PPA Valuation of lump sums

[FAPInJax]

It was my understanding that IF I was valuing a lump sum for a participant that the following methodology would apply:

1 Calculate the lump sum at each age between attained age and normal retirement age using 417e assumptions

2 Multiply by the probability that the participant actually leaves at that age

3 Discount the result back to current age using funding target assumptions

Is this correct OR should the first step be that the lump sum is calculated using funding target assumptions as well??

Thanks for any and all commentary.

[Mike Preston]

Well, *I* think your steps are correct as described, but there is some doubt out there.

Also, I think it depends on what year (pre-2009 effective date of proposed regulations or after the effective date) and whether you think the proposed regulations establish a requirement to do it the way you have described as your modified step 1 or whether there is even a different (third) way to go about it.

There was an article published on the web which I've cited before, that can be found here:

http://www.jpmorgan.com/cm/Satellite?c=JPM...l_Page_Template

In it, there is a section that has the language:

"Proposed regulations issued in December require an

approach consistent with the first interpretation above."

and note that the "first interpretation above" is the utilization of a single yield

curve such that the lump sum payable 20 or more years from now would be

valued as an annuity commencing at retirement age using the 3rd segment

rate for all presumed annuity payments. Note that this description just assumes that the 417(e) rates are entirely irrelevant and presumes that you are choosing whether to value the lump sum at, effectively, the third segment rate of the funding assumptions or "anew" with all three segment rates (again, of the funding assumptions) applying again from retirement date (or, in your case, at each duration between now and retirement age).

This is entirely at odds with what Jim Holland said to me, on tape, at a session in October 2007 at ASPPA. He indicated that the methodology you cite is the one to use and that the determination as to what the 417(e) rates would be at each duration is an assumption that must be made by the actuary.

[Andy the Actuary]

Our nearly New Years Eve proposed regs provide, "In the case of a distribution that is subject to section 417(e)(3) and that is determined using the applicable interest rate and applicable mortality table under section 417(e)(3), the proposed regulations would provide that the computation of the present value of that distribution will be treated as having taken into account any difference in present value that results from the use of actuarial assumptions that are different from those prescribed by section 430(h) only if the present value of the distribution is determined by valuing the annuity that corresponds to the distribution using special actuarial assumptions. Under these special assumptions, for the period beginning with the annuity starting date, the current applicable mortality table under section 417(e)(3) is substituted for the mortality table under section 430(h)(3) that would otherwise apply. In addition, under these special actuarial assumptions, the valuation interest rates under section 430(h)(2) are used for all periods (as opposed to the interest rates under section 417(e)(3) which the plan uses to determine the amount of the benefit)."

The principal difference between this and what you've outlined is (as Mr. Preston cited), only the 417(e) applicable mortality table applies and not the 417(e) applicable interest rates. I read the proposed reg is value everything as if an (deferred) annuity will be paid but from the annuity start date, use 417(e) mortality.

We all need to send Congress a letter thanking them for purifying the process!

[FAPInJax]

OK. So the reading of that cute paragraph would calculate the lump sum at the ages between current and normal retirement age. These would have probabilities applied against them. The result would then be discounted to current age using funding interest rates but the applicable mortality table instead of the optional small plan table (for example)

[Andy the Actuary]

I would interpret. Pre-annuity start date, use funding mortality; post-annuity start date, use 417(e) mortality. Use funding segment rates in all cases just as if benefit were payable in installments as opposed to a lump sum.

The difference between this interpretation and what you just wrote is what mortality table is applied prior to the annuity start date. If you are valuing a small plan and assume no pre-retirement mortality, the point is moot.

Best Regards,

andy t.a.

[Mike Preston]

Assuming that Frank is valuing a plan that has a pre-retirement decrement for employment, it most likely also uses a pre-retirement decrement for mortality.

It is still unclear to me. I can parse that final subsentence to mean that the annuity subject to 417(e) is valued using the funding segment rates; not that the lump sum is determined using the funding segment rates.

Remember that 417(e) applies to any distribution which is expected to be less than 10 years (it might be 10 years or less, but that distinction doesn't matter to this line of reasoning).

So, a lump sum is a period certain annuity of one payment. A 5 year certain annuity is a period certain annuity of 60 payments. Is that language allowing us to determine the period certain annuity using our own estimate of what 417(e) rates will be and then, once we have the period certain annuity amounts we must then determine their present value for funding purposes using regular funding rules, substituting 417(e) mortality for the period between annuity starting date and assumed distribution date of the specific payment in the 417(e) stream?

Or is that something that nobody else sees in the language?

[Effen]

If my assumed form of payment is a lump sum, I think they plan MUST use 417(e) mortality and i MAY make adjustments to reflect differences between the current 417(e) rates and the 430 Segment Rates.

1.430(d)...

(B) Substitution of annuity form

(3) Optional phase-in of section 417(e)(3) segment interest rates. --In determining the present value of a distribution under this paragraph (f)(4)(iii)(B), a plan is permitted to make adjustments to reflect differences between the phase-in of the section 430(h)(2) segment rates under section 430(h)(2)(G) and the adjustments to the segment rates under section 417(e)(3)(D)(iii).

© Distributions subject to section 417(e)(3) using other assumptions. --In the case of a distribution that is subject to section 417(e)(3) but that is determined as the greater of the benefit determined using the applicable interest rate and the applicable mortality table under section 417(e)(3) and the benefit determined using some basis other than the section 417(e)(3) assumptions, for purposes of applying paragraph (f)(4)(ii)(B) of this section, the computation of present value must take into account the extent to which the present value of the distribution is greater than the present value determined using the rules of paragraph (f)(4)(iii)(B) of this section.

I spoke to the IRS about this and was told you also need to consider when you expect the lump sum to be paid. For example, if you have a small plan and the principle is expected to retire next year, you might want to make a larger adjustment to reflect the relatively large difference between the 417(e) rates and the 430 segment rates. However, if you don't expect anyone to retire for 5+ years, using the 430 segment rates, w/ 417(e) mortality, might be acceptable. All of this is Post decrement only. Pre decrement needs to use the 430 assumptions.

I have a different question, that I think Mike touched on, if I expect to pay a lump sum at t=20, would the expcted lump sum be based soley on the expected 3rd segment rate then discount it back to today using the 3rd segment or do I assume the current segments will be in effect 20 years from now and value the expected lump sum using all 3 segments, then discount it back using the 3rd segment? Our software uses the first approach, and I convinced myself it was correct, but this new world of segment rate funding still doesn't fit in my one interest rate mind.

[mwyatt]

Also wonder about that point. Conceptually, what the segment rates are trying to get at is the yield curve reflecting current interest rates paid on bonds with varying levels of maturity. Hence, for example, the current rate 30 years out isn't a reflection of what the short term rate would be 30 years from now, rather what a 30-year bond right now would be crediting. Given that in the lump sum world, you're trying to model what would be in effect at retirement, I wouldn't think that using the 3rd segment rate for someone with 20 or more years out is an accurate guess as to what you would be looking at that time. Not sure if I'm explaining this thinking too clearly, but I'd be more inclined to your second interpretation than the first (using 3rd segment rate for lump sums 20 or more years down the road to determine the amount of lump sum payout).

[Mike Preston]

It seems that we are all saying the same kind of thing. Both of your comments are addressed in the cite I posted earlier. It indicates that there is at least one source out there that thinks the proposed regs are written in such a way as to require what mwyatt accurately described as something he "wouldn't think .... is an accurate guess". And that same methodology is described by Effen as what his software uses and what he thought was "correct" (not, necessarily, in the theoretical sense, but in the sense as to what the rules require us to do).

It wouldn't be the first time that a regulation requires us to do something which is theoretically less than sound.

I'm still holding on, though, to the possibility that mwyatt's description is found effective enough to thwart the interpretation that is used by Effen's software.

I think the rules are written in a manner that it is intended to be either:

1) intentionally unclear and thus give the IRS time to consider what they will think is acceptable and to communicate same through informal influences such as sessions at professional conferences

or

2) intended to be clear as to the intention to give the actuary leeway in

determining the future values to be used in the valuation.

I'd rather it were the second, but the cynic in me thinks it is, in effect, the first.

[Guest Grant]

It seems that we are all saying the same kind of thing. Both of your comments are addressed in the cite I posted earlier. It indicates that there is at least one source out there that thinks the proposed regs are written in such a way as to require what mwyatt accurately described as something he "wouldn't think .... is an accurate guess". And that same methodology is described by Effen as what his software uses and what he thought was "correct" (not, necessarily, in the theoretical sense, but in the sense as to what the rules require us to do).

It wouldn't be the first time that a regulation requires us to do something which is theoretically less than sound.

I'm still holding on, though, to the possibility that mwyatt's description is found effective enough to thwart the interpretation that is used by Effen's software.

I think the rules are written in a manner that it is intended to be either:

1) intentionally unclear and thus give the IRS time to consider what they will think is acceptable and to communicate same through informal influences such as sessions at professional conferences

or

2) intended to be clear as to the intention to give the actuary leeway in

determining the future values to be used in the valuation.

I'd rather it were the second, but the cynic in me thinks it is, in effect, the first.

Mike, et al.

My reading of the 12/31/07 regs is that we can just pretend the lump sum is an annuity and value that annuity with the funding segment rates, period. oh, and just substitue the 417(e) mortality table post-event.

I think the IRS chose this method to simplify the debate about "restarting" the segment rates. All along, I thought we would have to make some assumption about future lump sum segment rates, calculate a lump sum at event, then discount that with funding segment rates. I think this is much simplified, albeit not very "purist".

[Mike Preston]

I understand that is a bit simpler than restarting the segment rates at each point when a distribution is contemplated. But it is also a significant reduction to the liability, at least in today's interest rate environment.

[mwyatt]

One other point on the substitution of the (unisex) 417 mortality tables for post-retirement v. sex distinct if you're contemplating lump sum payments. If you stick w/ the regular sex distinct table and have a predominately male population (or 100% in a small plan), going to be underfunded at the end due to the approximate 3 year setback; or in the situation of a one-female maxed out plan, calling for required contributions under IRC 430 that are in excess of what can be recouped due to the age set forward.

One "silver lining" of this market collapse: has taken a few overfunded plans I had in the last year out of that situation.