Defined Benefit Minimum Required Distribution calculations for plans w

[Guest Victoria Pelletiere]

How are other administrators calculating the minimum required distributions under the new regulations for DB plans that allow for single sum distributions? In the past we have treated the benefit as an account balance and divided the PV by the life expectancy. The new regulations do not allow for this option. In addition, I read something that indicates a single sum is not available to the individual after he retires (i.e. individual reaches 70-1/2, continues to work, continues to accrue, retires at age 75). I wanted to use the new rules for those attaining 70-1/2 in 2002 so that there is not a change next year. Also, I question the transition from the old rules where we used the account balance to the new rules where we cannot use the account balance.

[AndyH]

I don't see a realistic alternative to using the old rules (or forcing a retirement election) until this mess is straightened out.

[David]

The DB part of the new RMD regs. were temporary and they have been delayed so I believe the old rules still apply. There is a recent thread on this subject on the Distributions message board.

http://benefitslink.com/boards/index.php?showtopic=17072

[RTK]

The temporary regulations have a specific provision at 1.401(a)(9)-6T, Q&A 1(d), on calculating how much of a single sum distribution made from a db plan is the required minimum distribution. One of the methods generally permits the single sum distribution to be treated as an individual account and the required distribution calculated accordingly. The other generally treats one year of annuity payments as the required distribution.

[David]

RTK, that Q&A addresses the question of how much of a lump sum cashout of a participant's entire accrued benefit needs to be treated as a RMD (and therefore not eligible for rollover). If I understand Victoria's question correctly, she is talking about using the account balance method to calculate the RMD for an individual that is not cashing out their entire accrued benefit, in other words, an ongoing participant. The tempoary regs, which are being delayed, removed that method as an option for an ongoing participant.

[RTK]

David, you are right. I misread her question. Thanks for putting me on the right track.

[MGB]

The IRS's position is that you never could use the DC method in this situation. The new proposed/temporary version did not change that.

The DC method was in the 1987 proposed regulations for determining if distributions that were not in an annuity form met the required minimum. They felt that there really weren't any situations where this was needed and dropped that language in the new version.

I assume the structure of the plan is to allow an annuity payment or a lump sum. Unless there is some other strange payout option described in the plan (and the participant elected that option), the DC approach was not available, even under the 1987 proposed regulations.

The proposed/temporary regulation is still effective 1/1/03. Although the service has said they will delay, it hasn't officially happened yet.

[AndyH]

Thanks for the clarification, MGB.

Now I have to add that this lunacy is another reason why I will never refer to the IRS as "the service".

[David]

MGB, you said that the IRS position is that you could not use the account balance method "in this situation". Is there a situation that the account balance method could be used in a DB plan, other than in the year of a total lump sum distribution?

[MGB]

It takes extreme imagination to come up with a form of payment that is not an annuity distribution (even a lump sum is considered an annuity distribution that is very short in duration).

Let's say that a form of distribution is defined as: 1000/mo for 3 years (regardless of the value of the accrued benefit), or for fewer years (the payments only go for as long as they are equivalent in value, this form of payment does not add value), then a lump sum of 50,000, then an amount for 3 years that causes the entire stream of payments to be the actuarial equivalent of the accrued benefit.

If this is a considered a non-annuity form of distribution (I can see an argument that it is still an annuity, but just varies in payment). Under the 1987 regulations, you would apply the DC rules to see if the 1000/mo. is enough. However, given that no one has this strange type of form of payment, the service decided it didn't need the DC rule at all.

The DC method does not apply in the year of a total distribution, other than to determine how much cannot be rolled over. It has nothing to do with determining how much needs to be distributed from the DB plan.

[David]

I don't have it in front of me right now, but I seem to remember the preamble to the temp. RMD regs explicitly stating that they were removing the account balance methodology for DB plans. That would seem to indicate that it did exist. I'm just trying to understand this issue.

[MGB]

Yes, it did exist. But it could only be applied to non-annuity forms of distribution, according to the 1987 proposed regulations.

I can imagine a plan stating that the amount of payment each year is the amount that is calculated as the minimum required, using the DC rules. This would have to be an explicit description of a form of payment that the participant elected in lieu of the normal and J&S forms. In this case, I could see it argued that it is not an annuity form, and that the DC rules would apply. However, if the only forms of distribution stated in the plan are annuities (including a lump sum), then the DC rules don't apply.

[David]

I have always thought of an annuity as a stream of payments which would make a lump sum a non-annuity, but then again, the IRS has never called me to ask how I define anything.

[MGB]

Ever look at 417(e)? The language is "nondecreasing annuities" are exempt from 417(e) rules. A lump sum is considered a decreasing annuity. It starts at a certain amount (the lump sum) and then decreases to zero.

Even if a lump sum were a nonannuity, that doesn't change anything. If you pay out the lump sum you have satisfied the RMD no matter how you calculate it.

The argument is whether you can use the "VALUE" of the lump sum (without actually paying it) to determine the RMD. That is what is not allowed unless the actual form of distribution is a nonannuity. They aren't actually paying the lump sum so whether or not it is an annuity is irrelevant.

[Blinky the 3-eyed Fish]

Anyone else confused?

[Guest Victoria Pelletiere]

I am! How about these examples for plans that allow LSD at termination?

#1. A/B = 40.00; PVAB = 4000; 100% vested; still employed; elected to receive MRDs; Is the MRD 480.00 per year? At actual retirement, can the plan cash out as a LSD? If so, is it 40 valued at the actual retirement date or a LSD reduced for the payments received to date (like an account balance)?

#2. A/B = 6,000; PVAB = 600,000; 100% owner; MRD = 72000 per year (if elected) or J/S equivalent? election made before RBD with spouse's consent if SLA; can the participant elect a LSD at actual retirement? If so, how is the initial election changed?

My biggest concern is that the participant be allowed to receive a LSD at actual retirement.

[MGB]

What do you mean by "elected to receive MRDs?"

Is this a stated optional form of payment in the plan? If not, what is it they are electing?

This is the whole issue that the IRS doesn't like.

[Guest Victoria Pelletiere]

The plan allows for non-owners to either defer distribution until actual retirement or start receiving MRDs at age 70-1/2. The individual elected to start receiving MRDs at 70-1/2.

[David MacLennan]

I came upon this old thread, and MGB's explanation of the proper use of the individual account method under the old regs was eye-opening. No matter how much you think you know this stuff, something always surprises you.

What if the plan document allows for partial lump-sums, and the distribution forms allow the participant to elect partial lump-sum annual payments determined under the individual account rules in 1987 regs. Would this be a proper use of the individual account rule for DB plans?

Also, assuming the answer to the above question is yes, can one use the new tables in the 2002 regs? Or, to be considered in compliance with the 1987 proposed regs, must you use the old tables (I'm thinking of the transitional relief mentioned in Notice 2003-2).