Another accounting question: G/L on assets

[bluekayaker]

If I'm trying to determine the gain or loss on investments (say, for the balancing equation for a reconciliation of the unrecognized accrued benefit/liability), do I use actual or expected contributions when calculating expected assets?

I'm practising on Edge exam #2, question 4. The expected assets are based on EROA (which would have been based on expected contributions) but also on actual contributions. I find this counter intuitive. Anyone able to explain why I'd use expected EROA and actual contributions?

Expected Assets = BOY assets + actual contributions + EROA (based on expected CFs) - benefit payments

Thanks!

[Sifto]

in practice, most clients (not all) do not like to adjust their expense from the estimated expense determined during the previous fiscal year. the estimated expense is determined from expected cash flow.

when you do the disclosure then any differences in actual and expected cash flows are caught in the gain/loss.

the EROA is just an assumption set based on the long-term return of the asset mix. i'm not sure what you mean when you say the EROA ies based on actual contributions.

[Kenny]

Your gain or loss is simply the difference between what you expected and what actually occured. It is as simple as that. Roll forward libaility based on expected benefit payments and take the difference between that and your actual liability at that date. Roll forward assets using expected benefit payments and expected contributions, take the difference between that expeced value and the actual value and that is your asset g/l.

Exam problems typically only provide one benefit payment number and frequently only one contribution number (although occasionally with settlements or curtailments actual contributions for the year differ from expected prior to the event). For this problem, my guess is only one contribution value was provided so you did not have a choice in what to use. If expected contributions were provided and they difference from actual contributions, then I would guess that the solution has an error.

[bluekayaker]

In this particular problem, expected contributions were (I think) $32000, while actual contributions were $0. The roll forward of assets was: Beginning assets + EROA + Conts - Ben Pmts where:

EROA was based on expected contributions of $32000 (among other items)
Conts were $0

I would have expected the EROA entry in the equation above to be based on actual contributions and the expected RoR.

[Sifto]

as in practice, it's normal to use expected contributions. you would then come up with an expected asset value and any difference with the actual assets (which has actual cash flow) will be a g/l item.

[bluekayaker]

OK.. I guess the Edge solution still puzzles me then. Perhaps it's an error?

[Sifto]

i just did that problem...it says there are no contributions during FY2007 and benefit payments are not affected by the curtailment/settlement. your pension expense during the year would be based on zero contributions and the benefit payments provided in the case study.

You're probably aware of this but note that the case study provides valuation numbers as at Jan 1, 2007. It provides cash flow for 2006 but not for 2007. The question tells you what cash flow to assume for 2007 which is zero contributions and expected benefit payments as per the previous year. The latter is a pretty standard assumption if nothing unusual has happened in the previous year.

[Kenny]

it says there are no contributions during FY2007
Agreed, see second to last bullet point of the problem.

I have no idea where you came up with Exp Contributions of $32,000

[Kenny]

I should also add, it is pretty common (at least in the US) to know the contribution schedule before they are made. FAS 87 requires cash accounting so there is no concern with receivables. Many companies contribute the minimum required and strictly follow the quarterly requirement schedule so even in practice expected contributions likely = actual contributions
.

[bluekayaker]

OK, I think I've got it now. Kenny, the $32k came from line (m) under "2007" column on the case study. I assumed, since the column heading was 2007, that these were 2007 amounts which were revised by actual experience of $0 in 2007. I guess it is a 2006 amount and nothing has been revised - in which case it all makes sense.

Sifto, you nailed my confusion on the head! By the way, did you find this question quite labour intensive and really long? If we get something like that with 3 rate changes, non-concurrent settlement and curtailment, it's going to take me more than half an hour!

[Sifto]

I should also add, it is pretty common (at least in the US) to know the contribution schedule before they are made. FAS 87 requires cash accounting so there is no concern with receivables. Many companies contribute the minimum required and strictly follow the quarterly requirement schedule so even in practice expected contributions likely = actual contributions
.

question kenny, nothing to do with the problem that we discussed, but i've hardly ever come across a case where actual contr = expect contr.

for a flat benefit plan, if the sponsor contributes on a per member basis, then normally there will be membership changes during the year which would change the contributions.

for a salaried plan, if the sponsor contributes as a % of salary, then membership changes and salary changes would disrup the contributions.

i understand the differences are normally not material but almost always differences exist.

[Sifto]

Sifto, you nailed my confusion on the head! By the way, did you find this question quite labour intensive and really long? If we get something like that with 3 rate changes, non-concurrent settlement and curtailment, it's going to take me more than half an hour!

i guess it was hard to do it in 24 minutes (8 point question) but i wasnt surprised..that's the nature of the accounting questions so you really have to know what you're doing without too much pause.

[bluekayaker]

I'm second guessing myself again.

When looking only at the rightmost column, the items in line (m) and (n) of the current case study (expense valuation pages), are these amounts actual 2006, expected 2007? solutions from what i can tell treat these as 2007 amounts (can be verified by recalculating the line (c) amount, EROA based on 32000 conts, 15000 ben pmts, and the jan 1 FV).

I also have been assuming all along that the items in section 2, pension expense, are for the year 2007 in the rightmost column.

[Sifto]

I'm second guessing myself again.

When looking only at the rightmost column, the items in line (m) and (n) of the current case study (expense valuation pages), are these amounts actual 2006, expected 2007? solutions from what i can tell treat these as 2007 amounts (can be verified by recalculating the line (c) amount, EROA based on 32000 conts, 15000 ben pmts, and the jan 1 FV).

I also have been assuming all along that the items in section 2, pension expense, are for the year 2007 in the rightmost column.

yes you've got that right. the pension expense is for 2007. the cash flows (32,000 and 15,000) are the actual 2006 cash flows and the expected 2007 cash flows. These were used for the 2007 expense which to be precise is an estimate. The question in the sample exam said there were no contributions in 2007 so to be precise, the pension expense for the first 6 months should be recalculated (since the EROA will be different) instead of halving the expense provided in the case study which is what the Edge solution does.

[Kenny]

question kenny, nothing to do with the problem that we discussed, but i've hardly ever come across a case where actual contr = expect contr.

for a flat benefit plan, if the sponsor contributes on a per member basis, then normally there will be membership changes during the year which would change the contributions.

for a salaried plan, if the sponsor contributes as a % of salary, then membership changes and salary changes would disrup the contributions.

i understand the differences are normally not material but almost always differences exist.
I don't see a question but I'll give an answer anyway. US funding rules require advanced funding of a portion of the minimum required contribution for the year if the plan does not meet a certain funding threshold in the prior year. We refer to these as quarterlies because they are required to be made once a quarter with the balance of the min made at a later date (also specified by law). It has become quite common in the US for sponsors to only make the required minimum contribution, thus they follow the contribution schedule we lay out as required by law. Because we typically prepare funding valuations (and this is common for most of the plans I have ever worked on) prior to or concurrent with the expense valuation we know the expected contribution schedule for the calendar year and can reflect that in our expense calculation. There is typically an asset g/l because actual return always varies from expected, but it isn't usually due to a difference in contributions.

[Sifto]

I don't see a question but I'll give an answer anyway. US funding rules require advanced funding of a portion of the minimum required contribution for the year if the plan does not meet a certain funding threshold in the prior year. We refer to these as quarterlies because they are required to be made once a quarter with the balance of the min made at a later date (also specified by law). It has become quite common in the US for sponsors to only make the required minimum contribution, thus they follow the contribution schedule we lay out as required by law. Because we typically prepare funding valuations (and this is common for most of the plans I have ever worked on) prior to or concurrent with the expense valuation we know the expected contribution schedule for the calendar year and can reflect that in our expense calculation. There is typically an asset g/l because actual return always varies from expected, but it isn't usually due to a difference in contributions.

interesting..and how are the going-concern contribution requirements expressed for a final-average plan in a valution? would it be $X for the year in which case I can see it is a fixed amount which wouldn't fluctuate during the year. Or would it be expressed as Y% of member's earnings in which it's surely going to be different based on membership and salary changes.

my point was that it's very unlikely that actual contributions equal expected contributions due to membership/salary changes.

[Kenny]

yes you've got that right. the pension expense is for 2007. the cash flows (32,000 and 15,000) are the actual 2006 cash flows and the expected 2007 cash flows. These were used for the 2007 expense which to be precise is an estimate. The question in the sample exam said there were no contributions in 2007 so to be precise, the pension expense for the first 6 months should be recalculated (since the EROA will be different) instead of halving the expense provided in the case study which is what the Edge solution does.I'm going to disagree with the bolded statement.

I missed this when I quickly glanced through the problem earlier, so blue it appears you are correct, the $32,000 is expected contributions when the expense is initially determined[b]. Thus the original expense calculation [B]does incorporate the $32,000 contributions into EROA. I disagree with Sifto because I think a simple proration is a perfectly acceptable method for calculating the partial year expense when the full year expense has already been provided on that basis (this is an even more useful method given the time constraints on the exam). It would be more precise to recalculate the expense strictly for that time period, however it is really a gray area as to whether or not you should update the expected contribution assumption after the fact. I likely would not.

However, it is reasonable to assume that at some point during the year the knowledge that no contributions are going to be made is known and it is reasonable to use this knowledge at one of the remeasurements. Please note the specific language of "will be" rather than "were not" (i.e. future vs past tense). However, if you used $32,000 for all periods and that were the only difference between your answer and the "official" answer I doubt it would materially affect your score, if at all.

[Kenny]

interesting..and how are the going-concern contribution requirements expressed for a final-average plan in a valution? would it be $X for the year in which case I can see it is a fixed amount which wouldn't fluctuate during the year. Or would it be expressed as Y% of member's earnings in which it's surely going to be different based on membership and salary changes.

my point was that it's very unlikely that actual contributions equal expected contributions due to membership/salary changes.
It "used to" be a going concern valuation such that the minimum was calculated using a reasonable funding method, subject to specific statutory constraints and plans started strictly making the minimum required for other business reasons. I say "used to be" because the Pension Protection Act changed the required assumptions and essentially limits the calculations to termination liability assumptions. In my experience US corporate plans no longer take a long-term, traditional actuarial funding view, which is probably why the FE approach, to valuing liabilities at least, is starting to gain traction.

ETA: sorry, I didn't actually answer you question. The min (and thus the contribution schedule) is expressed as a fixed $ based on beginning of year liabilities and assets.

[Sifto]

I'm going to disagree with the bolded statement.

I missed this when I quickly glanced through the problem earlier, so blue it appears you are correct, the $32,000 is expected contributions when the expense is initially determined[b]. Thus the original expense calculation [B]does incorporate the $32,000 contributions into EROA. I disagree with Sifto because I think a simple proration is a perfectly acceptable method for calculating the partial year expense when the full year expense has already been provided on that basis (this is an even more useful method given the time constraints on the exam). It would be more precise to recalculate the expense strictly for that time period, however it is really a gray area as to whether or not you should update the expected contribution assumption after the fact. I likely would not.

However, it is reasonable to assume that at some point during the year the knowledge that no contributions are going to be made is known and it is reasonable to use this knowledge at one of the remeasurements. Please note the specific language of "will be" rather than "were not" (i.e. future vs past tense). However, if you used $32,000 for all periods and that were the only difference between your answer and the "official" answer I doubt it would materially affect your score, if at all.

your point is appreciated regarding pro-rating the expense in light of the time constraint...but then how long does it take to recalculate the EROA based on zero contributions? it saves disclosing when the knowledge of zero contributions is assumed to be made available.

i would just calculate the expense based on the cash flow provided in the question...i think it's just "safer".

having said that, i don't think either method would result in a deduction of marks.

[Sifto]

It "used to" be a going concern valuation such that the minimum was calculated using a reasonable funding method, subject to specific statutory constraints and plans started strictly making the minimum required for other business reasons. I say "used to be" because the Pension Protection Act changed the required assumptions and essentially limits the calculations to termination liability assumptions. In my experience US corporate plans no longer take a long-term, traditional actuarial funding view, which is probably why the FE approach, to valuing liabilities at least, is starting to gain traction.

ETA: sorry, I didn't actually answer you question. The min (and thus the contribution schedule) is expressed as a fixed $ based on beginning of year liabilities and assets.

this is news to me and i find it very surprising. i know we're probably off the exam topic but i wanted to discuss.

when doing a valuation on a termination basis (i guess that would be more like a solvency vlauation), would you have a salary scale assumption for a final average plan? if so, then expressing the normal cost contribution requirement as a fixed $ will under/over fund the cost of benefits accrued if the average age remains constant (valid assumption for an open plan) but the salaries change (which is almost certain).

here's an example:
Time 0:
total plan NC = 200,000
sum of salaries = 2,000,000
salary scale (reflected in NC) = 3%

The NC can be expressed as a fixed $200,000 or 10% of earnings.

Time 1:
Suppose the salaries have increased as assumed: 2,000,000*1.03 = 2,060,000
Assuming the average age is constant, the cost of benefits accrued is: 200,000*1.03 = 200,600

But the sponsor has contributed only $200,000 so the 1 yr of service has been underfunded by $6,000.

Now, if the NC contribution had been expressed as 10% of earnings, then the amount contributed would have been: 10% * 2,060,000 = 200,600 which matches the cost of benefits accrued.

I find it surprising to express the normal cost as a fixed $ for a salaried plan with a salary scale. The above is even more exagerrated if say a bunch of people left the plan and so the time 1 salaries were 1,050,000. Now you're really overfunding. And remember, all of this is only on that 1 yr of service accrued.

When doing the gain/loss for the valuation report, I'm guessing you would then lump the NC salaried gain/loss with the AL salaried gain/loss.

[Sifto]

one last thing regarding the general discussion on whether to use actual or expected contributions in the pension expense. sorry kenny, i'm not trying to be difficult but i've had this discussion with actuaries in the past and they agree that technically it's correct to use actual contributions. using expected expected contributions in the pension expense poses the following problem:

- the gain/loss item captures the differences in the cash flow. The gain/loss on assets should really only reflect how the assets performed in comparison to the expected return assumption. By using expected cash flow in the expense we are also including the the differences in the cash flow in the gain/loss. Therefore, the gain/loss item can be siginificantly distorted depending on the differences between the 2 sets of cash flows.

- Same thing on the liability side. The disclosures require a reconciliation of assets and liabilities which include actual cash flow. You could put in expected cash flows in here instead and put the difference in the gain/loss but again, you're distorting the gain/loss item.

I would think that auditors usually know what the actual contributions were for a company. When they look at pension accounting disclosures (the rare cases they do) they would look for the actual contributions and see if the pension expense matches. Unless you disclose somewhere the expected contributions, it would be impossible to duplicate the expense calculation.

I understand from a client perspective, they do not normally like to restate the expense but as consultants I think we should try to encourage them to do so. Especially in this day and age where everything is moving to marked-to-market.

Ok this is way longer than I had hoped....sorry!

[Kenny]

From an accounting standpoint I do not see that happen practice. Perhaps it would be more appropriate, but I'm not convinced. I'll have to think about it, and while I would prefer to discuss things other than what is on the exam, I'm doing my best not to spend time on topics I don't need to pass on Friday.

Maybe we can come back to this at a later date.

[Kenny]

I started a thread to discuss our tangent in the Pension section http://www.actuarialoutpost.com/actuarial_discussion_forum/showthread.php?p=2851018#post2851018