MWeddell

Yes, you've got until the end of the GUST remedial amendment period (12/31/2001 for calendar year plans) to freely switch back and forth. Make sure your plan document correctly lists which method you used for which year. Note that the method you use for the 2001 plan year will constrain your choice for 2002, so in a sense the period when you freely switch back and forth ends with 2000. QNEC timing tends not to work well with prior year testing. If you want to make a QNEC now for a plan year ended 12/31/2000, then you need to use the current year testing method. Good luck!

The fact that the second list of disclosure items must be provided "upon request" has led me to interpret the first list of items that must be "provided" as really having to be provided, not just on request. That means that plan sponsors who wish to comply with ERISA 404© should make sure that prospectuses are given to participants immediately before or after they first invest in any registered fund. Keep in mind that the only time this will be relavant is in a lawsuit against an employer after things have already gone so sour that a plaintiff has hired a lawyer and filed suit. A judge is likely to interpret the 404© defense fairly narrowly in that context and try to let the case be decided on its merits. So if an employer thinks this 404© legal defense is worth having, I'd advise that it better make sure the prospectuses are provided, not just made available. IMO, there are an awful lot who think they are complying with ERISA 404© who are not meeting all the regulatory requirements.

I agree. Under Notice 89-23, the 401(a) plan may be used when performing general testing for the 403(B) plan. However, there's no authority that works in the opposite direction. When demonstrating that the 401(a) plan satisfies 401(a)(4), you can't aggregate the 403(B) plan contributions or benefits with the 401(a) plan.

That's right, you measure compensation during the 12 months prior to the beginning of the plan year, which means during all of 1999 in your case.

It'd be great if someone could confirm again that church 403(B) plans are subject to 401(m) testing. My client is a religious order that administers a college and health care facilities, so it does not meet the narrow definitions of a church or a qualified church-controlled organization under Code Sections 3121(w)(3). It looks other types of Code Section 414(e) church plans that sponsor 403(B) plans do have to comply with 401(m) testing. As best as I can tell, IRS Notice 2001-9 does not affect this conclusion. It extended the deadline for complying with regulations under sections 401(a)(4), 401(a)(5), 401(l) and 414(s). If the IRS intended to extend the deadline for 401(m), it would have expressly listed 401(m). Compare Sections III(A) and III(B) of Notice 96-64 for example. [Even if (contrary to my belief), Notice 2001-9 applies, it only delays the effective date of the regulation and the plan would still have to comply with a reasonable, good faith interpretation of 401(m) anyway.] Can anyone confirm for me that yes, my client would need to do 401(m) testing on the match in its 403(B) plan? It would also be subject to 410(B) testing as well.

I agree. Don't prorate the $170,000 to make quarterly limits unless the plan document specifically provides for this.

I could imagine two arguments for why the IRS would think this needs to be corrected. The employer has operated the plan in a manner that fails to comply with the plan document, because the document very likely gives the participant the right to change his or her contribution percentage election. The various EPCRS Revenue Procedures make clear that the IRS regards that as a qualification problem. I think that's very debateable, but prior threads on these Benefit Boards indicate my view may be a minority position. Hence, the IRS probably views this as requiring correction. A second argument is possible. Failure to implement the employee's modification to his or her contribution election may cause the plan to no longer have a "cash or deferred election" under Reg. 1.401(k)-1(a)(3) and hence no longer have a "cash or deferred arrangement" under Reg. 1.401(k)-1(a)(2). I don't think it's a strong argument because suppose the employer intentionally was trying to do this (presumably because it preferred 401(a)(4) testing to ADP/402(g) testing). I think the IRS would say failure to implement the modification doesn't prevent it from being a cash or deferred arrangement unless the conditions of Reg. 1.401(k)-1(a)(3)(iv) are met. After considering this second argument, I'd reject it. Note that pragmatically, I can see why the employer may choose to ignore the situation: it may feel that the employee hasn't really been hurt and that chances of IRS or DOL detection are nil.

I don't know of any deadline for non-ERISA 403(B) programs. Note that if there's a really extreme wait before the money is forwarded that the DOL might argue that holding onto this money is in effect giving the employer compensation and might accidentally convert this into an ERISA 403(B) plan.

Treas. Reg. 1.401(k)-1(d)(3) is pretty muddy on whether the restriction on having the same employees eligible for another 401(k) plan ends 12 months after distribution of all assets from the terminated plan (last sentence of the regulation) or 12 months after the date of termination (third from last sentence of the regulation). Reading the regulation for yourself is probably more illuminating than any discussion I can add here.

card, I agree both with your reason for rejecting my original idea and your proposed modification to it. Thanks for adding the citation to your argument. The new 401(k) plan would have to continue for 12 months before one could then merge the two plans together. Having to operate both 401(k) plans for 12 months adds to the transaction costs. For all but the largest employers, the same money that is spent on transaction costs for this idea might be spend on communications or QNECs in a manner that more cost effectively would solve the testing problem until the beginning of the next plan year.

Alf, Sorry, I misunderstood what rule you were referring to perhaps. I agree the cites in your last posting are the relevant restrictions that we're trying to work around. I think establishing a new plan with a short plan year (but still at least 3 months) can comply with the safe harbor rules and then one transfers in a trust-to-trust transfer the assets of the old plan to the new one. It's an aggressive position (like I said before, check with legal counsel) and involves LOTS of transaction costs, so it's not for everyone. I thought I'd mention it because it seemed responsive to the original question that started this thread.

There is a rule that prohibits distributions upon plan termination if a new 401(k) plan covers the same employees during the 24 month period beginning one year before and ending one year after the plan termination date. Treas. Reg. 1.401(k)-1(d)(3). That rule wouldn't interfere with a plan-to-plan transfer of assets from the old plan to the new plan.

No need to set up two plan documents. IRS regulations will treat the 401(m) portion as a separate plan for compliance testing purposes from the 401(k) anyway. If you want to run the 401(k) test without being hurt by early entries, there are two methods. The older regulatory method is to disaggregate the plan into two component plans, one of which covers those 21 or older with 1 or more years of service (with an entry date assumption if desired) and the other of which covers the remaining employees who could have been excluded. The newer statutory method is that you run only one 401(k) test, ignoring NHCEs who are < age 21 or < 1 year of service. Depends on the software you're using, but you may indeed have to run the 401(k) test separate from the 401(m) test and manually apply the multiple use limit.

One can start a new plan, say February 1 or March 1, 2001, covering the same employees and then merge the existing plan into the new plan. The new plan can meet the safe harbor rules. Because it is a brand new plan, there's an exception that allows the first plan year to be less than 12 months and still meet the 401(k) / 401(m) safe harbor rules. There's a LOT of transaction costs involved with this idea (new plan document, additional audit, additional Form 5500, new SPD, extra vesting if the 1,000 Hours method is used, etc.), so the idea would not appeal to anyone but very large employers. Obviously check with your legal counsel before proceeding.

This discussion has indeed made me question my position. Great discussion! The match is discriminatory if it stays in the plan. The same logic that's in the regulatory preamble applies. If the hanging match that remains after a 401(k) or 402(g) refund is discriminatory, the same reasoning applies if the hanging match is left after a 415 refund. The Announcement 94-101 provision cited by card concludes the same thing. The difficulty is how to fix the problem. Distributing the money to employees is prohibited by law and doesn't fix the discrimination problem. It seems like our choices are to either forfeit the money without statutory or regulatory authority, arguing only by analogy to the cites quoted in my prior post or to allocate extra matching contributions to enough NHCEs to fix the discriminatory benefit, right, or feature as advocated by card. I still hold to the former position, but I'm much less sure of it based on this discussion. I have a hard time believing that one is required due to a reasonable error in estimating compensation etc. which caused a 415 excess to now allocate extra matching contributions to a large number of NHCEs. Of course the forfeiture of a hanging match due to a 415 refund wasn't a widespread possibility before 1991, which helps explain why we don't have statutory authority for this refund.

I agree with Mr. X's comments that there's not a discrimination problem if no HCEs are affected by the 415 reductions and that other correction mechanisms are available. The 1994 revisions to the 401(k) regulations were published in the Federal Register on 12/23/1994, almost certainly later in the calendar year than Announcement 94-101 that card refers to. Here are the regulatory cites I referred to in my first post. While they don't expressly refer to match related to an excess annual addition, I believe the IRS would apply the same analysis. Reg. 1.401(k)-1(f)(5)(iii): "Matching contributions forfeited because of excess deferral or contribution. For purposes of section 401(k)(2)© and paragraph ©(1) of this section, a qualified matching contribution is not treated as forfeitable merely because under the plan it is forfeited if the contribution to which it relates is treated as an excess contribution, excess deferral, or excess aggregate contribution." Reg. 1.401(m)-1(e)(3)(vii): "No corrective distribution of matching contributions other than excess aggregate contributions. A matching contribution that is an excess aggregate contribution may be distributed as provided in section 401(m)(6) and 1.401(m)-1(e)(3). A matching contribution may not be distributed merely because the contribution to which it relates is treated as an excess contribution, excess deferral, or excess aggregate contribution. See §§1.401(k)-1(f)(5)(iii) and 1.411(a)-4(B)(7) regarding permissible forfeitures of matching contributions that relate to excess contributions, excess deferrals, or excess aggregate contributions." This is from the preamble to the 1994 final regulations: "Where a matching contribution relates to an excess deferral, an excess contribution, or an excess aggregate contribution that is distributed, a discriminatory rate of match may result, unless the matching contribution is distributed as an excess aggregate contribution. If the matching contribution cannot be distributed as an excess aggregate contribution, so that the discriminatory rate of match cannot be corrected by distribution, a plan may provide that the matching contributions are forfeited, as permitted by sections 411(a)(3)(G) and 401(k)(8)(E). See sections 1.401(k)-1(f)(5)(iii); 1.401(m)-1(e)(4); 1.411(a)-4(B)(7). Alternatively, the discriminatory rate of match can be corrected by additional allocations to the accounts of nonhighly compensated employees, under section 1.401(a)(4)-11(g)(3)(vii)(B)."

Sure, the 403(B) program could be organized as not subject to ERISA. See Labor Reg. 2510.3-2(f) for a description of the very limited role an employer must take. Having a previous ERISA 403(B) plan or a current 401(k) plan shouldn't affect whether the new 403(B) arrangement is subject to ERISA. Make sure payroll understands that both 401(k) and 403(B) deferrals apply toward the annual dollar amount limit but that the make-up election applies only to the 403(B) arrangement. Also, make sure that 403(B) vendors understand the situation as well, because they'll be enforcing individual deferral limits too.

The plan should provide that the associated match is forfeited. This is found in the 401(k) regulations, although the most readable explanation of it is in the preamble discussion to the 1994 revision of those regulations.

If an employee satisfies the eligibility conditions for being eligible to receive a 401(m) contribution but doesn't receive that particular year's contribution due to the last day of the plan year condition, this is treated as a 401(a)(4) issue, not a 410(B) issue. See Treas. Reg. 1.401(a)(4)-2(B)(4)(iii). Because there's not a similar provision in 1.401(a)(4)-4 means that one would have to do perform benefits, rights, and features testing on the match's availability without ignoring the last day of the plan year condition. However, if this is the only thing going on with the match, using a snapshot population testing method even with the appropriate adjustment suggested by the data substantiation rules of Rev. Proc. 92-34 should make this test pass easily. The threshold for passing the BRF test is not the 70% ratio percentage threshold but the much lower figure from the nondiscriminatory classification test. I didn't mean to get too technical but wanted to suggest a resolution to some of the testing concerns. I agree that although as a consultant I generally prefer the year-to-date method or a year-end true-up matching contribution, matching per payroll period is still more common. I don't know of a survey addressing this issue. I agree that many, probably most, plan documents are silent about what period is used to calculate the match but that to meet the definite allocation formula requirement, the document should address the issue. Note that there is always a certain arbitrariness when choosing what period is used to measure fairness. For example, take two hypothetical employees both earning the same amount during 1999-2000 with no bonuses, no raises, etc. Assume they participate in a calendar plan year that matches the first 6% of pay. Both A and B contribute 12% except A takes a hardship withdrawal with 12 month suspension on 7/1/1999 and B takes a hardship withdrawal with 12 month suspension on 1/1/2000. Both employees resume contributing at 12% immediately after the 12-month suspensions end. There is no policy reason to give B half of the match compared to what A receives, but that's what happens with a plan year match calculation compared to matching each payroll period. Hence, a plan year match calculation might seem fairer and handles better the interaction with the 402(g) limit on elective deferrals, but it's not perfect.

The 10/2/2000 issue of Pensions & Investments contains quite a bit of information on market share of defined contribution plan recordkeepers based on data reported by the recordkeepers themselves. One chart lists the number of plan sponsors that each provider has, broken down by the type of plan including 457 plans.

Yes, the allocation formula could contain the language you suggest, but to avoid having employer discretion, you'd have to specify in the plan all of the variables that affect how actual deferral ratios are calculated.

Yes, this is permitted by law, because the elective deferral requirement that the contribution be made before the cash is available to the participant does not apply to after-tax contributions. However, most plan documents do not permit this lump sum contribution.

I don't know how these can be treated as a single plan given that (with a very limited exception) a money purchase plan cannot have a qualified cash or deferred arrangement. It seems to me like you'd have to have 2 plan documents.

Yes, Treas. Reg. § 1.401-1(B)(1)(ii) requires that the plan state an allocation formula for allocating any type of profit sharing contributions, which includes QNECs. I've often seen the IRS issue favorable determination letters on documents which don't have a definite allocation formula, but that doesn't prove much. If you've got access to back issues of various periodicals, a good discussion is Rovell, “Legal Update: Reverse Compensation Allocation Formula Provides More Bang for the QNEC Buck”, Panel Publishers 401(k) Advisor (April 1996).

If no other contributions were made to that IRA other than rollovers from qualified plans, then you should be able to roll it back into a qualified plan. Most plans that accept rollovers don't place any restriction, so you probably can roll it back into your employer's plan.