Present Value of AB using Segment Rates

[JAY21]

2008 distribution. Is the present value for someone say 12 years out from NRA equal to the product of (a) annnuity factor at NRA (based upon the 1st segment rate) times (b) the discount factor (12 years out would be based on the 2nd segment rate) times © the accrued benefit ? Assume no pre-NRA mortality discount.

Or is it more complicated than that ?

[Guest SVP]

the PPA lump sum for someone 12 years away from NRA/BCD would have no payments being made in the first segment so the segment 1 rate is not used in the calculation. Payments start in the 2nd segment so the segment 2 rate is used for the payments that fall in the 2nd segment. The 2nd segment rate is the rate used to discount those payments to the current age. For payments expected in years 20 and beyond you use the 3rd segment rate for discounting payments to current age. That's my understanding.

So nothing at the first segment rate, an annuity payable for an 8 year temporary period deferred 12 years at the second segment, and a lifetime annuity deferred 20 years at the third segment rate.

[JAY21]

Glad I asked. I guess that makes sense. Did I translate the approach correctly on the attachment ?

PVAB.pdf

[Mike Preston]

Looks right. Annual benefits are a bit unusual, though.

[JAY21]

Agreed.

[zimbo]

Looks right. Annual benefits are a bit unusual, though.

I am struggling with a lump sum distribution methodology when calculated in partial years. I have a terminee who is age 47 and 7 months at the assumed 2/1/08 distribution date. The plan's NRA is 65. Plan does not discount pre-ret mort in PVAB. Do I use some sort of interpolated Ns when calculating my deferred temporary and deferred life annuities at the 2nd and 3rd segment rates, or can I simply interpolate the lump sums determined at whole ages 47 and 48, or in this new PPA world do I only use whole ages (nearest age or attained age)?

[tymesup]

I've been taking the brute force approach and laying out all the payments from attained age to the end of the mortality table. 12 payments a year times 120 years = 1440 rows of Excel later, I've got results.

In practice, I don't see how the feds can complain if you use N's and interpolate - they didn't in the past. There's another thread where we've come up with multiple values for some simple situations:

http://benefitslink.com/boards/index.php?showtopic=37263

[Andy the Actuary]

As we focus in on the absoluteness of the calculatin, we should keep in mind that a lump sum is a proxy (at least it once was) for an annuity premium and that you could be applying a stale interest rate that doesn't relate to present economic conditions (e.g., use August '07 rate to determine lump sum payable December 1, 2008). Unless the IRS delineates precise lump sum calculation methodology (which hopefully they won't), it would seem that any reasonable method applied consistently would be acceptable.

[JAY21]

Is there any written guidance for this approach that SVP and others here have stated is the correct approach on the 417(e) PVAB ? I'm not doubting it just wondering where we got it from (verbal ??? vs. written guidance), or is it just a "of course that's how you do it stupid" and obvious to everyone but me. It seems a fairly laborious approach as has been mentioned in other threads.

[david rigby]

Unless the IRS delineates precise lump sum calculation methodology (which hopefully they won't), it would seem that any reasonable method applied consistently would be acceptable.

Yes. Let's remember that we are entitled to use generally accepted actuarial methods.

[Guest mingblue]

Using the example given by JAY21 and expanded on by SVP would the answer be the same in a funding rather than current distribution context ? Or would it be the value , discounted back to the val date using the current segment 2 rate, of the following amount valued at NRA : a 5 year temporary annuity starting at NRA and using a projected segment 1 rate plus a 15 year temporary annuity deferred 5 years from NRA using a projected segment 2 rate plus a life annuity deferred 20 years from NRA using a projected segment 3 rate ?

[Mike Preston]

<br />Using the example given by JAY21 and expanded on by SVP would the answer be the same in a funding rather than current distribution context ? Or would it be the value , discounted back to the val date using the current segment 2 rate, of the following amount valued at NRA : a 5 year temporary annuity starting at NRA and using a projected segment 1 rate plus a 15 year temporary annuity deferred 5 years from NRA using a projected segment 2 rate plus a life annuity deferred 20 years from NRA using a projected segment 3 rate ?<br />

<br /><br /><br />

Ming, what you have described is the present value, today, of a single payment expected to be made at retirement. You have calculated the expected payment that is to be made at retirement the same way that I would calculate a lump sum to be paid on that date.

Hence, if your valuation assumption is that 100% of participants will elect a lump sum at retirement age, your description is spot on.

But if the actuarial assumption inherent in your funding calculation regarding form of payment is that something other than lump sums will always be elected, then you would value the expected payments (whatever they are) using the current segment rate structure.

I believe there is something in the regulations that were just issued that reference the projected segment rates of which you write. If I had time today I'd look up that reference to see if it mandates one thing or another. I don't. Maybe somebody else does.

[David MacLennan]

One minor note: the notation for a deferred annuity is not correct. 20|ax+8 is an annuity commencing at age x+8+20, not age x+8 (where a is an annuity due - please mentally add the dieresis, can't do it here).