Cross Testing with Permitted Disparity

[Guest SuzieQNEC]

A profit sharing plan has 401(k), qmac and discretionary employer contributions (there is no DB plan). 401(k) and qmac are tested using the adp test. Employer contributions are allocated based on 5% to TWB and 20% > TWB and is then cross-tested. As part of the whole cross testing, is it required to test first the employer contribution, and then to test the sum of all contributions under the ABP Test for coverage? Second question, the plan has never been tested with permitted disparity. Can the plan now be tested with permitted disparity? This year it allows the the cross test of the employer contribution to pass.

[Tom Poje]

ok, first some clarification:

1. a plan's allocation formula can use permitted disparity. you simply follow the terms of the document.

2. for nondiscrim testing, a plan is permitted to 'impute disparity' which looks very similar to permitted disparity, the concept is basically the same. to impute disparty is a testing assumption. it is not document driven. it is purely optional. you don't have to be consistent from year to year.

now, as for testing

if you need the avg ben % test, all contributions - deferrals, qmac and profit sharing are used. only for that test. repeat: only for that test.

any other testing (rate group testing) only uses non-elective contributions.

there is not requirement that you use the avg ben % test. if you can pass the 70% ratio % test (for each HCE rate group) then you do not needed it.

[Guest SuzieQNEC]

Thank you for that clarification on terminology. Here's where I'm still confused. Grasping the correct order of steps and tests required is certainly a challenge. I am using Relius Administration and running the General Nondiscrimination Analysis.

First I am running 410(b) to test coverage - the plan passes. Part two of the 410(b) test looks at the ABPT with all contributions. In this case, the plan passes both the Ratio Test and ABPT.

Then I am running 401(a)(4) to test nondiscrimination. The whole group passes the ABPT (same test as above). Is this enough to be finished or do we still need to do individual rate group testing? Relius tests every rate group. For each rate group, first there is a check of ratio percentage test. For each group that fails the ratio percentage test, the average benefits test is done. Relius does this test only on employer contribution. When you comment that rate group testing only uses non elective contributions, then is it correct that as part of this test, the average benefits % test only uses the non elective contributions?

[Mike Preston]

First I am running 410(b) to test coverage - the plan passes. Part two of the 410(b) test looks at the ABPT with all contributions. In this case, the plan passes both the Ratio Test and ABPT.

A little confusing, if you ask me. If you are running your 410(b) analysis and it passes the ratio percentage test, you stop. There is nothing further to do. Your goal is to satisfy 410(b) and you do so. Now, I understand the reason that software products print things which are not technically necessary. They can be wonderful planning tools. If you "just barely pass" the 410(b) test on the basis of the ratio test, it can be very comforting to know that, since you pass the average benefits test, the plan can use a much easier to pass threshold than the 70% ratio test. So, even though the ABT isn't required, it is sometimes extremely helpful to be able to look at it without expending extra effort to get it. Nonetheless, from your perspective, if it isn't necesssary, then try to ignore it. Your plan passes 410(b). Be happy.

Then I am running 401(a)(4) to test nondiscrimination. The whole group passes the ABPT (same test as above).

Don't you think it is a little strange that here you start with the ABT and above you made it the second item done? That is the key to understanding this whole thing, in my opinion. The purpose of the ABT is to allow a test to use lower percentages. Don't pass the ABT? If so, then you are stuck at 70%. Pass the ABT, then you can use a lower threshold (depends on the plan's concentration percentage, but certainly no higher than 45%). But you don't need the ABT if the ratio percentage for your test in question is 70% or more.

Is this enough to be finished or do we still need to do individual rate group testing?

Now that you know that the ONLY purpose of the ABT is to allow you to use lower percentages in your testing, you know that you must still do individual rate group testing, right? But knowing that you pass the ABT means that you know you will have a much easier (although certainly no guarantee) time of passing your rate group testing.

Relius tests every rate group. For each rate group, first there is a check of ratio percentage test. For each group that fails the ratio percentage test, the average benefits test is done. Relius does this test only on employer contribution.

Of course they test every rate group. That is required. Your description gets a bit confusing once you say that the ABT is done. You say that "Relius does this test only on employer contribution (sic)." What do you mean by that? Are you saying that they so the same ABT that you referred to above, when talking about 410(b) testing, but only including the employer contributions? I doubt that they actually do that, but is that what you meant to say? What I think they do is that they run the ratio percentage test using the EBARs that result from using only the employer contribution. If that test fails, then they run the very same ABT that is run under 410(b) - including all contributions, both employer and match and deferrals and, if they have them, employee after tax contributions - and they use that result to determine whether the rate group test they just ran can be considered to pass based on a lower percentage than 70%.

When you comment that rate group testing only uses non elective contributions, then is it correct that as part of this test, the average benefits % test only uses the non elective contributions?

In a word, no.

[Tom Poje]

If you are printing the 401(a)(4) nondiscrimination summary I could understand your confusion. I never use that that report, I find the info on that report baffling.

part of the problem is the use of terms on Relius.

In terms of the regs there is 1. coverage, which is 410(b)

and 2. nondiscrim, which is 401(a)(4)

in Relius both of these are listed under nondiscrim analysis, though you could have a plain vanilla profit sharing that you don't care about nondiscrim and are only testing for coverage, but you find the coverage report/testing results under 'nondiscrim'.

all that aside, if you are testing nondiscrim, then the 'overall report' is a good starting point.

each hce is listed (e.g. each rate group). on the far right are the pass/fail columns.

first coulmn is ratio % - if it says 'pass' then that HCE passes. the last column (for better or worse is titled 410(b) test) will also say 'pass' as a result.

the 2nd column is titled 'nondiscrim classification test'. the HCEs ratio % (which shows on this report) must be greater than the midpoint (which unfortunately does not appear on this report unless you have modified the report - it is not that hard to add) it does appear elsewhere, simply not on this report. This may say pass or fail. A 'pass' is only part one. the next column 'avg ben % test' must also say pass. in other words, if columns 2 AND 3 say pass, then coulmn 4 will say pass. column 4 will be the same for everyone since there is only one and only one avg ben % test (which includes all contributions) If column 1 OR 4 say pass, then column 5 will say pass and the HCE passes. all HCE must pass.

all that is probably confusing unless you are looking at a copy of the report and then it will should make a little more sense.

[Guest SuzieQNEC]

Mr. Preston, Mr Poje:

Thank you very much for your input. I have read and reread your responses and I'm still trying to figure it out. Mr. Poje - this is in response mostly to your reply, in regards to the details of the output for the 401(a)(4) test in which we run the 410(b) test for each rate group....

I have printed out the 'overall report' as per your suggestion and compared what you wrote to what is on the report. I think I see where my confusion has been all along. Step one of the 401(a)(4) rate group test, being the ratio percentage test, only looks at the employer contribution. As such, the first columns 'NHCEs benefiting' and HCE's benefiting' give the result of those participants whose ratio is greater than the HCE (or rate group) in question based only on the ER contribution. For those participants who fail the ratio test, we move on to the next step - the Nondiscriminatory Classification Test. In this step, we still compare that ratio test value, based only on employer contributions, to the midpoint. In column 3, the 'calculated avg ben percentage test', this is the column that would say Pass for everyone based on the one ABPT which has been done using all contributions and which allows us to run the nondiscriminatory classification test in the first place. This test is done only the one time for the group as a whole, not for each individual rate group, but is also the second part of the Average Benefits Test which must pass when the Ratio test fails for the rate group. When you say, "column 4 will be the same for everyone' then, I believe you mean column 3?

So if I am correct then, and to paraphrase Mr. Preston, there is never actually a comparison of the sum of all employer contributions because the ABPT is done only once for the entire group to allow the group to proceed past the ratio test within the rate group testing.

One more thing, the report you said you never use I assume is titled, "Nondiscrimination Test Under IRC $401(a)(4)." That test shows ebars based only on the employer contribution, and then at the end says, "Passed Calculated Average Benefit Percentage Test". Is that Pass based on the ABPT under 410(b) results page which shows ebars with all contributions and then the actual 70% results? Also I thought the ABPT excluded after tax contributions. That report may be the very source of my initial confusion.

[Mike Preston]

Who are these old men of which you speak?

[Tom Poje]

I believe you are correct on all points. There is one and only one ABT and it is indeed the same for 410(b) as for nondiscrim testing. if you undersatnd that much then you are well ahead of the game! very good!

my apologies if I incorrectly numered the column in my description, it is hard enough to understand some of this stuff.

and yes, employee contributions (after tax) are not included, unless they are roth deferrals. [and dont include catch up contributions either]

by the way, attached is my modified version of the overall report. however, it would have to be copied into the relius directory of reports as you can't run it out of custom. thus you would first have to rename the overall report and then copy this one to that folder. the modification is simply to show the total count of NHCEs and HCEs, the NHCE concentration % and the midpoint. since the report shows the ratio % for each HCE this gives an immediate comparison. also the report will show the avg ben % test numbers. I'm simply pulling the fields that show up on another report so I dont think anything can go haywire. besides, I didn't modify anything in the pass/fail columns so nothing is going to change there. [ report version at 12]

[Guest SuzieQNEC]

Hi Mike: okay and thank you for your help!

Tom: Thank you very much for your help too. I took a look at your report. Turns out we're stuck in version 11 here so it doesn't seem to show what you have inserted. I will hold on to it though because it certainly makes sense.