"Session PS-6 International Accounting Standards Séance SP-6 Norm es"
SEMINAR FOR THE APPOINTED ACTUARY (PS-6) 1 Session PS-6: International Accounting Standards Séance SP-6 : Normes comptables internationales Moderator/Modérateur: Lesley B. Thomson Speakers/Conféfenciers: Johanne Papillon Lesley B. Thomson ?? = Inaudible/Indecipherable ph = Phonetic U-M = Unidentified Male U-F = Unidentified Female Moderator Lesley B. Thomson: Good afternoon. I’m Lesley Thomson, I work for Sun Life. I sit on various CIA committees, but today I’m here to talk to you about what we’re doing at Sun Life. I know the program says that I’m going to talk about the Task Force on International Accounting and Actuarial Standards’ comments on the IAA’s Risk Margins paper, but I’m not going to do that. The Program Committee decided this Seminar should focus on IFRS Phase 1 rather than Phase 2. We want to give you a flavor for what’s going on right now in terms of the conversion to IFRS in 2011. I am a speaker at this session, together with Johanne Papillon, who you saw this morning. Johanne and I have similar roles (I at Sun Life and she at Manulife), coordinating the actuarial efforts of the conversion to IFRS in 2011 in our companies. We’ve been working together on an industry group Johanne chairs which is discussing classification of contracts, and it’s been a great learning experience to share some of our experiences. Now, I warn you, you’re going to hear quite a bit of repetition of what you heard this morning from Peter Martin and Nick Bauer and also Johanne, but we figure this is pretty important information and much of it is new so it won’t be too bad to hear some of it twice. Also, it’s important to remember that we’re coming at it from the perspective of what we’re doing within our companies rather than from the perspective of what the CIA or the profession is doing. So I’ll get started. I’m going to be concentrating the measurement of insurance contract liabilities under IFRS 4. Johanne will focus on the classification of contracts and how we decide which contracts are insurance contracts, investment contracts and service contracts, so so that will come later, but I will talk about the measurement of the insurance contract liabilities under IFRS 4. The agenda for my section covers three main topics. First, IFRS 4 outlines the criteria under which your current accounting policy can be used for the valuation of insurance contracts in IFRS (Phase 1). In my opinion, CALM meets these criteria, at least for life insurers (I’m not an expert in P&C insurance). Second, IFRS 4 allows a number of changes to the accounting policy for the valuation of insurance contracts, and I will give you some of the reasons we at Sun Life have decided not to make any of those changes. We’ve decided to continue using CALM for the valuation of insurance contracts. Finally, I’ll talk about a few of the other issues and options that are discussed in IFRS 4: discretionary participation features, unbundling and then discounting of tax cash flows. One other thing I’d like to mention is that we’re well into this project at Sun Life (and Manulife) and we’re making some of these decisions before we really know exactly where all the rules are going to come out, or what all the guidance will be. So, we’re on the edge here a bit but, hopefully, our efforts will help you. PROCEEDINGS OF THE CANADIAN INSTITUTE OF ACTUARIES Vol. 19, September 2008 2 COLLOQUE POUR L’ACTUAIRE DÉSIGNÉ (SP-6) IFRS 4 covers the measurement (or valuation) of insurance contract liabilities, including reinsurance contracts. It only covers Phase 1 for now; Phase 2 is expected in 201?. I’ve been told that Nick Bauer has a bet with Phil Arthur, an over/under bet for when Phase 2 will be in place, and the over/under year is 2013. Another thing to remember is that IFRS 4 applies in any country that’s adopting IFRS, so there’s a lot of material in IFRS 4 that we don’t really care about as Canadians, but is needed in order to cover every country’s accounting policies. First, IFRS 4 allows you to continue using your current accounting policy, which is CALM in Canada, for the valuation of insurance contract liabilities provided five criteria are met. The five criteria are listed in paragraph 14 of IFRS 4. I’ll go through the criteria one by one. The first is that there should be no provision for future claims on insurance contracts which are not in existence at the reporting date. Now, that sounds like it would be obvious, that if you didn’t have a contract you wouldn’t have a liability for it, but it’s not crystal clear. When you read the explanation and the basis for conclusions of why they needed this criteria, it’s clear that they really mean catastrophe and equalization provisions, neither of which are liabilities under CALM. Also, section 2130.03 of Standards of Practice says that we only recognize liabilities for contracts in existence at the reporting date, so CALM definitely satisfies this criteria. The only thing that might be an issue is when you’ve committed to entering into a contract but the contract isn’t in force yet. When should the liability be recognized? The second criteria is probably the most difficult to assess. It says that you must carry out what’s called a “liability adequacy test”, and there are two minimum requirements for this liability adequacy test. The first is that the test considers current estimates of all contractual and related cash flows, including embedded options and guarantees. The second is that if the test shows the liability is inadequate, the entire deficiency is recognized in profit or loss. So, if we can show that CALM meets these two minimum requirements, then we don’t have to do anything further about the second criteria. If we can’t show that CALM meets these two minimum requirements, then we would have to do a separate liability adequacy test in addition to CALM. This would be important if we were in one of the many countries that don’t have a liability adequacy test at all. IFRS would tell us that we have to do one, and also how to do it. In my opinion, CALM meets the minimum requirements for the liability adequacy test. It clearly meets the second requirement because the entire change in the CALM liability is recognized in profit or loss as stated in the CICA handbook. I also believe the first requirement is met, although it isn’t as clear. Under the Standards of Practice, Section 2130, CALM does make provision for all contractual cash flows and related cash flows, including those from embedded options and guarantees. Also, “current estimates” are used in the test, which normally refer to estimates before provisions for adverse deviations are added, so that gives us a safety net because CALM is more stringent than it would be using ‘current estimates’. Now, the one concern might be the limitations on the term of the liability under CALM, because the test says you have to include all contractual cash flows, which might be interpreted to include cash flows beyond the term of the liability. However, I believe that this is OK, because the restriction on the term of the liability almost always acts to increase the liability. The only time it doesn’t is when the insurer has an unconstrained right to reprice the contract, in which case the renewal could be considered a future contract. The wording about whether you’ve got a current obligation for something that you have a right to reprice in the future is very similar in IFRS to the term of the liability concept we have now in Canada. Ok, I’m going back to the list of five criteria for being allowed to use CALM. The third is derecognition, which says that the liability is only removed from the balance sheet when it is extinguished. Again, I don’t think that’s a problem, since section 2130 of the Standards of Practice say we keep a liability on the books until the contract is extinguished, or it might say ‘until the obligation is expired’ or something like that but it’s effectively the same thing. The fourth criteria is actually the one that CALM doesn’t meet. There’s a requirement that you not net your reinsurance ceded against your liability, but it looks like it will be OK to just change the presentation so we put Vol. 19, septembre 2008 DÉLIBÉRATIONS DE L’INSTITUT CANADIEN DES ACTUAIRES SEMINAR FOR THE APPOINTED ACTUARY (PS-6) 3 reinsurance ceded on the asset side of the balance sheet instead of as negative liabilities. So, you end up with higher assets and higher liabilities, but the net is the same. So that’s the one change we’re going to have to make to CALM going forward. I believe we need a reasonable way to allocate the net liability into the direct liability and ceded asset, because we only have assets to back the net. It’s only the net that we’re concerned with under CALM, so we need some way to split that net into a direct and a ceded. The split should be based on the underlying cash flows and some kind of reasonable assumption about the related assets, but most of us do that already anyway so it’s just a matter of dotting the i’s and crossing the t’s. The last criteria is that you have to consider the impairment of reinsurance assets or the possible inability of your reinsurer to pay claims, but that’s directly in section 2130 of the SOP. So, in my opinion, under IFRS 4, CALM can continue to be used for the valuation of life insurance contracts when we move to IFRS in 2011 with one change in presentation required (reinsurance ceded amounts must be shown as assets). We’re going ahead on that basis. OK, now for the flip side of this. We are allowed to use CALM, but there’s also wording in IFRS 4 that allows you to make certain changes to your accounting policies for the valuation of insurance contracts. There are restrictions on the types of changes that you’re allowed to make. A change is allowed if and only if it makes your financial statements more relevant and no less reliable, or more reliable and no less relevant. In other words, you can’t make things worse but you can make things better. There are some specific changes that are discussed and deemed either acceptable or not acceptable. For example, if you want to change to reflect current market interest rates, it would be permitted. However, if you want to change to measure liabilities on an undiscounted basis, it would not be permitted. If your current accounting policy doesn’t discount liabilities then you can continue to use it, but you’re not allowed to take a valuation approach that discounts the liability and change it to one that doesn’t. So, technically speaking, IFRS 4 allows us to use a method different than CALM, but Sun Life has decided not to make any such changes for a number of reasons. First, it’s only really during the transition period that the new method would apply. In Phase 2, we’ll have a whole new approach to the valuation of insurance contracts, so we’d rather just stay the same for now and only change once. Also, it would cause a lack of comparability with other companies, which gives us all kinds of headaches. Next, we believe CALM is appropriate, reliable, relevant, and we’re comfortable with it as a valuation approach. We don’t want to do anything that would make analysts suspicious or give OSFI any concern. We also heard this morning that OSFI might go as far as saying that we can’t make any changes anyway, and we don’t want to be on the wrong side of that decision. Discretionary participation feature (DPF) is another option under IFRS 4. It allows you to separately recognize your guaranteed components from your DPF. DPF is best thought of as traditional participating policyholder dividends. What IFRS 4 says is that you could hold the provision for future policyholder dividends separately, and also that you could put part of it in equity instead of the liabilities. IFRS 4 allows it, but doesn’t tell you how to decide whether you should or shouldn’t. Now, I want to mention IASP 7. IASPs are the International Actuarial Associations papers. They aren’t binding on us, but they are out there and we respect them and we would prefer to follow IAA Guidance whenever it makes sense to do so. Also, there’s a very good chance that Canada will adopt some of the IAA Guidance as our own, but that decision hasn’t been made yet, it’s up to the Actuarial Standards Board in the case of standards and the Practice Council in the case of educational notes. IASP 7 is the current IAA paper on the topic of DPF, and it says that constructive obligations should be classified as liabilities. Under CALM, we include DPF in liabilities to the extent required to meet policyholders’ reasonable expectations (PRE). At Sun Life, we think PRE is essentially the same thing as constructive obligations, so we aren’t going to change our accounting policy. PROCEEDINGS OF THE CANADIAN INSTITUTE OF ACTUARIES Vol. 19, September 2008 4 COLLOQUE POUR L’ACTUAIRE DÉSIGNÉ (SP-6) Also, we feel that shift the DPF from liabilities to equity would fail the ‘no less reliable test’. Moreover, we would have a lack of comparability with other companies and OFSI would be nervous. Those are all good reasons to keep doing what we do today. OK, unbundling is another option we have under IFRS 4. Unbundling is separating a contract into components or pieces, each of which is subject to a different financial reporting standard. There are three different types and I’ll go through each. The first is unbundling a deposit component from an insurance contract; the second is unbundling an insurance component from an investment contract; and the third is unbundling a service component from an investment contract. The first, unbundling deposit components from insurance contracts, is required under certain accounting regimes. (Remember, IFRS 4 applies to all countries around the world.) There are some countries where they will have to unbundle because their valuation method would otherwise ignore some of the obligations of the deposit component. But that’s not true for CALM. CALM takes into account all the obligations of the deposit component; therefore, this type of unbundling will not be required in Canada. But it will be permitted if the separation can be done reliably. In other words, if you can extricate a deposit component from an insurance contract, you would be allowed to separately report it under IFRS 4. The most probable candidates would be account values for universal life and amounts on deposit. If you do unbundle a deposit component, it will be subject to IAS 39 (the one that applies to investment contracts) and the residual contract, what’s left over, would be subject to IFRS 4, so your total liability would change. Sun Life has tentatively decided not to unbundle deposit components from insurance contracts. Why? Again, we prefer to wait until Phase 2. Also, we’re not crazy about deposit-type accounting (premiums aren’t revenue anymore), we believe that the CALM liability is appropriate in aggregate, and OFSI would be nervous. Next, unbundling insurance components from investment contracts is really just an administrative convenience. This is business where you’ve got a contract that’s got a little bit of insurance risk but is really an investment contract. We’ve found it’s sometimes easier to strip out the insurance and treat it with other insurance contracts. This is mainly outside North America. The last type of unbundling is unbundling service components from investment contracts. This is required under IAS 39 and IAS 18. The most common place will be when you’ve got investment management services and fees on investment contracts. If you’ve got an investment contract that includes the provision of investment management services for a fee, then you have to split the contract in two. You have to split the fees and you have to split the transaction costs. Transaction costs, by the way, is IFRS-speak for deferrable acquisition costs (DAC), but it’s quite a bit more restrictive than our current definition of DAC or the US GAAP definition of DAC. So you have to split the contract in two, into the piece related to the underlying investment contract (the financial liability) and the piece related to the fee for services (service component) and there’s a lot of guidance on how to do this in IASP 4. Now a few miscellaneous issues to clean up. There are some people who have said that they believe that discounting of future tax cash flows is prohibited under IAS 12, and that this would apply to insurance contracts even in Phase 1. We know that we’re not going to be able to discount tax cash flows in Phase 2, but some argue that we won’t be able to in Phase 1 either. The converse argument is that IFRS 4 allows you to continue to use CALM, and prohibits a change that would make the result less relevant. Sun Life is taking the latter position. We believe it will continue to be appropriate to discount future tax cash flows in a valuation of insurance contracts under Phase 1. I understand that the Insurance Accounting Task Force of the CICA has been asked for their opinion on this matter so we’ll have to change if they rule differently, but I’m hoping they won’t and I can provide a good argument if it would help convince them. There’s a similar issue for foreign exchange provisions. The IFRS would seem to suggest that you have to use current exchange rates forever; that you can’t make any provision for changes in future exchange rates, but I believe it is the same rationale as discounting tax cash flows. We are allowed to do it under our current accounting policy, so it should be OK. Vol. 19, septembre 2008 DÉLIBÉRATIONS DE L’INSTITUT CANADIEN DES ACTUAIRES SEMINAR FOR THE APPOINTED ACTUARY (PS-6) 5 Now I’ll pass it over to Johanne who is going to talk to you about contract classification. (Applause) Speaker Johanne Papillon: Good afternoon. For those of you who are a bit nervous because I did talk about contract classification this morning, I will be going over some of the same stuff but I will go over it quickly. This is a joint session this afternoon and this morning was just life, so for the benefit of people who didn’t attend this morning, I will go over a little bit of what I said, but I’ll try to do it quickly, and I will be talking as well this afternoon about embedded derivatives in a fair amount of detail and, hopefully, you’ll stay awake for that. I’ll also give a very, very high level overview of IAS 39. Lesley and I decided to talk about contract classification and the measurement of insurance contracts under IFRS. We are not today talking about how investment contracts will be valued under IAS 39, and one of the reasons for that is that we’re actually not fully prepared to talk about it; we’re still debating a number of issues and so maybe that can be a session for the next meeting. And finally, I’ll talk a little bit about our observations and conclusions, and I want to reiterate that the conclusions that I’m talking about here are those of Manulife or my research as part of Manulife. So this is the part where I just want to very briefly go over again, the three criteria for a contract to meet the definition of an insurance contract. The insured event must adversely affect the policyholder, and in this context, longevity does adversely affect the policyholder. Some might not agree with that, but the idea is that outliving your savings is an insured event. The payment must be significant, and the risk must be non-financial. The significance of insurance risk refers not to the probability of the event but rather to the amount of benefits payable. What you have to do is determine what additional benefits you would have to pay and compare it to the benefits that would be available if the insured event does not happen. I think the experience in Europe has been that companies consider additional benefits on the insured event of 5% to 10% as significant, but it is up to your company to set its own threshold. The other thing that I didn’t mention this morning and we are talking about in more detail this afternoon is that the risk that is covered by the insured event has to be pre-existing so, for example, lapse risk or like the risk of losing fee income if the contracts surrender, that’s a business risk that does not in itself constitute insurance risk. So what we did at Manulife is we provided our businesses with a decision tree to assist them to go though their inforce and product offerings to decide whether the contracts would meet the definition of insurance. I’m taking a very simple example here, but you identify first what the insured event is and, you then ask the question, does the insured event adversely affect the insured, and if the answer is no, then the contract is not an insurance contract. An example of an event that would not affect the policyholder adversely is, say, a toss of a coin so you’re more making a bet so, I’m the company and I agree to pay you $10,000 on a toss of a coin if you get heads say, but that, if I’m writing that business, that’s not insurance. If the answer is yes, then, you move on to assess the significance of the insurance risk by identifying a scenario, any scenario, where the insured event happens so, in this case, the policyholder dies, and you determine what benefits are payable so the face amount is $100,000, then you identify a scenario where the insured event does not happen, in this case, the survival, the policyholder to the end of the coverage period and what benefits are payable to the policyholder then, well, they might surrender, like they have access to a surrender value so that is the benefit you have to take into account. So then you compare the benefits and you determine if the scenario (a) benefits are significantly greater than scenario (b) and if the answer is no, then the contract is not an insurance contract, and if the answer is yes, then, it’s an insurance contract and IFRS 4 applies. So this is something that we didn’t ask our businesses to do for every single policy, but it is really to give them a flavor or a more in-depth appreciation for how IFRS works and how it defines insurance risk. Another consideration when you’re assessing insurance risk is, once a contract is insurance, it’s always insurance. So for example, if you have a life contingent annuity with a guaranteed period, and the annuitant dies PROCEEDINGS OF THE CANADIAN INSTITUTE OF ACTUARIES Vol. 19, September 2008 6 COLLOQUE POUR L’ACTUAIRE DÉSIGNÉ (SP-6) within the guarantee period, that contract continues to be valued as an insurance contract for the remainder of the term. The assessment is made at contract inception. Also, the contract is assessed as a whole, so when you’re looking at a contract, Lesley talked about deposit components within insurance contracts or service components within investment contracts; for insurance contracts, the assessment is made by looking at the whole contract. Once you decide you have an insurance contract, then you ask yourself, are there any other elements in there that have implications under IFRS such as embedded derivatives, deposit components or discretionary participation features? Another important consideration to take into account is that IFRS does allow these classifications to be made at the segment level, so you don’t have to go contract by contract for blocks that have relatively homogeneous risk profiles. There is that allowance in the IFRS literature (I don’t have the exact reference here). An example is you have a type of business where there is minimal risk in some of contracts and higher risks in some others, as long as the block is relatively homogeneous, you don’t have to split those out between investment and insurance, they can all be classified as insurance. I did mention I believe that lapse and expense risk do not constitute insurance risk, and this is true for the direct writer, but if an reinsurer takes on that risk from a direct writer, that would be considered insurance from the reinsurer’s standpoint because it is a risk that is pre-existing to the reinsurance arrangement being entered into and also it is a risk that would adversely affect the direct writer. And the final point is that, interestingly enough, we have found some situations where the assessment of a contract at the legal entity level might be different than at the consolidated level. Most of our work so far at Manulife has been on a consolidated basis but we’re aware we have a number of divisions (in Asia) which have already converted to IFRS for their local basis and we are now looking at whether their classification of contracts for their local IFRS will be the same at the consolidated level. One example, brought to me by Éric Jobin from Industrial Alliance, is if you have a mutual fund contract with an insurance wrapper which is sold from a different legal entity. Within the entity writing the mutual funds, that contract would be an investment contract as there’s no insurance risk, but once you look at the consolidated level and there’s the insurance wrapper, the contract is an insurance contract. So you have to think about how to deal with those situations. Here is a list of insurance contracts we’ve come up with and I did talk about them this morning, they are in your handouts and if you have any questions about it later, come and see me, but I will just skip over those for now. So that ends the contract classification part of my presentation and now we will talk about my favourite topic, embedded derivatives. I do not believe my kids know what embedded derivatives are yet, which is good news for me. Under IFRS, embedded derivatives within insurance or investment contracts must be separately measured at fair value under some circumstances. For those of you who might be familiar with US GAAP, you’ll be ahead of the game here because there’s been a lot of work in terms of measurement of embedded derivatives and assessment of how they should be treated under US GAAP and a lot of what we need to do under IFRS is similar although there are key important differences that we’ve identified. The idea is that if you have exposure in your contract to an external variable, the impact of that external variable needs to be measured and flow through your earnings at fair value each period. Now, there are some exceptions I will go over now. If the embedded derivative is an insurance contract itself or if the cash flows that the embedded derivatives actually affects are life contingent, then the embedded derivatives are exempt from this fair value measurement under IFRS 4. Another important rule is that if the embedded derivative is closely related to the host contract, then it is also exempt and I will talk about some of the examples of these two exemptions in a second. Finally, if you have an embedded derivative within an investment contract and you are measuring that investment contract at fair value under IAS 39, then you do not need to separately measure the embedded derivative. It’s only if you have an embedded derivative within an investment contract that you’re measuring at amortized cost and I will talk a little bit about those two options under IAS 39 for measurement of investment Vol. 19, septembre 2008 DÉLIBÉRATIONS DE L’INSTITUT CANADIEN DES ACTUAIRES SEMINAR FOR THE APPOINTED ACTUARY (PS-6) 7 contracts but you don’t need to separately measure the embedded derivative if your investment contract is measured at fair value. I’ll go over the definition of derivative very quickly. There are three key things. The value changes in response to an external variable; the derivative requires no initial net investment; and it is settled at a future date. This definition is quite similar to the US GAAP definition, though the US GAAP settlement requirement is actually that it is net settled as opposed to settled so … don’t ask me exactly what that means. One of the key differences in the treatment of embedded derivatives is that under IFRS, if the embedded derivative is an insurance contract, it’s exempt and that’s the same under US GAAP, but the definition of insurance under IFRS is, in our opinion, broader, which may result in different treatments of different embedded derivatives. That was the definition of a derivative but there is also a definition of an embedded derivative which means it’s a derivative that is part of a hybrid contract which modifies the cash flows of the contract in response to external market variables. Now I’ll talk about the two exemptions I mentioned. First, if you have an embedded derivative that is insurance itself, then it is exempt from separate fair value measurements. As I said earlier, if the cash flows that are affected by the embedded derivative are life contingent, then the embedded derivative is an insurance contract. So some examples of that are guaranteed minimum death benefits for seg funds, guaranteed minimum income benefits for seg funds, and what I call ‘for-life’ GMWB, guaranteed minimum withdrawal benefits that extend for say 20-years and life thereafter. Those definitely meet the definition of insurance and, therefore, are exempt. They are embedded derivatives but they are exempt from separate fair value measurement, and you would apply IFRS 4. One that we’re exploring right now, or two I guess, but they’re the same rationale, is the treatment of guaranteed minimum accumulation benefits. What we’re talking about here is just a simple minimum guaranteed return after a certain date. Under US GAAP, our treatment of those is to fair value them and they fall under FAS 157. However, under IFRS, because there is definitely a survival element to these features, we believe that they may meet the definition of an insurance contract and so would be exempt. I don’t have the answer, that’s why there’s a question mark there, and we’ve been in active discussions within our company as well as with the industry group that I talked about this morning and with our auditors to try to figure out if a position can be taken. If we can demonstrate that this is true, then those embedded derivatives will also be exempt from separate fair value measurements which will make our life a lot easier converting to IFRS. Another that is definitely exempt is a CPI indexing feature on disability contracts because the payments are life contingent or contingent on morbidity, and so meet the definition of insurance. For those of you who would like more examples on this, I would refer you to the IFRS 4 implementation guidance examples. There are lots of examples of embedded derivatives and whether they do or do not meet the definition of insurance and what the appropriate treatment is. The other exemption is embedded derivatives that are closely related to the host. An embedded derivative is closely related if the risks inherent in the embedded derivatives and the host contract are similar. The IAS 39 literature, which is where you would find the definition of closely related, goes to great lengths and has many examples of features in contracts, both insurance and investment contracts, that meet the definition of closely related. One explicit one is that if the instrument or the derivative is so interdependent to the host contract that it cannot be measured without considering the host contract, then that feature is deemed closely related and, therefore, is exempt from separate fair value measurement. The paragraphs are AG30 and AG33. Unit-linked deposit components within contracts are deemed closely related, there’s an explicit I paragraph in IAS 39 that talks about unit-linked contracts. What we are trying to do now is find out if the crediting feature of equity-linked UL deposits also meets the definition of being closely related. I’m not sure we can make this case that it is closely related but if it is not, I don’t believe there will be significant measurement implications. It might just be disclosure. We don’t have the answer on that one just yet. PROCEEDINGS OF THE CANADIAN INSTITUTE OF ACTUARIES Vol. 19, September 2008 8 COLLOQUE POUR L’ACTUAIRE DÉSIGNÉ (SP-6) And similarly if you have UL accounts where you’re guaranteeing a minimum return so you might have a return expressed, you know, as a percent of an external fixed interest and less a fee to you but you guarantee that it will never be less than 2% or 3%. Is that an embedded derivative that requires separate fair value measurement? Actually, I feel pretty confident that that one can be shown to be closely related, but I’ll update the group at some other time once I’m finished my research. Other embedded derivatives that are at risk, in our opinion, of requiring separate fair value measurement, and I do say ‘at risk’ because, again, we are not finished with our research on this, we started with our US GAAP list of embedded derivatives and looked at those closely and then tried to identify if we had other ones on ceded or assumed GMIB business. GMIB is insurance risk, it’s longevity risk, so it is exempt on the direct side, but if you cede out that business, the way the reinsurance arrangements are designed so they don’t necessarily have longevity risk, so we think those might be embedded derivatives that require separate fair value measurement. There are embedded derivatives under US GAAP - credit derivatives and ModCo Reinsurance of the DIG B36 - which I’m not very familiar with, which may also meet the definition of an embedded derivative under IFRS and require separate fair value measurement. Finally, any surrender values in a contract that are on a fixed schedule or based on a fixed amount plus fixed interest are exempt from separate fair value measurement, but if you have surrender values that are based on an external index, they would be embedded derivatives and might require separate fair value measurement. Now I’ll talk about measurement implications, but we don’t have the answers. We’ve started discussions internally and we have a position at Manulife that we would like to take, however, we’ve just started discussions with our auditors and we’re not sure where we’re going to end up. So, there are two possible ways to measure when embedded derivatives have to be separately fair valued. You either value the host contract, say it’s an insurance contract, you value that under IFRS 4, you value the embedded derivatives at fair value under IAS 39 and you add up the two and you come up with a number that is different than you would have had if you didn’t separately fair value the embedded derivative. The other possibility is to value the whole contract under IFRS 4, including the embedded derivatives cash flows, and you separately fair value the embedded derivative and you allocate the difference to the residual contract. So your total stays the same as it was before and it’s just really a presentation difference. Under current CIA Standards, section 2130, CALM requires that all cash flows of a contract be reflected in the valuation, so I guess one could argue that if we are going to continue to use CALM for insurance contracts, that the Standards do require that all cash flows be reflected and, therefore, the second approach would apply. But there will be some work I’m sure being done in the CIA and we’ll have to see how things evolve in those discussions. On to IAS 39, which I’ve said will be a very high level overview. A number of you might have been involved in the conversion to CICA 3855 in Q1 of 07. CICA 3855 was basically the asset side of IAS 39. For financial liabilities, you can elect one of two valuation methods: the amortized cost valuation method or fair value option. The amortized cost method is basically: at contract issue, you determine the effective rate of interest that exactly discounts your future cash flows to the initial fair value, and then you value the contract on that basis for its duration. The literature seems to imply that you have to go back to contract inception when you do that so, on January 1st, 2011, if you wanted to elect amortized cost for your financial liabilities, if you have to go back to contract inception and assess what you would have held at that time under the amortized cost method and what the effective interest rate would have been at that time; then you have to roll forward the contract to the date of conversion; and then you would take the difference between that and your current liability as part of your opening retained earnings adjustment. Now that’s the interpretation of our accounting corporate controllers area at Manulife and they assure me that the CICA has said that you would have to go back to contract inception to apply amortized cost. There might be transition rules under IFRS that allow you not to do that because, in our opinion, that is way too operationally complex and would not be feasible. Vol. 19, septembre 2008 DÉLIBÉRATIONS DE L’INSTITUT CANADIEN DES ACTUAIRES SEMINAR FOR THE APPOINTED ACTUARY (PS-6) 9 The other implication is that you might have to reclassify the assets backing the business as AFS so you can have proper matching of assets and liabilities going forward if you were to elect amortized cost. And we also found that if we did that, we would have an issue with some of our assets that are allocated to different segments. Just to go very quickly over fair value, the fair value option is similar to US GAAP fair value although there will definitely be differences. At a very high level, you should make maximum use of market inputs to the extent available plus a margin for bearing risk. If market inputs are not available, you have to use some valuation technique and you have to then develop your assumptions based on as much external information as you have. And there are DAC implications. For investment contracts, there is no such concept as a DAC under IAS 39, however, if you have a service component embedded in your investment contract, then you can unbundle that service component and then you can defer some of your transaction costs or acquisition costs. As Lesley mentioned, the rules are generally more restrictive than under current GAAP. So I will conclude quickly by saying that, what we have found is that the definition of insurance is quite broad under IFRS and it is unlikely that companies will have material amounts of investment contracts. We anticipate having less than 10% of our actuarial or policy liabilities as investment contracts and that would be generally similar to the experience that European companies had when they converted in 2005. For embedded derivatives, we may find that most of them are exempt as either insurance or closely related. They will not all be exempt but we’re still working through that. For investment contracts, we’re just starting to understand the implications of IAS 39 at a more granular level and there are potentially some asset segmentation issues with segments that have both insurance and investment contracts. In particular, for example, blocks of payout annuities where you might have a significant amount that are term certain annuities versus life contingent. For service contracts, we don’t anticipate having a lot of changes there in terms of revenue recognition but there will definitely be implications. That concludes my presentation. (Applause) Moderator Thomson: Thanks. We’ve got lots of time for questions, feel free to come up to the mike and give us your name. John Brierley: I find it interesting that GICs are not investment contracts. If you’re considering Canadian GAAP, the term of the liability is to the end of the interest rate period, and then you’re saying that the reason why you can use Canadian GAAP is that there’s some long-term annuity benefits that are ignored in Canadian GAAP. I just don’t understand how you can possibly make that connection if I look at the rules that I just looked at. Moderator Thomson: Well, most of our GICs have some sort of guaranteed payment on death, for example that pay book value on death, so that covers the majority of them. We have had to take a position on annuities that don’t have any death benefit guarantees, but our view is that the annuitization guarantee constitutes significant insurance risk. You could take the opposite approach if you argue that the contract ends at the end of the interest guarantee period. André Racine: Since this is a joint session, we’ll throw in just a little bit of P&C. Moderator Thomson: Our apologies. I didn’t know it was a joint session when I was asked to moderate it. André Racine: OK, I just want to comment that from your presentation, Lesley, I gather that not only the CALM method satisfies their requirements but also most of what is done in P&C also does. There is the potential exception of a premium reserve which, until today, I have always thought that it would disappear and be replaced by a current PROCEEDINGS OF THE CANADIAN INSTITUTE OF ACTUARIES Vol. 19, September 2008 10 COLLOQUE POUR L’ACTUAIRE DÉSIGNÉ (SP-6) estimate of the future cost on an expired portion of policies but from your representation, it seems that the only requirement is that this future cost be considered in determining the liability and what we do in Canada is that we look at the future cost and we look at the premium on a pro rata basis and we pick the higher of the two as the reserve so maybe that would satisfy their requirement. Maybe you are not necessarily aware of the uncertainty that but I’d like if somebody can tell me whether the premium reserve would actually have to disappear because of those new rules or whether it can still stay there until Phase 2 at least. Moderator Thomson: Right, so the question is whether that part of the valuation satisfies the minimum liability adequacy test requirement. André Racine: Yes. Moderator Thomson: Yes, it sounds like it would if I understand what you’re describing. André Racine: The answer is yes, it does. Moderator Thomson: Well, you’re hoping the answer is yes. Actually, the only people that have the right to say whether anything complies with IFRS is your auditor. André Racine: That’s what I mean. Moderator Thomson: So, part of what we are trying to do here is to put out explanations or descriptions or take positions that will make it easy for our auditors to agree with us. André Racine: The other point, the way we do it for premium deficiency purposes, we do recognize all the costs, in fact, what we don’t recognize is potential profit, so as a result of that, we have all the costs in our methodology and it would meet the test in question and I do have access to a few auditors and they do agree with me, so I think that on the P&C side we shouldn’t have any difficulty with our actuarial ability methodology as meeting the liability adequacy test. Although, in some jurisdictions, there’s no reserve on the P&C side, the liability is just a formula without necessarily any connection to reality and in those jurisdictions, that would not pass the liability adequacy test, you would have to calculate something and the matter is very similar to the one we calculate, the way we do it in Canada. Moderator Thomson: Yes, but the unearned premium is a proxy for the liability associated with remaining cash flows, right? André Racine: To the extent it’s profitable. Moderator Thomson: Yes, ok, thank you. Dave Congram: Just a small comment on the same issue. I think it’s important we link in the discussion that took place this morning where there was a reference to the fact that it’s your accounting policy that you are going to continue to use. One of the points that was made that the current accounting, the current CICA handbook makes direct reference to CALM in 4211; the other point was made that there is no reference to the Canadian Actuarial methodology in the CICA handbook for the non life side, so the first point I would like to make to be sure that you’re going to be on side, it would be very helpful if your accounting policy actually said that you apply CIA standards. I think the other point we should just accept is that there are a number of P&C companies that trade Vol. 19, septembre 2008 DÉLIBÉRATIONS DE L’INSTITUT CANADIEN DES ACTUAIRES SEMINAR FOR THE APPOINTED ACTUARY (PS-6) 11 on the stock exchange in Canada and they are allowed under the trading rules to trade on the basis of, let’s say, US GAAP as opposed to Canadian GAAP. Now, they are holding companies, but I think it’s important that we understand those distinctions so that you understand it when you’re signing off. I think from the actuarial side, I would think that the liability adequacy test that the CIA Standards of Practice require for non life do meet the test, but I just want to make sure that we all understand those kind of distinctions and that your accounting policy actually says that. Moderator Thomson: Thanks. Dale Mathews: Lesley, I wonder if you could just elaborate on what you said about foreign exchange rates and the continuation of current rates under IFRS and how that contrasts with our current standards which would be continuation of current rates but with a margin. Moderator Thomson: It’s the margin that would be in dispute under the change that CLIFR is expecting to make. Dale Mathews: That’s something that which we are still struggling with as you probably know. Moderator Thomson: Yes, I think we’ve had this discussion before. I know that CLIFR has taken a general approach of not making changes that move us further away from IFRS, and the foreign exchange change would be one that would move us further away. Dale Mathews: So it’s the margin that’s the issue. Moderator Thomson: Currently, it is just the margin. Dale Mathews: Ok. Moderator Thomson: It would suggest that no margin is allowed. Dale Mathews: Ok, thanks. Moderator Thomson: Anything else? Ok, thank you, everybody. (Applause) PROCEEDINGS OF THE CANADIAN INSTITUTE OF ACTUARIES Vol. 19, September 2008