Coverage/Discrimination Testing for Profit Sharing Contribution?

[SRNPEBT]

Client offers a safe harbor 401(k) plan (3% nonelective) which has a discretionary profit sharing contribution with no allocation conditions. Client wants to add a 1,000 hour allocation condition to the profit sharing contribution. If done, it is expected that only 2 HCEs will receive the profit sharing contribution (client is a physician's office which employs 13 HCEs and 8 NHCEs, most of whom work only a few days a month due to hospital commitments). Since all employees receive the 3% SH nonelective contribution, I believe we pass coverage testing (even if none of the HCEs receive the discretionary profit sharing contribution, because they are at least getting some nonelective contribution). But, do I need to test the discretionary profit sharing contribution for discrimination under 401(a)(4) as well? If so, do I have any chance of passing if 2 HCEs will benefit and no NHCEs will?

[Lou S.]

I'm going to assume you are talking about adding this condition for the next plan year and not the current plan year.

But yes this sounds like a fairly standard cross tested design with the safe-harbor as a floor in your general test. If you want the HCEs to get more than 9% total allocation though, NHCEs are most likely going to need a larger contribution to pass gateway testing whether or not they work a 1000 hours.

[SRNPEBT]

Yes - the allocation condition would be added effective 1/1/16, so we are kicking around options now. Historically the profit sharing contribution has been pretty generous, so I'll have to look into whether they are willing to cut it down. Otherwise, my choices are to make sure all NHCEs are getting at least 5%, or make sure the profit shariing contribution is no more than 3x whatever the NHCEs are getting, right?

[John Feldt ERPA CPC QPA]

Run the 401(a)(4) test. Just giving 3% or 5% is not a guarantee of passing the test.

If you want any HCE to get more than 9%, you'll have to give more (profit sharing gateway) to the all the NHCEs (except those you can carve out under the OEE rule, if any) in order to satisfy the minimum gateway, the lesser of 1/3 the highest HCE rate or 5%. But 401(a)(4) might require you to allocate more than that, depending on your demographics.

[Tom Poje]

not sure why they wouldn't put everyone in their own group with no hours requirement.

I guess conceivably you could give 1 NHCE the same as 2 HCEs and test on an allocation basis and avoid the gateway which in effect is the same as passing

the gateway on a 'broadly available' basis because 1/8 for NHCE and 2/13 for HCE is over 80%

but that type of situation lends itself to possible abuse.

[austin3515]

Does someone in your office have expertise in these designs? Your getting great advice from people in the know, but this is the kind of thing where the finer details can get you into a lot of trouble. If you want to be doing these designs I would strongly recommend getting at least your QKA from ASPPA. Not trying to be rude, I honestly have the best intentions, which is to not see you get burned.

[SRNPEBT]

Thanks for the feedback. I'm not in the TPA business so the granularity of coverage testing isn't something I need to dig into often - happy when I can punt to someone more in the know. In this case, it's a small plan that hasn't historically required any testing and as such is administered by the client's accountant. I wanted to confirm my suspicion that this type of design tinkering would require testing that is far beyond what their accountant can handle, and now I can make the argument that they will need to move to a more sophisticated TPA if they want to achieve their new design objectives.

[austin3515]

Now we're talking! Send me a PM if you would like to me propose!

[Bird]

is administered by the client's accountant

uh-oh.

now I can make the argument that they will need to move to a more sophisticated TPA if they want to achieve their new design objectives

Good thinking. Just be prepared for that TPA to find a bunch of stuff that was done wrong.