ED/2013/7 Exposure Draft: Insurance Contracts

Ian Laughlin Deputy Chairman 31 October 2013 Mr. Hans Hoogervorst Chairman IFRS Foundation 30 Cannon Street London EC4M 6XH United Kingdom Dear Mr. Hoogervorst, ED/2013/7 Exposure Draft: Insurance Contracts The Australian Prudential Regulation Authority (APRA) welcomes the opportunity to comment on the International Accounting Standard Board s Exposure Draft ED/2013/7 Insurance Contracts ( the ED). APRA is the Australian prudential regulator for all life and general insurance companies that operate in Australia. APRA strongly supports a number of the changes made in the ED. The specific changes supported by APRA include the unlocking of the Contractual Service Margin (CSM) and the changes made to transition. The approach to unlocking the CSM is similar to the unlocking of profit margins under the Margin on Services (MoS) approach that is used in Australia. The approach proposed in the ED is very similar to that adopted for reporting by insurers in Australia for a number of years now. Our experience is that this has worked well. Current value reporting gives a clearer indication of the impact of both economic and duration mismatches. Supervisors, accordingly, take a close interest in the impact on results of the market value of assets and liabilities, and our regulation paradigm is built upon current value reporting. Nonetheless, there are improvements that we strongly believe are still required if the proposed standard is to meet a variety of needs, including those of supervisors. APRA is concerned by the level of complexity introduced by specific aspects of these proposals. The standard is complex both in terms of implementation of the requirements and of the information to be presented. Complexity increases the operational risks and costs to regulated entities, and those additional costs may be passed on to policyholders. In some cases complexity may compromise the quality of information provided to users, including supervisors. Any simplification that can be achieved while maintaining the quality of financial reports should be seriously considered. Page 1 of 14

In particular, we believe the final standard should measure and present current values for insurance assets and liabilities on a consistent basis, and that it should include an option to achieve this. APRA observes that the proposal to mandatorily report discount rate changes through Other Comprehensive Income (OCI) materially undermines consistency and creates undue volatility in the Profit and Loss Statement for those who already value all supporting assets at fair value through P&L (FVP&L). Moreover, this approach substantially increases complexity, costs and operational risks for insurers. The use of OCI should be introduced only as an option available to those for whom it is appropriate, and should not be mandatory for others. Note that the option needs to be a package which includes accretion of interest which should only be at locked-in rates if OCI is also used, otherwise accretion should also be at current rates. If the use of OCI is made optional, but accretion remains at locked in rates, then not only will there be inconsistency but the problems identified above with the use of OCI will remain largely unresolved. Additionally, although the mirroring accounting treatment of relevant underlying assets in the insurance liability measurement is conceptually attractive, there are likely to be significant practical and conceptual challenges with this approach. In particular, what is proposed is very complex, with significant detail needing to be worked through, making application challenging for preparers. Moreover, as a user of financial statements, we believe this approach might limit understanding and comparability, especially in respect of insurers traditional participating business. An approach consistent with the Building Block Approach (BBA) and/or a widening of the ability to unbundle (such that business which would otherwise use the mirroring approach can be accounted for using other standards) would seem to us to be better solutions. Also, while the general approach to unlocking of the CSM is to be applauded, there are still matters of detail which need to be confirmed or clarified in terms of its application. For example, the risk adjustment should be treated on the same basis as other future estimates, with changes in the risk adjustment applying to future events also resulting in an unlocking of the CSM rather than being recognised immediately in P&L. Other details are mentioned in the appendix. Finally, we note that while the disaggregation of the deposit component in presentation is important, there can be significant practical challenges with disaggregation. For some products it is very complex (similar to the difficulty in unbundling) as the deposit component is highly interrelated to the other cash flows. We propose that the IASB consider a more operationally efficient solution for highly interrelated business to reduce unnecessary burden on the preparers of financial statements, with little associated reduction in benefits to users. The appendix to this letter provides more detailed responses to some of the questions set out in the Exposure Draft. If there is any way in which APRA can assist the IASB further, please contact David Rush (tel: +61 2 9210 3269; email: david.rush@apra.gov.au) or me. Yours sincerely Page 2 of 14

Appendix Note that APRA has chosen to comment on the main issues and so has not provided detailed responses to Questions 6 and 7. 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: (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? Summary APRA agrees with the concept that the CSM should be adjusted for changes in future estimates, but believes it should be related only to changes to estimates of future cash flows in respect of which coverage is still to be provided. Changes in estimates of cash flows in respect of which coverage has already been provided (e.g. for claim liabilities) should be recognised immediately in profit or loss. APRA also believes that: Changes in estimates of risk adjustments should be treated in the same way as the cash flows to which they relate; unlocking should not apply to changes in discount rate; and if losses have occurred, the rebuilding of the CSM for any subsequent improvements in future estimates should only occur after the loss has been reversed. Detailed Response Unlocking, and practical calculation of, the CSM APRA agrees with the general concept of unlocking the CSM. If the future estimate of cash flows changes as a result of changes in the level of assumed future experience, there will be a strain or surplus, which should not flow immediately into profit or loss (P&L), but should be released over the balance of the coverage period. This is done simply by equating the liabilities under the old and new assumptions (as is currently done under Margin on Services (MoS)), adjusting the CSM accordingly. If the following approach to determination of the CSM is allowable under IFRS, then calculation of the CSM and the CSM run-off becomes quite simple. For example, for simple risk insurances the CSM releases may be a fixed proportion of the fulfilment value of expected incurred claims. A similar assumption is made for other products. That proportion is set at issue to give the correct CSM at recognition. The CSM emerging in the period is then that same fixed proportion of claims expected to emerge in that period, allowing for discounting. If the future experience assumptions change (presumably at the end of the period), then that proportion is changed, so that the total pre-claim liability, discounted using current discount rates, as at the date of assumption change, does not change. Page 3 of 14

In terms of cash flows in respect of events that have already happened (claims) or are expected to have happened (Incurred But Not Reported claims), any change in cash flows, even if they are expected to arise in the future, must be recognised immediately in P&L. For a product giving rise to such cash flows, the service of providing coverage has been completed and so profit or loss associated with those cash flows should be recognised immediately. (Note that APRA considers that coverage is the insurance service, not payment of claims anybody else could provide payment of a cash flow, but only an insurer can provide coverage.) Risk adjustments As currently stated, the CSM is not affected by changes in the risk adjustment such changes must be recognised immediately. This is partly because the risk adjustment is not assumed to be part of future cash flows. This is not supported by conceptual analysis where risk arises because of uncertainty over the future cash flows (if you have no cash flows then you don t need a risk adjustment) and so the risk adjustment should be linked to the cash flows and treated consistently with them. It also opens up a window for manipulation of profit and loss in relation to perhaps the most subjective element in the valuation assumptions the entity s assessment of the value of risk. Management will be motivated to change the entity s assessment of the value of risk in order to generate immediate profit. This is of concern to users, particularly supervisors. It follows that changes in estimates of the risk adjustment should be treated in the same way as changes in the estimate of associated future cash flows. Note, that this implies that separate risk adjustments should be determined, and appropriately treated, for preclaim and post-claim liabilities. However, this is not difficult, is already done in Australia for general insurance reporting to APRA and reflects the different risks to which pre-claim liabilities and post-claim liabilities are exposed. Discount rate There is a lack of clarity in the ED in respect of changes in the discount rate used to determine the value of the liability at the end of the reporting period. Our understanding is that as, under the ED, all changes in discount rates go through OCI, such changes do not result in unlocking of the CSM and so are recognised immediately (albeit through OCI). While we agree with the result, we do not agree with the reasoning, and it leaves unanswered the question of what happens if the effect of discount rate changes does not go through OCI does this result in unlocking or not? APRA believes that the effect of changes in discount rate should not be absorbed by the CSM, but should (unlike changes in estimates of future cash flows in respect of pre-claim liabilities) be recognised immediately in P&L. This reflects a degree of matching with assets, for which movements in value (where FVP&L is used) would be recognised in P&L. Otherwise, the effect of asset value changes would be recognised in P&L but not equivalent changes in the value of liabilities, even though assets and liabilities may be perfectly matched. Another way of looking at it is that discounting is a factor that is applied now, based on rates applying now (or locked in at inception), albeit that it applies to cash flows that are expected to occur in the future. Hence, any change in discounting is something arising now, even for longer term rates, and so the effects should be recognised immediately rather than resulting in unlocking of the CSM. Page 4 of 14

Note that unlocking of the CSM and presentation of the effect of discount rate changes through OCI are two separate issues. APRA s view is that discount rate changes should not result in unlocking of the CSM, regardless of whether the effect of such changes is reported through OCI or P&L. Loss reversals Although there does not appear to be a clear IASB decision on this, the draft ED does not appear to allow for reversal of loss recognition should estimates of pre-claim cash flows subsequently improve. APRA believes that the correct treatment is to allow reversal of previous loss recognition as this creates an element of symmetry and reduces opportunities to artificially create future profits through the CSM. (Future profits could be artificially created by using conservative assumptions when losses are recognised and subsequently reversing the conservatism.) Without this symmetry, profit reporting (and capital calculations dependent on the liability valuation) will be distorted. Note that this is the current treatment in Australia and has worked well. Note also that the tracking of (negative) CSMs is no different from the need to keep track of positive CSMs on a portfolio, and so should not prove onerous. Page 5 of 14

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 (ie 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? Summary We do not agree. Although the mirroring accounting treatment of relevant underlying assets in the insurance liability measurement is theoretically attractive, there are likely to be significant practical challenges with this approach. In particular, what is proposed is very complex, requiring a significant level of detail to be addressed and thereby making application challenging for preparers of financial statements. Moreover, as a user of financial statements, APRA believes this approach might limit understanding and comparability, especially in respect of insurers traditional participating business. An approach consistent with the BBA and/or the accounting of much business which would otherwise be mirrored through other standards, through a widening of the ability to unbundle, would seem to us to be better solutions. Detailed Response Complexity At first glance, mirroring seems like a good idea ensuring that assets and liabilities are matched if the two are linked. However, APRA is concerned that there are likely to be significant practical and conceptual challenges with this approach. Page 6 of 14

In particular, the proposal creates complexity which imposes an obvious cost for both preparers in preparing the information and users in interpreting it, and it is not yet clear to APRA that the expected benefits from the approach outweigh this cost. Furthermore, the reality is that any mirrored contract will contain one or more elements that cannot be mirrored. The resulting interaction between these components where some are mirrored, some profit is recognised immediately in P&L, some profit is recognised through P&L as CSM is released and some profit is recognised through OCI is complex, confusing to users, and arguably not a faithful representation of the underlying economics of the contract. Consistency with local legislation While it might be appropriate for Participating business to be treated in the same way world-wide, there are differences in legislation applying in various jurisdictions which are not recognised by the IASB and taken into account in the measurement. In particular, from an Australian perspective, profit recognised in respect of Participating business through application of the ED won t match that required by the Life Insurance Act in respect of the same business, creating additional cost for preparers. In Australia, profit under the Life Insurance Act essentially reflects supported future distributions (similarly to fees less expenses on investment linked business), whereas the ED tries to measure what is theoretically earned. Accordingly, the ED recognises the full shareholder portion of actual investment earnings immediately, whereas the Life Insurance Act recognises profit which has already been smoothed. A possible solution would be for the IASB to focus on ensuring, for Participating business particularly, that the general purpose accounting is aligned with the accounting required under local legislation. The result is likely to be more readily understood by preparers and users within that jurisdiction. In contrast, under the approach proposed in the ED, the results are less likely to be understood and used, and while preparers and users in certain jurisdictions may be satisfied, others will inevitably be disappointed. Use of the BBA Another alternative to the approach in the ED is to not mirror Participating business at all, but to use the BBA (as for non-participating business). Such an approach would be capable of consistent application with other products. The cash flows, risk adjustment and CSM under this approach would still be measured on a current basis. Investment Linked business Mirroring also includes all Investment Linked business which is measured under the Insurance Contracts standard. In Australia, Participating business is treated quite differently from Investment Linked business (the latter is usually classed as Non- Participating). Profit emergence on the two is similarly quite different. In particular, mirroring of Investment Linked business under the ED is not the same as treatment under IFRS 9, and so similar contracts are likely to be accounted for differently. For example, Q2(c)(iii) puts movement in fees and expenses which are not directly linked to the underlying assets through the CSM, which is not the same as the treatment of fees and expenses on stand-alone investment contracts under IFRS 9. Additional insurance benefits and Unbundling There is the issue that additional insurance benefits, lapsing if the main contract lapses, are not distinct, and so must not be unbundled as required by local legislation. This requirement is too strong, as components which are effectively distinct can still not be unbundled. Obviously, if you lapse the main investment component then you must also lapse the additional insurance benefit. Page 7 of 14

That doesn t alter the fact that the policyholder has two benefits under the contract which are different in nature and should be accounted for separately. In fact, Australian insurers are doing that already, as local legislation requires such business to be effectively unbundled and reported separately. Furthermore, there are components which while not fully distinct are not highly interrelated and which can be practically unbundled, but which are not able to be. The ED introduces the concept of highly interrelated. A solution that utilises both the concepts of distinct and highly interrelated might therefore be: If components are distinct then they must be separated. If components are highly interrelated then they must not be separated. In order to be operational, application of the concepts distinct and highly interrelated would need to take into account the specific facts and practical circumstances of the insurer and would need to be applied in a principles based manner. A particular case in point are greater-of benefits. Such benefits are clearly not distinct from the account balance. But although the two components may be interrelated, they are not highly interrelated, and can be practically unbundled. The account balance is clearly a financial instrument and should be measured under IFRS 9. The additional insurance benefit is clearly an insurance contract and can be measured as the difference between the measurement of the whole contract under the Insurance Contracts standard, and the measurement of the account balance only under the same standard. The latter contains all the components of the Insurance Contracts standard (expected value of cash flows, time value of money, risk adjustment and CSM) and is relatively easy to do. What is more, if unbundling were allowed in such cases it would obviate the need for most, if not all, mirroring, reduce the level of complexity and increase comparability (as all such business would be consistently measured under IFRS 9). Conclusion A pragmatic approach to unbundling of Investment Linked business, combined with a Building Block Approach alternative for Participating business, would thus allow all of the complexity and confusion of mirroring to be done away with. Page 8 of 14

Question 3 Presentation of insurance contract revenue and expenses Do you agree that financial statements would provide relevant information that faithfully represents the entity s financial performance if, for all insurance contracts, an entity presents, in profit or loss, insurance contract revenue and expenses, rather than information about the changes in the components of the insurance contracts? Why or why not? If not, what would you recommend and why? Summary APRA sees the proposals as an improvement from the 2010 ED. Furthermore, disaggregation of the deposit component in presentation is important. However, there can be significant practical challenges with disaggregation. For some products it is very complex (similar to the difficulty in unbundling) as the deposit component is highly interrelated to the other cash flows. We propose that the IASB consider a more operationally efficient solution for highly interrelated business to reduce unnecessary burden on the preparers of financial statements, with little associated reduction in benefits to users. Detailed response APRA understands the IASB s desire to have a revenue presentation that is consistent with the principles set out in the 2011 Exposure Draft Revenue from Contracts with Customers. APRA also appreciates the IASB s efforts to present volume information in the Statement of Comprehensive Income. While separation of the investment component is important in presentation (ensuring a reasonable comparison between those only writing pure risk business and those providing investment to customers through insurance), determination of the investment component may be difficult and costly for some products where the provision of insurance is very much related to the provision of investment. Such products might include traditional products (such as whole of life or endowments) or lifetime annuities. The IASB should consider whether the cost of disaggregation outweighs the benefits in such cases, and whether a more operational solution should be allowed. (Note that the investment component is not the same as the surrender value under the contract - if one is payable under the contract. Apart from the fact that the surrender value is a contractually defined amount which is not necessarily the same as the amount the policyholder has invested, the implication is that the remainder is the value of the risk component, which is also not necessarily the case.) As revenue is not of itself comparable across industries, the loss of consistency from this practical expedient is not significant. Additionally, users may make more useful revenue comparisons through the separate disclosure in the notes of: earned premium for risk insurance products; fees for unit linked products; and written premium for traditional whole of life or endowment policies. Consequently, while APRA sees merit in an approach consistent with revenue recognition principles we have concerns as to how understandable the outcome of this will be to users. Page 9 of 14

Question 4 Interest expense in profit or loss Do you agree that financial statements would provide relevant information that faithfully represents the entity s financial performance if an entity is required to segregate the effects of the underwriting performance from the effects of the changes in the discount rates by: (a) recognising, in profit or loss, the interest expense determined using the discount rates that applied at the date that the contract was initially recognised. For cash flows that are expected to vary directly with returns on underlying items, the entity shall update those discount rates when the entity expects any changes in those returns to affect the amount of those cash flows; and (b) recognising, in other comprehensive income, the difference between: (i) the carrying amount of the insurance contract measured using the discount rates that applied at the reporting date; and (ii) the carrying amount of the insurance contract measured using the discount rates that applied at the date that the contract was initially recognised. For cash flows that are expected to vary directly with returns on underlying items, the entity shall update those discount rates when the entity expects any changes in those returns to affect the amount of those cash flows? Why or why not? If not, what would you recommend and why? Summary APRA believes that the final standard should measure and present current values for insurance assets and liabilities on a consistent basis, and that it should include an option to achieve this. APRA observes that the proposal to mandatorily report discount rate changes through OCI materially undermines consistency and creates undue volatility in P&L for those who already value all supporting assets at FVP&L. Moreover, this approach substantially increases complexity, costs and operational risks for insurers. The use of OCI should be introduced only as an option available to those for whom it is appropriate, and should not be mandatory for others. Note that any option needs to be a package which includes accretion of interest which should only be at locked-in rates if OCI is also used, otherwise accretion should also be at current rates. If the use of OCI is made optional, but accretion remains at locked in rates, then not only will there be inconsistency but the problems identified with the use of OCI will remain largely unresolved. Detailed Response Capital paradigm The IASB has, in the past, preferred a current value approach for the value of fulfilment cash flows for the statement of financial position. APRA s capital regime for insurers is based on a similar paradigm. Note that this doesn t just depend on the value of liabilities reported in the statement of financial position, but also the performance giving rise to that. We therefore don t support the idea that entities should be forced to adopt an approach in the statement of financial performance which differs from this paradigm. Page 10 of 14

Economic mismatches We understand that one of the IASB s aims is to provide information to users regarding economic mismatches - for example for liabilities that are supported by fixed interest assets recognised at fair value through OCI (FVOCI) which is key to understanding the economics of insurance businesses. While we support this aim, we believe the IASB s current proposal does not achieve it, primarily because accounting mismatches will arise for any liabilities that are supported by assets which are recognised at FVP&L (rather than amortised cost or FVOCI), which is commonly the case for Australian insurers writing long term business. If the assets and liabilities are appropriately matched, then it doesn t matter how they are measured, so long as it is consistent. The effect of any mismatching should flow to profit or loss. By requiring the mandatory use of OCI such consistency cannot be assured, and it will be difficult for users to distinguish between true economic mismatches and those created by accounting. Complexity By requiring unwinding using locked-in rates at inception, the IASB is mandating that records be retained of the interest rates applying when all new business was written. This is a substantial complication of the valuation task. In the extreme, since market conditions vary continuously, no two contracts would use the same yield curve (although, in practice, some degree of aggregation would need to be applied). While the necessary computations are feasible, they are onerous. APRA is particularly concerned about the operational risk that this complexity generates. Optionality We therefore propose that the use of OCI be an option, available to entities that use an amortised approach for asset valuation, in order to reduce accounting mismatches. (Of course, such an option should not be a free choice but should be constrained by appropriate conditions such as consistent application across an insurance entity, a class of business or a particular portfolio, and irrevocable designation.) APRA acknowledges that the Board may have concerns about introducing optionality on the use of OCI, including the potential reduction in comparability, and the possibility of circularity (as an option already exists on the asset side). The risk of reduced comparability is mitigated if the optional use is introduced with a requirement that the presentation of changes in both assets and insurance liabilities should be consistent and accompanied with additional disclosures. And the risk of circularity is mitigated if the default on the liability side is to report the effect of all changes in discount rate through P&L (i.e. opposite to the default on the asset side) with the use of OCI being an option. In this way only one option needs to be used (on the asset side or the liability side, but not both) to achieve consistency. Package Note that APRA believes that such an option should be a package, such that if the option is chosen then a number of things must be done, not just one. Conversely, the default approach should require a number of things relating to measurement and interest accretion, not just presentation of a particular amount through P&L or OCI. (In this regard, it is similar to the use of PAA instead of BBA.) Page 11 of 14

Thus, the default should not only be that everything (whether resulting in unlocking of the CSM or recognised immediately) is reported in P&L, but should also include: discounting of everything in calculating the liability (cash flows, risk adjustments associated with those cash flows, and contractual service margins to be released each period) should be at current interest rates, as this is consistent with measurement of liabilities at a current value; and Accretion of interest on that liability should be at the interest rate applying at the start of the period, as this is consistent with the way the liability is calculated at the start of the period. Recalculation of the liability at the current date but discounting using the rate locked-in at inception would not be required (as it is not needed). It follows that if the option is utilised, not only should the effect of discount rate changes be reported through OCI, but: accretion of interest on the liability should be at the locked-in discount rate, consistent with the reporting of discount rate changes through OCI; and the liability for accretion purposes should be determined at the start of the period using that locked-in discount rate (in order for the accretion to be determined correctly). In summary, if the use of OCI is made optional, but accretion remains at locked in rates, then not only will there be inconsistency but the problems identified with the use of OCI will remain largely unresolved. Page 12 of 14

Question 5 Effective date and transition Do you agree that the proposed approach to transition appropriately balances comparability with verifiability? Why or why not? If not, what do you suggest and why? Summary Yes, overall APRA agrees that the proposed approach to transition appropriately balances comparability with verifiability. Detailed Response APRA believes that the retrospective establishment of CSMs on transition should help to enhance comparability in the subsequent measurement of contracts in force at the transition date and contracts that are recognised after transition. Furthermore, APRA believes that will result in more meaningful profits than the significant sustained losses that would have resulted from the previous exposure draft. In this regard, the ED provides a practical approach to establishing CSMs when objective information is not reasonably available to do this exactly. The need for auditor review and disclosure requirements should provide added discipline on the process of setting the CSM at transition. Note that if mandatory accretion of interest at locked-in discount rates is retained then, on transition, the requirement to ascertain and apply the discount rates applicable at initial recognition would also be difficult and in some cases may prove impracticable. This would be a particular problem for altered contracts, where details of the original contract have not been adequately retained. Page 13 of 14

Other Reinsurance APRA is not convinced that reinsurance is properly treated in all cases under the proposals. From a regulatory perspective reinsurance is a form of risk mitigant (a negative insurance) such that the insurance risk is transferred from the direct writer to the reinsurer. The IASB needs to consider whether it is then appropriate for accounting to assume that the direct writer accounts for all the insurance risk, regardless of any transfer of insurance risk, or whether it would be more appropriate to view the accounting for the direct writer from a net position (i.e. net of reinsurance). Consideration of a 100% quota share reinsurance on exactly the same terms as the direct writer might be useful in these circumstances. If the original contract is onerous then, under the proposals in the ED, both the direct writer and the reinsurer will report a loss on writing the contract (even though there is only one risk insured in the market) and the direct writer would report a profit on the reinsurance in subsequent periods, even though it effectively had no business. Alternatively, if reinsurance is treated as a negative insurance, then the negative loss on the reinsurance (i.e. a gain) would cancel out the loss for the direct writer, so that only the reinsurer reported a loss on the onerous terms on which the risk is insured. From a supervision perspective, this is how it should be. (Note that the credit risk associated with the reinsurer is dealt with separately, as it is for other assets.) Page 14 of 14