Amortization Base in Year Following Early Retirement Window

[LIBOR]

A plan is amended to provide an Early Retirement Window (ERW) in 2004 but utilizing Rev Ruling 77-2 it's not recognized until 2005 ; the funding(cost) method was changed to FIL a couple of years ago and so it can't be changed to something like Aggregate this year ----- for those partcipants that qualify, the window amendment provides (1)enhanced accrued benefits (2) smaller early reduction factors and (3) the opportunity for immediate retirement and payment ---- some participants would not have been eligible for early retirement under the pre-window plan --- participants who opted for the window are "pending retirees" on the 2005 valuation date having made their benefit elections at the end of the 2004 plan year - many were lump sums.

My question involves the correct way to set up the amendment base - I'm thinking it would be A minus B where A is EAN-AL of the enhanced benefits for all those that elected to retire under the window.

My quandary is with B ?? Is it :

(a) the EAN-AL of the same window group assuming they retired under the

provisions of the pre-amended plan.

(b) the EAN-AL of the same window group assuming they terminated from

the plan with deferred benefits payable at NRD.

© something else ???

Any insights would be appreciated and I'd also like any thoughts on whether there should be any other base , e.g. assumption change - this doesn't seem likely since the assumptions for the 2004 valuation didn't include a consideration for a percentage of participants that would opt for the window.

Thanks again !!!!!!!!!!!!!!!

[david rigby]

Not sure if there is significant difference between (a) and (b), but I think © would be to assume the EEs did not sever employment at all.

But what is the correct answer? For example, it is never EAN-AL of A minus B. The proper formula is EAN-UAL(A) minus EAN-UAL(B). The difficulty is that (B) requires you to measure the liability under one of (a), (b), or © as above (easy enough), and to measure the assets at the same date, but that asset amount is hypothetical because it assumes different benefit payments than the actual pattern.

How do actuaries solve this problem? Verrrrry carefully. Often, it depends on convenience: you may have to "wing it", or hope that there is no significant difference among the various choices. In my own history, I usually choose whatever is the simplest to calculate, usually because it is not significant to the plan as a whole. BTW, I think (but am not absolutely sure) that FAS87 would tend to focus on option (a).

[LIBOR]

Thanks Pax - one of my colleaques suggested not setting up a base at all and just running the first "post-window" valuation using FIL ; in my situation those that selected the window had made their benefit elections by 6/30/05 ( i.e either annuity or cash) and the assets were still intact on my val date of 7/1/05.

Not creating a base has some merit since one of the many dilemmas/questions here is : "If the early retirement window amendment had not happened, would those opting for the window still be active?". And depending on the answer the UAL(before) , my B in the original post, could be your © or my (a). Staying with FIL pushes everything into the NC.

The only thing for sure it seems to me is the A in my post ; because if I was able to change to Aggregate, then the PVB would include the window group's liability as "pending retirees entitled to cash" except, of course, for the couple who elected annuities.

Another question : Does 412 require you to establish a base when there's an amendment ??

[Effen]

I think you need to establish a base, unless your assumption is that no one will take the window, but since you know people did, that would probably be unreasonable to assume they won't.

Since you know who is taking it, won't it be reasonable to create a base equal A-B, where A = EAL for those who took the window, including the value of the window and B - EAL prior to any change (not assuming they retired or terminated, but just whatever the EAL was). Isn't that really the impact of the amendment?

[LIBOR]

I'm starting to lean that way - Effen, the window amendment enhanced accrued benefits - some elected an annuity, others cash - for purposes of base creation w.r.t. those electing cash, would you determine EAL(after) on the cash assumptions or valuation assumptions ?

And let the liability difference between assumptions flow to the NC.

[david rigby]

Ah, there's the rub! Determining the "before" AL and/or the MVA will require you to make some "adjustments", such as adding back in some (but not all) payments. Awfully tempting to change to Aggregate (but as you state, not available here, unless you request IRS approval).

I agree with Effen that the defintions and structure of IRC 412 (et al) will direct that you create a base, unless you can demonstrate that it is de minimus. Not usually, but it is OK to ask yourself that question. However, my understanding of 412 is that you should use the Unfunded AL. Perhaps Effen was just using shorthand.

If you did a window 8-10 years ago, you probably found out that windows work very well in a plan with surplus assets; if you are really fortunate, it will have full funding, and you won't care what the base is. But I digress.

Technique (1) might define the AL and assets on a "before" basis by assuming all such employees did not sever employment. Thus, you will determine the assets by adding back all payments actually made to those participants (use your judgment about an interest adjustment, but I would be skeptical about that). Second, you will have to calculate liability by reversing all the retirements and make each one an active employee; that probably means you must impute a full year's pay. (Kind of makes you uncomfortable, doesn't it?)

After these machinations, you must then review the result, ie, use your actuarial common sense. If it does not work, then try technique (2): same as (1) but assume all affected employees retired/terminated without the window provisions. This technique is a bit simpler, mostly adjustment to the assets without adjustment to the census data.

Alternatively, perhaps you think (2) should be tried first. And maybe you're right.

[Effen]

Boy, maybe I forgot my actuarial funding 101, or maybe I never learned it, but that seems like a lot of work Pax. I have always thought of an FIL as basically a spread gain method, except that assumptions changes and amendments create bases and gain/losses do not. Therefore, I would take a simplistic solution and just take the difference of the AL's before and after the change. This assumes that Plan's UAL will be > 0 and the Plan is not in full funding. I have never thought of trying to create some hypothetical asset to calc a UAL, but that certainly doesn't mean your solution isn't without merrit, I just wouldn't have done it that way.

Regarding the lump sum question, I could probably argue both sides. Do you assume a lump sum payment in your funding assumptions? If not, then your method treats the payment form selection as a gain/loss, which doesn't belong in the FIL base. The theory would be that they elected the window, so you should value the window on the standard assumptions. If after they elect the window, they select a lump sum form of payment, then that piece is a loss and is not part of the base. On the other hand, if the lump sum is part of your assumption, than I think you can argue it should be part of the base. You could always add the assumption of a lump sum payment with your other window assumptions so that it becomes part of the base.

I just think you need to be reasonable and be able to defend either position.

[LIBOR]

Just a quick note to thank everyone for their time providing very helpful suggestions and insight - thanks again !