to full use of an internal model or to deal with instances such as the merger of two insurers, one of which uses an internal model, and the other which uses a standardised approach. Given the potential complexity of a full internal model, use of a partial internal model could be a satisfactory approach provided its scope is properly defined (and approved by the supervisor). Provided the reduced scope of the internal model is soundly justified, the use of a partial internal model could be allowed as a permanent solution. However, as discussed above, there could be a tendency for an insurer to adopt a “cherry-picking” approach in the use of internal models. This particularly applies where partial modelling is allowed. The supervisor should place the onus on the insurer to justify why it has chosen to only use internal models for certain risks or business lines. Where this justification is not sound enough, the supervisor should take appropriate action eg refuse or withdraw approval of the model or impose a capital add-on until the model has developed to a sufficient degree.

This ICP should be applied to both partial and full internal models. Partial models should therefore be subject, as appropriate, to the full range of tests: the “statistical quality test”, “calibration test” and “use test” (see Guidance 17.13.1 - 17.17.8). In particular, an insurer should assess how the partial internal model achieves consistency with the modelling criteria specified by the supervisor for regulatory purposes. As part of the approval process for regulatory capital use, an insurer should be required to justify the limited scope of the model and why it considers that using partial internal modelling for determining regulatory capital requirements is more consistent with the risk profile of the business than the standardised approach or why it sufficiently matches regulatory capital requirements. The insurer should clearly document the reasons behind its decision to use partial internal models. If, for example, this is to ease transition towards full internal models, the insurer should outline a transitional plan, considering the implications for risk and capital management of the transition. Such plans and use of partial internal models should be reviewed by the supervisor, who may decide to impose certain restrictions on the partial model’s use for calculating regulatory capital (for example, introducing a capital add-on during the transitional period).

The modelling criteria for internal models for regulatory capital purposes and the process for internal model approval that a supervisor establishes should require broad consistency between group-wide regulatory capital requirements and regulatory capital requirements of individual insurance legal entities.

Group-wide internal models can vary greatly depending on their group-specific nature. In allowing the use of group-wide internal models for regulatory capital purposes, supervisors should preserve broad consistency between insurance groups and insurers with broadly similar risks eg insurance legal entities and insurance groups operating through a branch structure in a jurisdiction. The supervisor should design modelling criteria and the process for model approval so as to maintain broad consistency between the regulatory capital requirements determined using internal models and standardised approaches.

The IAIS recognises that modelling criteria may differ among supervisors. For Insurance groups operating in multiple jurisdictions, the degree of consistency in regulatory capital requirements across group members may vary.

Each supervisor should set out for which group-wide regulatory capital requirements, corresponding to the solvency control level or levels which apply to an insurance group, the use of group-wide internal models is allowed.

In particular, when the supervisor considers allowing the use of internal models for the purpose of determining group-wide regulatory capital requirements at the MCR level, the issues relating to possible legal challenges may differ from those encountered in respect of individual insurance legal entities. For example, supervisors may need to work together to establish and co-ordinate grounds for legal action in respect of the different insurance legal entities within a group.

Where insurers may be permitted to use internal models for calculating regulatory capital requirements, the models used for that purpose should be subject to prior supervisory approval. The onus should be placed on the insurer to validate a model that is to be used for regulatory capital purposes and provide evidence that the model is appropriate for those purposes. The IAIS considers that an insurer should not need supervisory approval for the use of internal models in determining its own economic capital needs.

The supervisor may prescribe requirements which will allow it to assess different models fairly and facilitate comparison between insurers within its jurisdiction. However, overly prescriptive rules on internal model construction may be counter-productive in creating models which are risk-sensitive and useful for insurers. Therefore, although a certain level of comparability can be achieved by the calibration requirements, full and effective comparison across jurisdictions to align best practice may be best achieved by dialogue between supervisors and industry.

The supervisor should require that in granting approval for the use of an internal model to calculate regulatory capital requirements, it has sufficient confidence that the results being produced by the model provide adequate and appropriate measures of risk and capital. Although the supervisor may encourage insurers to develop internal models that better reflect their risks as soon as possible, this should not lead to models being approved until there is confidence that they are calibrated correctly. The supervisor may therefore feel it necessary to evaluate an internal model over a specified period of time, for example a few years, prior to approval. For supervisors, approval of an internal model could require considerable expertise (depending on the sophistication of the model) which may need to be developed. In addition, it may be necessary to introduce different supervisory powers to allow the approval of internal models.

The supervisor should use, at least, the “statistical quality test”, “calibration test” and “use test”, as the basis of its approval process. While a broad range of internal model approaches may be suitable for internal economic capital assessment purposes, and this should not be subject to supervisory approval, supervisors may want to place requirements on the internal model approaches that would be regarded as acceptable for regulatory capital purposes. In approving the use of an internal model for calculating regulatory capital requirements, the supervisor should consider the primary role of the model as part of the insurer's risk management procedures. Any requirements imposed by the supervisor on the approval of a model for use in determining regulatory capital requirements should not prevent the model from being sufficiently flexible to be a useful strategic decision making tool which reflects the insurer's unique risk profile. Consistent standards for the approval of an insurer’s internal model should be applied by the supervisor, regardless of whether the model is developed in-house by the insurer or by an external party.

The “statistical quality test” and the “use test” are envisaged to be more insurer-specific measures which should allow the supervisor to gain an understanding of how a particular insurer has embedded its internal model within its business. The “calibration test” would be used by the supervisor to assess the results from the internal model in comparison to the insurer’s regulatory capital requirements and to those of other insurers.

• modelling criteria (risk measure, time horizon, level of safety);
• valuation bases for regulatory capital purposes;
• the risks that have to be modelled;
• treatment of intra-group transactions;
• approach to group-wide capital adequacy (eg group level or legal entity focus); and
• recognition of diversification across the group

The involved supervisors of an insurance group that conducts insurance business in more than one jurisdiction may consider their joint and common interests for the joint approval of the use of a group-wide internal model for regulatory capital purposes. If so, it may improve the efficiency and effectiveness of the approval process if the supervisors agree on common requirements for the process eg standardised language or languages for the application process.

Alternatively, the supervisors may independently approve the use of a group-wide internal model. Therefore, an insurance group seeking approval for a group-wide internal model may receive permission from one supervisor to use the model in that jurisdiction, while not receiving approval in another jurisdiction.

Similarly, where an insurance legal entity operates in other jurisdictions through a branch structure, the supervisors in those branch jurisdictions will have an interest in the solvency of the insurance legal entity. If local branch supervisors in these jurisdictions are not satisfied with the capital requirements of the home supervisor, possibly because they are determined using internal models, the local branch supervisors may impose limitations on the branch operations. The home supervisor, however, does not need to have the approval of the local branch supervisors in order to approve the use of the insurance legal entity’s internal model for its own purposes.

The degree of involvement of different supervisors in the approval process depends on a number of factors as illustrated in Guidance 17.13.12 - 17.13.16.

In the simplest case, an insurance group operates in one jurisdiction only. Clearly only the supervisor in that jurisdiction needs to be involved in the group-wide internal model approval process. Where there is more than one supervisor in a jurisdiction, eg where different insurance activities of a group are supervised separately, then both may need to be involved depending on the scope of the model. Nevertheless, some liaison with supervisors in other jurisdictions may be mutually beneficial to facilitate convergence and comparability across jurisdictions in respect of internal model standards and practice.

In the case of an insurance group that operates in more than one jurisdiction but only applies to use its group-wide internal model for regulatory capital purposes in one jurisdiction, eg the parent’s jurisdiction, the group does not need group-wide internal model approval of other jurisdictions provided that it is using other approaches to meet the capital requirements of those other jurisdictions. However, the supervisor considering approval of the group-wide internal model may wish to consult the other supervisors about the relevant insurance markets, the group’s operations in those markets and the standard of modelling.

In the case of an insurance group that wishes to use its group-wide internal model in more than one jurisdiction (eg to calculate insurance legal entity PCRs), the supervisor of each of those jurisdictions should consider approval of the specific application of the group-wide internal model in its jurisdiction, having regard to the considerations in Guidance 17.13.15 - 17.13.18 below.

• its group-wide regulatory capital requirements;
• whether and the extent to which its jurisdiction allows the use of internal models for regulatory capital purposes (eg PCR or both PCR and MCR);
• how its jurisdiction interacts with the other jurisdictions potentially involved when supervisory intervention is being considered; and
• the arrangements for collaboration between the supervisors of the legal entities within the insurance group.

A supervisor may delegate the approval process to another supervisor or agree to be bound by its decision while retaining supervisory responsibility. Alternatively, a group-wide supervisor may have ultimate decision-making authority over some or all of the supervisors involved. If more than one jurisdiction is concerned, making such authority legally binding may require a treaty between these jurisdictions. To be effective, each arrangement requires a high level of collaboration between supervisors. To require the model appropriately addresses all categories of risk, the supervisor making the decision needs sufficient knowledge of the local circumstances in which the group operates.

Supervisors should require that the approval process for the use of a group-wide internal model for regulatory capital purposes is sufficiently flexible to achieve an approach appropriate to the nature, scale and complexity at each organisational level in an insurance group (group/sub-group/individual insurance legal entity). Risks which may have a large impact at insurance legal entity level may have much smaller significance at insurance group level. Conversely, risks that may have a small impact at insurance legal entity level may aggregate to have a larger impact on risk at the group level. The nature and complexity of risks may also vary at different levels in the insurance group.

Whether the group-wide internal model is appropriate for regulatory purposes given the nature, scale and complexity of the risks depends on the regulatory capital requirements of a jurisdiction. While the risk coverage by an internal model may look reasonable from a group-wide perspective, it may not be reasonable from the point of view of each member of the insurance group. For example, in a group of many non-life insurers and one small life insurer it may be appropriate from an overall perspective to place less emphasis on the modelling of the life insurance risks. However this may not be appropriate from the life insurer’s or supervisor’s perspective. In such circumstances, it may be necessary for the group to upgrade its model to include an adequate life insurance risk component or to set up a self-contained internal model for the life insurer in order to gain approval.

Given the importance of an embedded internal model to an insurer's risk management policy and operations, an internal model would generally be constructed to deliver a probability distribution of the required risk capital rather than a “point estimate”. A range of approaches could constitute an effective internal model for risk and capital management purposes, and supervisors should encourage the use of a range of different approaches appropriate to the nature, scale and complexity of different insurers and different risk exposures. There are several different techniques to quantify risk which could be used by an insurer to construct its internal model. In broad terms, these could range from basic deterministic scenarios to complex stochastic models. Deterministic scenarios would typically involve the use of stress and scenario testing reflecting an event, or a change in conditions, with a set probability to model the effect of certain events (such as a drop in equity prices) on the insurer's capital position, in which the underlying assumptions would be fixed. In contrast, stochastic modelling often involves simulating very large numbers of scenarios in order to reflect the likely distributions of the capital required by, and the different risk exposures of, the insurer.

The IAIS recognises that there are numerous methodologies which an insurer could use as part of its stress and scenario testing. For example, an insurer may decide to model the effect of various economic scenarios, such as a fall in equity prices or a change in interest rates, on its assets and liabilities. Alternatively, an insurer could consider a run-off approach, where the effect of various scenarios on a specific portfolio of business as it is run-off is examined. The insurer should use scenarios which it regards as most appropriate for its business. Where the internal model is used for regulatory capital purposes, the onus is on the insurer to demonstrate to the supervisor that the chosen methodology is appropriate to capture the relevant risks for its business. This includes testing of the model to require that it can replicate its results on request and that its response to variation in input data is adequate such as that corresponding to changes in base or stress scenarios. Overall capital requirements derived from an internal model can be highly sensitive to assumptions on the effect of diversification across risks. Supervisors and insurers should therefore give particular consideration to aggregation issues. Conducting stress and scenario testing to determine the effect of shocks may be a suitable tool to validate statistical assumptions.

Where an internal model is established to assess risks at a modular level, ie on a risk-by-risk basis, in order to conduct an overall risk assessment, the insurer should aggregate the results for each of these risks both within and across business lines. Several methods exist to aggregate the separate results allowing for diversification effects. The IAIS considers that an insurer would generally be expected to decide how best to aggregate and account for the risks to the whole of its business. The determination of overall regulatory capital requirements by the internal model should consider dependencies within, as well as across, risk categories. Where the internal model allows for diversification effects, the insurer should be able to justify its allowance for diversification effects and demonstrate that it has considered how dependencies may increase under stressed circumstances.

Internal models need high quality data in order to produce sufficiently reliable results. The data used for an internal model should be current and sufficiently credible, accurate, complete and appropriate. Hence, a “statistical quality test” should examine the appropriateness of the underlying data used in the construction of the internal model. A “statistical quality test” would include the examination of the aggregation of data, the modelling assumptions and the statistical measures used to construct the internal model. This could include an annual (or more frequent) review of the various items that are being measured (claims, lapses, etc.) updated for the additional data available together with a scrutiny of data from previous periods to determine whether this data continues to be relevant. Older data may no longer be relevant possibly due to changes in risks covered, secular trends or policy conditions and guarantees attaching. Similarly, new data may not be of substantive use when modelling items that require a long-term view of experience (such as testing the predictions of cash flows for catastrophic events).

An insurer may not always have sufficient reliable data in-house. In instances where an insurer lacks fully credible data it may rely on industry or other sufficiently credible data sources to supplement its own data. For example, a new company may lack its own historical data and so could use market data sources in constructing its internal model. Some supervisors have published jurisdictional data which may be of some use.

Another possible source of data may be from reinsurers - whose data pool is typically larger and covers a wider spectrum of the market. It is, however, important to consider that such industry data may not be entirely appropriate for all insurers. Reinsurers often only receive data in aggregated form and sometimes are only informed of larger claims or from smaller insurers whose market may not be applicable for all or many insurers. Therefore, any data not specific to the insurer would need to be carefully considered before deciding it was appropriate for use as the basis for an insurer's “statistical quality test”. Even where deemed appropriate, it may still be necessary to adjust the data to allow for differences in features between the data source and the insurer.

In assessing suitability of data and of other inputs, eg assumptions, to the internal model, expert judgment should be applied and supported by proper justification, documentation and validation.

As part of the “statistical quality test”, the insurer should be able to demonstrate that the base quantitative methodology used to construct its internal model is sound and sufficiently reliable to support the model's use, both as a strategic and capital management tool, and to calculate the insurer's regulatory capital requirements, if appropriate. The methodology should also be consistent with the methods used to calculate technical provisions.

A “statistical quality test” should also include a review of the internal model to determine whether the assets and products as represented in the model truly reflect the insurer's actual assets and products. This should include an analysis of whether all reasonably foreseeable and relevant material risks have been incorporated, including any financial guarantees and embedded options. Insurers should also consider whether the algorithms used are able to take into account the action of management and the reasonable expectation of policyholders. Testing should include future projections within the model and to the extent practicable “back-testing” (the process of comparing the predictions from the model with actual experience).

For use in determining the regulatory capital requirements of an insurance legal entity, a group-wide internal model should meet the same standards as applicable to a stand-alone internal model of that insurer.

For use for group-wide regulatory capital requirements, group members should be sufficiently engaged with a group-wide internal model and its application to their businesses (through their input to the model, local Board involvement, capital allocation, performance measurement etc.), even if the insurance group does not use the model to determine the regulatory capital requirements of individual group members.

The “calibration test” should be used by the insurer to demonstrate that the internal model is calibrated appropriately to allow a fair, unbiased estimate of the capital required for the particular level of confidence specified by the supervisor. Where an insurer uses different modelling criteria than those specified by the supervisor for regulatory capital purposes, it may need to recalibrate its model to the supervisor’s modelling criteria to achieve this.

See Guidance 17.14.10 and 17.14.11 for additional guidance for group-wide internal models.

In considering the use of an internal model for regulatory capital purposes by an insurer, the supervisor should not merely focus on its use for that narrow purpose, but should consider the wider use of the internal model by the insurer for its own risk and capital management.

The “use test” is the process by which the internal model is assessed in terms of its application within the insurer's risk management and governance processes. In order for the insurer’s internal model to be most effective it should be genuinely relevant for use within its business for risk and capital management purposes.

Where an insurer decides to adopt a higher confidence level than the level required for regulatory capital purposes for its own purposes, for example, in order to maintain a certain investment grade rating, then “calibration” testing should also be conducted by the insurer to allow the insurer to determine the level of capital needed at this higher level. The insurer should then assess whether holding this amount of capital is consistent with the insurer's overall business strategy.

The insurer should have the flexibility to develop its internal model as an important tool in strategic decision making. An insurer should therefore have the flexibility to use the most appropriate risk measure and modelling techniques in its internal models. It may be beneficial if the insurer is able to demonstrate why it has chosen a particular risk measure, and it should include in its internal model an appropriate recalibration or reconciliation, if necessary, between the modelling criteria used in the model for its own risk and capital management purposes and those set by the supervisor for regulatory capital purposes. Differences between the economic capital and the regulatory capital requirements should be explicit and capable of being explained by the insurer to its Board and the supervisor.

The “use test” is a key method by which the insurer can demonstrate that its internal model is integrated within its risk and capital management and system of governance processes and procedures. As part of the “use test”, an insurer should examine how the internal model is used for operational management purposes, how the results are used to influence the risk management strategy and business plan of the insurer, and how Senior Management are involved in applying the internal model in running the business. An insurer should demonstrate to the supervisor that an internal model used for regulatory capital purposes remains useful and is applied consistently over time and that it has the full support of and ownership by the Board and Senior Management.

The insurer's Senior Management should take responsibility for the design and implementation of the internal model, in order to ensure full embedding of the model within the insurers' risk and capital management processes and operational procedures. The methodology used in building the model should be compatible with the overall enterprise risk management framework agreed to by the Board and Senior Management. Although the Board and Senior Management may not be able to de-construct the internal model in detail, it is important that the Board has overall oversight of the model's operation on an ongoing basis and the level of understanding necessary to achieve this. The Board and Senior Management should also ensure that processes are in place to update the internal model to take into account changes in the insurer's risk strategy or other business changes.

Various business units within the insurer may be involved in the construction and operation of the internal model, such as risk management, capital management, finance and actuarial departments, depending on the size of the insurer. The experience and technical ability of staff involved in the construction and operation of the internal model should be an important consideration for the insurer. For a model to pass the “use test” it would be expected that an insurer would have a framework for the model's application across business units. This framework should define lines of responsibility for the production and use of information derived from the model. It should also define the purpose and type of management information available from the model, the decisions to be taken using that information, and the responsibilities for taking those decisions. The “use test” should also ensure the adequacy of systems and controls in place for the maintenance, data feeds and results of the model. The IAIS notes that internal models may require significant IT resources and costs, which should be a consideration for the insurer in developing its models.

The IAIS considers that governance processes and communication in respect of an internal model are as important as its construction. An internal model should be subject to appropriate review and challenge so that it is relevant and reliable when used by the insurer. The key elements and results from the internal model should be understood by the key personnel within the insurer, including the Board, and not only by those who have constructed it. This understanding should ensure that the internal model remains a useful decision-making tool. If the internal model is not widely understood, it will not be achieving its purpose and adding value to the business. The “use test” is key in ensuring the relevance of the internal model to the insurer’s business.

See Guidance 17.14.10 and 17.14.11 for additional guidance for group-wide internal models.

Over time an insurer's business may alter considerably, as a result of internal factors or events (such as a change in insurer strategy) and external factors or events (such as a change in interest rates), so that the internal model may no longer fully capture the risks to which the insurer is exposed unless adapted. The supervisor should reassess an insurer's internal model and the results that it produces on a regular basis against the criteria of the statistical quality test, calibration test and use test so that it remains valid for use, both as a strategic decision-making tool in the context of the insurer’s own risk and capital management, and as a means of calculating regulatory capital requirements where appropriate. In general only material changes to the model (such as changing the underlying model structure or the risk measure used) or to the risks faced by the insurer should require the model to be reassessed by the supervisor. A “model change policy” could be agreed between the supervisor and the insurer regarding the degree and timing of changes made to the internal model. This would enable the insurer to enact minor changes to its internal model without seeking prior supervisory approval (provided the changes are in accordance with the agreed policy), thereby allowing the model to be updated in a quicker and more flexible way.

The insurer should be required to notify the supervisor of material changes to the internal model and to properly document changes to enable the supervisor to assess, for continued approval, the ongoing validity of the model for use in determining regulatory capital requirements. Following any material changes to an internal model, the supervisor may give the insurer a reasonable amount of time so that the updated model is embedded in its risk strategies and operational processes.

The insurer should demonstrate that the data used in the internal model remains appropriate, complete and accurate for this purpose.

The supervisor should take care that its ongoing validation requirements do not unduly restrict the use of the internal model by the insurer for its own risk and capital management purposes and thereby reduce its ability to comply with the use test.

The insurance group should adjust the model for material changes in group composition and operations, including mergers, acquisitions and other structural changes of affiliated entities or jurisdictional changes.

The supervisor should require the insurance group to provide documentation of material changes in group operations and the reasons why continued use of the group-wide internal model would remain appropriate following the change. If such reasons cannot be given or are not sufficient the supervisor should require the group to propose appropriate model changes as a result of the material change for re-assessment of approval by the supervisor.

The IAIS considers that it is essential that supervisors are able to understand fully the insurers' internal models and be able to appraise their quality. To this end, the supervisor should have access to experienced personnel with appropriate technical ability, as well as sufficient resources. It is likely to take time for supervisors to acquire the necessary experience to appraise an insurer’s internal model. Without the experience and resources, the supervisor may be unable to reliably approve the use of an insurer’s internal model for regulatory purposes. The supervisor may wish to use external specialists that are considered to have the appropriate experience, such as actuarial consultants, accountancy firms and ratings agencies, to assist it in reviewing an insurer's internal models. In such instances, the supervisor should retain the final responsibility for review and approval of the use of the internal model for regulatory purposes.

It may be appropriate for a supervisor to consider transitional measures when permitting insurers to use internal models for regulatory capital purposes for their first time. Such measures will permit the necessary time for both insurers and the supervisor to become familiar with the internal models and their uses. For example, during a transition period, the supervisor could include the use of partial internal modelling, to allow the insurer to move gradually to full use of internal models or the supervisor could require parallel reporting of regulatory capital determined by both the internal model and standardised approach. The supervisor may also consider applying a minimum capital level during the transition period.

The supervisor may need to impose additional capital requirements (capital add-ons) or take other supervisory action to address any identified weaknesses in an internal model, either prior to approving the use of the model, as a condition on the use of the model or in the context of a review of the ongoing validity of an internal model for regulatory capital purposes. It may be necessary to introduce additional supervisory powers, to allow such supervisory actions and measures, when internal models are allowed for regulatory capital purposes by a supervisor.

Where an insurer which is a subsidiary of an insurance group seeks approval for the use of an internal model which itself is part of a broader “group model”, the supervisor of this subsidiary should conduct the approval process in close co-operation with the group-wide supervisor. In particular, the supervisor of the subsidiary should check the status of the “group model” and seek information from the group-wide supervisor about its own approval process.

For supervisory approval purposes, supervisors should require the insurer to submit sufficient information for them to be able to approve the use of the internal model for regulatory capital purposes and to give confidence to the supervisor that the insurer is appropriately carrying out its responsibility to manage its risks and protect the interests of policyholders. This should include the results of analysis conducted under the “statistical quality test”, “calibration test” and “use test”. While supervisors should have the power to determine the exact nature and scope of the information they require, supervisory reporting should be appropriate to the nature, scale and complexity of an insurer's business.