Non Discrimination Testing

[Gary]

I know this is a DB board, but wanted to submit this here, since many DB practioners/actuaries also administer DC plans.

A law firm has two plans.

One for Associates - 401k only

One for Partners and support staff and some other attorneys - 401k, 401m and PS

Under the ABT for the partners plan (for eg.) when computing the allocation (or benefit accrual for cross testing) my understanding is that all non-excudable employees, are given an employee benefit percentage and included in the ABT test, as opposed to just the participants in the partners plan, even if the partners plan is to perform the general test on an individual plan basis?

And furthermore, all allocations (or benefits) are combined from both plans, including elective deferrals?

And that the rate groups for PS non discrim testing are based on just the participants in the partners plan and exclude elective deferrals (just PS allocations included)?

For coverage testing the 401k, the 401m and the PS portions of the partners plan are all tested separately and must each pass?

Thanks and interested in any differences of opinion (if they exist).

[AndyH]

What you have said is correct, although a technical clarification to the comment below might be useful:

And that the rate groups for PS non discrim testing are based on just the participants in the partners plan and exclude elective deferrals (just PS allocations included)?

That is true only if you are not permissiviely aggregating plans for 401(a)(4), in which case you would include PS allocations from both plans. Since you have stated that their are PS contributions in only one plan than this is irrelevant to your example, however.

[SoCalActuary]

You did not disclose the ability of the two plans to stand on their own. If each passes without the other, then you are not required to aggregate then to pass non-discrimination.

Are any HCE's in the Associates plan? If so, does it pass the 410(b) coverage test? If this fails, then the Associates plan would only pass if combined witht the regular plan, so you have permissive aggregation.

Similarly, if the regular plan fails 410b, then you need the associates plan and must use permissive aggregation.

[Gary]

Thanks. Your comments help very much.

For purposes of my comments and example we are treating the plans separately and there are no HCEs in the Associates Plan.

I know I indicated that the general test rate group calculations are used based on plan participants of the one plan being tested. And we agreed on that. I may be overthinking now, but when doing the general test rate group calculation for the partners plan, I'm thinking (actually wondering) we still use the non excludable employees that are not in that plan for rate group testing ratios. That is, include all non excudable employees in the test? Or is it or must it be based on ratios of only the plan participants?

However, on second thought (or at least one whould hope), including the non excludable NHCEs not participating in the partners plan would only hurt the chances of passing ND tests and thus are not needed to be in the ND rate group tests.

Thanks and hope I am not adding confusion to the mix.

[AndyH]

All nonexcludables are in the test. Those with no PS, or those with a PS but in a different plan that is not being aggregated are treated as 0%'s for 401(a)(4) but their allocations and deferrals are counted for the ABT because aggregation is required for that portion of the test.

[SoCalActuary]

I agree w Andy. The excluded associates must be part of your rate group testing, along with any other employees in the affiliated service group, by the way.

[Gary]

Awesome. Thanks so much, that's a huge help.

Of course I will verify these points on my own, but it points me in the right direction.

[Gary]

Another thing cropped up and appears relevant.

In looking at last year's ND testing for the same client they show allocation rates "with PD" and "without PD".

So for example for the HCEs the allocation average benefit percentage without PD is 19.20 and with PD is 21.57. Since it is only 2% or so greater I don't see how it could mean participant deferral (pd) as the participant elective deferrals amount to over 6% on average.

Any thoughts?

Thanks.

[SoCalActuary]

Gary, you don't know about "permitted disparity", so I think you are swimming in water too deep.

Go get help from people who have done this before.

[Gary]

So Cal Actuary

Your point is well taken.

I still have much to learn.

Being that the plan ws not integrated I didn't think permitted disparity was a factor, but it appears that allocation rates can be adjusted based on imputed permitted disparity.

Regarding cross-testing, my understanding is that allocations can be converted to an equivalent accrual rate using the accrued to date technique, which uses the total profit sharing (for rate groups) and PS, 401k, 401m for ABT account balances, not counting investment income related to a balance such as a rollover or any other amounts.

Bottom line is that since the tests will likely not pass based on the annual method, I can request more data to buy time and then perhaps when I get the additional data I can perform the accrued to date testing and the imputed permitted disparity method.

Of course, the plan may be able to pass with just the PD and not having to do the accrued to date, but I don't think I can feel comfortable with PD in one day.

So regarding the accrued to date cross testing, I need to get account balance data and I guess prior two year's of compensation for average compensation purposes. I rambling now and am not able to think it through, but perhaps you see where I am trying to go.

Curious to get your thoughts if they are not too overwhelming.

Thanks

[AndyH]

No, Gary, that is not right. The accrued to date method is a phrase used to describe the measurement period (the current year and prior years) and has no effect on the people in the test or the type of allocations considered (still PS only-not K or M).

I would suggest searching for "accrued to date" on the cross testing board or finding some reference sources from one of the conferences as a study guide.

[Gary]

Going off in a slightly different direction. Though everyone's contributions to this post has helped immensely to this point.

Regarding Top Heavy.

Say the Partners plan passes 401(a)(4) and 410(b) by using the average benefit test (which of course uses non excludables, including those in the Associates plan), but otherwise does not aggregate with the other Associates plan.

By using the ABT to pass 401(a)(4), is it required that the plans be aggregated for Top Heavy testing, and thus minimum allocations?

Thanks.

Gary

[AndyH]

The ABT is not something you use to pass 401(a)(4). It is something you use to pass 410(b).

Remember that a(4) test starts with employer nonelective only, not the other stuff that is in the ABT such as deferrals.

The question is whether or not you are permissively aggregating two plans for 401(a)(4) testing purposes. If so, then yes, those are aggregated for top heavy testing purposes.

[Gary]

I am not sure if I totally grasp Andy's response so I will try and be more specific.

Say the one partners plan does not pass 410(b) by means of the 70% ratio test, but does so by means of the ABT test. I.e. passes safe harbor NDCT percent and ABP tests.

Would this require that the plans be aggregated for TH ratio and minimum allocation for both plans?

Or alternatively say the partner plan meets the 70% coverage ratio test on its own, but passes 401(a)(4) by means of each rate group passing by means of the ABT, would this require aggregation for TH ratio and minimum allocations for both plans?

Thanks, I hope I am more clear.

[AndyH]

I'll try and back up a step.

First, the same testing group applies for nondiscrimination - 401(a)(4) and coverage testing - 410(b). The same groups are either aggregated or they are not aggregated.

IF we permissively aggregate for a(4) and 410(b) then, yes, we aggregate for top heavy testing.

BUT, if we are testing for coverage under 410(b) as in Gary's question and we do not satisfy the ratio/percentage test, we still have the average benefits test available. Although this includes benefits and nonexcludables from other plans, this by itself is NOT permissively aggregating. If you can pass coverage without aggregating then you have left your options open.

BUT, you still must test for nondiscrimination. If you are NOT permissively aggregating, then the nondiscrimination "rate group test" will include all nonexcludables in the controlled group but they will be treated as if they are getting $0 or 0% (i.e. treated as if they are not benefiting).

If instead you are permissively aggregating, then the nonelective employer contributions are counted. That is the difference. No permissive aggregation = $0 or 0% in the rate group test for those nonexcludables not benefitting in this particular plan.

If you are doing the nondiscrimination test and you are not aggregating and you can pass the NCT test (all rate groups above the mid point but not above 70%) then you would proceed to the Average Benefits Percentage Test. This requires bringing in all bodies and mostly all allocations including non-elective PS contributions but this is not considered permissively aggregating. This is a mandatory part of the test regardless of what your choice is.

[Gary]

Andy's response is perfectly clear and is the result I was hoping for.

My concern arose because in part 416(g)(2)(A)(i) states that

(i) Required Aggregation means

(I) each plan of the employer in which a key employee is a participant, and

(II) each other plan of the employer which enables any plan described in subclause (I) to meet the requirements of section 401(a)(4) or 410.

So since the partners plan has a key employee and we use the Associates' plan by means of the ABT to pass 401(a)(4), since the rate groups do not all meet the 70% ratio test, then I was concerned that this meant that the other Associates' plan was enabling the partners plan to pass 401(a)(4) by means of the ABT.

However, from Andy's response, he does not consider using the ABT (as opposed to the ratio test which doesn't itself need the ABT) as meaning "that other plan enabling the partners plan to pass". While I totally understand what Andy says, I just was not able to come to his conclusion with certainty from what is stated in 416(g)(2) above.

However, I do understand the distinction between using the 0% accrual rate in the rate group test as compared to if we fully combined plans and used the accrual rates from the Associates' plan if they had any.

And lastly my interpretation is that since the plans are not being combined, then for purposes of the ADP and ACP tests, we only use the rates in the partners plan and ignore say the ADP rates of the Associates' plan in the ADP NHCE average?

Thanks.

[AndyH]

Right, except that the k and m portions are treated as separate plans each, so what you do for the PS "plan" does not require you to do the same for ADP, 410(b) and 401(a)(4) testing of the deferrals, or ACP, 410(b) and 401(a)(4) testing of the match.

You have a full range of options with each of these (except perhaps for top heavy aggregation purposes-I'm not sure about that).

[Gary]

OK, so are we still of the belief that using ABT (not combining for rate groups or ratio testing) alone for passing coverage or non discrimination does not require aggregation for TH testing, accruals?

Or is that the part we are not sure about?

Thanks.

[AndyH]

I don't think there is any uncertainty at all.

[Gary]

Thanks again, my misunderstanding. I think the puzzle is solved and resolved for me at this point.

This web site makes being a sole or solo practitioner more feasible.

[Gary]

Something new for me to consider!

A plan has 6 month eligibility for 401k and 1 year for profit sharing.

Say a person begins participation during the plan year for 401k, but as of the end of the plan year still does not participate for PS purposes.

My impression is that although the participant is not eligible for the PS allocations they still

1) are required to get the 3% TH minimum (assuming plan is TH of course).

2) The 5% gateway (assume that is the rate provided to meet gateway in this plan) is not required if the employee is not part of the cross testing aspects of the non discrimination testing and is required if the person is to be included in the cross testing to pass ND.

Providing the employee with 5% seems like a conservative approach and then the employee can be included in the ND tests on a cross tested basis.

Any in agreement? Other views?

Thanks.