Coverage and BRF Testing

[justatester]

I need some direction...

Here is the situation: Plan A and B are part of a Controlled Group. Each plan on passes Coverage.

Plan A: Match formula is 100% up to 3%

Plan B: Match formula is 100% up to 4% for Location X and 50% up to 10% for location Y

Since there are differenct formulas within Plan B a BRF test is needed. When running the BRF test, do I only consider the employees of Plan B? Or do I have to considered all employees in the controlled group.

I have always been under the impression that once a plan passes coverage ALL discrimination testing is done on the plan basis not the controlled group basis.

Any thoughts would be greatly appreciated as well as regulations for reference!

[Tom Poje]

something appears to be missing in your description on how the testing was performed.

1.401(a)(4)-9(a) is clear that if you permissively aggregate plans and treat them as a single plan under 1.410(b)-7(d) (coverage) then you MUST also treat them as a single plan for purposes of section 401(a)(4) [nondiscrim]

so it would certainly be incorrect to say

"once a plan passes coverage ALL discrimination testing is done on the plan basis not the controlled group basis." if you permissively aggregated the plans for purposes of coverage.

[justatester]

They are not permissively aggregated for coverage testing. Each plan passes coverage on their own taking the other plan into consideration. So the plans are treated separately for ADP/ACP. So does that mean the plans can be treated separately for BRF. For Plan A, a BRF is not required. For Plan B, since there are two match formulas, a BRF would be required.

I guess my quesition is when running the BRF for Plan B, do I need to treat Plan A participants as non excludable not benefiting?

[Tom Poje]

my logic says since the plans are not permissively aggregated, and you don't include the one group of employees in the ACP test then you wouldn't include them in BRF, though I don't know of any particular cite that actually says.

[justatester]

That is my logic, but I can't find anything to support it.