MCCSR Guideline Revisions for 2015

Document Properties

  • Type of Publication: Letter
  • Date: November 6, 2014
  • To: Federally Regulated Life Insurance Companies & Fraternal Benefit Societies

Revisions have been made to Guideline A, Minimum Continuing Capital and Surplus Requirements (MCCSR) which will come into effect on January 1, 2015.    

The revised guideline includes, among others, the following key items:

  • Revisions to Chapter 2 to reflect the revised qualifying criteria that will apply to capital instruments issued on or after August 7, 2014;
  • Adopt revised criteria for qualifying par policies that benefit from reduced capital charge;
  • Require that a portion the impact on retained earnings of using a morbidity improvement assumption, net of the mortality improvement assumption by product, be reversed from Available Capital;
  • Clarify that deferred income taxes be excluded in the Total Gross Calculated Requirements and the policy liabilities for segregated funds when determining the capital requirement for segregated fund guarantees under the standard approach;
  • Clarify that deferred income tax liabilities do not qualify as Available Capital/Margin.  Also clarify that associated deferred income tax liabilities are not to reduce the carrying amount of an item that is to be deducted from Available Capital except where explicitly permitted; allow for an additional exception in the case of intangible assets where specific conditions are met;
  • Apply the equity factor to mutual funds employing leverage;
  • Require that assets pledged in a reinsurance arrangement be in the form of cash or securities.

Questions concerning the Guideline should be addressed to your Relationship Manager at OSFI or, alternatively, to Mr. Henri Boudreau, Managing Director, Capital Division, (henri.boudreau@osfi-bsif.gc.ca).

Mark Zelmer
Deputy Superintendent

Summary of Public Consultation Comments

Comment OSFI Response

Morbidity Improvement (S1.2.5, S2.1.1.1)

  1. The draft 2015 Guideline requires future morbidity improvements be reversed in isolation of future mortality improvements.  The proposed approach would not be warranted as it does not recognize the link between morbidity and mortality improvements or the product adjustability.

     

    Future morbidity and mortality improvements both contribute to the cost of claims - future mortality improvements represent the lengthening of life expectations and future morbidity improvements can represent either lower claim incidence or shorter claims length, or a combination. We believe future morbidity improvements are interrelated with future mortality improvements - a by-product of mortality improvements is longer healthy life expectancies and this should result in later claim incidence as well as potentially shorter claim lengths. In the event that morbidity and mortality improvements do not materialize, premium rates could be increased to pass-through shortfalls. Accordingly, setting aside capital for potential losses which would be recovered is quite conservative.

    Health products should include the combination of the morbidity and mortality improvement tests allowing for offsets within a product.  The addition of the following wording to the paragraph at the end of Section 2.1.1.1 discussing the phase-in would address this:
    For health products (e.g. LTC, CI and disability), the determination of the impact of reversing the future morbidity improvement assumption in policy liabilities should include the reversal of the mortality improvement assumption.  In addition, companies may elect to phase in the impact on retained earnings of reversing the future morbidity improvement assumption in policy liabilities.  The phase-in will be made on a straight-line over the phase-in period.  The phase-in period begins on January 1, 2015 and must be completed by the earliest quarter-end occurring on or after December 31, 2017”

It is not necessarily the case that improved mortality would lead to improved morbidity.  It is conceivable that improved mortality could be associated with deteriorating morbidity, resulting in higher eventual claims incidence (policyholders hit with disability instead of death) and longer claim lengths (policyholders survive longer while disabled).

Based on our review, mortality improvement is supported by more research and stronger evidence than morbidity improvement, but not at the level of confidence required for regulatory capital.  Since mortality improvement is reversed from retained earnings in the MCCSR, it follows that morbidity improvement should also be reversed as the strength of the assumption is at a lower level of confidence than mortality improvement.

There are separate provisions in the MCCSR that make adjustments to capital requirements for product adjustability.

The MCCSR has been amended to allow an offset between morbidity improvement and mortality improvement at the product level if it is not used in the calculation of the insurance and annuity offset.

  1. Morbidity improvement is just one side of morbidity trends. Morbidity trends include both projected improvements as well as deteriorations at two levels: a) for separate components of the claims rate within the same insurance product resulting, in aggregate, in either a net increase or net decrease to actuarial liabilities for that product (e.g. CI, where heart attack incidence may be projected to decrease but cancer projected to increase over time) or b) for the overall claims rates within a product (e.g. a deterioration to all cause disability termination rates).

     

    It is the net aggregate impact of morbidity trends across all products that should be considered, specifically at the company consolidated level for all health business combined (i.e. CI, DI, LTC combined).

Aggregate claims could decrease for a variety of reasons and not necessarily as a result of improvements in morbidity trends – for example, a change in business mix towards younger ages.  In addition, it is difficult to isolate the impact of morbidity improvement from mortality improvement.

OSFI allows an offset for the net impact of mortality improvement between insurance and annuity business as there is a stronger relationship between the mortality improvement trends across products (due to greater homogeneity).  Morbidity is not homogeneous in the same way as mortality since it covers a wide range of conditions that may not necessarily exhibit the same trends.  In this sense, it is not appropriate to allow offsets between morbidity improvement and mortality improvement at a total aggregate level. 

However, if it is not used in the calculation of the insurance and annuity offset, OSFI will allow an offset between morbidity improvement and mortality improvement at the product level.

  1. Especially when viewed together with OSFI’s existing treatment of the reversal of the impact of mortality improvement assumptions for insurance and annuities, OSFI’s general approach to treating reversals of improvement assumptions is asymmetrical. The proposed construct only reverses the impact on retained earnings if improvement assumptions decrease policy liabilities and the determination is done at the product level of aggregation instead of at the highest level of aggregation. We maintain the germane issue is not having a degree of correlation between morbidity and mortality improvement as a necessary condition, rather one of symmetry. While we recognize the QIS methodology is also heading in this direction, we now wish to question the appropriateness of such an approach.

     

    Secondly, in the determination of morbidity risk capital for health businesses, the impact of any increases to policy liabilities due to mortality improvements should be incorporated. For example, if the morbidity trend assumptions decrease policy liabilities by $100 and mortality improvement assumptions increase policy liabilities by $10, then the reversal should not be 50% of $100, but rather incorporate the offset from the $10. To recognize symmetry and to incorporate the reversal of mortality improvement assumptions as well, one approach would be to amend the third bullet point under the sentence: “For MCCSR purposes, the following items are reversed from reported retained earnings” as follows:

    Item i) less item ii)

    •    50% of the net decrease in policy liabilities, net of reinsurance, resulting from the recognition of morbidity trend assumptions, offset by the impact of mortality improvement assumptions.
    •    The net increase in policy liabilities (for insurance and annuity business combined, net of all reinsurance) resulting from the recognition of future mortality improvement under CIA standard 2350.06 and additional future mortality improvement under CIA standard 2350.11;

OSFI allows an offset for the net impact of mortality improvement between insurance and annuity business as there is a stronger relationship between the mortality improvement trends across products (due to greater homogeneity).  Morbidity is not homogeneous in the same way as mortality since it covers a wide range of conditions that may not necessarily exhibit the same trends.  In this sense, it is not appropriate to allow offsets between morbidity improvement and mortality improvement at a total aggregate level. 

The MCCSR has been amended to allow an offset between morbidity improvement and mortality improvement at the product level if it is not used in the calculation of the insurance and annuity offset.

  1. There is a potential opportunity to misinterpret the intent of the draft wording added to 1.2.5. Currently the draft wording states: "... future morbidity improvement if the impact of such improvement has been reversed from retained earnings."

     

    Since at most 50% of the impact is to be reversed, this statement will never be true. Consequently we recommend instead:  if any portion of the impact...,

    Hence we recommend the following wording for Section 1.2.5:
    “In addition, the policy liabilities used should exclude fifty percent of future morbidity trends if any portion of the impact of such trends has been reversed from retained earnings in the determination of gross tier 1 capital under section 2.1.1.1.”

The MCCSR has been amended to clarify that the liabilities used should exclude the amount of morbidity improvement that is reversed from retained earnings.

  1. OSFI’s cover letter states a four year transition, but in 2.1.1.1 it states the phase-in period is January 1, 2015 to December 31, 2017. Also, consistent with prior phase-ins, it would be appropriate to include a quarterly phase in formula.

A quarterly phase-in has been added to the MCCSR, consistent with prior phase-ins.  The treatment of fifty percent reversal of the impact of the morbidity improvement assumption will be phased-in over a three year period; the treatment of the remaining fifty percent will be determined at a later date and will be consistent with the treatment of morbidity improvement in the new life capital framework.

Criteria for Qualifying Participating Policies (S1.2.6)

  1. Any demonstrations of the sufficiency of the present value of dividends should not be an ongoing (i.e. quarterly MCCSR) requirement. Instead it should be periodic based on a trigger. We recommend the trigger be a change in the dividend scale.

Demonstrations of sufficiency, as with many other MCCSR requirements, are on-going and can be based on the judgment of the Appointed Actuary. OSFI can request that the insurer demonstrates sufficiency or demonstrates that its on-going process is sufficiently robust.

  1. The requirement for dividend scales to be updated within two years from when the shortfall in experience is recorded is not necessarily in line with how companies manage dividends and would lead to more frequent and small dividend updates that would not be practical, in particular where an experience shortfall is insignificant and potentially transitory. This element of the draft guideline unnecessarily restricts par governance practices for insurers. We believe the frequency of dividend scale updates for experience shortfalls should be determined by internal policies for par products, which is backed by a solid internal governance framework that meets policyholder reasonable expectations.

     

    The following alternative wording is suggested:
    “Furthermore, it must be able to demonstrate that substantial shortfalls in actual overall experience with regard to the risk component are recovered through a level or declining reduction (i.e. front loaded dividend reductions or accelerated experience recovery) of future dividends, including the dividend scale, and the dividend stabilization reserve, if any. The elements used in the determination of future dividends must be updated within two five years from when the substantial shortfall in actual overall experience occurs.”

The rationale for providing a capital credit for qualifying participating policies is that when losses occur, there is an ability and willingness to pass these losses through to policyholders. This principle does not change with MCCSR 2015; however, the new provisions allows more flexibility for managing dividends, without compromising policyholder protection. The MCCSR 2015 allows insurers to directly influence the level of volatility of the dividend scale, from very stable to very volatile.

The change to the MCCSR moves from substantial recovery on a cash-basis within five years to a pass-through on a present value basis in dividend scales and reduction in surplus or dividend stabilization reserves over the lifetime of the policies. The change also extends the time insurers have to reflect experience losses in the elements used to determine the dividend scales. Whereas MCCSR 2014 implied that losses had to be recovered starting in the year immediately following the period of the loss the recovery, the MCCSR 2015 provides that the loss can start to be recovered  within two calendar years after the period of the shortfall.  

It should be emphasized that the use of surplus, dividend stabilization reserves or similar concepts is also now expressly permitted and may help to absorb minor shortfalls/deficits that would otherwise require a change in the dividend scale.

  1. For criteria #3, in the English version the word "substantially" was removed, while in the French version the word "largement" is still there.  Which is correct?

This omission has been corrected in the final version; the word "largely" was removed from the French.

Criteria for Qualifying Capital Instruments (Chapter 2)

  1. The new criteria will be effective immediately for new financial instruments issued on and after August 7, 2014. Financial instruments issued prior to August 7, 2014 that do not meet the criteria will be “subject to transitional measures in due course”. However, the transitional measures, including timing and amount, are yet to be defined. Grandfathering of all existing instruments that have previously received confirmation from OSFI would be our expectation and we would welcome an explicit statement to this regard in the Guideline.

It is OSFI’s intent to delay the determination and implementation of transitional measures for capital instruments that do not meet the new qualifying criteria until 2018 when the revised framework is in place. OSFI will determine the transition measures based on the impact of the change.

  1. Given the extensive changes made to the qualifying criteria for available capital instruments in the draft guideline, we suggest the deferral of implementing transitional measures beyond 2015 on existing qualifying capital instruments that do not meet the new criteria. We propose grandfathering all existing qualifying instruments until the inception of the new supervisory framework, at which point any transitional measures can be collectively applied with those of other changes. This would allow sufficient time for companies to resolve potential issues on existing instruments with OSFI and help reduce the risk of the external markets making incorrect inferences on the impact of the new criteria.

It is OSFI’s intent to delay implementing transitional measures for capital instruments that do not meet the new qualifying criteria until 2018 when the revised framework is in place. Until 2018, all previously issued qualifying capital instruments will continue to qualify for inclusion in the respective tiers.

Qualifying Criteria for Common Shares (S2.1.1.2)

  1. The fifth criteria states: "Distributions are paid out of distributable items (retained earnings included)".

     

    It is not clear how this would be demonstrated and incorporated into the terms of the instrument.  Money is fungible.  Furthermore, consideration should be given in a situation where an operating company is well capitalized but has negative retained earnings (e.g. the holding company raises capital and injects into the operating company and when the holding company wants to redeem capital, is the operating company restricted from paying a dividend to the holding company to fund the redemption if the retained earnings are negative?).

OSFI would not expect to see this requirement incorporated into the terms of the instrument. Rather, this criterion is with respect to the treatment of distributions for financial reporting purposes.  Since common shares are reported as equity, any distributions should be reported as a reduction in equity rather than a reduction in the current year’s earnings.

Qualifying Criteria for Tier 1 Capital Instruments Other than Common Shares (S2.1.1.3)

  1. On Page 16, in paragraph 7b “cancellation” should be “non-payment” and “credit event” should be removed (as non-payment of multiple discretionary payments may lead to a downgrade which may be considered a credit event).

Cancellation is used because the coupon or dividend payment must be fully extinguished (i.e. it is cancelled and not simply deferred). A “non-payment” would not meet this requirement.

A credit rating downgrade is not a credit event. Within an instrument’s terms and conditions, there should not be any consequences of coupon/dividend cancellation that would compromise the insurer’s full discretion over payments. For example, it would be unacceptable for an instrument to contain a feature that would be triggered in the event of a cancelled dividend payment. This feature would be viewed as an impediment to full discretion because the insurer may choose not to cancel the coupon/dividend payment in order to avoid the resulting unfavourable consequence.

Information Requirements for Capital Confirmations (Appendix 2-A)

  1. We believe it was not OSFI’s intention to require this capital confirmation on items already covered within the Guideline. This Appendix 2-A checklist appears the same as the list in the CAR Guideline under Appendix 2.2 regarding instruments with NVCC. Since there is no requirement at this time for NVCC within the construct of the MCCSR, we request that OSFI remove Appendix 2-A.

Capital confirmations are not mandatory; OSFI provides insurers with confirmations of capital quality to reduce the risk that new issues are later found, e.g. as a result of an OSFI review, to not qualify for regulatory capital. The intent of the Appendix is to provide insurers with a list of items that are required by OSFI to perform a capital confirmation.

Principles Governing Inclusion of Innovative Instruments in Tier 1 Capital (Appendix 2-C)

  1. Appendix 2-C (page 63) includes a statement at the end of the first paragraph: “Subsequent amendments to the principles, if any, will not disqualify approvals granted under this Appendix”. Does this statement imply that existing Innovative Tier 1 instruments that qualify as Tier 1 capital currently will continue to qualify as Tier 1 capital after the proposed changes? Please confirm this interpretation.

     

    To further clarify that pre-August 7 innovative instruments continue to be included as Tier 1, in section 2.1.1 on page 11, please add “or Appendix 2C’ after “Appendix 2B”.

To qualify as Available Capital, instruments must meet the criteria as outlined in the 2015 Guideline. The first paragraph to Appendix 2-C has been deleted. Until 2018, all previously issued qualifying capital instruments will continue to qualify for inclusion in the respective tiers.

Reference to Appendix 2-C has been added to section 2.1.1

Exclusion of Deferred Tax Liabilities (S2.1)

  1. The Draft Guideline states: “Unless explicitly stated otherwise in this Guideline, deferred tax liabilities may not be used to increase any component of Available Capital /Margin, and the carrying amount of any item required to be deducted from Available Capital /Margin may not be reduced by any portion of the associated deferred tax liabilities”.

     

    As Retained Earnings are net of DTLs and hence Available Capital is lower than otherwise, any deductions from Available Capital must be net of their associated DTLs to avoid double counting.

The treatment of deferred tax liabilities (DTL) will be further reviewed along with the treatment of deferred tax assets as we develop the new standard approach. The statement made that unless explicitly stated otherwise in the Guideline, deferred tax liabilities may not be used to increase any component of Available Capital /Margin, and the carrying amount of any item required to be deducted from Available Capital /Margin may not be reduced by any portion of the associated deferred tax liabilities does not represent a change in policy but rather a clarification of the current MCCSR which provides for specific treatment of DTL for a limited number of elements only.

Intangible Assets (S2.3)

  1. We suggest some alternate wording of footnote 45 to clarify applicability at a consolidated company view, and directly cite the Fair Value Differential that creates the DTL (for intangible assets) associated with the new goodwill on acquisition:

     

    “The fair value differential of the intangible asset is completely non-deductible for tax purposes in the tax jurisdiction in which it was acquired, so that the tax base for the intangible asset remains unchanged from its pre-acquisition amount.   In particular, no portion of any change in the fair value differential of the intangible asset due to amortization or impairment is allowed as a tax-deductible expense, and no portion of the fair value differential of the intangible asset is tax deductible if it is sold.”

The treatment of deferred tax liabilities will be further reviewed along with deferred tax assets as we develop the new standard approach. The footnote text suggested has a different meaning than intended. Therefore the suggestion was not adopted.

  1. There can be circumstances where an intangible asset can arise from a business combination that did not take the form of a share purchase.  We suggest that the first sentence of footnote 45 be modified as follows:

     

    “intangible asset arising from a business combination acquired by share purchase…”

The footnote has been revised to no longer include the reference to “acquired by share purchase” and to now include reference to IFRS 3 – Business Combinations.

  1. In the GIAS, it is stated that “deferred income tax liabilities are not to be included in Available Capital /Margin and associated deferred income tax liabilities are not to reduce the carrying amount of an item that is to be deducted from Available Capital.” However, footnote 45 has been added to the effect that intangible assets can be used to reduce their associated deferred tax liabilities under certain conditions.

The draft GIAS provides a general overview of the proposed changes and will be updated to reflect the existence of exceptions.

Negative Reserves (S2.4.1)

  1. In the French version, the paragraph immediately preceding sub-section 2.4.1.1 should begin: " Pour une société étrangère, la réduction maximale…", we must add the word "étrangère" so that the French version matches the English version.

This has been corrected.

Unsolicited Ratings (S3.1.1)

  1. We request that unsolicited ratings be extended to corporate exposures, similar to the treatment of sovereign exposures.  The distinction between a solicited and unsolicited rating does not seem essential given that rating agencies would issue ratings only when they are comfortable with their validity.

     

    We request that OSFI revisit its position on this matter.  We feel that the added ongoing expense as well as the additional operational complexity to comply with the Guideline of excluding unsolicited ratings is unwarranted.

OSFI does not recognize unsolicited ratings for capital purposes. We have been informed that efforts will be made to reduce the burden for insurers of complying with the requirement (e.g. by publishing a list of unsolicited ratings on a quarterly basis).

Short-Term Obligations Threshold (S3.1.2)

  1. OSFI is shortening the threshold from one year to three months since it is “consistent with their other time frames”. We disagree with this change, as OSFI’s position for making the change, based entirely on wording in IAS7, is not compelling enough. At a minimum, the criteria should be based on the remaining term, not the original term.

There is no change in the 2015 MCCSR with respect to this requirement – OSFI has not shortened the threshold from the 2014 MCCSR.

Leveraged Funds (S3.1.9)

  1. The draft guideline indicates mutual funds that use derivatives to amplify returns are considered to be leveraged. We believe the guideline should further clarify whether funds that use derivatives for risk mitigation or security replication purposes are to be considered as leveraged.

For the purposes of MCCSR, derivatives used for amplifying returns make a fund leveraged, but derivatives used strictly for hedging do not; derivatives that are used for security replication (e.g. futures or bond-equity swaps) make a fund leveraged, since the initial cost is zero.

  1. The flat factor of 15% applied to all leveraged funds under the draft guideline does not take into account the degree to which a fund may be leveraged. As such, the resulting capital requirement may be incongruent with the underlying risks of a fund – slightly leveraged funds and heavily leveraged funds would be subject to the same factor. An alternative approach could be to apply a multiple (scaling factor tied to the maximum amount of leverage allowed within the prospectus) to the C-1 charge applicable to an unleveraged, but otherwise identical, fund.

     

    We suggest the following wording could be added to the existing section of mutual funds and similar funds to appropriately handle funds that employ or allow a limited amount of leverage.
    “For funds that use or allow leverage, the weighted average risk factor determined using the method stated above should be grossed-up by a factor equal to 100% plus the leverage percentage allowed by the fund, to a maximum leverage-adjusted risk factor of 15%.

    • Ex: if a bond fund can only invest in investment grade bonds (BBB or higher) and it can employ up to 25% leverage, the initial weighted average risk factor would be 2% (for non-qualifying business).  The leverage-adjusted risk factor would be 125% of 2%, or 2.50%.
    • Another example is a balanced fund with an initial weighted average risk factor of 8% and a maximum leverage ratio of 20%.  The leverage-adjusted risk factor would be 120% of 8%, or 9.60%.
    • However, if another balanced fund had an initial weighted average risk factor of 12% and a leverage ratio of 50%, the leverage-adjusted risk factor would be 15%, as the maximum factor of 15% would be applicable instead of 150% of 12%, or 16.00%.

The MCCSR will apply a binary treatment to leveraged/unleveraged funds similar to the binary treatment of substantial investments in leveraged/unleveraged funds without control (S2.7). 

The flat factor approach is simple and appropriate for non-equity funds using the strategy because the investor becomes an owner of the residue assets (i.e. equity position) since the lender providing the leverage has priority and is paid first.  The option of using a scaling factor tied to the maximum amount of leverage permitted within the prospectus was explored and rejected. There are other factors that can produce a leverage effect (such as high stock beta) that should be considered as well, which could make the calculation too complex for the standard approach.

Recognition of Equity Option Hedges (S3.8.2)

  1. The Draft Guideline states: “Option hedges of segregated fund guarantee risk may not be recognized in the segregated fund guarantee capital calculation without explicit approval from OSFI. The form and amount of any such recognition will be specified by OSFI at the time of approval. Option hedges of segregated fund guarantee risk that receive recognition in the segregated fund guarantee capital calculation cannot be applied towards other equity risks”.

     

    The 2014 Guideline states: “Option hedges of SFG risk may not be applied towards other equity risks simultaneously”.

    We would appreciate OSFI’s insights into the necessity of the amendment.

The amendment is the result of a review of specific and potential cases which require additional details in the MCCSR to resolve with respect to recognition of option hedges of segregated fund guarantee risk.

Segregated Funds (S8.3)

  1. Segregated fund MCCSR required capital is calculated as the difference between the Total Gross Calculated Requirements (TGCR) and the reported reserve. As such, in the case where the reported reserve is negative, segregated fund capital requirements would be higher than the TGCR as the absolute value of the negative reserves would be fully included within required capital. This is in contrast to non-segregated fund products, where negative reserves would generally result in lower capital requirements.

     

    As such, we believe that the seriatim negative reserve adjustment to available capital should not apply to segregated funds, as there is adequate provision for the risk within segregated fund required capital. Said another way, the total capital requirements for segregated funds should be capped at the TGCR.

This issue is under review and may be addressed in the 2016 MCCSR.

Assets Pledged in a Reinsurance Arrangement (S10.5.1)

  1. In light of the proposed change from “cash” to “currency” in section 3.1.2, should the word “currency” be used in this section, or does OSFI wish to distinguish between these two terms and, if so, what is the distinction?

Cash is usually understood as comprising currency (i.e. banknotes and coins) in addition to cash equivalents such as demand deposits, checks, drafts, etc.  The 0% asset charge is intended only to apply to currency, whereas cash equivalents may be acceptable as reinsurance collateral if a ceding company is able to perfect a valid first-priority security interest in them under applicable law.