New FASB calculations

[FAPInJax]

I attended the ASPA webcast and had several conversations with other actuaries regarding the 'benefit payout' information that is supposed to be a new disclosure.

Does anyone have an idea HOW this information is supposed to be calculated??

How detailed the calculations have to be - for example a joint and survivor annuity for a retiree - possibilities abound that 'could' be valued - both alive, participant/spouse dies, both die???

The presentation appeared to imply that these calculations are NOT present values at the valuation date BUT the value in the year paid. Does anyone agree with this or disagree?

Just trying to get a handle on how these calculations are to be performed since no direction was provided by FASB.

[Guest DBtech]

Frank, the following is from the FASB Q&A on 132:

Q14. Should estimated future benefit payments be calculated on an open- or closed-group basis?

Consistent with all other measurements pursuant to Statements 87 and 106, projections of benefit payments should be calculated on a closed-group basis (include only current employees).

I've been basing the projections on the plans' normal annuity form, even in cases where most employees elect lump-sums.

[MGB]

The payouts should be EXACTLY as you are valuing them in your assumptions. Using a single life annuity for a future decrement is incorrect if you have assumptions that they will take another form (either some or all of the time).

For example:

Each year for the next five years, you have assumptions about the probability that a person will retire without staying active after five years. In each of those years, you have probabilities of each form of payment.

For one decrement in the future, let's say you are assuming 50% take a lump sum of X and the other 50% take a single life annuity of X/12 per year. In the first year after decrement, the benefit payout projection is (prob. of living until decrement)*(.5*X + .5*X/12). For every year thereafter, the benefit payout projection is (prob. of living to decrement)*(prob. of living to each benefit payment age)*.5*X/12.

These are then each multiplied by the probability of decrement and summed over all five possible decrement dates.

This is straight-forward actuarial mathematics. The more complicated your PV assumptions, the more complicated this is.

By the way, the Board feels that if you are not doing this in your valuation system already, then you are not applying SFAS 87 correctly, given that you must recognize the time to each payment in determining the discount for every cash flow in the future. Using commutation functions or PV factors with one discount rate is not following the rules correctly, according to the Board.

Therefore, they saw this as a simple piece of information already available and no new programming should be needed, except perhaps to capture a summation of each future possible cash flow.

[david rigby]

This is straight-forward actuarial mathematics.

I love that phrase.

[rcline46]

All I can say is- WHAT A BUNCH OF MALARKY.

In the small plan market place annuities are extremely rare, so the PV of lump sums due in future years is the actuaries best estimate. And mortality discounts are also rare, so single value decrements are common.

FASB just cannot take the actuary's best estimate out of the picture to conform to some fantasy they have. FAS 87 is already unreal enough to the actual health of a plan. I will not make it worse. PV of lump sums is what will be provided in my reports.

[Belgarath]

Pax, I agree! Following is an excerpt from a letter I wrote to a client last year, who had several questions on calculating funding for "Regular" DB plans and 412(i) plans:

"...in a very general way, IRC Section 412 concerns the minimum funding requirements for a DB plan (the actual calculation of which is a combination of mathematics, necromancy, and artistry more commonly known as "actuarial science"..."

[david rigby]

Now, hold on! I take issue with "necromancy" (I had to look it up).

[Chester]

The FASB Board might think that is how actuaries are calculating the numbers, but I am sure that the vast majority of actuaries are using commutation symbols and discount rates instead of the cash flow method FASB envisions. Am I right?

[Guest flogger]

First let me say, at the risk of stirring up a whole new conversation, that I agree completely with rcline46! Just like the entire IRC, ERISA, etc. and ALL the promulgations, FAS is just another insanity that we (actuaries) have bought into. Of course, we must abide by the rules, but there is never an uprising by the pension community to say "let's blow this whole thing up"! It's just not the nature of the actuary (includes me) to be anything other than compliant, non-creative and factual. We pile complication upon complication, always to the negative of our clients, and then proceed with our work like the whole mess makes sense.

That being said, I use the multiple decrement approach to FAS only when it makes sense, and that is very seldom for my clients. Rcline46 is right about the small plan market, as it is once again ignored by rule makers. One saving grace we have in our pocket is the amortization of loses or gains that result from our (always) inaccurate assumptions. As long as that delta between actual and assumed is not too large, then ignoring the multiple decrements is a "no harm no foul" situation.

[SoCalActuary]

When I work in large plans, the cash flow calculations have to be projected for each of the decrements built into the FASB valuation. If you have decrement assumptions, each decrement creates a matrix of future payments by year.

FASB appears to want you to disclose the total of all such future payments by year for all decrements used.

In small plans, I usually assume that lump sums will be taken, but not always. When the plan has a history of annuity payments, then I feel constrained to use the same complex model as I use for large plans.

This is not a problem for people who use large plan software, but none of the small plan packages that I use (Relius, ASC, Datair) have built the reports needed.

[flosfur]

I attended the ASPA webcast and had several conversations with other actuaries regarding the 'benefit payout' information that is supposed to be a new disclosure.

....

Is there a new FAS # or modifiation to an existing FAS that requires this new disclosure?

[david rigby]

See FASB statement no. 132, as revised in December 2003.

Try www.fasb.org