Session PEN-22 : SERP Pension Session Séance PEN-22 : Séance sur les RSRA

Transcription

1 SERP Pension Session (Session PEN-22) 1 Session PEN-22 : SERP Pension Session Séance PEN-22 : Séance sur les RSRA Moderator/Modérateur: Speakers/Conférenciers: William T. Moore Douglas P. Chandler Lyle Teichman (?? = Inaudible/Indecipherable; U-M = Unidentified Male / U-F = Unidentified Female; ph = phonetic) Moderator William T. Moore: of Session Pension 22. I guess if this is not where you wanted to be you can leave for sure. We re talking about supplemental pension plans. I think we all know what supplemental pension plans are, SERP is a pension plan, either defined benefit or defined contribution that operates on salary above the limits imposed by the Income Tax Act. It is effectively a supplement to an underlying registered pension plan. Generally speaking, you wouldn t see a supplemental pension plan that didn t operate on top of a registered plan. At least in Alberta if you did that, or if you didn t have an underlying registered pension plan the supplemental plan would actually be subject to the Alberta pension legislation. So I m going to very quickly go through some background on supplemental pension plans. Lyle Teichman, if he was here, would be doing a session on the Jobs Creation Act in the United States, which perversely impacts taxation of members of Canadian supplemental plans. I ll do Lyle s piece for him. I just had a brain dump over the phone and, but it ll be at a 50,000 foot level. For those of you who have clients with American citizens who are participating in the plans, this is quite a serious issue unless proper action is taken they will end up being taxed on their accruals under even an unfunded supplemental pension plan. Anyway, quickly then, what the typical, and typical has to be taken in a broad context here, the typical SERP is a defined benefit pension plan. Some of them are defined contributions, but typically established by Board resolution and in the best practices case at least you would have a written document. They are generally unfunded. We ll talk more about that. They re, again, subject to the previous caveat, they re not subject to federal or provincial pension legislation and properly constituted they don t create a constructive receipt issue. In other words, the executor or the employee participant isn t taxed on the accruals normally until they start receiving the benefit. They are of course subject to the normal accounting rules, as any other pension plan, and they re prevalent not just in the private sector, they re actually quite common in the public sector. For example, the federal employees who participate in PSSA have a supplemental pension plan on top of that if they make a sufficient salary and they contribute on all of their income. Alberta set the same system up in 1999 I guess and has a formal funded RCA arrangement. The next two slides are really taken from a Towers Perrin report on supplemental pension plans that was published in Lyle, if he was here again, was one of the primary authors and project manager of that report. And one of the questions in there, why do SERPs exist? And in the priority recorded by the survey participants, clearly the pension plan is there primarily to restore the I use the word restore in the public sector because prior to 1992 there was no effective limits on how much pension you could provide in a public sector plan. So it is to restore in the public sector, but in the private sector it s more of an internal equity issue to provide benefits on whole salary. Clearly, they re as important or even more important maybe from an attraction and retention perspective. There s a large number, or at least a growing number of basically third tier supplemental plans for senior executives that are designed specifically with some very strong retention hooks in them. Clearly they want to be competitive within their industry. Providing adequate pensions is sometimes discussed, but pretty low priority normally. And from the executive s perspective, and again, you know to properly constitute a plan it is tax effective compensation, deferred compensation. A lot of boring numbers here, but again, from the Towers Perrin report, it covered 280 employers and a fairly Proceedings of the Canadian Institute of Actuaries, Vol. XXXVII, No. 2, June 2006

2 Séance sur les RSRA (Séance PEN-22) 2 significant number, this is primarily large employers. I think something like 65% of them had over 1,000 employees, so it s the bigger Canadian companies. So 65% of the surveyed employers had a supplemental plan on a defined benefit basis. Another eight had a defined contribution arrangement and clearly a strong correlation with the size of the employer. I think the numbers are probably higher than this because I m aware of a number of circumstances or have come across them over the years where the supplemental pension plan is only defined in the employment contract with the executive. So I think 65 is maybe a little low. 53% of them are broad based plans, which means everybody who s in the registered plan and who has a high enough salary will accrue benefits under the supplemental plan. Remarkably, only 60% have a formal plan document. There s usually a board resolution around, if you can find it. But to not have a plan document is quite problematic at times. I came across one with a very major corporation that was a board resolution with a two sentence description of the plan, which probably would work fine as long as everybody was in a non-litigious mood, but once that got to court to try to figure out what the executive was actually entitled to was a challenge. 57% of plans provide for a 2% unit, but a number of them, or I think it s what I call super SERPs, actually provide more than a 2% unit. In the oil plans in Alberta it s entirely common to have a third level of supplemental plan for the CEO and the SVP levels. And again, as I say, usually with very strong retention hooks, but extremely rich plans like 3-4-5% units with the objective of accelerating toward a 50-60% pension over a fairly short period of time. But again, only for executives. The preponderance also recognizes incentive pay in the formula, the effort here to leverage on whatever incentive is created by the underlying incentive compensation. The broad based plans tend to be best at the same rate as the registered plan. Again, getting into our super SERPs if you like. 33% of SERPs actually provide an enhanced benefit to the executive group and that could be defined either shallow or deep, but a significant number of basically third level supplemental plans, which the investing public and the regulating public have become very aware of in recent years. And 41% of the plans are now funded or secured in some way. I can t remember the exact breakdown, but in my experience, typically they are secured with letters of credit. There are not perhaps all that many that are actually funded with hard capital assets. The current issues from a regulator s perspective I used to wonder why, with the growing number of people that are covered by supplemental pension plans, why the regulators aren t more interested in bringing these plans under the control or the aegis of their legislation. But so far I ve seen no evidence of that. I think by and large regulators have enough on their plate now. The perception of overly generous executive pensions, I guess the most public example of that was about the chairman of the New York Stock Exchange with his $150 million plus pension. I think since that time, or in some cases even before that, boards of directors are becoming more and more aware of just how big the liabilities are. The regulatory bodies are requiring a lot more disclosure and some boards, or I guess many boards now have been demanding a lot more information from HR and from the actuary. Big impact on corporate earnings. Of course, it was just a couple of days ago or last week there was a note in the paper that the supplemental pension plans or the executive pension plans at General Motors were a billion dollars. I ve lost, well the liabilities were a billion, I ve lost track of what the market cap at General Motors is, but it s a fairly small multiple of a billion dollars I think, so huge impact. Now the two other things that were, let s call them issues I suppose, the American Job Creation Act that I m going to briefly skip through is one little wrinkle that has arisen. And Doug is going to talk more from an operational perspective. I always enjoyed working with supplemental pension plans because there was no rule book, there was no legislation as long as you took care of the constructive receipt issue, it was pretty much a clean sheet of paper. I thought it was good and it was fun, but on the other side of the coin, you couldn t go to the legislation and look up what you could or couldn t do, you couldn t really go to the actuarial standards either. So Doug is going to get into those issues a bit more. So now if you ll excuse me for a minute. Lyle Teichman is a lawyer with Towers Perrin in Toronto. He s had the pleasure or pain, I m not sure which, of working with actuaries for about ten years or so, or more than ten years. Part of that Lyle was with what was then called Revenue Canada. So he s admirably qualified to talk about this new piece of legislation, it s primarily tax legislation. It was passed in October of 2004 and became effective January 1, And the occasion for this was basically the ENRON and the WorldCom types of corporate failures. Part of the failure process if you like was the insider executives who were very much aware of what was happening, triggered their early retirement rights under their supplemental pension plans and walked off with quite a bit of money that they wouldn t have received if they had rode out the insolvency and the bankruptcies. So Lyle calls the legislation a bit of a knee jerk reaction. It is intended to remove the discretion that a U.S. citizen has in triggering their retirement benefits. NQDC is non-qualified deferred compensation plan, basically what we would call a SERP, but there s probably all sorts of plan Délibérations de l Institut canadien des actuaires, Vol. XXXVII, n o 2, juin 2006

3 SERP Pension Session (Session PEN-22) 3 types that are not exactly defined benefit plans, lots of diversity here. But the issue is if you don t now set these things up right, the accruals are going to become taxable to a U.S. citizen working in Canada and participates in a SERP. The U.S. tax law is a little bit unique in that it taxes every citizen irrespective of where they live and taxes them on their world income. And there are certainly a large number of Americans working in Canada who are participating in supplemental plans, pretty much the same way that a Canadian citizen is. So to those of you who are in the consulting field, it s important that you make sure that these people are taken care of because they re typically at a pretty high decision-making level. If these people were in a funded supplemental plan in the past, the accruals were always taxable. But now this new law, if you don t set your plan up properly, will impose taxable income on the accruals even of an unfunded plan. Now the tax penalties are significant. The cost of the current year s benefit would be taxed at 35% with a 20% surtax on top of that. There are additional interest-based charges that are a little fuzzy, but they seem to be more than just interest on the payment, the total potential tax rate being greater than 55%. But the hook in all of this of course is that once the benefit becomes payable in Canada then it s taxable under Canadian tax law. Now you would get an offset for that tax in the U.S., but you get no offset for the initial 55%, so as Lyle says, ouch. So, in order to remove the discretion that executives have had in the past, there are very stringent rules about what are permissible payment dates, what kind of an election you have to make up front in order to get around this tax penalty and certainty elections regarding the forms of payments. Now there is a grandfathering provision, the second major bullet point says that any benefits accrued, well will only apply to benefits accrued after 2004, provided that you don t make any material amendments that would jeopardize that grandfathering position. So, you can only pay benefits on separation from service. Now, some quirks here. If you re a key employee, and key employee is pretty much defined as the top 50 officers of a company, there s a salaries hurdle that you have to be over, I think it s $140,000 U.S. But if you re in that group and you are a key employee and even when you terminate you ll be ordinarily entitled to a lump sum or whatever, it has to be postponed for six months to see whether the company survives that long. Very strict definitions on disability and unforeseeable emergency change in control and so on. And the problem is, or one of the problems, is that most SERPS in Canada would not include language that would be acceptable under the IRS code. And Lyle in his somewhat dry manner points out that the usual definition of death applies. Only lawyers can do that. So one way to do this would just be to write your plan so that all of the discretion is completely removed from the executive, basically you say this is a SERP, your pension is payable at 65, you ve got no early retirement rights and so on and so on. That would probably work. In the real world there are plans with options, either options from a participation perspective or options once you re inside the plan. What the law requires the plan sponsor and then the executive to go through is a process where the executive determines in advance of each year of service what his benefit, or the timing in the form of the benefit. Now I must admit, I don t know how employers are dealing with this. We ve certainly got a lot of U.S. citizens in the Alberta Pension Plan, so There is an exemption once you reach retirement and you ve followed all the rules, you can then determine what the form of your pension is, within limits. If it s a life only you can convert it to joint and survivor and so on, but you can t convert it to a life annuity with a guaranteed term, that would create some certainty. The previous slide talked about the initial election, the election that either is taken away from you in plan design or that you make before you perform the year of service. There are circumstances where re-deferral, for instance, if you define the normal, well let s say you define the normal retirement age as 55. The executive reaches 55 and he wants to continue on till age 65 or age 60 let s say, the law says he can change his election, but the next payment trigger date can t be less than five years in advance. And somewhat quirkily, that decision doesn t become effective until 12 months have gone by from the election date. Well in Canada, under Canadian tax law we have the same constructive receipt doctrine as they do in the U.S. and in that 12 months the individual legally is entitled to trigger payment of that benefit. And since he can t do so, Lyle is of the view that potentially this would trigger constructive receipt. So the executive would have to make sure he makes his decision before age 54 so there wouldn t be that year where he is entitled to receive a benefit. Now you can t accelerate payment either. And the only really exception here is that you could change the vesting schedule, but you can t bring forward the date that the individual will receive his benefit, neither from the employer I guess or from the employee s perspective. A great number of supplemental plans let s say the employer wants the employee basically to have the option of either taking a pension or receiving a lump sum, but the employee can t really make that decision because that would create constructive receipt and would be taxable. So typically the plan document would say something that the basic benefit is a life annuity, whatever. But the employer in his sole discretion has the right to force it in a lump sum payment. Now that gets you around constructive receipt, as far as we know, but that would Proceedings of the Canadian Institute of Actuaries, Vol. XXXVII, No. 2, June 2006

4 Séance sur les RSRA (Séance PEN-22) 4 conflict with the IRS code. Now there is an exemption, which when I left the consulting world I don t think this existed, but there is an exemption for something called foreign broad based retirement plans. Now basically this would be a plain vanilla supplemental pension plan, a broad based plan that everybody s entitled to participate in. And for a U.S. resident living or working in Canada, probably this provision gets you around Section 409a, provided that you don t break the U.S. maximum pension limits, either on the defined contribution side or on the defined deficit side. But it seems quite clear that this exemption does not apply to any circumstance where there is a super SERP or executive or where the executives have any better provisions. I can certainly think of some plans where this exemption is probably applicable. But most broad based plans I think still have some sort of an executive kicker attached to them. So this may not be all that helpful. The transitional really, as I pointed out, the law is already enforced with effect from January 1, There will be the deadline for demonstrating compliance has been postponed until December 31, Now in the interim, the legislation talks about a good faith operational standard being in place, you are to administer as though your plan documents have been changed and so on. Now, Lyle points out that the IRS actually requires you, or may come to require you, to demonstrate that you have been even handed in your treatment. If you ve had a lot of executives retiring in this period you need to have a rough balance of decisions that are in favour of the executive versus decisions that are in favour of the IRS. Clearly he thinks it would be best to have most decisions in favour of the IRS. Final regulations in September of 06, no significant change is anticipated, but Lyle says he is getting some rumours out of the U.S. that they aren t going to be able to get the legislation passed in time. But he cautions that he has no real evidence for that. So the compliance dates are here, but the important one is that anybody in this situation, which includes a large number of Canadian supplemental plans, you need to get it addressed and documented through the Board by December 31 of this year. And with that I ll turn it over to Doug. Speaker Douglas P. Chandler: Thanks Bill. I ll spend most of my time talking about SERP commuted values and maybe a little bit at the end about some of these other SERP actuarial issues. Bill used the word typically a few times in his presentation and I will use it as well because in the absence of regulations of SERPs there s a lot of diversity out there. Typically only monthly income is specified in the plan document. There s scanty documentation of a lot of the plans, particularly around the area of commuted values, partly because at the outset they may not have wanted commuted values. It may be just as Bill alluded to in a third tier plan you wouldn t want the CFO being able to make a decision to get security for the plan when it could jeopardize the solvency of the company. So there are some issues around whether you want to offer lump sums at all in a supplementary plan. But very often a situation arises where suddenly they do want to settle up a commuted value, maybe because of a change of control or because the employer is tired of the balance sheet liability or as part of a severance negotiation or just as an affordable form of post-retirement security. The typical plan sponsor has never thought about whether this is an issue. The typical plan member and the sponsor just trust the actuary to get the job done. It s our business to come up with commuted value and it hasn t occurred to them that it s anything other than what they do in the registered pension plan. And unfortunately there s not much help in the Standards of Practice. Section of the commuted value standard says that it does not apply to the determination of commuted values of pensions and deferred pensions payable from arrangements that are not registered under the Provincial Pension Benefits Standards Act. That rules out most but not all SERPs, there are some seamless plans out there where the commuted value standard does apply. And it goes on to say in , where an actuary uses assumptions or methods described in the standard to calculate commuted values in a situation where the standard does not apply the actuary should not state or imply that the commuted value has been computed in accordance with this standard. Now if you re doing a SERP commuted value you can t say in your report that this value was calculated in accordance with the commuted value standard. You can maybe say using the same interest rate but it doesn t apply. Well I ve taken that approach before, but it s simply saying the commuted value standard is no help to you, you have to go back to the general principles in the consolidated standards of practice. And what those principles say to me is that this is a situation that calls for an unbiased calculation, that you re required to make assumptions about the relevant factors and clearly tax rates are one of the things that the CSOP covers as a relevant assumption. It is okay to have the client tell you what assumption to use, to stipulate the assumption, but you have to disclose that the assumption has been specified by the sponsor. And Section says that if the stipulated assumption is appropriate but near the end of the accepted range that it may be useful to report the result of an alternative assumption near the other end of the accepted range, especially in an external user report. So if you decide to make a tax adjustment of zero, Délibérations de l Institut canadien des actuaires, Vol. XXXVII, n o 2, juin 2006

5 SERP Pension Session (Session PEN-22) 5 then that would seem to me to be a situation where you might want to give an indication of how much the tax adjustment might be at the other end of the range. So how do you go about this? Well you re going to have all the usual commuted value issues to deal with, what mortality rate to use, what discount rate to use, when are some people retiring if they re not retiring right away, possibly marital status and the future retirement date and a couple of other new issues that don t crop up for registered plans, the tax adjustment and the fact that this is a plan integrated with the registered plan. And I ll talk mostly about those two issues. I guess in terms of the other issues, I have found it reasonably safe to fall back on the commuted value standard for those first four issues. It seems to be non-controversial when the numbers are looked at by other actuaries. Despite the fact that I know that high income individuals live longer than low income individuals and there might be other differences there that, you know, in a particular situation. First of all, we make no question about it, tax adjustment can be a very big deal. Here s an example of somebody who s drawing a pension of $100,000 a year and without tax adjustment the commuted value would be $1.4 million with a tax adjustment that drops the expected after tax insurant assets from 5% to 3% and you end up with a $1.8 million commuted value. And that s about the best case. You re going to get a bigger spread if it s an index plan, you re going to get a bigger spread if it s a deferred pension. So let s just look at the principles around this tax adjustment question. From the employer s perspective, if it s an unfunded SERP then paying the lump sum means that they get the corporate tax deduction right away rather than having to wait and get the corporate tax deduction in little pieces over the lifetime of the pensioner. So if you like, the cost of paying the lump sum is their after tax cost of capital not their before tax capital cost. From the employee s perspective, they re going to have to pay tax on the entire lump sum up front instead of having to pay it year by year during their retirement. They re going to have to pay tax on their investment earnings, but of course, dividends and capital gains income get special treatment so they might have some different opportunities than they would have in a registered plan. Or different opportunities they would have with monthly payments. And as they draw the money, to the extent that they re getting a return of capital, there will be no tax on their monthly income over their lifetime. And the key issue here is, and the key motivation for taking a lump sum in some cases, is that it does give you a broader range of tax strategies. First of all, you can turn ordinary income into dividend or capital gains income. Secondly, you can time the incidence of the income a lot better. And this is a really important point when it comes to what sort of tax adjustment is appropriate. I ll take you through this a little carefully. This is back to my example of somebody s getting $100,000 annual pension. On the left-hand side is before you commuted, the right-hand side is after you commuted. So before you commuted you got $100,000 of pension and maybe $50,000 of registered pension, so the total income before tax to this executive is $150,000, all taxable at the regular rate. And in this example, I m just using a 40% tax rate everywhere, so the $100,000 consists of $60,000 of after tax income and $40,000 before tax income. The registered plan is $50 and $30, so in total the executive is getting $90,000 of after tax income year in and year out they re paying 40% tax on all of it. If you commute both the registered plan and the supplemental plan then the executive would be left with $1.4 million without tax adjustment or even more with a tax adjustment in their cash account with their investment dealer and $700,000 in their registered income fund. And the first tax strategy that will come to mind with that money is use the non-registered money first and the registered money last. And if they do that to generate the same $90,000 of after tax income then you see what happens is they only pay tax on their investment earnings from the cash account during the first few years and they don t start paying tax on the registered plan money until they turn 69 and start drawing the money out of the RIF. And these numbers work out, you see it s kind of bumpy and the break in between. That s because the registered retirement income factor and minimum withdrawal factors sort of jump up from age 69 to 71. And then eventually all of the non-registered money is gone and the individual is living off of their registered plan money only. Now if this is what s going on, as you can see, they re paying a lot less tax after having paid the initial hit than you might think. And working this out conceptually, even with a 40% tax rate, you don t end up going from $1.4 million to $1.8 million in terms of the tax adjustment, you end up maybe going from $1.4 to $1.5. So if this is your perspective for the commuted value, if this is your model assumption for what s going on in terms of tax adjustment, then you re talking about a much smaller tax adjustment than it might at first appear. And another tax strategy that you have to think about in terms of what s an appropriate tax adjustment is what sorts of marginal tax rates apply. These are just some numbers I pulled down a couple of years ago, I haven t updated them. And be very careful with these numbers, different numbers apply in different circumstances. But just because in this case in Ontario, which is sort of Proceedings of the Canadian Institute of Actuaries, Vol. XXXVII, No. 2, June 2006

6 Séance sur les RSRA (Séance PEN-22) 6 a middle of the road province, the top marginal tax rate might be 46%, that doesn t mean that somebody s going to pay 46% on all of their retirement investment earnings. I show not just the top tax bracket but also the other ones because even somebody who s making say $200,000 a year in the few years before they retire may not have been making that level of income their entire career and may be retiring on an executive pension much smaller than say $100,000 a year. And also there are situations that supplementary plans apply to other than because of the tax limit they can apply to top up survivor income benefits or bridge benefits, that sort of thing. So it s not just the top tax bracket that matters. And so you re not necessarily looking at the same tax adjustment in all situations. Certainly if we re talking about one of those third tier plans that Bill alluded to, the situation s going to be a little bit different. I don t think it would be safe to assume that say an executive is going to put all their bonds into the registered plan and all their equities into the non-registered cash account and have nothing but equity tax rates if the supplementary plan is say 90% of their total income. But in a lot of cases the supplementary plan isn t that big a piece and you can t look at these things. By the way, in terms of most efficient strategy, it s not always the same. Most efficient strategy in the bottom tax bracket, at least before the last federal budget, I think the most efficient tax strategy was dividends and the top one, at least the most efficient strategy is capital gains. So again, be a little bit careful about this. But the point to this slide is just to make you aware of the fact that it s not always 46%. So this brings you to the question of whether you should be tax adjusting at all and this is, I guess the way we got here was there was a bit of a discussion of this in the Canadian Institute of Actuaries general list server a number of months ago and some people argued for tax adjustment, some people argued against it or some people said this isn t an actuarial issue at all, it s an issue for the lawyers. And that s a great way to avoid the issue if you can do it. Specify the commuted value basis in the plan document or make it so that you re in a position to take a stipulated assumption from your client and maybe or maybe you don t need to explain the effect of taxes in your report. But if you can get it dealt with in the plan document then you re not left with making this decision. And as I say, that s worth a try, but in the absence of that, and very often we are working in the absence of that, I think it s an issue that actuaries have to face head on and prefer to make an assumption about it just like the other things we make assumptions about. If you want to specify the basis, I ve seen it done a number of different ways, some of them work better than others. The most common one I ve seen is to say that commuted values and all actuarial factors in a supplementary plan will be the same as in the registered plan, which means no tax adjustment, and that s just the benefit. It may not be, in our view, enough to replace on an after tax basis the level of income, but that s the basis. I ve seen plans that specify the purchase price of prescribed annuity and that could be a very generous basis, particularly since the executive usually cries poverty in that situation and argues that they are going to be in the $30,000 tax bracket after the initial one and that really bumps up the cost of prescribed annuity. So if you re going to do that you have to think about issues around what other levels of income are you going to consider in determining what tax bracket the executive is in. So I guess I m saying that and some others, there might be some booby traps in terms of they think they prescribed the basis but then when you get into a real life calculation you re still faced with making some assumptions. I guess I haven t seen fixed rates, except for in more complicated SERP arrangements where they re somehow tied to a DC plan or have a DC plan with a fixed rate in it or something like that. There s usually some other reason for a fixed commuted value rate that doesn t reflect the market. And as I said, if you re making assumptions you re going to have to have some sort of a model in mind in terms of what asset mix the money is being invested in. And even if you don t think that you should be taking an equity risk premium into account in calculating commuted values, you should take account of the fact that there are ways to get other types of investment income and use derivatives inside the RPP or somewhere else to make the tax issues turn out differently. There are very creative things that people can do around tax when they re working with a million dollars. The second issue I wanted to talk about commuted values was integration with the registered pension plan. And all I can really say about this is that it s really treacherous. Now Bill commented to me when we were preparing for this that when he set up SERPs he always tried to write them as stand alone documents so that you could determine the amount of the supplementary plan looking only at the supplementary plan document. But I ve seen an awful lot of supplementary plans or two line directors resolutions that just say the executive will get whatever the registered plan would pay but without regard to the tax limits and then the registered plan benefit will be subtracted from the total. And this is an example that I don t think is all that uncommon where the total benefit is say $5,000 a year for a year of service, so $50,000 after ten years of service say. And the plan reduction for early retirement is 5% a year, which brings the executive down from $5,000 a year at age 65 to say $3,750 at age 60. And Délibérations de l Institut canadien des actuaires, Vol. XXXVII, n o 2, juin 2006

7 SERP Pension Session (Session PEN-22) 7 the registered plan pays whatever a registered plan can pay, which at age 60 is $2,444 plus indexing plus a bridge benefit. And if that s the case, then the supplementary plan turns into a temporary pension not a lifetime pension. And this is not the stuff that the average pension administrator can do very quickly, it s really tricky. This is also a good point to talk about, best age issues. In this sort of a situation probably, you know, if you re talking $5,000 a year with 5% per year reduction, you might find that the best age is not age 60, the best age is age 61 or 62, I don t know if the factors work out for the total benefit, but clearly the best age for the registered pension plan taken in isolation is age 60. And probably the best age for the, I don t know what the best age is for the SERP taken in isolation cause it s kind of hard to take it in isolation. I guess my view on that issue is that, as I read the Consolidated Standards of Practice around best age issues, the best age, taking all things into account, which means if the employee is required to elect a commuted value for both the supplementary and registered plan or either the supplementary or registered plan, then you should calculate the best age on the unlimited benefit, even when calculating the commuted value limited there. I guess I ve also run into the odd case where a plan sponsor will provide retiree health benefits to deferred pensioners and that leads me to believe that maybe the best age is going to be the earliest age because they can claim the retiree health benefits when they take the pension. So just remember that best age is all factors taken into account if the member s allowed to make the decision without restriction. And that s another little treacherous point in the calculation of supplementary pensions. And again, remember that the Consolidated Standards of Practice don t apply to the supplementary plans, you re not obliged to use best age. And again, I wouldn t call that an advised best estimate, so maybe you shouldn t. I guess that s just about all I ve got to say about SERPs. Bill, just carry on with the other stuff? Okay. Other types of SERP work, and I ll concentrate on the first couple. We do funding valuations of funded supplementary pension plans and that usually means a retirement compensation arrangement or RCA. And we are from time to time in a letter of credit situation required to do a valuation to come up with the face amount of a letter of credit. And other types of actuarial work we do are plan design studies, costing studies and corporate account. All of which fall within the technical definition of work in consolidated standards of practice. So just a few comments particularly aimed at people who do most of their practice in the registered pension plan world and haven t thought about the special issues around supplementary plans. And of course the big issue with the funded retirement compensation arrangements is that dreaded 50% refundable tax. It was developed and announced by the government in the mid-1980s with the intention to prevent a tax advantage for the funding of supplemental plans, but nobody ever said it was particularly intended to be a penalty. It s just that in the 20 years since that time corporate tax rates have drifted down, personal tax rates have drifted down and nothing s been done about the refundable tax rate. Some calculations that Watson Wyatt did last fall preparing a submission to the federal government and they suggested a tax rate in the low 30s would be neutral from both the employer s perspective and from the employee s perspective. And after the federal budget I suspect it s even lower, it s probably down around 30% now. And at the time we did that we got input from a number of employers. It s clear that if they could, if the tax penalty weren t so great a lot more employers would fund their RCAs. And there s a link down at the bottom if you bring down the website you can probably link to that directly, you can see the statistics around that and a bit more about how we did our arithmetic. It begs the question of whether the government were to drop the 30% tax rate whether it would actually cost them anything in terms of revenue. I have a hunch that there would be so many more funded RCAs at 30% that it might actually increase the short-term revenue for the government. But I know I m talking to the wrong crowd here, so I ll move on from that. In terms of doing the valuation, most of us would just automatically say okay, the expected return on the assets in the RCA is 50% of what you d get in a registered pension plan. But sometimes it s worth working at that a little bit more carefully if the numbers are significant. You can do better in an RCA. You don t get the dividend tax credit or the capital gains inclusions credit that an individual gets, but you don t pay tax on your capital gains until they re realized. So you d have some opportunity for tax to flow within an RCA. And you get to keep the money each year from the time you earn investment earnings until the next spring when you file your RCA tax return, so that helps a bit. On the other side of it, when you get into the benefit payout phase you don t get the refund of the 2% tax on the benefits paid until six months afterwards. So depending on the circumstances, the asset mix, the turnover, you might end up back at 50%, you might be able to do a little bit better. And over the long term in terms of cost of this plan, in particular plan design studies comparing cost of this to cost of some other security arrangements, these issues might be worth thinking about Proceedings of the Canadian Institute of Actuaries, Vol. XXXVII, No. 2, June 2006

8 Séance sur les RSRA (Séance PEN-22) 8 If you re doing a funding valuation, the first question is why are you doing it and why is the employer funding the plan at all? And employers have a lot of different reasons for getting mixed up with RCAs and refundable tax accounts. And depending on the reason, you might end up with a different valuation basis. If it s just about eliminating the balance sheet liability they normally won t even ask you to do a separate valuation, they ll just look at the accounting valuation and do their funding that way. If the objective is to come up with a stable contribution rate then you re probably looking at some sort of going concern valuation. If it s a seamless plan or it s in some way a registered provincial authority then you re going to be subject to all the provincial rules, which means you re going to be doing both the going concern and the solvency valuation. If it s just about getting a tax deduction for the employee contributions to the SERP, and I m looking at one like that now, it s just a matter of getting enough contributions in because under the RCA rules in order for the employees to get a deduction for their contributions to an RCA, the employer has to at least match the employee contributions. And if the objective is to provide the employees with benefit security, which might be the most common reason for setting up a funded RCA, then obviously what you re looking at is the settlement cost. And we ve talked for a while about what settlement costs might be in a commuted value basis or might be in a prescribed annuity basis or something like that. I guess the problem here is you need to know what the settlement basis is. And in the absence of that knowledge you might be looking at some very difficult problems coming up with just what to fund. I ll talk a bit more about that, but that s more of an issue for a letter of credit valuation, so I ll talk about it then. I guess the issue here is that if you re doing a funding valuation that are standards of practice for pension fund valuations, which do apply in this case, say the objectives of funding a pension plan in accordance with accepted actuarial practice or the systematic accumulation over time of dedicated assets, which without recourse is the employer s asset, secure the plan s benefits in respect to member s service already rendered. And the second objective is the orderly and rational allocation of contributions in long time periods. That leads you to a going concern type valuation, but it doesn t necessarily lead to the way we fund RCAs. If you re just funding on a settlement basis and the settlement cost is much lower than the going concern basis, it makes no sense to pay the extra tax penalty to bring it up to a going concern basis. So that s okay to do that, just be careful about what you say in your actuarial opinion as to whether you re funding in accordance with accepted actuarial practice. And having told you what s bad about RCAs, I ll tell you what people have found as a solution to RCAs, which is letters of credit. And I m assuming there s some familiarity with what they re all about. They re designed to be tax effective, at least in the short term, as an alternative to RCA funding. Basically what you re doing is getting a bank to guarantee that the benefit will be paid if the employer doesn t make good on it. But, you know, that letter of credit has to have a specific dollar face amount, which means that the actuary has to provide an estimate of what it would cost to settle the benefits. Same problems as commuted values. If there s no settlement basis specified in the plan documents, you don t know how to do the valuation. And optimal tax adjustment depends on some choices that can be made by the plan member. So yes the plan member might be able to do pretty well if they invest in income trusts and dividend bearing stocks and that sort of thing, but the plan member s under no obligation to do so, so if a letter of credit is called you re into this dispute over how much is enough. And you could end up using the 46% tax adjustment basis and not the smaller tax adjustment basis that you might think is more appropriate and probably more the way it s going work out just because it wasn t specified in advance. And you re into an adversarial situation when it happens. So if you re doing one of these valuations, the first thing you have to do is be aware of the annual cycle involved. There s a whole bunch of layers, there s the plan sponsor, there s a trustee who is going to hold the letter of credit, there s the bank that s going to issue the letter of credit, there s somebody who s administering the plan, which is probably the same person as the sponsor. And it might be annual, well it s always annual with an annual renewal period, but the date of renewal is not the date necessarily the date that you have to have your valuation done. Very often the valuation has to be in the hands of the trustee a month in advance or something like that in order to prevent triggering the previous year s letter of credit. So just be aware of who all the players are, who your client is, who you re delivering the report to and who s relying on it and what the annual cycle is. And keep in mind that when you re doing one of these valuations, it s sort of like appraising a house for fire insurance, you re telling them how much insurance to buy if things go really badly. And keep in mind that it s not how much it would have cost. Well, you may say in your report that a letter of credit of $10 million would be sufficient to settle all the benefits of the plan as of January 1st, But if the purpose of your report is to provide for renewal for the period from September 2006 to September 2007, what really matters is whether or not it s enough insurance during that entire year. And the amount that it would actually cost to settle the benefits depends on interest rates at the time, which fluctuate. You might have some plan provisions in there like median cliff vesting, which in a SERP is much more common than a registered plan. No benefit if you leave before you re 60, Délibérations de l Institut canadien des actuaires, Vol. XXXVII, n o 2, juin 2006

9 SERP Pension Session (Session PEN-22) 9 so no letter of credit if you re 59 and a half, but a huge benefit if you leave when you re 60 and a half. So if those sort of median cliff vesting things happen during the year covered by your letter of credit the trustees and the beneficiaries aren t going to be very forgiving if you didn t pay any attention to them. Even something like bonuses or new members can really bump up the value of the letter of credit. Now, I m not sure what the best defence of these is. I don t think the best defence is provisions for adverse deviations in margins in your valuation, I suspect a better defence is good plan to the documentation that says either that a triggering event is if they bump up in the membership that the plan would cost they would have to do a new valuation or that says that the value of the settlement will be whatever the letter of credit is if you determine to allocate it in some way amongst the plan members. So again, and I think this will be a recurring theme in supplementary plans, you ll do better with documentation than with actuarial assumptions. Just a quick reminder that in a letter of credit there is a refundable tax account out there and if it s a pure letter of credit arrangement with no funds building up at all, you can t get the refundable tax back from the government until the very end of the world, which means that it doesn t matter much the day you set it up, but if the thing runs for years it s going to be a pretty big number. And in fact, in this little illustration of a single person letter of credit starting at age 40, by the time they re 95 or 96-years-old you don t need a letter of credit anymore because the refundable tax account has more money in it than you need for the whole plan. This is one reason that some people consider letters of credit to be temporary solutions, not permanent solutions because just over time they build up too much money in that noninvestment bearing account. And again you ve got the same sorts of issues around integrating with the registered pension plan complexity of what s the registered pension plan going to pay, particularly in the situation where the supplementary plan pays whatever the registered plan doesn t pay. Keep in mind that this is, you know, the event you re insuring against is probably a wind up of the registered plan as well as a wind up of the supplementary plan. So if the hypothetical wind up scenario, if I can use the actuarial terminology, is a wind up of both plans, that begs the question if it s an Ontario plan with legislated grow in and your planned document doesn t say anything about grow in, does that mean that the extra values paid from the registered plan because of grow in are an offset against the total benefit promised by the SERP or in addition to the total benefit paid by the SERP? If the registered plan has surplus and the administrator of the registered plan decides to distribute the surplus by improving the registered plan benefits, does that reduce the supplementary benefits or does that increase the supplementary benefits? And these are real issues. I ve had to deal with these issues in real life. And again, the solution is not to try to make assumptions about these things but to get them into the plan documentation. I am going to skip over these last couple of slides, they re in the handouts and just, you know, for question period if you want to talk about them you can. These cost comparisons are here not because they re accurate numbers. They re very volatile dependant on what assumption s used and that sort of thing, just to give you an idea of what s out there. And again, lead to a bunch of questions. And there s going to be lots of questions around corporate accounting because SERPs are special, they re typically very small plans, they have some funny provisions that you don t typically run into in registered plans. So I ll leave it at that and turn it over to Bill. Moderator Moore: Well thanks Doug. I mentioned I couldn t remember the proportion of plans that are actually funded with hard capital assets, but according to the Towers Perrin survey for defined benefit plans, 17% are funded with real money and 25% with letters of credit. On the commuted value issue or question, by an accident of history, I ve put in most of the public sector supplemental plans in Alberta and basically the way we ve done all of those is to define the commuted value basis as being the accounting basis so that the accounting numbers, the funding numbers, if they are funded, and the cash out values are all the same. Working with public sector clients keeping it simple is very desirable. So any questions? We ve got lots of time. I didn t think that there was this many people in Canada who were interested in SERPs, but While you re getting to the microphone, the Towers Perrin survey I think is available on their website if I m not mistaken. Or if not I m sure you can get a copy, it was in I know Wyatt has done surveys, Mercer s have done surveys so there s lots of prevalence material out there. Sir? Robert Rosenblat:?? microphone not turned on [55:55 56:24] I ve never had a problem with projection of my voice here. Rob Rosenblat. I ve got a question, Doug you mentioned something in your presentation that I tend to agree with about trying to define what the commuted value basis should be so you avoid all the negotiations later. But then you went on to say that typically you re often seeing the same basis as what we re seeing for commuted values of registered plan, not withstanding standards don t apply to it. Doesn t that contradict some basic actuarial principles like unisex mortality versus sex distinct? And what makes the Proceedings of the Canadian Institute of Actuaries, Vol. XXXVII, No. 2, June 2006