Imputing Disparity in a Non-Cross Tested 401(k) Plan with 3% non-elect

[Guest Dave Danziger]

My issue is both technical and arcane. If my fears are correct, the following fact pattern results in reverse discrimination against HCEs. I don't think that was Congess' intent in directing IRS to write the non-discimination rules, so I'm hoping some of you can provide a word of encouragement. Here's the deal:

I'm running the General Test for non-discrimination on a 401(k) plan. Cross-testing does more harm than good, so the test does not employ cross-testing (i.e., the test is based on contributions, not EBARs). I'm concerned about the proper methodology for imputing disparity under the following scenario:

The plan is subject to a 401(k) Safe Harbor Notice/Election which calls for a 3% non-elective contribution FOR NHCEs ONLY. (The employer's reason for not giving a SH contribution to HCEs, is the desire to avoid giving a contribution to HCEs who terminate prior to year-end.)

This year, the employer wants to contribute the following amounts beyond the 3% Safe Harbor for NHCEs:

1. 3% of comp for HCEs who are still employed at year end; plus

2. For HCEs & NHCEs alike, a contribution equal to 3% of all comp plus 3% of comp in excess of 100% of the SSTWB.

Logic (and IRS Notice 98-52) tells me that this is a non-discriminatory allocation, however, I have the following concern:

I know I must not adjust the NHCEs' 3% Safe Harbor contribution for imputed disparity, but MUST I adjust the 3% profit sharing allocation (under step 1 above) for HCEs?

If I must adjust the 3% for HCEs, but not for NHCEs, the adjusted total allocation for HCEs (step 1 & 2 above) will exceed the adjusted allocation for NHCEs (unadjusted 3% SH contribution plus adjusted step 2 amount from above).

This result would defy logic, and result in reverse discrimination! Afterall, if the HCEs got their first 3% pursuant to a Safe Harbor election, everyone would agree that it is not subject to adjustment, and Notice 98-52 indicates that it would not be discriminatory. But, if I give HCEs a less generous 3% (i.e., subject to vesting and last day requirement) AND am required to adjust this 3%, then the adjustment will cause the HCEs' 3% to be valued as something greater than the 3% for NHCEs.

(Note: This concern only arises if the employer's allocation is greater than 3% across the board (i.e., I wouldn't impute disparity if a level 3% was the only contribution). However, any additional allocation, if skewed to integrate with social security, will come out looking discriminatory until the total non-Safe Harbor allocation for NHCEs exceeds 5.7% (i.e., above that level, imputed disparity will have no further impact).

Has any one encountered this problem, and better yet, has any one received word from IRS to the effect that you don't have to adjust the first 3% for HCEs?

[AndyH]

I agree that if you have to general test on a contributions basis, normal testing will fail because of the disparity situation you've described. I think that within one "plan" you either impute disparity or you don't.

First, wouldn't the allocation you've described satisfy the safe harbor requirements? It's not clear to me exactly what the allocations are.

Second, you could separate the group into component plans. I think you could then inpute on some groups but not others. But each subplan must satisfy 410(B), so one couldn't be all HCEs and the other all NHCEs. I don't know if component plan testing would work in the case you've described, but it's an option to consider.

[Guest Dave Danziger]

Great observations AndyH!

1. You're right in thinking that the contribution I had descibed would satisfy the safe harbor. In my case, there willl be additional contributions for participants in other classes. So, I must use the General Test.

2. I'm intrigued by the notion of restructuring. Do you (or any other reader) think I could test the exact same population as part of 2 separate component plans? That is, assume my original scenario would benefit 2 HCEs and 5 NHCEs. Could I put that entire population into a component plan in which the benefit is 3% accross the board (and demonstrate non-discrimination without imputing disparity); Then designate a second component - consisting of the exact same population, with a benefit equal to 3% of all comp plus 3% of comp in excess of the SSTWB?

I've assumed you can't put the same people into more than one component, but I've never really looked in to the concept. Any thoughts?

[AndyH]

That's right. Each person must go into one and only one component plan.

[Tom Poje]

no - in component testing you do not split the contribution, you split the people . a person can only be in one plan.

It would sort of like you having a controlled group with 2 companies, and not aggregating for testing purposes.

for example, test 1 hce with 2 NHCE using allocation method

and test 1 hce with 3 nhce using accrual method.

test 1

NHCE ratio = 2/5 = 40%

hCE ratio = 1/2 = 50%

ratio % = 80%

Plan passes ratio % test.

test 2 would have an even higher rati %

now, since each component plan passes 410(B) you can try your cross testing. component testing is especially useful when you have one young ee whose EBAR goes through the roof when using accrual method.