AndyH

1. Why not a 412(i)? (only kidding!) 2. Get rid of the MP plan 3. Real high maintenance and ego tending. 4. Looks like you have all the technical stuff considered. 5. Maybe include a 401(k) provision?

It is true only in very limited situations. For example, assume it is 2002 and you have two people, each in their own class, one earning $150,000 and the other earning $50,000. You give the guy earning $150,000 what happens to amount to 12.4738% of pay. He gets $18,710.70. The other person gets 10% or $5,000. If you test each on a contributions basis in 2002 and impute permitted disparity, each will have an allocation rate of 15.70%. Now add a third person to the mix and put him in either group. If he is in the higher group, his contribution will not be the same as in an integrated setting unless his comp is the same. But if he is in the second group, as long as he earns below the SS Wage base he will still get the same 10% he would have received in an integrated plan.

Wow is all I can say.

How can you get an integrated allocation if you have a discretionary class document with people in the same class at different comp levels, some above, and some below the integration level? Is that not impossible, or at least virtually impossible? Maybe this helps. If you give each class the same percentage of pay, the result is a safe harbor and you do not need to test. If instead you have two people, each in their own class, it should be possible to reach the same allocation result as if you had a safe harbor integrated plan. Then do you need to test? I suppose you might technically, but it would be a no-brainer to pass by simply imputing permitted disparity in the test (and testing on a contributions or allocations basis, not cross-testing).

Sure, that is true. But I'm not sure that was the question. The question may have been, if the formula in the document fails, can they simply revert to a safe harbor formula, i.e. ignore the document. No is the answer to that question (unless of course the document specifies that as the correction for a failure). Or, if there are two people above the SS Wage base and in the same allocation class, and the allocations are discretionary, it may be possible to achieve the same result as if the plan was an integrated safe harbor. So, I was trying to more clearly identify the question.

Yeah, agreed. My suggestion makes no sense unless either the DB accruals or the DC is tiered to begin with.

Thanks, this is all helpful. In the one I'm looking at, which is a proposal which supposedly duplicates another existing plan's design, the offset is restricted to .50%, resulting in no benefit for the rank and file, but an equal 4% for the chosen few. Does anything smell wrong with that? Actually, it smells wrong;I just can't figure out what if anything is being violated. Blinky, wouldn't that make yours more efficient, limiting the offset to the gross benefit of the staff?

Got it. Thank you!

Merlin, I remember a post here about Schultz' comments, but I haven't been able to find it on a search. If you were part of that discussion, do you recall the approximate time frame?

Right, that gets to the two questions, 1. What is meaningful, and 2. What is meant by "reasonable and uniform basis" for the offsetting allocation? Does this mean that the DC allocations must be a "uniform" percent of pay?

The question is vague. What does the document say? Does the document allow you to produce allocations that are identical to that of an integrated safe harbor plan, or does it not? If it does, then you can do it and probably be done; if it does not, then that is another set of facts.

General tested Floor Offset arrangement provides that Class A participants (almost all HCEs) get a 4.50% x Years of Participation DB accrual and rank and file get .50% x Years of Participation. DB benefits are offset by a large (around 9%) DC contribution. The maximum pension offset is .50% of pay, which in effect means that Class A people get 4%, others 0%. DB plan also provides for a minimum accrual of $40 per year (not per month) x years of participation, maximum 10 years, so a very small pension is paid to all participants. Will this pass 401(a)(26) if Class A is less than 40% of the population, and not more than 50 people?

Annuity rates are comprised of an interest rate and a mortality table. They would change if either were changed. It depends upon what you are using them for. For cross-testing, the regulations say you can use from 7.50% to 8.50% and one of several "standard" mortality tables. I don't know of sample rates being "published", but they can be generated from virtually any retirement plan administration software system.

What is the best way to get an answer to something like this, a PLR or is this an FDL issue?

tonjer, if the plan is general tested under 401(a)(4), there is no need for the 414(s) test, because the general test must use comp that complies with 414(s). So this is an alternative to 414(s) testing. And filing requirements, or lack thereof, are essentially the same.

Kirk, your point is a very good one, and yes, we are concerned about that, and don't know how to get beyond it. Any elaboration would be welcome. The regulation raises the standard of whether there is a "Legitimate business purpose" for the treatment of the imputed compensation. In my situation, yes, there absolutely is. But the fact is it is being paid by an unrelated employer or employers.

Everything Blinky said is right, but I'd just add that you cannot arbitrarily decide not to give a gateway contribution to certain people. You must of course follow the document. I mention this because of the amendment that you referenced; what that amendment stipulates is unclear.

Everett's cite seems to indicate that this can be done as an alternative definition of compensation, subject perhaps to 414(s) testing of some sort, but the fine print is not easy reading.

Employer with final average pay (5 year) DB plan allows employees to work for unrelated employer for one year as an "externship". Employee is paid 1/2 salary by employer and 1/2 salary by temporary employer. Employer is concerned that "externship" will reduce final average compensation in some cases, would like to credit compensation at full rate in effect prior to "externship", or at full rate counting both 1/2's. The plan is integrated and it is preferred that the plan remain a safe harbor. Can this be done as a safe harbor? If not, can it be done outside of safe harbor rules, and if so, what is tested? The program would cover almost exclusively NHCEs, but there is a possibility that a HCEs might qualify. Thanks for any thoughts.

Don't forget the DB/DC combo gateways that may apply 1/1/02 if you are testing them together.

Thanks, Merlin. Right on point.

Not that Tom needs backup, but Jim Holland was adamant. No comp means not in the test. Not a $0. Not a 0%. Not in any test.

Does a regulation under 1.401(k) have anything to do with a regulation under 1.401(a)(4) if it is not referenced by 1.401(a)(4)?

It is subtle. Look at 1.401(a)(4)-2©(3)(ii). "(ii) Application of Nondiscriminatory Classification Test" A rate group satisfies the nondiscriminatory classification test of section 1.410(B)-4 (including the reasonable classification requirement of 1.410(B)-4(B)) if and only if the ratio percentage is greater than or equal to the lesser of [the midpoint or the r/p of the plan]. The key is the section in parentheses. The DB general test section 1.401(a)(4)(3)©(2) says that "the same rules apply as in section 1.401(a)(4)-2©(3)" You may be right about Benefits, Rights, and Features, however, as I don't see the same "free pass" past the reasonable classification. So I guess your allocation groups don't have to be reasonable but your BRFs do? Seems weird.

Sorry, but I still completely disagree. Where's Blinky when you need him? Tom, you on vacation? Calling EF Hutton... I'll explain when I get a minute if nobody beats me to it.