American Academy of Actuaries C3 Life and Annuity Capital Work Group Response to Comment Letters regarding September 2009 C3 Phase III Report Presented to the National Association of Insurance Commissioners Life Risk Based Capital Working Group December 2009 The American Academy of Actuaries is a 16,000-member professional association whose mission is to serve the public on behalf of the U.S. actuarial profession. The Academy assists public policymakers on all levels by providing leadership, objective expertise, and actuarial advice on risk and financial security issues. The Academy also sets qualification, practice, and professionalism standards for actuaries in the United States. C3 Life and Annuity Capital Work Group Peter Boyko, F.S.A., F.C.I.A., M.A.A.A., Chair Simpa Baiye F.S.A., M.A.A.A. Nancy Bennett, F.S.A., M.A.A.A. Martin Claire, F.S.A., M.A.A.A. Arnold Dicke, F.S.A., M.A.A.A. Robert L. Ellerbruch F.S.A., M.A.A.A. Jean J. Forrest F.S.A., M.A.A.A. Jason Kehrberg F.S.A., M.A.A.A. Sandeep Kumar A.S.A., M.A.A.A. Shawn Loftus, F.S.A., M.A.A.A Jim Lamson, F.S.A., M.A.A.A. Bob Meilander F.S.A., M.A.A.A Craig Morrow F.S.A., M.A.A.A. Chris Olechowski F.S.A., M.A.A.A. Hubert Mueller, F.S.A., M.A.A.A. Dave Neve, F.S.A., M.A.A.A. Keith Osinski, F.S.A., M.A.A.A. Marianne Purushotham F.S.A., M.A.A.A. Larry Seller, F.S.A., M.A.A.A. David Smith, F.S.A., M.A.A.A. Kenneth Vande Vrede, F.S.A., M.A.A.A. Bill Wilton, F.S.A., M.A.A. The C3WG would also like to acknowledge the work of Susan Christy and Jason Alleyne F.S.A., F.C.I.A., F.I.A. Page 1 of 8
This document has been prepared in response to the NAIC s Life RBC Working Group s request to provide written comments on the issues raised in the comment letters regarding the C3 Work Group s C3 Phase III report (September 2009) discussed on the November 12, 2009, Life RBC Working Group conference call. The C3 Life and Annuity Capital Work Group (C3WG) would like to thank the authors of the comment letters for their input and provide the following comments. Scope (ACLI and UNUM letters) It has been suggested that certain product types be excluded from the proposed C3 requirements, and it has also been suggested that scope be limited specifically to Universal Life (UL) with Secondary Guarantee products. The C3WG considered the concerns related to scope, but continues to believe that a risk-based determination of capital requirements that requires all life insurance products to be included is superior to a product-based inclusion/exclusion approach. Any practical issues associated with a comprehensive stochastic method are addressed by the stochastic exclusion test and the alternative amount. C3WG s charge from the Life RBC Working Group is to develop a methodology for capturing C3 risk that is applicable to all life insurance products. A fundamental premise of that charge is that the C3 methodology should be more directly tied to the risk exposures created by the unique characteristics of an insurer s product portfolio. Determining risk capital by assuming a generic product type is not consistent with a principle-based approach and can result in failure to capture material tail risks that could be present in a company s particular product design or risk management. In developing our proposal, the C3WG formulated a process that is commensurate with the degree of risk. In that context, the Stochastic Exclusion Test is included to serve as a filter, identifying blocks of business as having material tail risk and therefore requiring stochastic modeling. The C3 required capital for those blocks identified as not having material tail risk would continue to be determined according to the existing C3 factor basis. Further, to lessen the burden where stochastic modeling might be required, the Alternative Amount option is provided as a possible simplification tool, facilitating the use of existing risk management processes. C3WG does not feel it is possible to positively, and in all cases, a priori define segments or products within the industry that are low risk. For example, it is entirely possible to have a UL product with a nominal secondary guarantee, say for one or two years, which has less C3 risk than some other product that might be exempt based on product definition. Page 2 of 8
This demonstrates that the key issue is not the product type, but rather the risk exposures created for the company by the product. The risks for the company depend not only on the nature of the liability, but also on the supporting assets and the investment and risk management strategies. Further, any methodology for establishing C3 capital must be based on the company s aggregate portfolio of retained risks, where the methodology establishes capital requirements based on the combined effect of a company s business decisions product, investment, and risk management strategies. Any methodology based on a portion of a company s risks has the potential for manipulation and unintended results. Limitation of scope to a product or group of products creates a number of issues that should be considered. For example, there may be an issue when the company manages UL with Secondary Guarantees together with other products. The company may decide to do this because the other products provide some form of natural hedge, have a similar profile, or for other business reasons. If the other products are carved out of the C3 requirements due to scope limitations, then the company would need to allocate items between the two product groups, such as: Reinsurance treaties -- may not be product specific. Hedging -- may be done for the block as a whole. Asset investment strategy -- may not be appropriate for the UL with Secondary Guarantee product alone. The scope limitation may introduce complexities and may result in a distorted picture of the risks for the block as a whole. In summary, C3WG continues to believe that a risk-based determination of capital requirements that requires all life insurance products to be included is superior to a product-based inclusion/exclusion approach for the above reasons. Any practical issues associated with a comprehensive stochastic method are addressed with the stochastic exclusion test and the alternative amount. Misplaced Reserve Adjustment (Aegon letter) In its comment letter, Aegon states that the C3WG proposal is based on the premise that existing reserve standards are incorrect and the proposal attempts to fix reserves. We strongly disagree with this assertion. The C3P3 methodology accepts the statutory reserves as determined in accordance with the valuation standards. The C3WG proposal takes a Total Asset Requirement (TAR) approach to determining capital requirements. The proposal identifies how much additional resources are necessary in addition to the reserves already being held. We believe that this is the correct approach to take. At the outset of our work, we considered two possible formulations, both within the context of a Total Asset Requirement approach. The first approach takes the excess of TAR over the reported reserve (our recommended approach), and the second approach Page 3 of 8
takes the difference between two risk levels, for example CTE90 and CTE70 with an adjustment for the difference between the reserve held and the CTE70 amount. We evaluated the pros and cons of these two options in concert with some initial modeling results and concluded that the recommended approach was the better of the two. The second approach could potentially have had a large impact on capital beyond C3 due to differences between statutory reserves and the CTE70 amount. This was considered too dramatic a change and so we continue to recommend a TAR reserve approach. The comment letter presents a way of looking at the C3 risk charge algebraically. It should be noted the C3 Phase III proposal is to apply to policies with both formula-based reserves and principle-based reserves (PBR) in the future. The algebraic breakdown will differ between policies with formula-based reserves and policies with reserves determined under PBR. It should be noted that for policies with reserves determined under PBR the issue of the so-called reserve adjustment term goes away and the two approaches that had been considered are the effectively the same. Also, it is very important to note is that under our recommended approach, the marginal effect on the resulting capital requirement is limited to reducing the existing C3 requirement of 0.5% of statutory reserve to zero. This limited effect by no means could be construed as an attempt to fix reserves. In fact, the limited impact is obscured in the algebra as presented because the C3 amount is subject to the maximum of zero, and the excess of TAR over reserve. A second issue related to the reserve adjustment that was raised in the Aegon letter is that of the impact of increases in reserve levels on RBC results. Aegon notes that an increase in reserves increases a company s RBC ratio. We do not see this as an issue given the level of disclosure required. While a given company may hold reserves in excess of the required minimum, any attempt to manipulate the amount of excess reserves from year-to-year would show up in the required disclosures. Aegon also noted that the potential to eliminate C3 risk by increasing reserves may encourage C3 risk-taking. Once again, we believe the existing level of disclosure would be sufficient to identify those that were trying to game the proposal and do not anticipate this will be a huge problem for the profession. However, in order to avoid any such manipulation, one could consider a methodology where the statutory minimum reserve is deducted from TAR. This is not an attractive solution as this would disadvantage those that hold additional reserves against specified risks and not for purposes of manipulation. Therefore, we would propose that a better solution would be to require, whenever there is a change in the basis for determining the additional reserves, a separate disclosure of the reserve between minimum and add-on components, the rationale for the add-on, and the amount of the change since last year. This would identify those companies that were attempting to manipulate the C3 requirements without disadvantaging others holding additional reserves for appropriate reasons. Page 4 of 8
Double-counted Diversification Benefit (Aegon letter) The C3WG proposal provides for the C3 amount to be allocated between C3a interest rate and C3c equity risk components. This was done for consistency with C3 Phase II and with the intent of aligning the type of risk of a block with the appropriate RBC component. In the case of a block of business backed solely by fixed income assets, there is no need to allocate the C3 amount it is all interest rate risk. Similarly, in the example of a block of business solely backed by equities there would be no need to allocate the C3 amount it is all market risk. Where both interest rate and equity risks exist, we do not provide a rule or formula for allocation. Rather, we rely on disclosure and guide the actuary to consider the covariance effect of any allocation. Inclusion in the bucket with least covariance impact would be a conservative and acceptable allocation method. We would like to raise a point about the algebra presented in the Aegon letter and the impact of diversification as presented. We believe that one needs to consider the effect in the context of the entire formula inclusive of all risks. It has been stated a number of times by industry that the C3 risk is a small portion of the total industry RBC. So, when considering the impact of any C3 allocation, all risks of the company (i.e., C1 to C4) need to be considered. In doing so, we believe the overall allocation impact on the industry would be much less than presented. To the extent that this remains a significant issue, possible solutions would be to either preclude allocation and ascribe the entire C3 amount to C3a, or conservatively allocate the amount to the bucket with the lowest covariance. Our group does not feel either of these solutions is necessary since the existing requirements for disclosure and justification of the allocation method are sufficient. Altered Definition of C3 Risk (Aegon letter) The Aegon letter states that C3WG has taken a new and novel approach in defining and measuring C3 risk. In fact, the approach is neither new nor novel. C3WG intentionally extended the existing C3 Phase I and C3 Phase II approaches to life insurance. The C3 measure is determined based on actual assets supporting the liabilities, reflecting the greatest present value of accumulated deficiencies in any given scenario. The C3WG approach takes into account that in the case of products with a cash surrender value (CSV), the company needs to have on hand assets in excess of the CSV. The November 2008 American Academy of Actuaries Life Capital Adequacy Subcommittee Report on the Working Reserve for the C3 RBC Calculation discusses the issues relating to working reserve and the pros and cons related to the various possible working reserve definitions. It was concluded that for the purposes of this project the use of CSV was the best of the options. The choice reflects the view that the purpose of capital is to have Page 5 of 8
sufficient assets on hand so that the regulator would be able to offload the block to another company to manage. CSV was considered to be the best choice in that regard. In the context of this discussion on methodology we must note that our charge was to fit the proposal within the existing NAIC accounting framework. As such, projected assets need to be valued in accordance with their statutory book values. Although it might be desirable from some perspectives to re-state all values to market values, this is not the system in which we currently operate. Scenarios (ACLI and UNUM letters) It has been suggested that a small scenario set be made available as an alternative to the stochastic scenario set. In a sense, a smaller scenario set has been provided in the form of the Stochastic Exclusion Test. In this case, the company is limited to using 16 deterministic scenarios and may determine its C3 requirement using the existing C3 factor if it passes the test. Note that the Stochastic Exclusion Test does not directly establish a C3 amount but rather serves as a basis for determining whether a block of business has tail risk beyond a certain threshold and whether stochastic modeling should be performed. C3WG has discussed the feasibility of defining a small scenario set with the Academy s Economic Scenario Work Group. It was concluded that due to the diversity of products inforce, supporting assets, and investment strategies, it would not be feasible or desirable to develop a small scenario set for purposes of determining the C3 requirement. It has also been suggested that a fixed number of scenarios be stipulated as a safe harbor in order to lessen the work burden and ensure that additional scenarios beyond those scheduled to be run need not be run at year-end and no further testing would be necessary beyond that fixed number of scenarios. It should be noted that the Academy has recommended an interest rate generator for use in stochastic reserve and capital calculations. A company can use the Academy generator or a proprietary generator to develop an appropriate number of scenarios that satisfy the recommended calibration criteria. Further, the actuary can apply scenario reduction techniques to determine the appropriate number of scenarios to use in the stochastic modeling. A very important point to remember in recommending calibration criteria and allowing for the use of scenario reduction techniques is that the appropriate number of scenarios will depend on a company s unique risk profile. It should be noted that a fixed number of scenarios applied to a fat tail risk could result in a CTE level that is different for a mix of different risk distributions (with and without fat tails) potentially resulting in differing CTE levels across companies. Under the C3WG approach, the number of scenarios needed is determined by the actuary and deemed appropriate for a specific company by the actuary. The actuary is required to Page 6 of 8
document and justify the choice of number of scenarios used in the determination of C3 amount. Placing the requirement on the actuary to determine the scenario set will result in the same CTE level for all risks fat-tail and otherwise. The C3WG recommendation is consistent with both the existing requirements under C3 Phase II and the proposed requirements under VM-20. This approach recognizes that, depending upon the distribution of the risk, an arbitrary number of scenarios may not materially capture the risk. It should also be noted that the company may perform the scenario analysis using data based on a date prior to the valuation date, subject to making appropriate adjustments to the resulting C3 amount for changes between such date and the valuation date. This option should lessen the year-end burden and the unanticipated need to run additional scenarios. Lastly, one of the concerns expressed about the C3WG approach is that the only way to determine whether a sufficient number of scenarios have been run is to run more. This is not true. Through variance reduction and other sampling techniques, it is possible to improve the accuracy of an estimate more efficiently than simply increasing the number of simulations. 1 Materiality Test (UNUM letter) The UNUM letter suggests that, in addition to the Stochastic Exclusion Test, some additional form of materiality test for exclusion be developed based on overall size of volatility of RBC (for example less than 5% of RBC). C3WG considers development of such a materiality test beyond its capability as C3WG members do not have the necessary data or the regulatory insight to effective develop such a test. The C3WG continues to believe that the Stochastic Exclusion Test and Alternative Amount options are appropriate simplification methods based on risks rather than size--a much less sophisticated and potentially inappropriate method. Documentation (ACLI and UNUM letters) C3WG has reviewed the suggested changes regarding documentation and agrees that documentation requirements could be streamlined and efficiencies could be gained by integrating the C3 documentation requirements within the Actuarial Opinion Memorandum. The requirement to disclose the number n of scenarios used and the methods used to determine the sampling error of the CTE (90) statistic when using n scenarios could be 1 See Manistre, B.J. and Hancock, G. (2003), Variance of the CTE Estimator, 2003 Stochastic Modeling Symposium, Toronto, ON, September 2003. Page 7 of 8
less specific. Simply requiring disclosure of the number of scenarios and rationale for such number should suffice. We agree that the requirement for a description of the internal controls and procedures used to ensure the appropriateness of the actuary s judgment when permitted is excessive and does not belong within the scope of developing C3 requirements. We also agree that the requirement to provide a list of the key risk measurement tracking tools that the company uses as an early warning of changes in experience between valuation dates is excessive and does not belong within the scope of developing C3 requirements. The requirement for descriptions and results of material sensitivity tests performed was included primarily to support the development and choice of prudent estimate assumptions. This may be adequately addressed in the Actuarial Opinion Memorandum. The requirement for disclosure of material changes in assumptions from the previous year is consistent with the existing requirements under C3 Phase II and included for consistency. Page 8 of 8