October 25, Mr. Hans Hoogervorst International Accounting Standards Board 30 Cannon Street London EC4M 6XH United Kingdom

Transcription

1 K 333 S. Wabash Ave. Chicago IL October 25, 2013 D. Craig Mense Executive Vice President and Chief Financial Officer Telephone Facsimile Internet Mr. Hans Hoogervorst International Accounting Standards Board 30 Cannon Street London EC4M 6XH United Kingdom Re: Exposure Draft ED/2013/7, Insurance Contracts Dear Mr. Hoogervorst: CNA Financial Corporation (CNA, we, or our) appreciates the opportunity to comment on the International Accounting Standards Board s (IASB or Board) Exposure Draft ED/2013/7, Insurance Contracts (ED). We are also separately providing comments to the Financial Accounting Standards Board (FASB) on its Proposed Accounting Standards Update, Insurance Contracts (FASB s Proposed Update), and that response is attached to this letter for your reference. CNA is the eighth largest commercial insurance writer and the thirteenth largest property & casualty (P&C) writer in the United States and has insurance operations in both North America and Europe. In addition, CNA s non-core business includes life and group (L&G) lines of business that are in run-off. Loews Corporation owns approximately 90% of CNA s outstanding common stock. In light of the need for a robust, comprehensive standard for insurance contracts under international financial reporting standards (IFRS), we support the IASB s stated objectives for this project of improving financial reporting by providing a consistent basis for the accounting for insurance contracts and making it easier for users of financial statements to understand how insurance contracts affect an entity s financial position, financial performance, and cash flows. We also believe, as a secondary objective, that international convergence of the financial reporting requirements for insurance contracts would benefit both preparers and investors. While we acknowledge these objectives and the theoretical merits of several individual aspects of the ED, we believe the proposed accounting model would increase complexity, which will reduce clarity and comparability of insurer financial statements. Specifically, comparability between periods and between peer companies would become exceedingly difficult, if not impossible. The cost to implement and maintain the proposed model will be significant due to this complexity. We currently apply U.S. generally accepted accounting principles (U.S. GAAP) and IFRS in our operations in North America and Europe, and we believe that U.S. GAAP represents a comprehensive accounting model for P&C insurance contracts that is well understood and provides investors with decision-useful information. We do not believe the proposed accounting model would provide more relevant, transparent, and comparable information to investors as compared with existing U.S. GAAP. For life insurers, we believe the proposed model, with certain exceptions noted below, does meet the Board s objectives.

2 Page 2 The ED would require significant changes (both to current accounting requirements and those proposed in the Board s Exposure Draft ED/2010/8, Insurance Contracts (2010 ED)) that, in our view, will increase complexity, reduce transparency, and provide less precise financial reporting results for P&C contracts due to layering of estimates and judgments, including timing of cash flows, discount rates, and portfolio composition. In certain situations, these changes may introduce significant accounting-driven volatility into the income statement. While we appreciate that the Board has attempted to address criticisms raised against the 2010 ED, we believe there should be a balance between the desire for theoretical purity and overall complexity. Each of the major changes introduced by the ED results in an increase in complexity, which in some cases may be warranted, but in others it is not. Coupled with the voluminous proposed disclosure requirements, the financial statements will be overwhelming to users and difficult to understand and interpret. In addition, traditional measures of P&C performance, such as the loss ratio and combined ratio, will not be apparent to financial statement readers. For P&C contracts, several aspects of the proposed accounting model would require significant and complex changes in operational and actuarial processes that are not consistent with how we manage the business. We therefore expect implementation to be very difficult and costly. Such costs can only be justified if the new accounting model provides financial statement users with significantly improved transparency into the financial performance of the business, the underlying key judgments, and the risks and uncertainties related to those judgments. As stated above, we do not believe this objective has been achieved. Furthermore, the Board has not achieved the secondary objective of convergence with the FASB s Proposed Update. Although the accounting models proposed by both the IASB and the FASB (collectively, the Boards) have some similar fundamentals, significant differences remain regarding: 1. One model versus two models The IASB designed the premium allocation approach (PAA) as a proxy for the building block approach (BBA). Under the ED, the PAA would be permitted, but not required, for contracts with coverage periods shorter than one year or where it is a reasonable approximation of the BBA. Alternatively, the FASB views the PAA as a separate and distinct model that represents a simpler approach to the BBA. 2. Margin composition The ED includes an explicit risk adjustment in the measurement of the insurance contract liability and a contractual service margin that is recognised on a systematic basis in line with the pattern of services provided under the contract. In contrast, in the FASB s Proposed Update, the margin under the BBA is viewed as unearned profit that is ratably recognised as insurance contract revenue as the associated cash flows become more certain. 3. Margin unlocking In the ED, a net increase in expected future cash outflows relating to future coverage or future services is offset against the remaining contractual service margin (other than for the effect of changes in discount rates), and a net decrease in expected future cash outflows is added to the contractual service margin. The risk adjustment is remeasured each reporting period with changes recognised immediately in net income. In contrast, in the FASB s Proposed Update, under the BBA, all changes in estimates (other than the effect of changes in the discount rates) are recognised immediately in net income and as an adjustment to the insurance liability.

3 Page 3 4. Allocation period for recognising margin In the ED, the contractual service margin would be recognised in net income as part of insurance contract revenue over the coverage period. The risk adjustment would be recognised in net income over the coverage and settlement periods. In the FASB s Proposed Update, under the BBA, the margin would be recognised in net income as part of insurance contract revenue over the coverage and settlement period. 5. Acquisition costs In the ED, the acquisition costs included in the measurement of the liabilities would include all acquisition costs directly attributable to obtaining a portfolio of insurance contracts, including costs related to unsuccessful efforts. In the IASB s BBA, the fulfilment cash flows would include expected acquisition costs. In the FASB s Proposed Update, acquisition costs included in the measurement of the liabilities would be limited to acquisition costs directly attributable to a portfolio of insurance contracts that are successfully obtained. In the FASB s BBA, the margin would reflect expected profit after deducting acquisition costs incurred. 6. Participation features The IASB and FASB reached a similar conclusion that participating policies that do not have discretion as to the amount to be credited to policyholders should be measured on the same basis as the related assets and changes in the measurement should be presented in the same statements (i.e., net income or other comprehensive income) as the assets. This was referred to as mirroring during Board deliberations. However, in the ED, that conclusion is extended further to contracts where the insurer has a practice without a strict contractual linkage as to the percentage participation. 7. Reinsurance For reinsurance contracts, the ED would require application of the classification criteria (i.e., use PAA only if it is a proxy for BBA) rather than defaulting to the classification of the underlying direct policies. In contrast, the FASB decided that the cedant should account for a reinsurance contract using the same approach used to account for the underlying direct policies (e.g., if the cedant accounted for the direct policies using the PAA, the reinsurance contract would also be accounted for under the PAA). As an international company with operations subject to IFRS and U.S. GAAP, the cost of implementing the proposed changes to insurance contracts guidance, as issued by the IASB and FASB, will be significantly higher in the absence of true convergence. We understand that the IASB is seeking input only on the significant changes made in response to the feedback received on the proposals in the 2010 ED, and does not intend to revisit issues previously deliberated. However, given the complexity of the proposal and the significant passage of time since the issuance of the 2010 ED, we feel that it is appropriate to view the proposal as a whole, rather than commenting on specific items outside the broader context of the ED. As such, we identified the following concerns and proposed refinements to the guidance. One Model vs. Two Models The ED proposes a single model for recognition of all insurance contracts, with a simplified version of that model provided if certain criteria are met. In contrast, the FASB s Proposed Update retains separate accounting models for contracts with different economic characteristics. Under the FASB s Proposed Update, most life, annuity, and long-term health contracts would apply the BBA and most property,

4 Page 4 liability, and short-term health contracts would apply the PAA. We fundamentally believe that P&C insurers and life insurers operate different business models that require different accounting approaches to appropriately reflect the economics of these business models. As such, we support the FASB s view that the BBA and the PAA represent two distinct measurement models. While we have significant concerns with specific aspects of the PAA as it applies to P&C contracts, as noted both throughout this letter as well as in our response to the FASB s Proposed Update, we believe the BBA, as it applies to life-type contracts, with certain exceptions, does meet the Board s primary objectives. We feel that U.S. GAAP for P&C contracts is well established and has been used globally without posing significant issues with regard to application, auditing, and analysis. Generally, users understand the accounting and reporting for P&C contracts, and the information (including performance metrics) that is reported to users is consistent with how management reviews performance of the business. Additionally, the uncertainty regarding the estimation of the claims liability is well understood with significant disclosure regarding the risks and uncertainties in the Critical Accounting Estimates section of the Management s Discussion and Analysis section of Form 10-K of Securities and Exchange Commission (SEC) filers. We strongly encourage the Board to consider dividing the proposed model into two separate, distinct models to be applied to contracts with different economic characteristics. In addition, we encourage the Board to consider aligning the requirements for the PAA more closely with current U.S. GAAP requirements for P&C contracts, making only targeted improvements to the U.S. GAAP model. These improvements may include: Broadening the scope of the guidance to cover noninsurance entities that issue insurance contracts; and Clarifying the risk transfer requirements for insurance contracts. Based on our comments above, where appropriate throughout the rest of this letter, we have differentiated our concerns as they relate to either the PAA or the BBA as separate models, as well as how those models relate to our P&C and L&G segments. Expected Value Under the guidance in the ED, for contracts accounted for using the PAA, the liability for incurred claims would be measured as the present value of the expected fulfilment cash flows (expected value) reflecting all available information on the amount, timing, and uncertainty of the remaining future cash flows. The application of this expected value approach to P&C contracts accounted for under the PAA represents a significant departure from the deterministic actuarial reserve measurement approach currently in use. Under current U.S. GAAP, carried reserves reflect management s best estimate (MBE) of the unpaid claim and claim adjustment expense liabilities of the company. For P&C companies, MBE is typically based on an actuarial central estimate (ACE) which is defined in Actuarial Standard of Practice No. 43, Property/Casualty Unpaid Claim Estimates (ASOP 43), as an estimate that represents an expected value over the range of reasonably possible outcomes. The detailed analyses performed use a variety of accepted actuarial methods and techniques to produce a number of estimates of ultimate loss. Our actuaries determine a point estimate, the ACE, of ultimate loss by reviewing the various reserve estimates and judgmentally assigning weight to each estimate given the characteristics of the product being reviewed. As this process may not include all conceivable outcomes, it produces a conceptual mean

5 Page 5 rather than an unbiased probability-weighted statistical mean. There is no current requirement (or recommendation) for the utilisation of probability distributions to calculate the ACE and, in practice, most actuarial reserve analyses do not create explicit probability distributions. To force the creation of a distribution of outcomes and of payment patterns or cash flows will imply a level of sophistication and certainty that does not exist in the process. This could create a false impression of the ability of actuaries to quantify the uncertainty inherent in any reserve estimate. Each quarter, the ACE, in conjunction with the results of the detailed reserve reviews, is discussed with our senior management to determine management s best estimate of reserves. This group considers many factors in making this decision. The factors include, but are not limited to, the historical pattern and volatility of the actuarial indications, the sensitivity of the actuarial indications to changes in paid and incurred loss patterns, the consistency of claims handling processes, the consistency of case reserving practices, changes in our pricing and underwriting, pricing and underwriting trends in the insurance market, and legal, judicial, social, and economic trends. Our recorded reserves reflect our best estimate as of a particular point in time based upon known facts, consideration of the factors cited above, and our judgment. The carried reserve may differ from the ACE as the result of our consideration of the factors noted above as well as the potential volatility of the projections associated with the specific product being analysed and other factors impacting claims costs that may not be quantifiable through traditional actuarial analysis. However, the carried reserve is still within the range of reasonably possible outcomes as determined by the actuaries. This process results in MBE. The removal of management judgment from the reserve estimation process would generally result in a reduction in nominal loss reserves and increased volatility in earnings. We do not believe this outcome would be in the best interest of users of the financial statements or policyholders to the extent that it adversely affects the insurer s solvency. Without full stochastic modelling, which is costly and not always practical or prudent, we believe the expected value approach outlined in the ED suggests a level of certainty in the model that should not be reasonably relied upon. We believe the final guidance should set forth a broader principle and allow the actuarial profession to set the standards on how that principle will be met. We do not believe that preparers should be restricted in their approaches and we do not believe accounting standards should govern the actuarial profession or how actuarial estimates are derived. We believe the two-model approach proposed by the FASB provides the required flexibility to differentiate between appropriate expected value calculation approaches for both the PAA and the BBA models. In addition, we believe that management is ultimately responsible for the appropriateness of the estimates recorded in the financial statements; therefore, management should have the latitude to adjust the ACE reserve estimate, if deemed appropriate. Time Value of Money In general, we agree with the theoretical merits of reflecting the time value of money in the measurement of the fulfilment cash flows. However, with respect to the P&C industry, we question whether the significant complexity introduced by this requirement is supported by improved decision-useful information. With the exception of reserves that are fixed and reliably estimable, the introduction of discounting is inconsistent with how we manage our business. Underwriting performance is typically measured internally, and analysed externally, on a nominal basis. This is the same basis upon which the claims are settled (i.e., on a nominal versus a discounted basis).

6 Page 6 As discussed in our Expected Value commentary above, there is inherent uncertainty involved in the estimation of the ultimate loss reserves on a nominal basis. Uncertainty also exists, to an even greater extent, in estimating the timing of the payment of losses and all of the other contract fulfilment elements. A liability accounting measure that is based on measuring cash flows rather than the ultimate liability introduces more judgment, subjectivity, and variation. While true for all policies, this issue is compounded for longer-tail exposures, such as workers compensation and asbestos and environmental pollution liabilities. The requirement to discount loss reserves will introduce significant accounting-driven volatility into the income statement. Accounting-driven volatility will arise in situations where the total nominal loss reserves are not adjusted, but the allocation of those reserves by policy year changes. The re-estimation of loss reserves between policy years, which occurs frequently, would require the application of different yield curves to the loss reserves transferred. Additionally, the re-estimation of loss reserves between expected payment dates would require the application of different points on the same yield curve for income statement measurement. The net impact on a present value basis will be reflected in the income statement as underwriting income or loss. The interest expense would also be impacted on a go-forward basis due to changes in accretion rates. We do not believe this volatility, which would need to be explained by management, is meaningful to users. We understand that current U.S. GAAP guidance produces a mixed-attribute accounting model with most investment assets being measured at fair value and P&C insurance contract liabilities being measured on a nominal basis. However, this should not be a cause for concern. As noted above and throughout this comment letter, the current accounting model is well understood by users as they evaluate the insurer s performance and provides decision-useful information. The measurement of investment assets is consistent with other industries, while the measurement of P&C loss reserves on a nominal basis reflects the manner in which we manage our business. The existing business model separates underwriting from investing, as investors typically invest in P&C insurers based on their ability to produce underwriting income versus investment income and they typically apply a different multiplier on each profit stream (higher for underwriting and lower for investing) to determine their enterprise valuations. The ED would integrate investing activities into the underwriting model, which will likely cause investors to back out the impacts of discounting to get back to reserves on a nominal basis. Under the ED, insurers will be required to select and retain a yield curve for each insurance contract or group of contracts, which would serve as the interest accretion rate for the interest expense calculation and as the reference point for recognising changes to fulfilment cash flows due to changes in discount rates in other comprehensive income (OCI). It is not clear how we would apply this guidance given that cash flows are identified at the portfolio level, but yield curves are identified at the contract level. The Board does not provide adequate guidance to assist preparers and users in reconciling the two units of account. The ED requires that discount rates reflect the characteristics of the insurance contract liability but acknowledges that discount rates may not be observable in the market. An entity should maximise the use of current observable market prices of instruments with similar cash flows but should adjust those prices to reflect the characteristics of the insurance contract liability. Although the ED does not prescribe the method for making those adjustments, it does outline two possible approaches: the bottom-up approach and the top-down approach. We believe that such methods introduce further complexity into an already complex general requirement to discount and may hinder comparability among insurers. With respect to

7 Page 7 the bottom-up approach, the calculation of an illiquidity premium, which is not market observable, could produce application inconsistencies that may not be overcome through expanded disclosure. We therefore suggest, as a practical solution to improve comparability among insurers, that the Board propose that nonlinked contracts (i.e. non-participating contracts) be discounted using a high quality corporate bond rate, similar to the requirement of International Accounting Standings 19, Employee Benefits. As the concept of time value of money applies to the PAA, we do not agree with the proposed requirement to discount the liability for remaining coverage if the contract has a financing component that is significant. We understand that the Board is seeking to separately record an interest expense to match investment income; however, we believe the liability for incurred claims is a bigger driver of this interest expense, as typically the payout period for incurred claims would exceed the period of the financing element for most policies. We acknowledge that a practical expedient exists that allows an entity to not adjust the liability for remaining coverage to reflect the time value of money if the entity expects, at contract inception, that the time period between when the policyholder pays all or substantially all of the premium and when the entity provides the corresponding part of the coverage is one year or less. However, we do not believe that the additional complexity introduced by this requirement is supported by improved decision-useful information. Although we currently do not have a significant number of policies that would fall outside the practical expedient, changing economic conditions may result in more P&C contracts being structured over a period longer than 12 months. Risk Adjustment We are concerned with the complexities in determining an appropriate risk adjustment and the inability to properly define and measure risk with a single number. While we recognise that the two-margin approach may be a more theoretically pure approach to viewing the economics of insurance contracts, we are concerned it will almost certainly result in incomparable results based on the amount of subjectivity involved. As such, we believe users of financial statements will remove any income fluctuation due to a risk adjustment from earnings and effectively add to capital any embedded risk adjustment. For example, Standard & Poor s Risk-Based Insurance Capital Model explicitly adds back any prudential margins included in reserves and equalisation/catastrophe reserves to Total Adjusted Capital. We believe the same approach would likely apply to any risk adjustment embedded in the statement of financial position, and other financial statement users will likely adopt similar approaches should risk adjustments be included in estimates of net cash flows. Insurers would be required to incur significant costs and resources to estimate risk adjustments each reporting period. As outlined above, these costs would be incurred to estimate an amount that does not provide incremental decision-useful information. We believe the composite margin approach is a more suitable alternative. Level of Aggregation The ED specifies that an entity should aggregate insurance contracts into a portfolio of insurance contracts when determining the contractual service margin at initial recognition, estimating the expected fulfilment cash flows (including the determination and recognition of acquisition costs and other overhead costs), and performing the onerous contract test. The ED defines a Portfolio of Insurance Contracts as a group of contracts that (1) provide coverage for similar risks and priced similarly relative to the risk taken on and (2) are managed together in a single pool. Despite the importance of the portfolio concept, we do not believe that the ED provides sufficient application guidance for preparers.

8 Page 8 Specifically, the Board provides little guidance regarding the interaction of fulfilment cash flows, determination of discount rates, and subsequent recognition of the contractual service margin with the concept of the portfolio. For example, under the BBA, the IASB proposes a principle that when recognising the contractual service margin, reporting entities should use a level of aggregation that ensures that the contractual service margin is recognised in line with the pattern of services provided under the contracts to which they relate. In the Basis for Conclusions, the Board acknowledges that, in practice, this may result in a smaller unit of account than the portfolio that entities would generally use to manage contracts under the same guidance, and may require entities to group together contracts that have similar contract inception dates, coverage periods and service profiles (i.e. at a sub-portfolio level). Alternatively, an entity could determine the recognition of the contractual service margin at an individual contract level, but the IASB acknowledged that requiring that approach in all circumstances would place a significant burden on preparers, and would be onerous. We appreciate that the IASB has chosen a more principles-based approach for the determination of portfolios. Based on this approach and analogising to the discussion of the recognition of the contractual service margin as noted above, we believe that the interest accretion rate could be determined at a subportfolio level, defined as the portfolio of contracts entered into during a specified period of time for example, the contracts within a portfolio written in a particular quarter. However, it is not clear whether our interpretation of the guidance in the ED would be shared by others or would be appropriate. As such, illustrative examples regarding the application of the principles for determination of portfolios would be helpful. In addition, it is unclear whether the IASB and FASB intended to diverge in application of the definition of portfolio, given that the criteria for determining a portfolio are not completely converged. We encourage the Boards to work toward a converged solution on this issue. Onerous Contracts PAA The ED requires entities to recognise an onerous contract in the precoverage period, but only if the portfolio of insurance contracts to which the contract will belong is onerous. Based on the discussion in paragraph BCA214, this results in recognition of adverse changes in circumstances without the need to track the contract individually before the coverage period; rather, high-level reviews of portfolios of contracts may be sufficient. We appreciate that the IASB listened to concerns raised in response to the 2010 ED that the high costs of implementing the necessary systems changes in order to account for the contract before the coverage period would outweigh the benefits of doing so. However, it is unclear how the high-level reviews of portfolios of contracts would be conducted without tracking individual contract cash flows at some level, and the Board does not provide any examples of how to conduct this analysis within the ED s Illustrative Examples. As such, illustrative examples in this area would be greatly appreciated. Under the onerous contract model, entities would be required to consider expected costs related to future events that have not yet occurred, and include those costs within the expected probability-weighted estimate of the future cash outflows. This requirement is in direct conflict with the underlying concept of the liability for incurred claims, which estimates future cash flows relative to a claim that has already been incurred and becomes problematic when possible catastrophe events are contemplated prior to a period end date but are not expected to occur, if at all, until subsequent to the period end date. The guidance as written requires entities to assign a probability of loss occurrence to the event and also to estimate claim and claim adjustment expense liabilities in the event of occurrence. We do not support this

9 Page 9 guidance. We believe it to be inconsistent with how we manage our business, and we believe that probabilistic-based measurement of the liability arising from low frequency, high severity events would lead to non-comparability as there is generally insufficient data available to develop and corroborate the probabilities of these types of events. We suggest that the proposed guidance for onerous contracts exclude low frequency, high severity events unless the event is probable and reasonably estimable at the balance sheet date, using only information that theoretically existed at the balance sheet date. This approach would be consistent with current U.S. industry guidance and U.S. GAAP, would provide decision-useful information, and would be operationally easier to implement. Chapter 3 of the American Institute of Certified Public Accountants Audit and Accounting Guide, Property and Liability Insurance Entities, notes that it would be rare to have a catastrophe event that would meet both criteria of being probable of occurring and being able to reasonably estimate the extent of the damage, using information that theoretically existed at the balance sheet date. For example, potential losses from a hurricane sitting off the coast of Florida at period end that hits the coast and causes damage shortly thereafter in the subsequent accounting period would rarely meet the criteria of being both probable and reasonably estimable, using information that theoretically existed at the balance sheet date. As a result, the hurricane should not be included in a premium deficiency calculation. Furthermore, we believe this view is supported by both ASC 450, Contingencies, and International Accounting Standard 37, Provisions, Contingent Liabilities and Contingent Assets (IAS 37). ASC Subtopic , Loss Contingencies, requires that before an estimate for a loss contingency can be accrued, (1) it is probable that the loss has occurred and (2) the amount of the loss can be reasonably estimated. IAS 37 provides slightly different criteria for recognition of a provision than ASC 450, but does require that it is more likely than not that there will be an outflow of benefits and, if so, that the amount can be estimated reliably. Due to the nature of many catastrophe events, it is not possible to know in advance with a high degree of certainty the following: When, or if, the catastrophe event will occur; The exact location of the event. A catastrophe event hitting a densely populated area would result in a significantly greater expected loss than an event hitting a less densely populated area just a few miles away; The magnitude of the event; Types of damages subject to coverage; Existence of deductibles, self-insured retentions, reinsurance coverage that could materially vary depending on the above variables. More often than not, the above variables only become resolved in the days and weeks after a significant catastrophe event. We are also concerned that the ED does not provide clear guidance as to appropriate sources of market data for the calculation of probabilities. For example, if a meteorologist releases a report at the beginning of the year that predicts ten severe weather events during the year, would all companies consider this report in the same manner? Also, how would companies reconcile conflicting data regarding the probability and magnitude of future catastrophe events? Due to the above listed concerns, we do not believe the proposed guidance for onerous contracts under the PAA represents an improvement to the guidance for premium deficiency testing under existing U.S.

10 Page 10 GAAP. We believe the inclusion of an onerous liability in the financial statements, caused by a potential catastrophe event occurring in a subsequent period, introduces unnecessary volatility. By the time the financial statements are issued, it is more likely that entities will know whether the catastrophe event did occur, and if so, have a more reliable estimation of the liabilities incurred. A disclosure of such information would be more meaningful than the inclusion of an onerous liability based on a probability weighted estimate of the impact of the catastrophe event. Unbundling Under the proposed guidance, an insurer would be required to separate its insurance contracts into noninsurance components and insurance components and measure the non-insurance components under other standards. In addition to unbundling certain embedded derivatives and investment components, an insurer would be required to unbundle distinct performance obligations to provide goods and services. The ED states that a performance obligation to provide a good or service is considered distinct if either of the following criteria is met: The entity (or another entity that does or does not issue insurance contracts) regularly sells the service separately in the same market or jurisdiction; or The policyholder (or its beneficiary) can benefit from the good or service either on its own, or together with other resources that are readily available to the policyholder. As this concept relates to claims processing services, we believe the Board intended that performance obligations would only be considered distinct if (1) they were offered as a standalone service by the insurer to the policyholder; and (2) were for periods during which the insurer did not have insurance risk, due to the nature of the contract e.g. a stop-loss coverage with a high limit. If the intent of the Board is to require insurers to unbundle claims processing services on all policies, if such services were offered on a standalone basis for a subgroup of customers or a subgroup of policies, we do not support this requirement. It would be an overly complex exercise to bifurcate the insurance premium between the insurance coverage and the claims processing services based on what may be limited market data on the pricing of such services. It introduces an information burden on the reporting entity as market participant information may not be readily available, or may be available at a cost that is not justified by the benefits of the information. In addition, we see little benefit with requiring this unbundling, as insurance contract revenue will be recognised on the majority of PAA contracts over a one-year period in the same pattern as the revenue would be recognised for the claims processing services under current revenue recognition guidance in ASC 605/606, Revenue Recognition, and IAS 18, Revenue. Further clarification of the guidance on this issue would be appreciated. Premium Allocation Approach As proposed, the PAA would be permitted, but not required, (1) for contracts that have a coverage period of one year or less, or (2) if doing so would produce a measurement that is a reasonable approximation to those that would be produced when applying the BBA. The application of the PAA is presumed to produce a reasonable approximation to the measurements that result from application of the BBA if, at contract inception, it is unlikely that during the period before a claim is incurred, there will be significant variability in the expected value of net cash flows. We do not agree with the restrictive nature of these criteria primarily due to the concern that certain similar contracts could have different measurement models and presentation. This would not be efficient or produce useful, consistent information for users of the financial statements. We believe the PAA should be available for application for all short-duration

11 Page 11 insurance contracts as defined under current U.S. GAAP in ASC 944, Accounting and Reporting by Insurance Enterprises. We are also concerned that the guidance provided with respect to determining whether a contract has significant variability in the expected value of net cash flows is not clear and may lead to inconsistent application. The Board does not provide any illustrative examples of (1) how to determine whether the PAA is appropriate for a contract or (2) how to apply the PAA once a contract is determined to meet the criteria for the simplified approach. Absent additional guidance, it is unclear how to apply the guidance as proposed. In looking to the FASB s Proposed Update for further direction on the application of the PAA, we are still concerned that there is a lack of clear guidance in this area. Of the 13 examples provided in the FASB s Proposed Update for the determination of the measurement model to be applied to an insurance contract, only two examples are provided for contracts with coverage periods greater than one year, and these do not adequately explain the concept of significant variability in the expected value of net cash flows. We have specific concerns regarding surety contracts that have similar economics, but varying lengths of coverage. For example, under the FASB s implementation guidance, a three-year surety contract is considered to not have significant variability in the expected value of the net cash flows during the period before a claim because the contractor s financial position is considered stable, and the contractor historically completed projects on time. However, some of our surety contracts extend up to ten years, and the guidance is not clear regarding at what point (i.e. length of contract) a contract has significant variability in the expected value of the net cash flows during the period before a claim is incurred. Notwithstanding our general concerns with the application of the PAA to P&C contracts, if the proposed criteria to account for insurance contracts under the PAA are not revised to incorporate all similar contracts, and absent further clarification on the application of the guidance, we agree that the PAA should be permitted, but not required. Some entities may find it operationally easier to apply the BBA model to all of their contracts, or to all contracts for a particular product, rather than to have two different accounting methodologies and we do not believe they should be precluded from doing so if they so choose. Reinsurance We disagree with the Board s view that a reinsurance contract should be accounted for separately from the underlying direct contracts to which it relates. We believe that there should be symmetry between the recognition and measurement of reinsurance contracts and the underlying ceded contract. In the absence of such symmetry, there could be situations where a reinsurance contract is accounted for using the BBA and the underlying contract is accounted for using the PAA. This may result in a reinsurance receivable balance being recognised prior to the recognition of the related reserves on the underlying ceded contract. This resulting outcome would produce non-decision-useful information and would be confusing and difficult to analyse. We recommend that the Board reconsider the FASB s approach to reinsurance such that a cedant would account for reinsurance contracts using the same approach that the cedant uses to account for the underlying direct policies. We also disagree with the proposed guidance concerning the accounting for ceding commissions. The ED requires that ceding commissions paid to the cedant that are not loss sensitive be treated as a reduction in premiums paid by the cedant. Under current U.S. GAAP, ceding commissions represent recovery of acquisition costs that reduce capitalised acquisition costs. We believe this is a better reflection of the economics of the transaction. From a presentation perspective, it makes sense for the insurance and reinsurance premiums to mirror each other. Under the ED, for a contract that was 100% reinsured, netting

12 Page 12 the ceding commissions with the reinsurance premium would result in a residual net premium being presented in the income statement. Presentation and Disclosure We have significant concerns with how understandable the proposed presentation and disclosure requirements will be to users. We are specifically concerned that the proposed presentation requirements ignore traditional measures of premiums and claims (for example, volume information such as net written premium), which are a focus for many preparers and users of financial statements today. We predict that this information will continue to be a focus for preparers and users, and as such, we urge the Board to provide specific disclosure requirements around such measures. However, should the Board move forward with the guidance as proposed, we believe that the proposed presentation requirements in the ED are preferable to those in the FASB s Proposed Update. Specifically, we appreciate the one line presentation for insurance contract assets and liabilities, respectively, without distinguishing between the BBA and the PAA measurement models. Similarly, we prefer the IASB s proposed statement of comprehensive income presentation, which is more streamlined than that proposed in the FASB s Proposed Update. The proposed disclosure requirements are more detailed than current requirements and will significantly impact system requirements. The voluminous nature of these disclosures will be overwhelming to users and difficult to understand and interpret. We believe clear, concise disclosures produce the most beneficial and useful information for users of the financial statements. Therefore, as an alternative to the proposed sensitivity analysis disclosures, we would recommend a disclosure that provides transparency around the range of acceptable outcomes produced by the insurers actuaries and discloses the methods and assumptions used to determine such ranges. Summary The Board has stated that it is particularly interested in whether the revisions to the 2010 ED create extra complexity for users of financial statements. In addition, throughout its deliberations leading to the issuance of the ED, the IASB has considered whether the costs of implementing the revised proposals exceed the costs of implementing the 2010 ED, and whether the resulting additional benefits would justify the costs. We do not believe that the guidance in this ED, as it applies to P&C insurers, yields a sufficient overall improvement in decision-useful information to investors and other users of financial statements, as compared with the information provided by a model based on the requirements of existing U.S. GAAP, to justify the significant overall costs and complexity it will add. Several aspects of the proposed accounting model would require significant changes in operational and actuarial processes that are complex and are not consistent with how we manage the business. However, we do believe, subject to certain exceptions noted in this letter, that the ED, as it applies to life insurers, meets the Board s objectives. U.S. GAAP for P&C contracts is well established and has been used globally without posing significant issues with regard to application, auditing, and analysis. Generally, users understand the accounting and reporting for P&C contracts, and that information, including the performance metrics that are reported to users, is consistent with how management reviews performance of the business. As previously noted, we acknowledge the theoretical merits of several individual aspects of the proposal. However, when considered in totality and in conjunction with overall feedback provided to the Board by preparers and users, we believe that the magnitude of the costs and complexity that the P&C model would add is not

13

14 Page 14 Questions for Respondents Question 1 Adjusting the contractual service margin Do you agree that financial statements would provide relevant information that faithfully represents the entity s financial position and performance if differences between the current and previous estimates of the present value of future cash flows if: (a) differences between the current and previous estimates of the present value of future cash flows related to future coverage and other future services are added to, or deducted from, the contractual service margin, subject to the condition that the contractual service margin should not be negative; and (b) differences between the current and previous estimates of the present value of future cash flows that do not relate to future coverage and other future services are recognised immediately in profit or loss? Why or why not? If not, what would you recommend and why? Response 1: We disagree with the approach proposed within the ED, which allows for unlocking of the contractual service margin. We understand that the IASB believes that unlocking the margin avoids the incentive to overstate estimates of expected cash flows at inception so that an entity can subsequently recognise favourable development and a gain. However, we believe that remeasuring or recalibrating the margin is inappropriate because it would provide a potential opportunity to re-establish margin previously earned in subsequent periods based on changes in the variability of cash flows. One of the primary criticisms of existing U.S. GAAP is the understandability (or lack thereof) of the amortisation of deferred acquisition costs for certain products that require entities to unlock their assumptions and record a cumulative catch-up adjustment in net income. We are concerned that unlocking the margin would result in similar confusion. We are concerned with the fact that adjusting the margin to reflect changes in the expected cash flows provides a buffer for potentially smoothing either favourable or unfavourable results. This treatment would have the effect of not adjusting net income until there was a trigger of loss adequacy resulting in loss recognition when the margin becomes negative, which we do not believe provides users with transparent information on a timely basis. In addition, showing current results that are affected by changes in expectations made in prior periods may be confusing to users of the financial statements. We believe that recognising the changes in expected cash flows in net income as they occur would be more explainable to users and would not skew future results, thus providing increased transparency on a realtime basis. Similarly, we are concerned that it may be difficult to determine what constitutes an experience adjustment as opposed to changes in expected cash flows. An inconsistency may result depending on whether the change in estimate is identified in advance (and therefore treated as a change in estimate) or subsequently (and therefore treated as an experience adjustment). We believe that effect is incongruous given that one of the key reasons to unlock the margin is to avoid an inconsistency between measurement of the insurance contract on day one and subsequent measurement.

15 Page 15 We believe that two of the key objectives of this project are to (1) reflect updated assumptions in the measurement of the liability and (2) increase transparency to the users. Neither of these objectives would be achieved if the margin is unlocked due to the inherent difficulties in isolating changes in cash flow assumptions from experience adjustments, as noted above. Instead, we believe that recognising changes in expected cash flows in net income as they occur, as proposed by the FASB, would more clearly reflect updated assumptions about an insurer s core business, which in turn leads to increased transparency to users of an entity s financial statements. Question 2 Contracts that require the entity to hold underlying items and specify a link to returns on those underlying items If a contract requires an entity to hold underlying items and specifies a link between the payments to the policyholder and the returns on those underlying items, do you agree that financial statements would provide relevant information that faithfully represents the entity s financial position and performance if the entity: (a) measures the fulfilment cash flows that are expected to vary directly with returns on underlying items by reference to the carrying amount of the underlying items? (b) measures the fulfilment cash flows that are not expected to vary directly with returns on underlying items, for example, fixed payments specified by the contract, options embedded in the insurance contract that are not separated and guarantees of minimum payments that are embedded in the contract and that are not separated, in accordance with the other requirements of the [draft] Standard (i.e. using the expected value of the full range of possible outcomes to measure insurance contracts and taking into account risk and the time value of money)? (c) recognises changes in the fulfilment cash flows as follows: i. changes in the fulfilment cash flows that are expected to vary directly with returns on the underlying items would be recognised in profit or loss or other comprehensive income on the same basis as the recognition of changes in the value of those underlying items; ii. changes in the fulfilment cash flows that are expected to vary indirectly with the returns on the underlying items would be recognised in profit or loss; and iii. changes in the fulfilment cash flows that are not expected to vary with the returns on the underlying items, including those that are expected to vary with other factors (for example, with mortality rates) and those that are fixed (for example, fixed death benefits), would be recognised in profit or loss and in other comprehensive income in accordance with the general requirements of the [draft] Standard? Why or why not? If not, what would you recommend and why? Response 2: We appreciate that the Board has attempted to identify a solution for accounting mismatches in products that are linked. However, we are concerned that the proposal introduces significant complexity, which will create substantial cost for both preparers in producing the information and users in interpreting it, and it is not yet clear whether the expected benefits from this approach outweigh the substantial costs. In practice, it will be difficult to segregate cash flows related specifically to contracts that are linked, and we do not believe that the concept aligns with the Board s objective to simplify insurance contract accounting. As proposed, the guidance related to linked contracts represents a departure from the main principles of the BBA, which complicates the operability of the ED as a whole.