403(b) failing ACP test

[Santo Gold]

50 life 403(b) plan with 2 HCEs. They failed the ACP test pretty badly and both HCEs will need to take distributions of the excess match contributions.

But the match is a 100% fully vested match. If this were a 401(k) plan, I could test the match together with 401(k) contributions in the ADP test and maybe get a combined passing result.

Is that option available with the 403(b)? Since there is no ADP test for a 403(b) test, I simply combine the 403(b) elective deferrals with the fully vested match. Now the match is not subject to testing and voila, no 401(m) test or problems.

I'm sure this won't fly but was hoping someone could confirm one way or the other.

[ETA Consulting LLC]

You're talking about "shifting" or "borrowing". This is typically done in 401(k) plans "after" the ADP test as been satisfied. Since there is no ADP test within a 403(b), I wouldn't imagine that to be an option. Keep in mind too that there may be different eligibility requirements for the deferrals and match within the 403(b). The deferrals are subject to universal eligibility while the match is not. Trying to 'shift' or 'borrow' may change your testing population. Without going in researching the exact language, I wouldn't count on this being a viable approach.

Good Luck!

[Belgarath]

I just had the same question come up. I agree with ERISAtoolkit. See 1.401(m)-2(a)(6)(ii).

[BG5150]

Side question: does the document have to specifically allow for the shifting method? Or is it just an allowable mechanism for testing, like separating out otherwise excludables?

[ETA Consulting LLC]

Side question: does the document have to specifically allow for the shifting method? Or is it just an allowable mechanism for testing, like separating out otherwise excludables?

I would think that the document should not specifically preclude it. I would, however, look for language that would support it (e.g. vague references to the Code itself as opposed to specificity on each step). I normally look at the task of proving a plan is non-discriminatory as a mathematical test outlined in the Code and Regulations that may be applied unless there is document language precluding its use (i.e. a fail-safe option for a failed coverage ratio test). I am sure this is a matter of huge debate.

Good Luck!