AndyH

No, I don't agree. Testing under 401(a)(4) gets a free pass past the reasonable classification requirement.

I don't know if there is a right or wrong answer, but I'd be inclined to separately calculate the percentages as permitted under 1.401(a)(4)-9 and add the results together, even though they have different testing ages, provided it wouldn't affect PASS/FAIL. How can somebody prove that to be "wrong"? Or I suppose you could normalize the older one to age 65 instead.

Fred, if you have a plan with a discretionary contribution, deferrals, and a match, for (a)(4) and 410(B) you have three entirely separate plans. Just think of it as three physically different plans. You test all three separately because they are MANDATORILY disaggregated. You have separate testing options because you have separate tests for each separate plan. The only thing you would pull them together for would be the average benefits percentage test if needed because the regs say that you must include all plans that could be aggregated, with the specific condition that for this purpose you IGNORE the exception for k and m disaggregation, so for this purpose you treat them as if they COULD be aggregated. But again, ONLY for the ABPT, only if needed. And you normally don't need to do this test when you are testing the K or M "plans". You would only use ABPT then if you have a problem with the ratio percentage test. The ABPT is normally pertinent only when you are testing the discretionary contribution under the general 401(a)(4) test, i.e. cross testing or general testing on a contributions basis.

Correct.

Bingo, Merlin. But Fred may need some more help.

Right. The cross test includes the otherwise excludables OR it can be done by separately testing the otherwise excludables. Either way regardless of how ADP testing was done. What complicated my thoughts was coverage testing of the disaggregated deferral component. If you separately test otherwise excludables for ADP, does that require that you separately test excludables for 410(B) testing of the deferrals. I think so but I'm not 100% sure without looking it up since I don't do that often either.

Right (at least your last sentence). I'm not clear on your other sentences. Maybe you can try responding to Merlin directly and I'll see if I agree.

If they don't get a contribution, then they are not benefiting, and if they are not benefiting, they don't need to get the gateway.

Mike, I don't understand your second. Isn't that just saying that if aggregation makes a zero not a zero, then it is not a zero? Why is that a revelation? And whether or not the IRS is "correct", is it not at least reason enough for caution?.

So it sounds like you are testing the discretionary, which is mandatorily dissagregated from deferrals, so the ADP test has nothing to do with the test of the discretionary. So if you have a rate group that falls below 70% and you go to the ABPT, that is still part of the a(4) test for the discretionary, even though you must pull in deferrals. Again, the ADP test has nothing to do with the cross test, including the ABPT part, which is still part of the test of the discretionary. I suppose in the coverage test for the deferrals then you would need to be consistent, but certainly not the a(4) or 410(B) test of the discretionary.

Merlin, what are you testing under the average benefits test, deferrals? If you are testing a discretionary contribution, then that is a separate plan, and you have all options available to you.

PJaeger, if you lay our how many are in each contribution level I'll try to tell you how to restructure it to pass. Include anybody getting nothing that is not statutorily excludable as well, i.e. 0%, so you have three groups, 18%, 5%, and maybe 0%.

Q&A #3 here. I lied. It was 2001. http://www.abanet.org/jceb/2001/qa01irs.html

It was from a 2002 Q&A, ALI-ABA I think. I'm having trouble finding it at the moment but I will.

But you better pass coverage by the ratio percentage test, as the IRS has taken the position that a 0 rate group is tantamount to an unreasonable eligibility classification, making the average benefit percentage test unavailable for coverage testing.

Mike, yes, you are right (what a shock). I was expressing my conservative opinion on setting these up and should have specified that. I automatically set as a criteria the requirement that a client commit to funding it for at least 5 years, and this looked iffy on that criteria. But a DB could work at a scaled down deduction level with some careful planning and, more importantly, effective communication. And thanks, Dom. I know of only one company and there is no reference to 412(i) on their website. And yes, I'm sure there are many others. But it's funny how the larger companies have no references to them on their websites, either. That could of course simply be due to a lack of volume.

There is not a sustained revenue stream sufficient to justify a db plan of any type, IMO. This is a profit sharing plan candidate only.

Either way; one way or the other. But, dmb, I can't imagine a situation where you would want to cross test both an entire db and an entire dc plan. Unless you are restructuring into component plans, one usually works much better than the other, and it's usually on a benefit basis.

The answer either way is yes, except for top heavy purposes, where you have a minimum based upon 415 comp.

BNA 2/4/2003. I'm surprised that Benefits Link did not have a link to the comments, which were at the LA Benefits Conference. Maybe somebody here attended that? It was apparently a panel on "aggressive tax practices". Maybe somebody can find a story online.

My situation is a mutual insurance company. I assume that it would not fall under the SEC's jurisdiction, but insurance companies have their own funny FAS rules, so maybe they are towing this line as well. I can look at it from that angle. Thank you.

Certainly one point is that the IRS has recently stated (1/31/03) that they were giving "very high priority" to providing guidance on abusive 412(i) plans, and would "not be gentle" and also used the word "criminal" within this context. Other phrases used were "of paramount importance". They are considering issuing a warning to practitioners, then something more substantive to follow. Lori Hughes, unless you can answer Mike's 10 point plan, you might consider sticking to a traditional db plan. Although year 6 is still a concern, it sounds like there will be sufficient legitimate business reasons to terminate it at that time. Plus, investing in the stock market now may be the best "springing cash value" available within a 6 year time horizon!

MGB, are you saying, for example, that one of the big 4 (or whatever number it is) auditing firms can no longer can do FAS actuarial work? For example, we have a client that used to have their auditor do FAS#106 calcs while we did FAS#87 calcs. Now they've asked us to quote the FAS#106 work also. We're not sure why. Is this why? Can you point me to something in print? Thank you. And p.s., does this apply only to a publicly traded company?

Ok, this looks like a team effort; now I'll chip in. You said that both pass R/P. Great. But then you say that plan a must be cross tested since it's not a safe harbor. Wrong. It must be general tested, not necessarily cross tested. It could pass on a contributions basis. Assume it does not pass on a contributions basis. Then, yes, try cross testing. But in doing the cross test, you have the option of testing only plan A, or testing both plan A and B together (aggregating). You can try it either way. If you include only A, then you need not provide the gateway to B. But if you include both, both must get the gateway. And either way you try testing it, if each rate group doesn't have an R/P of at least 70%, then you must also satisfy the Average Benefits Percentage Test, which requires that the average EBAR (including deferrals and match for the ABPT only) or allocation percentage for the NHCEs at least equals 70% of the HCE Average. Nothing requires, for purposes of the Average Benefits Percentage Test, that the percentages for NHCEs to exceed HCE percentages, just that the average is at least 70%. See if that helps.

Thank you. Even more insightful than I expected. How about 10 being if the client understood the rating of the insurer and the risk that the insurer might become insolvent at some point. Somebody told me that the insurers offering these things tend to have lower ratings. Anybody know if that is true or not?