Gateway to otherwise excludable

[BG5150]

Having a brain cramp today.

I have a plan that has 3 month wait, entry is date of event.

Cross-tested 3% SH plan.  Do I have to give the otherwise excludable folks the gateway if I am testing them separately?  GW will be 5%.

[Tom Poje]

recall, if you are splitting the testing you sort of have 2 plans

if you are 'cross' testing you have to provide the gateway

now, if you test the otherwise excludables on an allocation basis then you would not need the gateway - it is its own 'plan'

[BG5150]

1 hour ago, Tom Poje said:

now, if you test the otherwise excludables on an allocation basis then you would not need the gateway - it is its own 'plan'

So, nonexcludables, 3% SH + 2% GW, cross tested.  Otherwise excludables, 3% SH, contribution tested.  OK?

[Tom Poje]

that sounds ok

[CuseFan]

that is absolutely OK

[BG5150]

Do I have to pass my cross-testing with the ratio test?  Or can I use the ABT?

[Tom Poje]

that is an incomplete question.

I assume you mean your plan has each person in their own group.

Therefore, for coverage you have, in effect, people classified by name, and therefore, yes you have to pass ratio percentage test - but that is for coverage only. for nondiscrim testing, whether by allocation basis or accrual basis no such rule exists. It was proposed a few years ago, but never passed.

Usually the otherwise excludable group has no HCEs so that is not a problem anyway, but I guess if you have some HCEs and you have more than 30% of the otherwise excludable NHCEs somehow not eligible for safe harbor you could fail, or as Gandalf would say "You shall not pass"

[BG5150]

That leads me to the next question. 

What if I am breaking my plan into component plans (restructuring).  Do I have to pass non-discrim testing using ratio test and not ABT?  Coverage passes ratio test.

I've asked this before. And I was told there is only one ABT for a plan, and therefore, if I'm arbitrarily creating two plans, with no business reason, I cannot use the ABT to pass non-discrim.  Again, coverage passes the 70% test.

[duckthing]

Well, there are really two ABT results: one for OEEs and one for non-excludables. Your component plans can each pass nondiscrimination testing at midpoint provided the ABT for the appropriate "plan" (OEE "plan" vs. non-excludable "plan") passes. Since you don't need to satisfy reasonable classification for nondiscrimination testing, it's not an issue here.

What you can't do -- and I think this is what you're getting at -- is run a separate ABT for each component plan.

[Tom Poje]

well, under 401(a)(4) you can pass nondiscrim by either ratio % test or avg ben test

 

component plan info is under 1.401(a)(4)-9(c)

1.401(a)(4)-9(c)(1) ...if each of the component plans of a plan satisfies all of the requirements of section 401(a)(4) and 410(b) as if it were a separate plan

 

1.401(a)(4)-9(c)(3)(ii) adds (paraphrased) "YOU CAN"T RESTRUCTURE TO AVOID THE GATEWAY MIN BY TESTING ON GROUP ON AN ALLOCATION BASIS"

1.401(a)(4)-9(c)(4)(I) is the spot the regs say there is one and only one avg ben pct test amongst the component plan (but you still could have 2 if you are testing otherwise excludables separately)

1.401(a)(4)-9(4)(ii) adds, (basically) even though you are splitting things into component plans you can't use this for purposes of satisfying 410(b) itself

[BG5150]

410(b) is NOT an issue with my component plans.  Plan 1 is at 70.59% and Plan 2 is at 109.8%  It's 401(a)4, I'm concerned about.

EVERYONE in both component plans are getting the 5% GW.

Component plan # 1 passes on a contributions basis.  The 1 HCE is getting 5% of pay, too.

The other plan has 3 HCE.  I am maximizing the owner, and truing to figure out how much the wife can get.  At, say, $40,000 rate group passes the ratio test.  Above that to about $45,000, it fails the ratio test,, but the ABT test, for the component plan passes.

Is this ok, or not? 

[duckthing]

Depends what you mean by "the ABT test, for the component plan passes". If you mean that the ABPT for the relevant (OEE vs. non-excludable) 'plan' passes and every 401(a)(4) rate group in both component plans passes at midpoint, then yes, you're all set.

[BG5150]

1 hour ago, duckthing said:

Depends what you mean by "the ABT test, for the component plan passes". If you mean that the ABPT for the relevant (OEE vs. non-excludable) 'plan' passes and every 401(a)(4) rate group in both component plans passes at midpoint, then yes, you're all set.

When I run the testing in Relius for the component plan only, these are the results for the 3 HCE:

Ratio test:  Pass, Pass, Fail

Nondiscriminatory Classification test:  Pass, Pass, Pass

ABPT: Pass, Pass, Pass

ABT:  Pass, Pass, Pass

410(b):  Pass, Pass, Pass (I know this one is irrelevant)

Again, 410(b) passes ont he ratio test.  My population is:

Total HCE: 4, NHCE 17

Plan 1:  1 HCE, 3 NHCE, 70.59%

Plan 2:  HCE 3 NHCE 14, 109.80%

[Tom Poje]

I'm unclear on your statement "for the component plan only"

since you should have 2 component plans (I guess you could have more but I have never seen anyone split into more than 2)

so step 1, run an avg ben pct test on all ees to determine if that test passes, since there is one test and it applies to both component plans

now, at least how I have been running component plans, I process the people in 'component plan 1' (only include those people, but I also plug in numbers under 'non benefiting non excludable' employees from component plan 2 because that treats them as 0 on the test. I run this on an allocation basis, and code 'pass/fail avg ben pct test

then I do the same for the component plan on an accrual basis (changing my nonexcludable non benefitting employees)

[BG5150]

I wasn't sure if I had to run the ABT on both plans together first.  I just did, and it passes.   Both mid-point and ABPT.

(So, why am I doing component plans?)

 

[Tom Poje]

because you have a young HCE receiving the minimal and the plan won't pass if tested all together?

[pmacduff]

I have a question again on "otherwise excludable" but this time in relation to cross-testing.

3 month wait for deferrals, 12 months 1000 hours for match and ps.

There are people who have never worked over 1000 hours but have received top heavy contributions over a few years.  They are receiving the top heavy for 2018.  Are these people still otherwise excludable?  The system is only excluding those who haven't yet met the 1 YOS, 1000 hour requirement.  Those who have been there a few years are not showing as otherwise excludable even thought they have never met the 1 YOS/1000 hours allocation conditions for employer match or regular ps.

Since they are receiving  a top heavy allocaiton do those folks need to get the extra 2% to bump them up to the gateway?

[Tom Poje]

I would check census and verify if they are showing with a year of service or if somewhere it is 'buried' to override something and include them.

gremlins have been known to do things like that

[pmacduff]

Thanks Tom.  I usually clear overrides when I update and create a new plan year because that seems to be cleaner and I don't end up with unexpected things ?

I did check and under the "otherwise excludable" option in the census for these people and it shows as "No".  I reviewed their hours/history and at no time did they exceed 1000 hours in a plan year.  I figured I could override the option in the census but didn't want to do that if I was missing something and they could only be otherwise excludable for the first 18 mos after hire.

 

[BG5150]

Go into their census screen on the Service tab.  In the Years box, check the Service entry.  Be sure it says zero.

[duckthing]

3 hours ago, pmacduff said:

Since they are receiving  a top heavy allocaiton do those folks need to get the extra 2% to bump them up to the gateway?

And just to answer this: no, unless there's an HCE in the OEE "plan", which is uncommon but possible. The minimum gateway requirement is determined separately for the OEE plan vs the non-excludable plan, so they can have two different requirements -- and if your OEE plan has no HCEs, that requirement is 0%.