Can QMACs count toward the gateway?

[Guest Dave Peckham]

Reading from the "Explanation of Provisions" in the Federal Register, Vol. 66, No. 126, 6/29/01, Section B, 3rd paragraph:

"The general rules and regulatory definitions applicable under section 410(b) apply also for the purposes of these regulations. For example, these regulations do not change the general rule prohibiting aggregation of a 401(k) plan or a 401(m) plan with a plan providing nonelective contributions. Accordingly, matching contributions are not taken into account for purposes of the gateway."

In other posts, it has been argued that ANY type of nonelective employer contribution, including QNECs, can count toward the gateway. How are QMACs treated for purposes of 410(b)? As a nonelective employer contribution, or as a matching contribution?

I'm afraid it's the latter, and therefore, QMACs cannot count toward meeting the gateway minimum. But I would love to be wrong. Was there anything on point in the final 401(k) regs? I haven't read them thoroughly yet.

[Tom Poje]

not that I know of

a match is a match is a match, whether it have a Q in front of it or not.

[Guest Dave Peckham]

Thanks, Tom.

After I posted the original message, a colleague from another firm claims that not even QNECs can count toward the gateway. What do you think?

[Belgarath]

My understanding is that under 1.401(k)-2(a)(6)(ii), the testing must be passed both with and without the QNEC's being taken into account. I take this to mean that for gateway purposes, you can't count the QNEC.

Seems like a valid shortcut would simply be to test without taking the QNEC into account - if you pass, then I have a hard time seeing how you would then fail if you include it, since the QNEC is going just to the NHC. But I'm undoubtedly missing some crucial point.

[Tom Poje]

well, the gateway rules 1.401(a)(4)-8(b)(1)(vii) says allocation rates are determined under 1.401(a)(4)-2©(2)....which says ...allocation rates including ALL employer contributions and forfeitures allocated to the account for the employee UNDER THE PLAN for the year...

now, as pointed out, you have to run the a(4) test twice - it has to pass both including and not including QNECs.

so lets say I don't include them, and there are no other nonelectives allocated to this individual. when performing the test did he receive any nonelectives? no, so no gateway required.

now I test with QNECs. did he receive a nonelective? yes. so he needs the gateway. does ithe QNEC count toward the gateway? I would say yes, since it is a nonelective.

now the problem is if he receives any gateway, then the answer in step one (testing without QNEC) is wrong ong, that he reecived no other contributions. if I go back and look at step 1, it seems he received some nonelectives, therefore needs the gateway. now can he count the QNEC? I think the answer is unclear.

all the regs say is that an ee must have an allocation rate = gateway minimum before performing cross testing. did the person in question receive that much? Yes. now I test and follow whatever rules are allocable. I think.

certainly SHNECs have been given a blessing by the IRS to be used in testing and count toward gateway, and a SHNEC is a type of QNEC under the new regs.

[Belgarath]

Tom - I'm finding this a bit challenging. If I test without including the QNEC, and there are no other nonelectives or contributions, then I would think the result is exactly the opposite - that now gateway contribution IS required because for testing purposes, the NHC are getting zero, which cannot possibly satisfy gateway. But if there were other contributions that get the NHC's to 5% (I like to use 5% for conversation because it is easier for me), then assuming the QNEC's are allocated in an equal percentage to all NHC, then how could the (a)(4) testing fail, and no additional gateway contribution required.

Where am I going wrong on this?

Thanks!

[Tom Poje]

I'm not sure you are going wrong. as I said I think the answer is unclear.

I view a(4) test as saying I must do the test twice, one with and one without QNECs.

ERISA Outline Book simply says you have to meet the gateway test as a PRECONDITION to demonstrating satisfaction of a(4).

ok I check to make sure all received the gateway. I then run my a(4) test, in this case twice.

The other argument would be, I need to run an a(4) test. In this case twice. One with QNEC and one without.

Before I run the test without QNEC I need to check for the gateway.

I don't see it that way, but I can understand looking at it that way.

(thank heavens I don't have anything that allocates a QNEC and so I don't have to worry about this particular situation)

[Belgarath]

I think we may possibly be looking at this the same, or getting the same results but saying it differently. Maybe a small example will help to clarify my thought process on this.

Suppose due to the ADP testing failure, it is determined that a QNEC of 4% must be allocated to all NHC. And the document provides that all NHC receive exactly the same percentage - no "targeted" or "bottom up" QNECS. The employer makes an additional discretionary profit sharing contribution, allocated on a cross-tested basis.

So now you have to do your (a)(4) testing, and you must pass both with and without considering the QNEC.

Situation 1: You test without the QNEC. The employer discretionary contribution, tested by itself, passes the rate group testing and gateway. If that is the case, and all NHC receive an additional 4%, it would seem like there's no way that testing could fail when you now test WITH the QNEC. Agree/disagree?

Situation 2: So now let's assume that the the employer contribution does not, on its own, pass gateway. First you test without the QNEC. You fail. So the employer must make an additional contribution sufficient to pass gateway. Now, you test WITH the QNEC. Again, if you already pass without the QNEC, and the QNEC is allocated equally to all NHC, I don't see how you can fail testing WITH the QNEC. Agree/disagree?

Even if you agree with my conclusions in the above situations, there may be holes in this methodology where it WON'T work, and if so I'd greatly appreciate having it pointed out. Thanks again!