Failing Avg Benefits Test - Return 401(k)?

[Guest gopher2378]

3% safe harbor cross tested plan going along nicely for 5 years. Dad (55) owns business and he and his CFO (59) hit 415 maximum each year while one other HCE only receives 3% safe harbor. Along comes 25 year old son of the owner in 2007 and defers 15% of pay in 1st year of eligibility. Now with only the 3% safe harbor contribution allocated the average benefits test fails miserably since sons' EBAR is 59.182%. No way I can allocate any additional profit sharing to the Dad and the CFO this year, UNLESSS, I can return 100% of the 401(k) to the son. I can't find any authority to make this corrective distribution.

Question #1 - can I return 401(k) contributions to an HCE due to failure of 410(b)?

Question #2 - if not, any other ideas?

Thanks!

[Tom Poje]

you must have done a good search, because you can not return deferrals once made.

you did not rpovide enough other info to determine what else the plan could do.

for example what is the avg ben % test when done on an allocation basis? will that pass?

once the kid is in assuming he is the youngest and received the SHNEC then I think you are left to test on a component basis - that is, the son and the older NHCEs tested on an alloaction basis, and the older HCEs tested with the younger NHCEs on an accrual basis.

if all you are doing is providing the 3% shnec to the nhces, then I guss the assumption is that the HCEs are capped at 9%?

[Guest gopher2378]

Due to a downturn in their business the Dad and the CFO will not be looking to hit 415 maximums this year so as a starting point I have them capped at 9%. Can you point me to a link or resource that discusses component testing rather than have you explain it here which I'm sure you have done more than you would like. The majority of our plans are "super integrated" but they are all done the same way - all NHCEs receive the same contribution and then each HCE is his or her own group and receives extra profit sharing as declared by the Board of Directors of the company each year. Does this component testing have to be written in the document?

Thanks for your help Tom, you are the super integrated guru!

[Tom Poje]

component plan testing is simply an option, so it doesn't have to be in the document.

Treas. Reg. §1.401(a)(4)-9©(1) describes it as Plan Restructuring

basic rules are as folows:

1. plan passes 410(b) before restructuring

2. split plan into 2 (or more) component plans

a. each individual must be in one and only one plan (though technically they show up in the other plan as a big fat zippo, zero nada, nothing, number of eyes blinky has if he lost 3 of them.

b. each component plan satisfies 410(b)

c. each component plan passes 401(a)(4) - usually you test the young HCE and old NHCEs on an allocation basis and the old HCEs and young NHCEs on an accrual basis. an ee 'not' in a component plan still shows up in that plan as a big fat zippo, nada, zero, nothing

d. if the plan as a whole passes avg ben % test, then the component plan passes avg ben % test.

hope that includes the basics.

if using Relius, you can put people into into divisions and then test the divisions (but fill in the # of HCEs and NHCEs not benefiting in the little grid or whatever you call that data entry screen)

[Guest gopher2378]

Thanks Tom. Off I go to research this a little more and to get down and dirty with Relius.

[Tom Poje]

good luck. I have actually run it that way on Relius.

It is almost like you want an initial 'dummy' run that shows you total number of NHCEs and HCEs, and then when you run component plans the number of total NHCE and HCEs is the same, even though you only see the names of fewer people on the report. The avg ben % test is the tricky part, but they have added a toggle on that to indicate 'avg ben % test passes' (assuming that it does. or if you get lucky, you wont need it and can pass ratio % test for each component plan.

[Mike Preston]

One thing that is frequently overlooked when restructuring is that unless the individual component plans satisfy the reasonable classification test, you are "stuck" using the 70% ratio percentage on each and every component plan. And since you are randomly (or not so randomly, but certainly not predicated on reasonable classifications) moving people from one component plan to another, you almost never satisfy the reasonable classification test.

Now, this usually isn't a problem because most of the time you can find wildcard NHCE's that can be put into a specific plan (and removed from another one) until the 70% level is satisfied in all components. Think of it as a big chess game and you move the pieces around the board until it all just works.

[Guest gopher2378]

I'm so confused. I've spent the last 2 hours with my face in Sal's ERISA outline book and the reg's trying to understand how to go about this, but I still keep coming back to the fact that the plan as a whole (without component plan's) does not pass the Avg Benefits Test due to the son's EBAR being so high. Is there anyway to make the component plan's work if the plan as a whole fails average benefits?

[Mike Preston]

Usually, but not always. Gotta either post your numbers or a spreadsheet.

[PensionPro]

One thing that is frequently overlooked when restructuring is that unless the individual component plans satisfy the reasonable classification test, you are "stuck" using the 70% ratio percentage on each and every component plan. And since you are randomly (or not so randomly, but certainly not predicated on reasonable classifications) moving people from one component plan to another, you almost never satisfy the reasonable classification test.

Now, this usually isn't a problem because most of the time you can find wildcard NHCE's that can be put into a specific plan (and removed from another one) until the 70% level is satisfied in all components. Think of it as a big chess game and you move the pieces around the board until it all just works.

Let's say the plan passes the ABT on an aggregate basis (ABPT > 70%), and the component plans pass the 410(b) ratio percentage test. Since the reasonable classification test is part of the ABT, are we required to apply and pass the ABT to the individual component plans in order to use the lower midpoint for the rate group test? Or are we required to apply only the reasonable classification test to the component plans?

Thanks so much for helping us navigate these choppy waters!

[Mike Preston]

I'm not sure what you've presented is the correct either/or. You get a free pass on the rate group test vis-a-vis the reasonable classification test. You do not get that free pass on the 410(b) test. Hence, if we assume that the selection of the individuals in the component plans are random (they usually are) we find that the 410(b) test for each component plan must satisfy the 70% ratio because it matters not what the ABPT result is (that is, whether it is over or under 70% doesn't matter - you can't use it). When we then move to the rate group test, now we find that merely passing the ABPT is sufficient because you get a free pass on the reasonable classification test. Hence, if the ABPT is satisfied, you can use the midpoint when doing your rate group testing.

[John Feldt ERPA CPC QPA]

And, if I understand Tom correctly, you can test one rate group on a contributions basis, and another rate group can be tested on a benefit accrual basis.

[Bird]

And, if I understand Tom correctly, you can test one rate group on a contributions basis, and another rate group can be tested on a benefit accrual basis.

Well, one group of employees on each basis. I think someone needs to put up some numbers to make it less ethereal; I'll try to do something simple here...

let's say you were sailing along with 2 older owners and 7 NHCEs. One of the owners brings his son in and messes up the ABT. The son gets the same employer contribution as the NHCEs. The idea is that you carve this into two plans - one with the 2 owners and 5 NHCEs (it's arbitrary, but you pick the youngest ones) and the other with the son and the 2 other NHCEs. "Plan" 1's ratio percentage test is 5/7 divided by 2/3, or 107%, so that's ok. "Plan" 2's ratio percentage test is 2/7 divided by 1/3, or 86%, so that's ok. Then, as long as each rate group test within each "plan" is greater than 70%, you don't have to run the ABT. The second "plan" will be 100% because everyone is getting the same contribution and you're testing that on a contributions basis. I'm guessing that the first "plan" will be ok too, tested on a benefits basis.

I don't know how you get Relius to run the tests separately.

Edit - I just re-read the original post and noticed that there was only one NHCE in the case presented. So the above is a more general discussion but not all that helpful to the OP. I'm not quite sure if there's a solution, other than increasing contributions to the NHCE.

[Tom Poje]

Bird-

so on Relius you put people into 2 divisions -

the Cruel and the not so cruel.

I mean

accrual and allocation

the accrual group consists of the 2 HCEs and the youngest 5 NHCEs

the allocation group consists of owners kid (HCE) and the remaining 2NHCEs.

now, when you test the accrual group, you fill in # of nonbenefitting HCEs=1, NHCEs=2 and tell the system to include only the accrual division. this will treat the kid and the 2 NHCEs as big fat zeroes

now test the allocation group, but plug in 5 for the nonbenefitting NHCES and 2 for the nonbenefitting HCEs. test on an allocation basis and only that division.

by the way, if test still fails, you could swith one of the NHCE to a different division. I always run a 'fake' avg ben % test just to make sure I have the correct number of HCEs an NHCEs in total.

[Mike Preston]

Bird, where do you get the idea that there is only one NHCE?

[Bird]

Bird, where do you get the idea that there is only one NHCE?

I mis-read the first post...

"never mind."

[Mike Preston]

OK, I thought one of us might have misread the post. I just wanted to make sure it wasn't me. No big deal.

[buckaroo]

Just read this and had a thought. You said that the son is messing up the test in his first year of eligibility. Before you go through the work of splitting this into component plans, can he be excluded under the Otherwise Excludable Employee rules? If so, he will get the same allocation as other people in the OEE group. As long as it passes coverage, you get an auto pass on 401(a)(4) because they are all getting the same thing. Then the upper group would be cross tested and, if the pop has not changed very much from the prior year, odds are you should be able to pass.

[AndyH]

Just read this and had a thought. You said that the son is messing up the test in his first year of eligibility. Before you go through the work of splitting this into component plans, can he be excluded under the Otherwise Excludable Employee rules? If so, he will get the same allocation as other people in the OEE group. As long as it passes coverage, you get an auto pass on 401(a)(4) because they are all getting the same thing. Then the upper group would be cross tested and, if the pop has not changed very much from the prior year, odds are you should be able to pass.

The OEE group is not excluded, it is separately tested. How is it going to pass coverage and a(4) with one of the 3 HCEs? We don't have specific numbers here, but that is iffy at best.

[buckaroo]

I believe that I misspoke in my original post. I understand that the OEE group needs to tested for coverage and non-discrim, but that group is done so separately, without regard to the Statutory group. They are considered excludable employees and not in each others coverage group. See 1.410(b)-7©(3). Also 2007 ERISA Outline book, page 8.4. Since this is the case, from your comment, there are two HCEs in the Stat group and the coverage would be 2/2. There would be one HCE in the OEE group and his coverage would be 1/1. As long as 70% of the NHCEs in the OEE group are benefitting, then coverage for the OEE is passed. IF those same OEE people are getting a comp to comp allocation, then the (a)(4) test will pass and, essentially, there is no need for the test for the (a)(4) test. For the Stat Excludable group, I would think you would have a similar test to last year. Obviously there many be some differences, but nothing like a young HCE to muddy the waters.

I hope this helps clarify.

[AndyH]

Yes, thanks. Not every buckaroo out there knows this.

[buckaroo]

Glad I could help.