Zero entry date comp, but gets TH minimum

[Kevin C]

21 & 1 with semi-annual entry

3% Safe Harbor using entry date comp

New Comp PS allocation, entry date comp, HCE’s receiving >15% of comp.

Plan is Top Heavy

NHCE Participant A hired 3/1/2014 and worked over 1,000 hours in first 12 months. A's work schedule changed to “as needed” 4/1/2015. Employed on 7/1/2015, and entered the plan 7/1/2015. Still employed beyond 12/31/2015, but no compensation during 7/1/2015-12/31/2015. A has $6,000 of compensation in 2015, all of it prior to 7/1/2015 and receives $180 of PS as a TH minimum since she is a participant and employed on the last day of the plan year.

Anyone had this situation come up before? If so, what did you do for gateway and (a)(4) testing?

[BG5150]

TH + Gateway (at minimum). May need more if testing is not passing. 11(g) amendment if that's the case.

[Tom Poje]

haven't heard of something like that before

if it was an HCE and 0 comp on the whole year the IRS leans on saying don't include in test even though 'active'.

the person hasn't performed work for at least 6 months.

I could see an argument the person, while yes, "on call", has terminated since hasn't worked for at least months and is not eligible for top heavy.

what happens if they perform no service in 2016? would you still think they deserve a top heavy in 2015?

[Mike Preston]

Gets gateway, included for a4.

[Kevin C]

BG, the gateway is 5% of entry date comp. 5% of zero is zero. A receives $180, so A is receiving more than the gateway. My gateway testing issue is that our valuation system freaks out and lists A with zero allocation, zero comp, a zero % allocation rate and says the gateway test fails. $180 / 0 is infinity, which is well over .05 (5%).

Tom, That would make life easier, but the client insists A is still employed. The TH minimum goes to participants employed at year end, so we don't have any choice. No services in 2016 would affect the 2016 TH minimum if A is still employed at 12/31/2016, but I don't see it having any effect on the 2015 TH. If the full year comp was zero, I agree A wouldn't be in the test. Unfortunately, that isn't the case here.

[Kevin C]

Gets gateway, included for a4.

Mike, I agree. I think A is getting more than the gateway, but how do you treat A for a4? We are testing using entry date comp and for Participant A, it's zero. An infinite allocation rate, while mathematically correct, doesn't seem like the right way to test.

[Mike Preston]

Gateway is as gateway does, so you need to ensure the document doesn't use comp for entire year even though it could be drafted to use comp while a participant.

Infinity is, by definition in my mind, unreasonable. Torn between using zero in a4 testing and treating as excludable. Can ignore issue by using full year comp for testing.

[Kevin C]

Mike, thank you for the quick response.

The plan document specifically says to use entry date compensation for determining the gateway.

If we treat A as having a zero allocation, doesn't that create a gateway testing issue? That's what the system is doing and it says we fail gateway because a benefiting NHCE receives a zero allocation.

We tried giving A entry date comp of $3,600 so that her allocation rate would be 5% like most of the NHCEs. She is about the same age as the owners, so my hope was that she would not be in any of the rate groups. But, she ended up being in 2 of the 9 rate groups. We had several rate groups that were barely failing until we increased contributions to an NHCE, so I'm thinking excluding her may make it easier for those rate groups to pass. It seemed a bit unreasonable to exclude her if it looked like it would help the testing.

[Mike Preston]

I think you are comparing the wrong things. First, using zero in a4 testing leads to your software thinking you have a gateway problem. You don't. You fix that by fixing the software. Worst case you run the test on a population that completely excludes this person. Yes, I know you can't use that data for any other purpose. What you don't do is put a fake compensation into the existing data because as you can see, it doesn't fix the problem. Second, it isn't that it is easier to pass by excluding this person. That's not possible. If you included the person at the calculated ebar (infinity!) I promise the test would be easier to pass. So, by excluding him you are by definition being conservative (making it harder to pass).

[Tom Poje]

Kevin -

will plan pass if you use total comp in testing (even though gateway is based only on comp from DOP) in other words, give the person a minimal comp as date of participation just so it knows the gateway was satisfied, but test everyone on total comp.

[BG5150]

Person is getting 3% of total pay for TH. So, b/c this person is getting an ER nonelective contrib, he has to get the gateway, no?

Applying the gateway formula seems to say zero.

11(g) amendment to get them to 5%?

We are talking $180 here. Just give him the $180 buck. What's the worst that can happen. Plan gets audited and they tell you he got too much. Take it back. Or, if he took a distribution, write him a letter asking for it back.

[Kevin C]

We use Relius software. Fixing the software isn't a practical solution due to timing issues. The client wants their contribution number this week. Having A included in the testing with an infinite allocation would definitely make the testing much easier. They do not pass average benefits with A excluded.

This plan has 18 NHCE's including A. Four of those are mid-year entrants. They won't pass if we test using full year comp unless we increase contributions to several NHCEs. Besides, I have a feeling the IRS would frown on using one comp definition for gateway and a different one for the rest of the cross testing.

There's never been any question that A gets the $180 TH minimum. Yes, that triggers the gateway, but A's gateway amount is zero, so she is receiving more than the gateway. The plan uses individual rate groups, so we can give A any allocation amount we want. But, the issue is the same whether A gets $180 or $1,800. With zero entry date comp, it's an infinite allocation rate if A gets anything.

We're going to test with A excluded. Given the options, that seems the most reasonable. Thank you all for your input.

[Tom Poje]

bully for you.

I do like the fact you have a unique situation, no one really knows for sure how to handle, but it's the software that needs fixing to handle it. Though I'm surprised the client want the numbers this week

usually they want the numbers 2 weeks ago, even though they didn't give you the info until this week!

[Lou S.]

Someone divided by zero. Oh NOES!

As others have said, interesting situation.

If you test for a4 using his full year compensation and actual 3% TH minimum does the plan pass testing. I mean other than the unusual situation that says he's not getting the gateway when clearly he is getting 5% of $0 = $0.

If the answer is yes than I think you are done. If the answer is no it's not as clear. Has some of the $6,000 been earned as a participant up to $3,600 you would have satisfied Gateway and TH with the 3%. I'm not saying this is correct but I think a reasonable approach might be to consider $2,400 earned before he was a participant and $3,600 earned after he was a participant. This would allow you to have compensation for a4 testing that satisfies both TH minimum and Gateway. Again not sure if this the best approach but it's an idea I throw out there as a possible reasonable compromise. This being under the theory that this creates the smallest EBAR for the participant that satisfies both TH & Gateway give his 415 comp for the year.

[Kevin C]

This plan did not pass the general test with all NHCE's getting the gateway. It wasn't solely due to this participant, but, this participant did affect who we had to increase to get it to pass.

We tried what Lou describes before we tried testing with participant A excluded. Between the two methods, it works out slightly less expensive testing with A excluded.

[Mike Preston]

Are you SURE you fail the average benefits test? One option that I rarely see applied is to substitute the average EBAR for the annual EBAR where the average can be for the year of testing and the prior year or the year of testing and the prior two years (a 2 year average or a 3 year average). There is disagreement amongst practitioners as to whether this averaging can be applied on a participant by participant basis (for example, use 2 year average for one participant, 3 year average for another participant and no averaging [1 year average?] for all others). I am of the strong opinion that it can be applied on a participant by participant basis. If folks disagree then the conservative choice in their opinion would be to apply the averaging to the extent possible (certainly somebody who just entered the plan would have a "1 year average" in any event.

I know, I know. Your software doesn't have the ability to do this easily. Welcome to my world.

[Belgarath]

" I am of the strong opinion that it can be applied on a participant by participant basis."

I agree. And if you are interested, Derrin Watson agrees - I only know that because I watched a webinar where he said so!

And no, our software doesn't handle it either.

[Mike Preston]

I am very interested. Thanks.