1 Company in 2 Controlled Groups

[Jennifer D.]

I have 4 companies - A, B, C, and D.  A, B, and C are a controlled group, and C and D are a controlled group.  A and C are not a controlled group and they do not qualify to be an affiliated service group.  Who do I test together?  My thinking was I test A, B, and C together, AND then test C and D together, but that could be redundant for C, so maybe I'm missing something?

[RatherBeGolfing]

There are differing opinions in the industry (and here on benefitslink).

It has been discussed in detail in this thread

While I agree with the ABC & CD method over ABCD method, the unofficial IRS response is that neither method is unreasonable 

[ERISAAPPLE]

I think it depends on which plan you are testing.  I have looked at this and I have concluded that if a company is not in a controlled group with another company, that's it.  I understand the different positions, but I don't agree with the idea that the IRS can just say it is so.  If you look at 1.414(c)-2(e), Example (4), it says X, Y, and Z are in a controlled group.  It also says GHI, X, and Z are in a controlled group.  Then it says Y is not a member of the GHI, X, and Z group.  If you automatically combine overlapping groups, Y would be in the GHI, X, and Z group, because Y is in the X, Y, and Z group, and X and Z are in the GHI, X, and Z group.  In light of that Example (4), I can't see how, when testing the GHI plan, the IRS could argue the GHI plan has to count Y's employees.  Using the example above, if D is not in a controlled group with A (under the rules in 414, not under the IRS "because we said so" rules), then if you are testing A's plan, you don't count D's employees. 

In your example above, it seems to me that, when testing C's plan, you have two choices.  One is count all of A, B, C, and D's employees, because all of those employees work for a company that is in a controlled group with C, even if they are different groups.  The analysis would be that, when testing C's plans, you ask "what companies are in a controlled group with C," and by that I mean "what companies are in any controlled group with C." 

The alternative would be to test C's plan twice: once with ABC to see if the plan of that "employer" passes, and a second time with CD to see if the plan of that "employer" passes.  If CD passes but ABC fails, for example 410(b), you would not fix the ABC plan by adding more D employees. 

I don't have any citations for my analysis (except as described above and of course the statute itself).     

 

[CuseFan]

On 6/20/2018 at 6:01 PM, ERISAAPPLE said:

I think it depends on which plan you are testing.  I have looked at this and I have concluded that if a company is not in a controlled group with another company, that's it.  I understand the different positions, but I don't agree with the idea that the IRS can just say it is so.  If you look at 1.414(c)-2(e), Example (4), it says X, Y, and Z are in a controlled group.  It also says GHI, X, and Z are in a controlled group.  Then it says Y is not a member of the GHI, X, and Z group.  If you automatically combine overlapping groups, Y would be in the GHI, X, and Z group, because Y is in the X, Y, and Z group, and X and Z are in the GHI, X, and Z group.  In light of that Example (4), I can't see how, when testing the GHI plan, the IRS could argue the GHI plan has to count Y's employees.  Using the example above, if D is not in a controlled group with A (under the rules in 414, not under the IRS "because we said so" rules), then if you are testing A's plan, you don't count D's employees. 

Does this work on affiliated service groups as well? If AC is an ASG and BC is an ASG, but A & B have no relation or ownership overlap, do I have two ASGs or one ASG = ABC?

[ERISAAPPLE]

ASGs are completely different animals.  You just have to go through the analysis.