Control Groups: adopting employer vs non-adopting

[doombuggy]

OK, I have a plan that we found out when they were audited earlier this year that the person we thought owned 100% of all 11 companies really doesn't (he does hold at least 33% ownership in each, mostly 50% or more). So he wants 3 of the 11 companies to be excluded from the plan. We are going to move him to Datair's non-standardized prototype, as this will accomidate his request.

My actual question is shouldn't he be reporting the census for these three companies who are not participating in the plan? Shouldn't they be tested under 410(b)? In 20 years, I have a couple of weird ones, and this one is right up there....thanks for your thoughts.

[12AX7]

Are they all members of a controlled group of companies?

[doombuggy]

It's complicated.

so the 11 companies break out like this:

Group #1 is Co A which "T" owns 100%, Co B and Co C which he owns 67% of each, but his mom owns the other 33%, so the attribution rules apply to him.

Group #2 is Co D & Co E, which he owns 50% of each, "D" owns 47% and "J" owns 3%.

Group #3 is Co F and Co G, which "T" owns 57%, "D" owns 10% and "M" owns 33%; Co H in which "T", "D", and "M" each own a third; and Co I where "T" owns 48%, "M" owns 48%, with 6 others sharing the remaining 4% (these others do not own anywhere else). Co I is a company that he wants to exclude from the plan prior to 10/1/11 - they have expanded from a plain P/S and added a 4019k) and match feature effective 10/1/11, their new fiscal plan year.

Group #4 is Co J and Co K, which he wants excluded (he told us the employees of these companies were terminated effective 9/30/05, but this inaccuracy came out in the 2007 PY audit - let's not go there...). "T" owns 50% of these, "D" owns 10% and "G" owns 40% of each.

the plan name is "Company A Profit Sharing Plan."

Should the census data for companies I, J & K (the companies he wanted to be excluded/non-adopters) be given to us for the plan year ending 9/30/10 to be counted towards testing?

[12AX7]

If Abbott and Costello did a routine on controlled groups, it would probably be similar to what you have written above. You can't make this stuff up ! I'll try to look at it in the morning.

[Guest Sieve]

A & B are members of one controlled group. D & E are members of another controlled group. F & G & H are members of yet another controlled group, but I is not a member of that group--I is a member of a controlled group with F & G, but not with H.

But J & K are not memebrs of any of the controlled groups. So why would you include them in any Section 410(b) testing? (Remember: the 1st test of controlled group for pension purposes is not a 50% test, but an 80% test: IRC Section 1563(f)(5).)

But, Co. I would have to be included in 410(b) testing with F & G.

If all of the included entities are in one plan, you do not have a single employer plan, but a multiple-employer plan with a number of separate & distinct controlled groups--and a non-standardized prototype cannot be used for a multiple-employer plan. Either each controleld group will have to have its own plan (thus each is a single employer plan), or else you'll have to use a VS plan for a multiple-employer plan.