cross-tested plans

[HarleyBabe]

Have a plan that is cross-tested, 8 different groups. Because of turnover, cross-tesing really doesn't work very well and it's too late to amend the allocation method for the year we are in . Does anyone agree or disagree, that you could run the calculation integrated and then plug the contribution figures into the participants within the groups? As a side note, if we do that, it fails AVB Test and General Test which I know we have to do as part of cross-testing. Some feel that this is acceptable because we are essentially testing on a contributions basis and it passes ratio. I seem to not be able to get past that being okay but can't prove it, and don't we have to amend to state the Integration Level in the Adoption Agreement and state that the contribution will be allocated on an Integrated Basis?

Thanks for any input.

[Archimage]

Two points:

1. Participants in a given group cannot generally cannot get different allocations. So no, you cannot integrate within a given group.

2. Do you have a last day rule? If so, amend the plan to an integrated formula.

[HarleyBabe]

Two points:

1. Participants in a given group cannot generally cannot get different allocations. So no, you cannot integrate within a given group.

2. Do you have a last day rule? If so, amend the plan to an integrated formula.

I wouldn't be allocating differently within the group. They would all be receiving the same percentage of pay, except for those whose pay is in excess of the SSWB and those participants are in Group1. Guess the questions is, do we have to amend? or is this considered testing on a contributions basis.

There is a last day rule on some of the groups but not all. Basically they are all getting a pro rata allocation, I would just be plugging it in to the groups. The amount to each group ins't the issue, it's, 1. it fails AVB, 2. it fails 401(a)(4), but it passes ratio. I just thing the doc needs amendmed to get rid of the groups. The master doc does only states that the allocations will be made by group.

[wsp]

1. Participants in a given group cannot generally cannot get different allocations. So no, you cannot integrate within a given group.

If the stars aligned and the groups only had individuals in it, with the exception of the final group having no one exceed the SSTWB, then could you do this?

Doesn't seem right to be able to, but just tossing it out there as I have a few plans that fit the criteria mentioned above and also on occasion have issues with terminations fouling up the tests.

[J Simmons]

Because of turnover, cross-tesing really doesn't work very well...

In general, if employee turnover is causing cross-testing problems, you might have too restrictive of eligibility and/or benefit accrual rules in your plan design. If you impose a vesting schedule, you might consider relaxing those eligibility and benefit accrual requirements. More of the newer employees will be brought into your testing pool, and if they are younger, lesser earning employees than the owner-employees and others you are trying to favor, this will help you pass testing. At the same time, the dollars put in for these newer employees won't be leaving the plan if those employees quit after a short stint.

Another design angle to consider rather than perhaps abandon cross-testing altogether would be to amend the contribution provisions to allow for a contribution to be designated, when made, to be an "integrated contribution" rather than for any cross-testing grouping. If for a year the testing shows that integration would be better than cross-testing, make no contributions for any cross-tested grouping per se, just make an "integration contribution" for that year that is allocated across all those eligible for the plan and for whom the plan year is an accrual year. This way, no amendment is required. During the GUST II restatement process, 7 or 8 different IRS district offices challenged this design when applications were made, claiming it did not meet the definitely determinable allocation formula requirement. However, each challenge was dropped and the design approved after pointing out some of the Service's internal memos on the topic. Also, I've incorporated this design into my DC prototype for EGTRRA and the Service did not challenge this when they gave me their punchlist of changes for the prototype a few months back.