Cross-Tested Plan - Does this example need to be tested?

[DLavigne]

A plan uses a volume submitter document and is set up as comparability with two groups, 1) one named individual and 2) everyone else who is eligible. The employer is a partnership and the person in group 1 is one of the partners (HCE & Key). He gets a 0% PS contribution and everyone else gets 11%. Does this scenario need to be tested for gateway and the classification/average benefits tests? It's as if the formula is salary ratio with one HCE not getting a contribution, but since the document is comparability, does it need to be tested anyway?

Would the answer be different if the partner was getting 11% too?

Thanks.

[movedon]

This seems like a trick question. If it's tested it will pass, so what's the problem?

Wait, let me guess - is there a nonkey getting a top heavy minimum? If so, and if the plan doesn't pass 410(b) without considering the "top heavy only" person/people as benefiting, then yeah, it needs to be general tested.

[DLavigne]

Well, in our situation, there is a lot more to the story but I thought I could ask the general question of whether the contribution has to be tested just because the document says the formula is comparability.

In this case, the plan has dual eligibility (1 YOS for deferrals & SHM and 2 YOS for PS), and the plan is top heavy. So, we very often have people who either get the TH minimum in the SHM or as an additional PS contribution. Since the HCEs who benefit get 11%, in order to satisfy gateway the people getting the 3% TH minimum all in PS would have to get an extra .67% to satisfy gateway IF we have to satisfy gateway. If we can say that the contribution is really salary ratio (with the one HCE exception) then we don't need to look at gateway. At least that's what I'm hoping.

[movedon]

I don't think it matters whether the plan is new comp or not - if everyone is getting the same percentage of comp as a ps contribution, then the plan satisfies the design-based safe harbor (and would satisfy the general test, anyway). If people are not all getting the same percentage of comp as a ps contribution (and not satisfying some other design-based safe harbor), then you may need to general test. If you have a few people getting a top heavy minimum ps contribution and everyone else getting a regular ps contribution of 11%, you should start with a coverage test considering the top-heavy-only-people as "not benefiting." If that passes (and all the top-heavy-only-people are NHCEs), then you can stop there. If it fails, then you need to general test the ps contribution. That would be true even if the normal ps contribution was a straight pro rata contribution.

[Mike Preston]

I disagree that the plan, as written, satisfies the design-based safe harbor. However, I agree that in this instance the plan can be proven to pass the contributions based general test with a sentence or two and doesn't require full-blown "test".

Keep in mind that if the plan passes on the basis of contributions, there is no cross-testing and therefore no gateway.

[movedon]

...I disagree that the plan, as written, satisfies the design-based safe harbor...

True, what I should have said was that like a design-based safe harbor, if everyone's getting the same percentage, there's nothing to general test. Well, there's a general test, but it's an allocation-based general test that you can do in your head.

[chc93]

Just as a note... many years ago, we had a volume submitter plan document(s) from an attorney that had contribution allocation groups, and the document specifically required passing general test on a cross-tested benefits basis. One of the sponsors of one of the plans looked at providing a uniform percentage of comp, and in such a case, would have failed cross-testing on a benefits basis. However, it was a uniform safe-harbor, and so also passed general test on a contributions basis. The attorney amended the plans to provide that if the the contribution allocation was a uniform safe-harbor, cross-testing on a benefits basis was not necessary (and passed on a contributions basis).

[DLavigne]

I appreciate the replies but I'm still a little unclear. Are you saying that we would need to run the general test on a contributions basis, and if so, wouldn't we still need to pass gateway? Or are you saying we don't need to run the general test at all, and therefore, also no gateway?

[movedon]

I appreciate the replies but I'm still a little unclear. Are you saying that we would need to run the general test on a contributions basis, and if so, wouldn't we still need to pass gateway? Or are you saying we don't need to run the general test at all, and therefore, also no gateway?

Not sure who you're asking, but I'll offer my opinion. Regardless of whether or not the plan is new comp, if you have a year where some people are getting one rate of ps contributions (11%, say), and some others (nonkey NHCEs, I'm assuming) are getting a lower rate attributable to a top heavy minimum (3% or whatever), then I think you start with a coverage test that considers the "top heavy only" people as nonbenefiting. If that passes, then you're done. Technically, if the plan is new comp, depending on the language in the document, you may need to general test the 11% group (without considering the 3% group, since the 11% passed coverage without them), but that test will always pass on an allocation basis because everyone in the test is getting the same contribution. I think this is what Mike was getting at above, and it's what I was addressing in my third post.

In any event, again regardless of whether or not the plan is new comp, if the above coverage test fails and will only pass if you consider the 3% people as benefiting, then you'll need to general test the whole group (including the 11% folks and the 3% folks). Depending on the census, that might pass on a contribution basis (and no, there's no gateway if you're testing on a contribution basis). If not, you'd then move to cross-testing, which would require the gateway. If it gets to this point, you're going to be comparing the various failed tests looking for the one with the cheapest fix (gateway will require at least an additional 2/3% for the 3% folks). I realize the 3% group might not all actually be at 3% because of the match, but assuming they're all NHCEs, I don't think that affects the process.

EDIT - as chc93 said, you need to watch your document and make sure it doesn't tie your hands on any of this. If it does, then an amendment may be necessary.

[Mike Preston]

I appreciate the replies but I'm still a little unclear. Are you saying that we would need to run the general test on a contributions basis, and if so, wouldn't we still need to pass gateway? Or are you saying we don't need to run the general test at all, and therefore, also no gateway?

Because of the design of the plan, you need to "run" the general test. But as poster after poster has told you, the test will pass on a contributions basis if the people getting the bigger percentage of pay satisfy 410(b). Hence, "running" the test can be either you going through all the motions to get something from your system that says: "See, we pass 401(a)(4)" or you can just demonstrate mathematically that you have passed the test with a few sentences that describe the population and the benefit rates.

The gateway is keyed to whether you use benefits testing. Since you will pass using contributions testing, there is no gateway.

[Mike Preston]

A plan document that forces benefits testing is outside of standard of care in this industry in 2012. I won't speak to whether it was when it was drafted, as there were years where the IRS took the position that the 401(a)(4) test needed to be described in full in order to avoid a charge that the plan didn't provide definitely determinable benefits. The the IRS backed off of that position and any document drafter after that point in time which forced benefits testing is pretty close to, if not over, the malpractice line.