Mike Preston

PS. I'm assuming that 5/31 is the limitation year.

The purpose of the compensation information is to allow the IRS to both check the reasonableness of the benefits and to ensure that non-discrimination isn't an issue. If you really can't get the information, I would submit it without it and annotate the 6088 with an explanation as to why the information isn't being provided. Don't be surprised if the IRS accepts it. Don't be surprised if the IRS rejects it.

Assuming compensation is at least $44,000, the answers are yes to both questions.

Egads. Something tells me this comment was in a different form in SoCalActuary's head and it morphed into the above somewhere between the brain and the keyboard. Under no circumstances can you release the employee predicated on their potential accrual as being too expensive. That is, on its face, a violation of ERISA 510. Don't do it. But whatever you decide to do, PAX has the right idea.

I didn't think that a plan formula that was based on service but pro-rated on participation qualified as a safe-harbor? Then again, from the description given, I'm not sure that this is a plan that bases the benefit on service at all.

Most people think that following what the IRS tells us at conferences is at least arguably the correct thing to do.

Not "similar to", "precisely like".

It is almost impossible in most cases to determine, as contributions are being made, whether they are catch up contributions or not. Sure, for contributions in excess of the 402(g) limitation it might be possible, but for all other sources of catchup contributions, such as amounts contributed in excess of the adp test limitations or amounts contributed in excess of a payroll limitation (say, 6% of pay), you really can't tell whether the violation has taken place until the end of the year. Hence, the IRS has stated in their regulations on this subject that there is no need to keep track of contributions as they are made as being normal or catchup, since any determination being made during the year might (in fact is most likely to be) modified at the end of the year when the full facts are known. Further, the IRS has made it acceptable to meld catchup contributions into regular contributions as far as account balances go. For things like top-heavy testing and other account balance determinations, they have said that catchup contributions are included to the extent they aren't ejected from the plan. So there really is no reason I can think of why a plan sponsor would want to keep track, separately, of catchup contributions.

I think you are looking at the ramifications of this issue in a myopic manner. If the only issue was adp/acp testing, I might agree with you that using the higher threshold is very likely to help the testing, as those between the higher threshold and the lower threshold are likely to be high level contributors. But once the plan sponsor has determined HCE status, that is used for much more than just adp/acep testing. It is random as far as direction goes when we are dealing with the non-discrimination tests (other than adp/acp, although even those are not set in concrete). That is, in my prior message, I could just as easily have described a situation where the use of the lower threshold caused the plan to fail non-discrimination and the higher threshold would have allowed the plan to pass the tests. Whichever way you go, there is the possibility that had you gone the other way, the plan sponsor would need to provide larger contributions for the NHCE population in order to satisfy the 401(a)(4) or 410(b) tests. I will not buy into the red herring that the administrator making a determination as to which is the correct interpretation falls into the trap of a fiduciary violation by making the decision which reduces an employee's ability to defer. That is just bunk. There is a correct and an incorrect way of doing things. The IRS is inconsistent in their communication as to which is the correct way of doing this particular thing. I hope that once the IRS recognizes that their communication has been faulty in this area they issue something which gives plan sponsors comfort no matter which way they have done it through a given time frame and then they clarify which way they want the determination to be made from that point forward.

Boy, talk about mixing issues. It sounds like you want confirmation that they above satisfies 410(b). Sure it does. But it would fail 410(a). See the example in the regulations. And to answer your prior question, whether the plan has the average benefits test available to it for 410(b) (which, for the life of me, I don't understand why it would need it), it would depend on the provisions of the plan as to how the exclusion of the HCE and the 2 NHCE's was accomplished. Excluded by name? No way. Excluded by way of the description above, sure, but remember, that would fail 410(a) so it isn't much of a victory. If you want the average benefits test available, you need to ensure that the plan document provision which operates to exclude the ineligible participants does so on the basis of what the IRS calls a "reasonable classification".

"Just when I think I'm out, they d...r...a...g me back in!" What makes you think this is a "no harm, no foul" situation? Take 2006 for example. Let's say my plan sponsor decides that everybody earning more than $95,000 in 2005 is an HCE for 2006. Your plan sponsor decides that the threshold must be $100,000. Let's say there is an individual that makes $97,500 in 2005, and this person, by reason of the class he is in, receives no contribution (note: the class is not based on him or her being an HCE or an NHCE). On my non-discrimination test, this person is an HCE and the test passes. On your non-discrimination test, this person is an NHCE and the test fails. How is this a "no harm, no foul" situation?

Especially if you use compensation while a participant as your testing compensation! Note that I'm not crosstesting so the age means nothing.

I also suspect that there is a third way you can pass the average benefits test. If you run the test by excluding all people who are otherwise excludable, I think you will also pass. I didn't run it this way because there was a non-controversial (famous last words?) method available to you. But just as an exercise, you might consider breaking your population into two groups, one of which is statutorily excludable and the other isn't. Then run the average benefits test on those that are not statutorily excludable and see if it passes. As I said, I *think* it will. There are a number of people who have heard the mantra that there is always one and only one average benefits test and that you, perforce, include all benefits in that test. While this is usually the case, there is a somewhat complicated analysis that has been published on BenefitsLink before that makes it clear (to me, anyway) that when you are testing compliance with 410(b) by permissively disaggregating the statutorily excludables from the non-excludables, you also end up performing two ABT's in that circumstance. Note that we didn't even begin to get into issues such as using the accrued to date method!!!!!

As I suspected, it looks like the plan passes. You even have a choice as to which method to use. First, you can use a non-controversial method, if compensation of those eligible purely for the TH contribution is pro-rata for the year. Re-run your average benefits test on a contributions basis. Change the testing compensation to "compensation while a participant". Make sure your system has that information. In this case, assuming somebody was hired on or about 7/1/2005, they would be a participant for 3 months, but they would receive a TH allocation of 3% of entire year compensation, which results in the Average Benefits Test treating this person as if they were getting a 6% of pay contribution. Then, when you impute permitted disparity, they get an additional 5.7% on top of that. I am editing this message to highlight the fact that you need to determine actual compensation for the period of time that the individuals were participants in the plan, and that using a pro-ration (as I did) is not appropriate in a "real" test. If my numbers are correct based on what you sent me, you should find that the average benfeits test is something like 9.74%/12.72% = 76.57%. Therefore you pass the average benefits test and you therefore pass 410(b) with only 9 out of 16 NHCE's "benefitting" in the profit sharing portion of the plan. Second, you can use the method that I think works, but that the Nut says doesn't. That is, you test on the basis of benefits in your average benefits test (you can even not use "comp while a participant" if you want) and it will pass. My numbers are 4.34%/5.77% = 75.2% using full year compensation and 5.22%/5.77% = > 90% using comp while a participant. Are we having fun, yet? If you run the average benefits test on the basis that I have specified and still don't pass, send me your report output and I'll try to identify where your results are different from mine.

I can see no good reason for precluding a non-U.S. citizen from being the Trustee of a qualified plan. However, and this is just conjecture on my part, just like the assets of a qualified plan must be, in certain respects, domiciled in the United States, I would think the same would hold true for a Trustee. That is, it is necessary for qualified plan assets to be subject to the jurisdiction of the US district courts, unless the regulations carve out an exception. Hence, investing in foreign assets is pretty much, but not completely, limited to those securities which are traded on a US based stock exchange. Mutual funds can be established for this purpose. At the far and of the spectrum a special class of investment known as an "ADR" (I think I've got that right, but I might be off a bit) can be established to give the US Courts some jurisdiction. In any event, this requirement is set forth in ERISA Section 404(b) and the regulations thereunder. So my guess is that as long as the individual's actions are subject to being overseen by the US district Courts, I would think that person could be a Trustee. As the appointment of a Trustee is, in and of itself, a fiduciary action, I would think appointing a Trustee whose actions are not subject to being overseen by the US courts is a dangerous fiduciary action to take. I couldn't find, after a quick look however, any specific mention of the requirements for being a Trustee, though, so the above is just a guess.

Where did I mention DB/DC combo? Thanks for the chuckle. I suggest you contact the IRS in Washington to determine whether your interpretation is correct. Either that, or Mr. Tripodi directly.

Perhaps you should think about the process and imagine why an exact analysis on the figure provided would be impossible without knowing the earning piece and then jump in and answer it yourself; since you think you know everything. I would be surprised, based on the way the question was put, if the amount you assumed had earnings in it, was representative of anything other than deferrals to be refunded. I will leave it to others to determine whether your assumption was appropriate or not.

Yes, it must be reclassified; and leaves him a deferral limit of $20,000 (less the reclassified catchup amount) for 2006. This would be different had he deferred during the first part of 2006, as his deferral limit for the year would have remained at $20,000. I am scating around the numbers as I do not know the "earnings" portion of your figure. Can you explain how the "earnings" portion of the figures makes any difference, at all, to the determination of remaining catchups? Or how said portion might be relevant to the poster's question at all? She gave a figure of 1101.50. I didn't know whether that figure included an earnings calculation or not. If that figure was $1000 in contribution plus 101.50, then you must reclassify the $1000 as catchup for 2006 and the earnings remain. Hence, I scated around the numbers while the principles remain in tact. Didn't we have a discussion about cherry picking statements? If you (or anybody else, for that matter) make a statement that I think needs clarification, I will ask for that clarification. You can choose to provide it or not.

If the plan is so inclined, they can assist the participant in escaping the withholding. First, have the Trustee borrow from the policy such that its surrender value is essentially zero. Then have the participant obtain a participant loan from the plan in the same amount, and on the same terms, as the plan arranged with the policy (if possible). When it comes time for a distribution (one day later?) the plan will distribute the $80,000 as a rollover and a life insurance policy with little or no value, along with a participant loan in the amount of $20,000. The participant can then repay the insurance loan, restoring it to the condition it was in before all this took place. The distribution from the plan was not subject to withholding since the participant loan is a qualified offset.

Count me on the side of those who believe the directed trustee that does nothing more than notify the plan administrator of the potential default is at risk. Let's start with a threshold question: are the loan provisions available to the directed trustee? Outside the ERISA area, a directed trustee would be hard pressed to make a loan without seeing a copy of the documentation and maintaining a copy in their files. If we assume that the directed trustee has a copy of the loan, then I think the Trustee is required to apply to provisions in the absence of intervening instructions to the contrary from the plan administrator. I say the directed trustee should notify the plan sponsor of impending deemed distributions and follow through, including the issuance of 1099's, unless they get instruction to the contrary from the plan administrator.

I agree with the majority opinion that this is a "no big deal" and that the course of action should be a simple notification to participants with a method for them to obtain a new amortization schedule. Just one word of caution: the rules require that the date of the last payment of the loan under the new schedule not be later than the last payment that would have been allowed had the original loan been made for a 5 year period (in the case of a loan not for the purchase of a residence). Take a person whose 5-year loan is currently scheduled for a final payment on 12/31/2008. If you find yourself with Friday paychecks such that your last payroll in 2008 falls on December 19 and that the next payroll is January 2, you must ensure that the last payment on the new schedule is made on December 19, 2008, not on January 2, 2009. If you reschedule the loan such that the last payment is to be made on January 2, 2009, I think the IRS could rightfully say that you extended the loan beyond the 5-year repayment cycle and that the full extent of the loan became taxable when the reamortization was initiated.... sometime in 2006. I know this seems ticky-tack, but it is really a simple matter to ensure that this rule is followed.

Vicki, unless there is a reason to classify a portion of the deferrals made from 5/1/05 through 12/31/05 as catchups as of 12/31/2005, they aren't catchups as of 12/31/2005. You might want to get a copy of Sal Tripodi's wonderful worksheets on determining catchups in non-calendar year situations and then put them into Excel. You will have a fantastic tool at that point that can really bring clarity to questions such as this.

Yes, it must be reclassified; and leaves him a deferral limit of $20,000 (less the reclassified catchup amount) for 2006. This would be different had he deferred during the first part of 2006, as his deferral limit for the year would have remained at $20,000. I am scating around the numbers as I do not know the "earnings" portion of your figure. Can you explain how the "earnings" portion of the figures makes any difference, at all, to the determination of remaining catchups? Or how said portion might be relevant to the poster's question at all?

The average benefits test needs to parallel your 401(a)(4) testing methodology when dealing with entry date provisions. This means that there are, indeed, two average benefits tests in the rare circumstance that you are permissively disaggregating the statutorily includables from the statutorily excludables. This has been discussed at length on BenefitsLink before, and given that it is well settled, I'm confident that the NUT will find some reason to say that everybody here is wrong on that score, too. In the vast majority of cases, it is in your best interest to run a single average benefits test, though. This is because you end up with those 3 late 70's birth dates in your test, which when applied on a crosstested basis, the test almost invariably passes. But that brings us full circle to whether doing such causes the gateway requirement to kick in for you. I couldn't quite understand all of the numbers you posted, so I'll elaborate here, if you don't mind. You say that there are 22 non-excludables, 16 of which are receiving an allocation, and 6 of which are terminated with more than 500 hours. You also say there are 7 people in the TH group and 4 in the not eligible group not otherwise included in the 6 who are terminated with more than 500 hours. That means you have 22 plus 11, which is 33, but you say you have 36 total employees. Can you tell me what I've missed? And while I'm at it, and I know this is going to sound sort of condescending, but there really isn't any way to mention this without doing so, you say that the Average Benefits is 4.12/6.73, for a percentage of 61.2%. This *was* run by including the deferrals, right? If not, please rerun it with deferrals and let me know what it looks like. Sorry, but I had to ask. Oh, one more thing. Your plan seems to have wording that appears to allow for a th minimum in excess of 3%. I think you will find if you look into it, that the language you are looking at is intended to cover a circumstance that you aren't in. For example, if the plan were part of a DB/DC combo you might find that 5% is the top heavy minimum. I don't think you can, arbitrarily, increase the th minimum to something higher than what it really is. So if you only have one plan, 3% is the max. Now, can you amend your volume submitter to define the TH minimum as 4% notwithstanding the fact that this will be an amendment which is outside the approved volume submitter thresholds? Sure. And if the IRS considers it a non-substantive change, you can even keep your volume submitter status. However, you would lose your reliance on safe-harbor because I don't think the safe-harbor exception extends to "just any old formula". I think it only extends to the TH requirement. I might be wrong on this, though, so somebody should look that up. Losing your safe-harbor isn't the worst thing in the world, though, because as we have seen, your plan will satisfy 401(a)(4) easily once you satisfy 410(b).

Which brings us to the real world application where the business owner hires the beautiful young gals (legal age of course) to be his receptionist and payroll clerks and then tells his spouse that he has to otherwise they wouldn't be able to maximize their retirement plans. I'm sure it's not the intent of the IRS to do that...but who am I to argue with the results. But I digress....Mike, what you are saying is that as you get smaller and smaller doesn't it become necessary to become flexible with entrance requirements to ensure 410(b) passage, correct? flexible as in immediate eligibility? Actually, there is a real world case where somebody, somewhere put themselves in the business of arranging for the temporary hire of high school age individuals for the sole purpose of satisfying the tests. This is, even in my opinion, abusive and is what the previously mentioned (by me) Gold Memo targets. However, in your case, if the receptionist is a legit employee, the husband is making a strong case, in my opinion. Great example. I'm not sure I agree with the rationale, but I agree with the conclusion. Certainly, since this is a demographically influenced test, as the numbers get smaller one has increasing variance. If just one participant is different from the prior year, the testing results can be dramatically different. So, yes, if the employer is willing to entertain the notion that it may be to the employer's advantage to relax the statutory entrance requirements in a given year solely for the purpose of satisfying the tests, I think that flies. The only restraint being that one can not make an amendment which is discriminatory. For example, you can't amend the plan to relax the eligibility requirements in the year that an individual buys into the practice (so that they are an HCE) and bring that person into the plan at a low rate to help your test. The way I understand the IRS' interpretation of a discriminatory amendment, it is always viewed on its own (and now I'm talking about a regular amendment, not an amendment made after the end of the year - those are called 401(a)(4)-11g amendments and they have a separate threshold to satisfy). Hence, bringing in additional NHCE's to pass your test will never be considered discriminatory. Of course, the Nut will argue with this, but that is his problem. I will agree with most of what he has said so far, though, because he has not touched on the real issue we were discussing, and other than that issue, his comments are relatively accurate. I'll respond to the ones I find questionable in another post. I'll hit a number of your individual questions here, though. Your insight into the ramifications of juggling benefits for the sake of satisfying the tests is crystallizing, based on your questions. This is a good thing. I'm still a little unclear on the plan's design and without clarity, conclusions are always suspect. I know that you have a 3 month eligibility for deferrals, but I'm not sure whether the eligibility for a PS contribution continues to be the statutory 21/1 or whether a participant who entered the plan in the year in question would be eligible for a profit sharing contribution if they had 1000 hours. I'm guessing that it is the former, not the latter (however, it rarely works that way in practice, because of entry dates - take the full time person hired 12/1/2004 and a 12/31/2005 plan year end - even if the document calls for 21/1 most still make this person eligible for the ps, even though if the statutory provisions were applied, entry wouldn't take place until 2006). For now, let's assume that there is nobody that upsets the apple cart like my hypothetical 12/1/2004 hiree. So, getting back to your insights, you are correct that your hypothetical 2080 hour employee hired in the first month of the year would indeed be eligible for the TH minimum because they were active participants in a TH plan on the last day of the year. The NUT wants you to bring in employees who are the most expensive first, in some vain attempt to satisfy his own ethereal definition of what is discriminatory. While that will most assuredly work (no way will the IRS object), unless you are bringing in people who are short service (see the Gold Memo previously referenced), you really have nothing to fear. Bring in those that are the least costly and you should be fine. If that doesn't sit well with you or the plan sponsor and the cost to put on a belt and a pair of suspenders is low, I certainly wouldn't argue against something more generous. As to your ability to "discriminate in that regard", you are free to design what you want. The requirements depend on whether the amendment is in advance of the plan's year end or in arrears. If it is in advance, it merely must not be a discriminatory amendment. If it is in arrears, it must not only be non-discriminatory, it must meet the requirements of -11g (which basically means that it must pass 401(a)(4) on its own). If you have any doubts about the propriety of such an amendment, submit it to the IRS for approval. This is where it is a judgement call. In your case, you have so few participants required to receive an additional allocation that it may not make sense to do anything other than what the NUT has suggested: be conservative, add more money to NHCE accounts than you might otherwise be required to do and sleep well knowing that the plan sponsor has saved money by not having to go through the effort of submitting. I agree with the NUT that pushing two participants from TH minimum only to full bore participation in the employer's contribution is one way to skin this particular proverbial cat. But I resubmit that I am not convinced that your plan fails 410(b). If it passed 410(b), then it is a safe harbor design and you don't have any correction to make. To complete the analysis, just make a list of the 16 non-excludables who are benefitting under the regular profit sharing allocation and list their: 1) Year of birth 2) Compensation 3) Allocation (including forfeitures, if any) 4) Deferrals Do the same thing for the the others who are in the plan at any point in time during the year (whether or not terminated), but add the following information: 5) Hours worked during the year I think your case is a great case study and I'll be happy to run the real test for you. I may be wrong, but my gut is telling me that you will pass without need of further contributions. Famous last words, right? Note that even with the above information I can not run all of the tests that I would normally run. To do that I would need 12/31/2005 account balances and compensation and allocation history for 2004 and 2003. That would allow me to also run the accrued to date test. I'll leave that one on the sidelines, if you don't mind. Don't be concerned about the flap here on BenefitsLink. It is certainly not your fault. The NUT is looking to me to be more and more dangerous, because much of what he writes is indeed correct. That lends credibility to his silly concepts of ignoring the regs and code in favor of some imprecise standard that only he can determine. I consider that extremely dangerous and will rail against it at every opportunity (as I'm doing now). If you want to send me the information in a private message rather than post that information on here, I can understand that. If you don't want to send out that data, even though it is unattributed (without participant identifying information) I can understand that, too. Whatever you decide is fine with me.