Affiliated service groups - what would you do?

[Belgarath]

Wondered what you'd do in this situation. I don't have a lot of details yet, so I'm making several assumptions to a possibly hypothetical, or possibly real, situation. And I'm hoping that I'm overthinking this, and it is simpler than I think.

1. You have four medical PC's, each owned 100% by a doctor.
2. You have partnership "x" - owned 25% by each of the 4 doctors. Partnership "x" does, let's say, all of the blood drawing and throat cultures for each of the 4 medical practices - that's all they do.
3. So, all of these are service organizations (health) and each doctor is an "FSO" and partnership "x" is an "A-org" for purposes of determining ASG status, as it regularly provides services to the PC's of each doctor.

Now, it appears to me that this means there are 4 separate ASG's - one different one between "X" and each PC. Agree/disagree?

I'm not finding solid guidance (well heck, not any guidance) regarding what has to be done once you determine that there are, in this case, 4 separate ASG's. The proposed regulations under 1.414(m)-2(g) give an example of determining that multiple ASG's exist, but that's as far as it goes.

So, if PC "A" for Doctor Killjoy sponsors a qualified plan, for coverage, testing, etc., he must consider all employees of partnership "x."

How the heck do you handle it when each PC has sponsored its own plan, doing their own thing, without considering the employees of partnership "x?" How would you even go about this? Do you have each PC test their own plan with partnership "x" employees only - and make sure coverage/nondiscrimination testing pass, and then somehow allocate the appropriate contribution amounts internally in partnership "x" to coordinate them for deduction purposes, based upon the relative amounts attributable to each partner's PC? For a simple example, suppose 3 of the 4 PC's contribute 5% to a straight PS plan. But PC 4 contributes 10%. In order to pass testing, let's say all of the employees of Partnership "x" have to receive 10% in order for PC 4 to pass. When it comes to allocating the PS expense in partnership "x" I presume any such expense would be allocated according their partnership agreement.

Blech...

As an aside out of curiosity - ASG plans I've run into have the plan sponsor being the FSO, do you see the A-orgs (or B-orgs)as sponsors very frequently?

[Belgarath]

Another thought - it appears to me that is this situation, assignment of "FSO" status and "A-org" status is arbitrary. Is there any reason, if you were trying to simplify the situation, why partnership "X" couldn't be considered the FSO, and therefore make all the businesses one ASG under Proposed Reg. 1.414(m)-2(g)(2)?

[K2retire]

I'm accustomed to seeing the partnership as the plan sponsor and each doctor's corporation or LLC as an adopting employer in a single ASG.

[Mike Preston]

You are correct, A-org and FSO are arbitrary. You have to run through the analysis twice. In this case I, like K2retire, will end up administering one plan based on a combined census as a single ASG.

[Belgarath]

Following up on this. So, suppose we assign FSO status to the partnership "x", and make all the individual doctors participating employers at some as yet unspecified point.

Has anyone actually processed a VCP correction submission, where each of the partners currently sponsors a plan independently but there is no plan for the FSO, or if they have one it is a SIMPLE or a SEP? This could turn out to be a pretty off-the-wall correction process, and I'm wondering if anyone has a methodology that they have submitted and gotten approved. Since all the plans have different provisions, how do you reconcile these?

For example, trying to be "reasonable" - do you think the IRS would approve something where each Doctor's plan, for prior years, would be allowed to stand alone, while the "most favorable" provisions would be applied to the Partnership X retroactively, and going forward all the plans would be merged into one plan sponsored by the FSO (the partnership X) as of (for example) 1/1/2016?

Or are you aware of another suggested methodology?

Thanks!