AndyH

pension222, I had the same question and reaction, but I've been convinced that what you suggest makes sense but does not comply with the regulations. I don't think that the regulations anticipated the 417(e) rate as being the most valuable accrual rate because interest rates were so high back then. If you have access to ASPA exam reading material (C4) or conference summaries you will see that it is not commonly (if at all) done your way.

Right, Merlin, except that I don't think that the plan needed to exist 12/31/2002. It appears to me that you could use an accrued to date measurement with a divisor of 6 if those prior 5 years "are taken into account" under the plan's benefit formula. But if you add the new DC plan then you have a consistency issue, because if you add one year DC and 6 years of DB and divide that by 6 you have a result that makes no sense, so unless you separately determine the rates as in 410(b), it would seem that one year is the only option. I don't know what you would do after one year of DC experience. Maybe you can argue that the effective date of the DC plan is a "fresh start" date and always have that period available??

Merlin, I don't see Blinky's answer, so here are my thoughts. I could be wrong, so I welcome comment. Under 1.401(a)(4)-8, the Equivalent accrual rate is defined as the normalized "increase in the participant's account balance during the measurement period, divided by the number of years in which the employee benefited under the plan during the measurement period". So, in the case of a new cross tested DC plan, the measurement period is clearly one year. But under the DB general testing rules, it seems clear that the number of years of service "as defined in the plan for purposes of applying the benefit formula" is acceptable, i.e. 6 years. But under the consistency rule of 1.401(a)(4)-9 (b)(2)(iv), it says that rates "must be determined in a consistent manner for all employees for the plan year. Thus, for example, the same measurement periods and interest rates must be used, and any available options must be applied consistently, if at all, for the entire DB/DC plan. Consequently, options that are not permitted to be used under section 1.401(a)(4)-8 in cross testing a a defined contribution plan or a defined benefit plan (such as measurement periods that include future periods, ......) may not be used in testing a DB/DC plan on either a benefits or contributions basis, because their use would inevitably result in inconsistent determinations under the defined contribution and defined benefit portions of the plan" So, for purposes of 401(a)(4), can you use a different measurement period for a db component of a db/dc aggregated plan? I can read the above as prohibiting it. In contrast, the 410(b) regulations allow the employee benefit percentages to be the sum of separately determined rates, but I don't see this option as being available under 401(a)(4). Again, this is just one person's interpretation. Dissenting views are welcome.

No problem, Merlin; I have a plan that does exactly that, because the salaried plan is richer.

pension222, because you did not interject a cross tested DC plan into this question, i.e. if this is only a DB plan, then I think you could test it as either $60/1/comp or $60/6/comp. This is because the measurement period is a little fuzzy in terms of whether it is years benefitting for 410(b) purposes, or years taken into account under the formula. It is the addition of a new cross tested DC plan (from a prior discussion) that I think limits you to the one year measurement period due to the consistency requirement, as I read the rules anyway.

I agree, but then again I almost always agree with Blinky, so that is not news. I've seen (too) many plans that have done it each way; I don't recall one being more prevalent than the other.

Well stated, but you also must be careful because there are subtle differences between applying the "tools" for one purpose versus another. For example, for nondiscrimination testing under 401(a)(4) you get a "free pass" past the requirement that the eligibility criteria be reasonable for purposes of satisfying the Nondiscriminatory Classification Test for nondiscrimination (401)(a)(4) purposes. For example, if your plan excludes people who are left handed then you will never satisfy the average benefits percentage test for COVERAGE purposes because you cannot get by the reasonability standard of the NCT part of the test. But if only 20% of NHCEs are left handed and you then cover 80% of the NHCEs, then you can pass coverage under the ratio/percentage test, and the average benefits percentage test would be available to you under 401(a)(4) testing because the reasonability criteria does not apply for purposes of the 401(a)(4) nondiscimination test. And there are other differences as well.

Right, the two independent requirements (coverage and nondiscimination) are being confused.

looks that way.

Lots of good questions. I don't have all the answers without researching them but I have a couple of answers and a couple of educated guesses. First, for 410(b) purposes, you are not applying the average benefits percentage test to the aggregated DB/DC plans. You apply the ratio/percentage test to the aggregated DB/DC plan. That result is 100%. That is how you pass coverage. You are applying for 401(a)(4) purposes, the general test to the aggregated plan, and yes, it sounds like you are right about the plan being primarily defined benefit in nature. Regarding death and disability benefits, I don't think those would be included in the MVAR calc but I would not state this with certainty without reading all the cross references provided in 1.401(a)(4)-(3) with respect to the MVAR. If the plan had a general lump sum upon retirement or termination, that would be included including the 417(e) calculation (this was controversial but I think it is now universally accepted), but if only upon death or disability, I don't think so. I believe that you would need to use a measurement period equal to the current plan year, and that you would need to include the full DB accrual into one year's calculation. I think this is true only because of the consistency requirement of 1.401(a)(4)-9(b)(iv). I could be wrong, but that is how I read it. But don't accept this without reading all the cross references. I have not run across your particular situation before. And, yes, you might try posting the MVAR and measurement date questions in the DB forum as well. If anybody out there interprets this stuff differently I'd like to hear it as well. Mike, are you on vacation?

Merlin, yes, you are right, if the eligibility criteria is reasonable and you can pass the NCT, and yes, your descriptions would seem ok to me. But Tom is also right of course; if there are no NHCEs in the PS plan, it won't pass anything by itself.

Just some initial comments/questions. 1. If you pass coverage separately, you can test separately under 401(a)(4) and if they are each safe harbors, then you are done. 2. If you aggregate for coverage, then you must aggregate for 401(a)(4), and then you have non-uniform benefits that would require the general test. And you may be subject to the DB/DC combo gateway rules. And I have to wonder how you get these eligiblity exclusions. If the criteria is not reasonable, for bonafide business reasons as defined in the 410(b) regulations, then you would be required to pass the ratio/percentage test for each plan for coverage. Do you pass that? I assume not since you mentioned the average benefits percentage test. I think you need to clarify these issues first before getting into the measurement period question.

I can understand KJohnson's IRS reviewer's comments about the need to limit coverage testing to the ratio percentage test if the plan has a failsafe. The average benefits percentage test is not unlike the general test in that there are infinite "additional testing options", interest rates, etc. so the question becomes when do you quit and say you failed. Per Larry Deutch, Ed Burrows says the general test never fails; the client just runs out of money.

It seems to me that he is electing to receive it (subject to whatever limits) or he is not electing to receive it now. I don't see any argument for a middle ground. Why would he make an election if he doesn't want some of it now, to lock in the low rates?

Magically well stated Merlin. Thanks for the clarification.

Sorry guys, Merlin this time, but I must again dissent on Merlin's second paragraph. I think that the allocation groups are completely irrelevant for purposes of testing the plan as a whole for 410(b). What matters is the eligibility criteria for the plan or aggregated plan. Where's Blinky?. We see eye to eye on this matter, and that's awful tough because he has 3 eyes!

Tom, I don't agree about the ratio percentage test needing to pass. I think this is a potential issue only if one of the people in their own class get a 0 contribution, in which case it could be arguably be the same as having an unreasonable eligibility provision, making the average benefits percentage test unavailable.

Tom, welcome back from vacation. Dug out yet? Two comments I'd like to make on the SSRA issue. First, Ed Burrows a couple of years ago told an actuary in my office that testing by using SSRA was a gray area. So that muddies the waters somewhat in my opinion. But also I recall in what I believe was the Conference Committee Report on Tax '86 where the statement was made, and I am paraphrazing from a bad memory " Congress believes that it is appropriate to encourage and allow the use of SSRA". It was a pretty clear comment about intent. I know that such comment and $2 might get you a cup of coffee, but it has stuck with me. If I still had that "Blue Book" I would quote it precisely. I always thought I'd go back to that if I needed it.

Well, Michele, based on how you described the opportunities, now I will volunteer. I actually tried to before but I had problems with the volunteer form on the website. Not a good excuse, just a good one that day. Thanks for your comments as well as the others. Good discussion.

I suspect his question was whether or not he has to give himself a notice. It kind of reminds me of a Meeting of a 1 man Board of Directors Resolution. Quite eventful. And our lawyers like to add to the beginning, "a meeting at which there was considerable discussion about .....". I love to find those resolutions in the file for one person operations that have meetings with themselves at which there is considerable discussion. That way I know I'm not the only one who talks to himself.

I think that it does. I'm just saying it might be tough to find a DB document with the necessary MP provisions.

Blinky, I don't think so. A rollover is an elective, voluntary action that includes the knowing forfeiture of any protected rights or features that were held under the prior plan in favor of whatever provisions the new plan has. A merger, restatement, or transfer is an involuntary action that requires that all such rights and options that are protected under 411(d)(6) be preserved.

Lisa, back to your original question, the answer is yes. But, in your extreme example you have a full ADP failure so you need to follow the document in terms of what the ramifications of that are. If the full deferral must be refunded, do you have a deferral that has a match to test? You need to look to the document for corrective action.

I've never seen this done, but I don't know any prohibition on it. I'd think, however, that you'd need a pretty complex document to account for all the MP account balance features (assuming you have balances). Since it would not be a voluntary rollover, you'd have to watch for 411(d)(6) issues. The document costs might outweigh the savings. How about transferring the money into a deferral only k plan instead?

Since you've had no responses I thought I'd add my two cents. You understand of course that this is the responsibility of the plan administrator, not the prior actuary? And a second point is that you should be sure that the first payment was in fact restricted. There seems to lots of room for interpretation about how recent the data must be to make that calculation. I've seen commentary, for example, about it being acceptable to use the most recently filed Schedule B. That could mean the data is quite old, so it is not impossible that it is restricted now but was not before based upon the data used in the prior calculation.