MWeddell

This sounds like a bad idea to me for at least 2 reasons. 1) The 401(k) regulations list the only permitted methods of correcting a test failure. (I'm assuming the plan year has ended.) Transferring excess contributions to a 403(b) plan is not one of the method listed so it is not permitted. I think if you review the structure of the regulations, you'll see that's how they work. 2) There was not a cash or deferred election to contribute these dollars to the 403(b) that was made prior to when the dollars became available in cash to the employee. Therefore putting them into the 403(b) plan now won't work.

Take a look at the portion of Code Section 72(t) cited earlier in the thread and I think you'll find that to satisfy the age 55 + separation from service exception to the 10% early distribution excise tax that the distribution must be a lump sum. FYI, at one point this provision did depend on how "retirement" was defined in the plan document but it was changed to age 55 during the 1980s.

jpod, I think the plan document in your situation doesn't expressly address whether contributions automatically resume or whether the participant must affirmatively elect to have contributions resume. In that situation, either interpretation, if made by the plan administrator or whoever has discretionary authority to interpret the plan document, likely would not be arbitrary and capricious. More generally, I believe that there are plenty of documents that are unclearly worded where contributions are suspended at least 6 months but don't resume without the participant taking affirmative action.

The group that is not eligible for a match does not affect the ACP test arithmetic at all -- but this assumes that the plan does not permit employees to make after-tax contributions.

Because the plan is tested separately for 410(b) and ADP testing, yes it is tested separately for 401(a)(4) testing including the BRF testing. You want to compute the (nonexcludable NHCEs to whom the BRF in this plan is currently available / all nonexcludable NHCEs in the controlled group) / (nonexcludable HCEs to whom the BRF in this plan is currently available / all nonexcludable HCEs in the controlled group).

I agree with the above answer. It makes no policy sense, but that's the answer I get looking at the regulations too. The fix (if one had considered this possibility last November assuming one is working with a calendar plan year) is to have the safe harbor notice provide for 100% match on the first 4% of pay deferred and that a discretionary match, expected to be 100% on the next 2%, may be allocated, which would bring to total to 100% on the first 6%.

There is no requirement that prevents you from allocating QNECs to just the NHCEs covered by the test for the otherwise excludable employees. However, you will have to follow the QNEC allocation formula that is in your written plan document, and often it will not exclude the carved out employees from the QNEC allocation.

It is not necessary to run a 414(s) test for the plan's definition of compensation in your situation. If you have access to the 1992 Enrolled Actuaries gray book, you might want to read Q&A 40 for confirmation of this.

The goal is to put the plan and the participant in the position they would have been in but for the operational error. Refunding the match to the employee does not advance that goal. It should be forfeited and used to fund future employer contributions.

It strikes me as highly unlikely that a company the size of GE will not allow you to change your catch-up contribution election other than annually. Typically one can change contribution elections at any time.

Are there any documents recording the union negotation that might constitute an amendment done by 12/31? As you know, discretionary amendments needed to be adopted by the end of the plan year in which they are effective.

There may be a problem here. One would like to characterize the BRF to be tested not as owning an insurance policy in the plan but having the opportunity to own an insurance policy in the plan. However, this doesn't seem to fit the facts because even those employees who had account balances back when the insurance feature was discontinued can't now buy insurance. I can't get the rules from Treas. Reg. 1.401(a)(4)-4(b)(3) to fit.

I agree with zimbo. The plan document must state a "definite predetermined formula for allocating the contributions made to the plan." Treas. Reg. 1.401-1(b)(1)(ii). A change to that formula typically is a discretionary amendment, which must be made by the end of the plan year in which it is effective. Rev. Proc. 2005-66, Section 5.05(3). Given that you are running your final year-end set of ADP/ACP tests, presumably the plan year has already ended, so you must allocate the QNECs in accordance with the written document. Plan documents often state multiple methods for allocating QNECs. It's questionable whether that complies with the definite formula requirement mentioned above, but it sure is handy! Hence, especially if you already have a favorable determination letter approving the written document as constituting a qualified plan, you may have some choices within the plan document for how to allocate the QNECs.

No, contributions by HCEs to the 403(b) plan will not be included in the 401(k) / 401(m) testing done on the 401(k) plan. You do have to coordinate both contributions when applying the 402(g) limit on elective deferrals ($16,500 in 2009 for non-catch-up eligible employees), but you seem to be aware of that already.

Why the different treatment just because the testing method is different? Is this a word of mouth or a written guidance? If written, do you have the cite? I agree with the prior responses. Regarding why the regulations provided that rule, the IRS did not permanently want to require recordkeeping systems to keep catch-up contributions (and investment earnings on them) as a source separate from the regular pre-tax deferrals.

Yes, I think it will mess up both the ADP and the ACP safe harbors. I've been quoting 1.401(m)-3(f)(1), but the 401(k) regulations have a parallel provision in 1.401(k)-3(e)(1). I think under the first sentence of those regulations, the ADP and ACP test safe harbors are lost. I'm hoping though that the second sentence of those regulations are inapplicable so that I am allowed to perform ADP / ACP testing. I've been trying to distinguish between the two sentences. Maybe the better argument is that I've satisfied the first sentence too, that safe harbor provisions were adopted before the plan year began and remained in effect for the entire 12-month plan year. I read the first and third sentences Reg. 1.401(m)-2(a)(5)(iv) as applying to a situation where the plan meets the ACP safe harbor and the second and third sentences of that provision as applying to a situation where the plan meets the ADP but not the ACP safe harbors. I don't see that that regulation contemplates an outcome where some of the match satisfies the ACP safe harbor but other matching contributions don't satisfy the ACP safe harbor.

Here's what I'm intending. The plan provisions adding the additional match were not adopted before the plan year began, so the plan would not meet the ACP test safe harbor under Reg 1.401(m)-3(f)(1)(first sentence). However, the amendment layered on an extra match but didn't amend the existing safe harbor provisions, so (I'm arguing) the second sentence of that regulation is inapplicable. The result is that the plan lost the ACP safe harbor status but that it may run an ACP test.

Thanks for the many thoughtful responses. My initial post was so long and technical that I wasn't sure whether it would draw responses, especially since I've not been active on these message boards for a long time. Summarizing what you've said so far ... It sounds like the objections focus on: - The fact that the notice didn't describe the possibility of a match increase. Perhaps this could be fixed before the notice went out, but if one knows of the possibility of an additional match way back then, then amending the plan to allow a discretionary match is a better approach. - The fact that an HCE might have a higher match rate available to him/her than what is available to an NHCE when the match increase takes place partway through the year. In that case, perhaps one makes the match retroactively effective for the full year of excludes HCEs from the extra match for the remainder of the first plan year. Another view is that simply implementing the match partway through the year violates the first sentence of Treas. Reg. 1.401(k)-3(e)(1) so the plan isn't safe harbor anymore (because an HCE may have a higher match rate for the year than an NHCE) but doesn't violate the second sentence of that paragraph (the one I quoted in the original post) so that ADP/ACP testing is permitted. - Concerns that the IRS objects to this as a matter of policy, based on comments in the preamble that Kevin quoted and informal statements such as what Tom Poje reports. No one has really shouted down the technical argument I advanced that we've not amended the provisions that comply with the safe harbor rules, but there are some other tricky obstacles, which I've enumerated.

Employer's 401(k) plan complies with a matching safe harbor formula of 100% match on the first 3% of pay deferred + 50% match on the next 2% of pay deferred. Employer decides to increase the match during the plan year to 100% match on the first 5% of deferrals. Employer is willing to comply with any applicable notice requirements. My initial impression was that a mid-year increase of matching contributions would cause the plan to no longer comply with the safe harbor requirements for that plan, instead necessitating that the plan be subject to ADP / ACP testing. However, upon closer inspection (and following discussion with legal counsel), the ramifications might be even worse. Treas. Reg. 1.401(k)-3(e)(1)(2nd sentence) states: In addition, except as provided in paragraph (g) of this section, a plan which includes provisions that satisfy the rules of this section will not satisfy the requirements of § 1.401(k)-1(b) if it is amended to change such provisions for that plan year. In other words, if one amends the plan to change the provisions of the plan that satisfy the safe harbor rules effective for that plan year and if the change is an increase, not a suspension or reduction of the match permitted by 1.401(k)-3(g), then the ADP test is not satisfied. I had read the preambles to both the proposed and final versions of the 401(k) / 401(m) regulations and do not see anything that addresses mid-plan year increases in matching contributions. I also found nothing relevant in recent IRS gray books. We can't avoid dealing with the regulations as they are written. Help me decide between two possible interpretations: (1) Employer cannot increase the matching contribution in the middle of the plan year. There are some other possible solutions described below. (2) Amend the plan to add in the additional layer of match (50% on deferrals in excess of 3% of pay but not exceeding 5% of pay) but don't amend any of the plan provisions describing the safe harbor match already in the plan. Now, the sentence in the regulation I quoted doesn't apply, right? [Query: might this interpretation go too far? Does the plan still meet the safe harbor rules so that ADP / ACP testing isn't required? Although it would be prudent to notify employees, is that technically required?] Other work-around solutions: - Wait until the beginning of the next plan year obviously. - Change the plan year in compliance with the rules in the safe harbor portions of the 401(k) / 401(m) regulations. - If one anticipates this happening in advance, use the discretionary match rules in the safe harbor portions of the 401(k) / 401(m) regulations, although to avoid having a higher rate of match available to an HCE than is available to any NHCE, the discretionary match probably has to be made for the whole plan year or exclude HCEs from the discretionary match. FYI, I searched and found some prior threads on related topics, but none that addressed my question exactly: http://benefitslink.com/boards/index.php?showtopic=29756 http://benefitslink.com/boards/index.php?showtopic=9518 Thanks in advance for any replies.

That's a worthy objection, ERISANut. I have received favorable determination letters with this type of provision, not that that proves much at all. Looking back, the standard I used was "reasonably determined to be NHCE at the time the match was made" which is tighter than the "good faith" standard. The problem to anticipate is that one usually won't know exactly what the HCEs are until late in January, and often the match is made before then. I'm more uncomfortable with taking match away from some employees, especially because this isn't a flaw in the procedures subject to the Self-Correction Program but rather a completely foreseeable problem.

Yes, it is that easy. Note that the discrimination would be in the form of a higher match to the for-profit employees. If you give a profit-sharing or other type of employer nonmatching contribution to the for-profit employees, it isn't covered by this year. If you give a high match rate to only a select group of the for-profit employees, you'll have a benefit, right or feature to test, which also will cause a problem. Without this special rule, controlled groups that used a 403(b) plan for their employees in not-for-profit locations and 401(k) plans for their employees in for-profit locations often had coverage testing challenges with their 401(k) plans.

Consider having the plan provision also allocate the higher match rate to HCEs who in good faith are erroneously believed by the plan administrator or its delegatee to be NHCEs. This means that occasionally the client will need to run a 401(a)(4) test demonstrating that the available match rate doesn't favor HCEs instead of automatically passing that test. However, it also eliminates the possibility of having to retroactively change the match for a stray individual in accordance with the IRS' Self-Correction Program. This type of provision might especially be helpful if the match is allocated each payroll period but one doesn't have prior year compensation figures used for determining HCEs finalized until late January following a plan year.

All match for that plan year will be subject to ACP testing, along with any employee after-tax contributions.

There's no excise tax, assuming the 402(g) corrective distribution was timely made. See Treas. Reg. 1.402(g)-1(e)(8)(i).

The best advice is to tell the client to hire a lawyer who specializes in employee benefits law to fix the problem and next time don't be so penny wise and pound foolish. Probably the course of action is: 1) Employer contributions to the executive's account in the 403(b) program are subject to coverage testing under IRC 403(b)(12). 2) If the part-time employee ever earned 1000 hours during an eligibility computation period, then the coverage testing fails. 3) Treas. Reg. 1.401(a)(4)-11(g) corrective amendment deadline is too late. 4) EPCRS needs to be examined. Sounds to me like a significant failure but your client might still be eligible for self-correction under the two-year rule. 5) 403(b) plan probably isn't a non-ERISA arrangement. Back Form 5500s are owed, fiduciaries need to review whether the investments are prudent, etc. 6) ERISA requires a plan document but there's not a firm deadline on that, so preparing one ASAP now probably isn't a compliance problem. I might have missed something, but if so see my first paragraph.