Tom Poje

yes, there is no exception under the safe harbor rules that apply to catch-ups.

Is it just a coincidence that if you combine 'The IRS' you get 'theirs'?

hopefully you exclude key employees from top heavy minimums. my concern would be that someday some IRS agent would possibly point to the IRS memorandum from a few years ago regarding "short service employees and other meaningful benefit schemes and abuses" while the context of that memorandum dealt with including young short term NHCEs to help the cause, I could easily see that including HCEs and giving them 0 somewhat a similar deal.

if the HCE is age 50 or older, then you still get the catch-up, so something could be deferred. $5,500 is better than nothing. no matter which route you choose, be careful if top heavy rules apply.

that is another issue indeed. I am assuming you document allows you to make a QNEC. Lets further assume, that the QNEC goes to NHCEs only. therefore, I don't think you can simply reclassify part of the nonelective as a QNEC, because in that case you have given the HCE 4% nonelective and lets say the NHCE 3% nonelective and a 1% QNEC. I don't think documents read that way. on the other hand, if your document allocates a QNEC to all employees, then there is certainly nothing alloacting 3% nonelective to all and a 1% QNEC to all. My understanding is that you don't have to use the QNEC in the ADP (therefore you don't use the HCE QNEC) however when you test a(4) you test all nonelectives and also test all those nonelectives NOT used in the ADP test. That is where the HCE would end up with a slightly larger contribution than the NHCEs. or at least that is how I understand it to work.

according to Revenue Ruling 2004-13, if there are no other contributions besides deferrals and safe harbor, the plan is not top heavy for that year, so it shouldn't matter if the safe harbor starts mid year. there was never any exception written for new safe harbor plans that start mid year.

coverage shouldn't be a problem because a nonelective is still a nonelective whether it is in the form of a QNEC or discretionary. however, if a QNEC is used in the ADP test, then you have a(4) testing. you must run 2 tests. you must pass with and without the QNEC. 1.401(k)-2(a)(6)(ii). in addition, the QNEC can not count toward the gateway minimum if you have to cross test.

for plan years beginning 1/1/2008 ther is no longer a 'requirement' to include GAP period income. I suppose that implies you can continue to add on GAP income, and I will be glad to offer your clients a better deal on refunds if you choose to continue to charge this extra amount. note: the change is for plan years beginning 1/1/2008. since testing is not completed until 12/31/2008, for all practical purposes GAP period income is eliminated in 2009 since that is when you would be issuing refunds. WRERA eliminated GAP under the same conditions. one such comment and I am sure a search would find many others, this one was from Relius "the IRS stated that employers will not need to adopt an amendment to reflect gap period income on excess deferrals until plans are updated for PPA"

coverage should not be a problem in regards to the nonelective since a safe harbor is still a non elective, so that will pass at 100%. however, you would have some receiving 5% nonelective in total, and others receiving an additional 5% ps. that would require nondiscrim testing. you are correct that coverage for match could be a problem since there would be a group not receiving. if all 18 are NHCE, then you would have a ratio % of 25/43 = 58.14%. if you passed the avg ben pct test then you pass coverage. and you are correct, you aren't suppose to change conditions mid year for a safe harbor, so while you can add new employees, I believe you are correct, you would not be able to change the allocation conditions until next year anyway.

this is how EPCRS reads: (3) Early Inclusion of Otherwise Eligible Employee Failure. (a) Plan Amendment Correction Method. The Operational Failure of including an otherwise eligible employee in the plan who either (i) has not completed the plan’s minimum age or service requirements, or (ii) has completed the plan’s minimum age or service requirements but became a participant in the plan on a date earlier than the applicable plan entry date, may be corrected by using the plan amendment correction method set forth in this paragraph. The plan is amended retroactively to change the eligibility or entry date provisions to provide for the inclusion of the ineligible employee to reflect the plan’s actual operations. The amendment may change the eligibility or entry date provisions with respect to only those ineligible employees that were wrongly included, and only to those ineligible employees, provided (i) the amendment satisfies § 401(a) at the time it is adopted, (ii) the amendment would have satisfied § 401(a) had the amendment been adopted at the earlier time when it is effective, and (iii) the employees affected by the amendment are predominantly nonhighly compensated employees. since it says amend to include 'only those ineligible employees wrongly included' I take that to mean it doesn't matter if there were other employees who weren't given a chance, just those that were included. Note: if the only person wrongly included was an HCE you have some real problems since the amendment is supposed to include predominantly NHCEs. since EPCRS offers no other solution (e.g. simply return the deferrals as if it didn't happen) this could imply you should amend to bring in other NHCEs as well, and since they weren't given a chance to defer, then provide them with a QNEC.

I know years ago I had written a report (I think it was before one was available in Relius) - I know that report won't work for me now because I was looking at category codes and stuff and those were zap-mi-fied awhile ago. But I've been using the Relius report (with modifications - put an individual's data on one line instead of a couple, supressed ineligibles, etc) all to cut down on the number of pages that were printing. that report seems to work fine. I don't even remember what what Relius 11 had. at 13 right now and soon to migrate to 14, so don't know how much help I can be. I've actually modified my one page plan spec report to print the top heavy %. maybe someday when you really want to get frustrated you'll update and lose 'Z' = paid out and all those other status codes many of my reports had been using! all that being said: is the overly large balance and exact multiple of the actual balance - e.g 8 times? if so, then sounds like something is wrong with the linking (or grouping on the report) or something similar, and the report is looping through a couple of times (e.g. it knows total end bal = 1000, but then it loops back because of another source or investment and so the value gets double, quadrupled, etc) but I wouldn't expect that to all of a sudden happen unless something was changed.

lets say the person deferred 16,500 not catch up eligible. so 1000 in excess deferral lets say there was a $100 loss, so the 1099 will show a distribution of only $900. however, the IRS has the W-2 which shows deferrals of 16,500, so the irs would expect 1000 in excess deferral on the 2008 tax form. ince the 1099 only shows 900, there must have been a loss of 100 which should show up - though interestingly enough, the way things read is that the person 'may' claim the loss, though I guess they don't have to.

since the amount someone defers is listed on the W-2, the IRS knows how much excess deferrals someone has, therefore that is the amount they expect to see for tax purposes on the 2008 form. according to the ERISA Outline Book (which is basing its comments on Notice 98-32) [11.303 2006 edition] you provide 1 1099R. The amount reported and taxable amount = distribution amount "A seperate statement provided with the form must state the entire amount of excess deferrals (unadjsuted for the loss) must be reported in income for the deferral year, and that the loss may be taken for the distribution year." note the loss under "Other Income" on your tax form. of course this statement is a year after the fact, but hopefully this year you have told the individual you need to claim this as income and next year we will send you a 1099R I guess this implies that the IRS can tell looking at the 1099 (in this case it would be a 2009 form - so 12 months from now) based on the coding that the excess deferrals have already been taxed. The IRS can then compare the 1099 with the prior year's tax form, realize there is a difference and quickly realize that the amount found on the 2009 tax form under 'other income' is indeed negative and matches to the nearest dollar.

interesting. so the gateway is satisfied using the gradual age or service schedule. haven't seen many of those. (well, granted the minimum contrib is already above 5%, but the formula itself meets the gradual service conditions)

you have a document it says the plan is safe harbor not giving a notice does not take the plan out of safe harbor status. at least according to IRS officials at numerous conferences. what you have is a failure to follow the terms of the document, namely, if you have a safe harbor plan you provide a notice. thus you have a possible disqualifying event. (at some conferences the IRS has said providing the notice ASAP under EPCRS is probably ok for the SHNEC, but they indicated a safe harbor match was problematic. one such response is e.g. American Bar Association Committee on Employee Benefits Q and A May 9, 2003 Company A adopts a safe harbor 401(k) Plan. IRS insists that each year that the safe harbor election is used, the employer must amend the plan to provide that the safe harbor contribution will be employed for that year. Is this correct? Proposed Response: If the plan contains a default provision, annual amendment to employ the safe harbor is not necessary. The acceptable default provision provides that in any year where the required advance notice that the safe harbor fails to be given, the Plan is subject to the standard ADP test. The employer can file a copy of the safe harbor notice with the form 5500. This procedure cuts down unnecessary paperwork and is consistent with the statute providing for the safe harbor. IRS Response: The IRS disagrees with the proposed answer. Notice 98-52 requires a notice to participants before the beginning of the year indicating the plan may be amended during the year to provide for a safe harbor nonelective contribution, and Notice 2000-3 provides for some flexibility by providing a supplemental notice to participants and amending the plan to provide for the nonelective contribution by December 1 of the plan year. There is NO DEFAULT OPTION under existing IRS guidance. ......... as indicated, a plan using the 'maybe' option provides some flexibility, but even then some type of notice has to be provided.

the preamble to the final regs clearly says: "Additionally, a plan that uses the safe harbor method must specify whether the safe harbor contribution will be the nonelective safe harbor contribution or the matching safe harbor contribution and is not permitted to provide that ADP testing will be used if the requirements for the safe harbor are not satisfied. The safe harbors are intended to provide employees with a minimum threshold in benefits in exchange for easier compliance for the plan sponsor. It would be inconsistent with this approach to providing benefits to allow an employer to deliver smaller benefits to NHCEs and revert to testing. Accordingly, if, at the beginning of the plan year, a plan contains an allocation formula that includes safe harbor matching or nonelective contributions, these regulations clarify that, except to the extent permitted under §1.401(k)-3 and §1.401(m)-3, the plan may not be amended to revert to testing for the plan year." one of the requirements for a safe haror is to issue a notice, thus I don't see how such document language is possible ..e.g. you are reverting to ADP testing. It is certainly possible the IRS missed something in a prototype review or they read it to mean something else. The language you cited sounds like the language that was in documents when safe habros first came out, but since then I thought all such language was done away with. of course I could be wrong.

based on the regs looks like you are correct, but then the wording in the ERISA Ouutline book makes little sense to say "if 2 plans are permissively aggregated the ADP safe harbor is not available unless the plans treated as a single plan satisfy the safe harbor requirements."

it doesn't matter what the 'average for the ADP test' is. what is important is what the highest allocation rate for any 'key' employee is (not an average of all keys). keys are different from hces. in addition, if there was a match, that would also be included in determining what a key employee received. now, if some of the return for a failed ADP test is treated as a catch-up, those amounts would not be counted toward determining what an individual received, otherwise the returns are counted.

in a controlled group scenario you have 'one' employer so you have 20 employees combined. 18 nhce and 2 hces now, looking at plan A you have 13/18 NHCE % and 1/2 HCE % easily passes ratio % test and plan B 5/18 NHCE % and 1/2 HCE % = 55.55% fails ratio % you have 18 / 20 NHCE concentration % = 90% or a safe harbor of 27.5% since 55.55% > 27.5% you pass nondiscrim classiifcation test. if you can pass the avg ben % test, you wouls pass coverage and not need to aggregate plans for testing.

if one plan is safe harbor and the other is not you can not aggregate and take advantage of the safe harbor. it's one of those all or nothing deals. thus, if you aggregate, you have lost the advantages that a safe harbor plan provides - in addition you have vested people 100%, etc.

below ground: I believe we are speaking on the same terms. I think you make a good point (whether you intended to or not) there is verbage used from time to time that may describe something that is not used by others. so lets say I have a plan with immediate eligibility. 1/1/08 - 12/31/08. 2 ees hired 5/1/08, so they enter the plan immediately. one ees is the owners kid, so is an HCE because of attribution. the other is an NHCE. 1. testing can be run including all people in the test. 2. testing can be run treating the NHCE and HCE kid as otherwise excludable (thus 2 ADP test are performed - in addition, 2 coverage tests are performed) 3. only the NHCE is 'carved out' and treated as otherwise excludable. in this case there is only 1 ADP test, because the otherwise excludable group consists of NHCEs only. this is a special rule found only under the 401k regs, so it does not apply to cross testing, nor is this rule found under the coverage rules.

I suppose one way of explaining it would be as follows: when testing otherwise excludables, the carve out rule only applies to NHCEs. in other words, an HCE can never be an otherwise excludable, no matter how few hours or age they are. They are always includable (once they have met the plan's eligibility conditions. note:this only applies to the ADP/ACP test, not to coverage and not to cross testing. since there are no HCEs in the otherwise excludable group, there is no need to perform a ADP/ACP test for that group.

hopefully you would use consistent logic. ignoring the safe harbor for the moment. consider an ee who terminates 12/28/08. he receives his last paycheck on 1/4/09. do you include him on your ADP test? what if he deferred? remember his final paycheck will indicate he had deferrals in 2009 even though he performed no services in 2009.

if the particpant receives a match they are not entitled to (e.g. comp wasn't limited) that match would be forfeited. if you fail ACP test, then you have an excess aggregate match, and that is distributed to the participant. if the person is not 100% vested, then you can either: 1. distribute the vested protion and forfeit the unvested portion or 2. distribute the amount and then you would need to track the vested portion portion properly. e.g. if after the distribution the person had $1000 balance and was 60% vested, the vested balance would be less than $600 becuase the person, for lack of a better term, received some unvested $. This is a good deal for the person, because as they continue to vest eventually the unvested protion disappears. if the match is a 'related match' due to a distribution of deferrals, then it is forfeited, but there is nothing to stop you from fixing the ACP test before the ADP test, and thus the match might not be forfeited.