Tom Poje

I haven't seen this addressed before, but it produces an interesting anomaly (probably not thought about by the IRS). for example lets suppose the person defers (though it doesn't matter, but it probably helps when thinking about it) for tax purposes, he has a W-2 in 2013 with deferrals : implication he worked in 2013! he will also show up on the ADP test. If the plan was safe harbor he would receive a safe harbor in 2013. In fact most plans are written a person receives a profit sharing in the year they retire, regardless of any hours condition. Over the years I've come to hold onto the idea that while a plan is a calendar year, the determination of when someone quits (consider someone who quits 12/27/2012) is based on the compensation that falls within the plan year, and thus, in this case, though the person physically retired in 2012, for all other plan purposes he retired in 2013, so it would actually kick things back a year. of course even that depends on how you checked your 415 amendment whether to count the comp in the prior or current year sample language: (I) Compensation Earned in Limitation Year but Paid in Next Limitation Year. If this paragraph (I) is checked, then effective as of the first day of the first Limitation Year beginning on or after July 1, 2007, Code §415©(3) Compensation for any Limitation Year will include any amounts earned during that Limitation Year but not paid during that Limitation Year solely because of the timing of pay periods and pay dates if: (i) these amounts are paid during the first few weeks of the next Limitation Year; (ii) the amounts are included on a uniform and consistent basis with respect to all similarly situated Employees; and (iii) no Code §415©(3) Compensation is included in more than one Limitation Year.

as a real simple example. let's say you defer $100 (pre tax). you immediately turn around and take a loan for $100 and stick it in a drawer at home, ready to spend the next day but you forget about it. Let's suppose somehow there was no interest charged. one day months later you trip across the $100 in the drawer so you pay back the loan. Certainly no double taxation on the $100. It's because you put $ in pre tax and then took thenm out without tax, it just doesn't seem that way when you eventually pay the loan back. technically there is a double taxation on the interest, though if you find one of the links rcline mentions it is supposedly minimal. I hadn't thought about the Roth issue, nor have I read anything about it, though I guess since its a Roth account you would avoid the issue of double taxation on the interest. I think the reason for a negative attitude toward borrowing from Roth is if you fail to pay things back (e.g. you quit) then you have the penalties that accompany early withdrawal from a Roth.

This type of question makes my brain hurt, so lets make sure of the details. It sounds to me like you said you have 2 companies, each with its own formula. In addition, there is a 3rd company you have no details on, and whoever processes that plan actually has at least the census details on your two plans, because the other TPA did some type of testing. well regardless. and it sounds like the other firm has put its stamp of approval that the whole thing passes coverage, but it was not indicated if this was done with disaggregate testing or combined testing. If it was tested disaggregately for coverage, then it had to pass either the ratio test (treating other employees as includable and not benefiting) or it had to pass the avg ben test. this includes the avg ben pct test, which means they had to know what the contributions were from the plans you run. There is one and only one avg ben pct including all employees and all contributions regardless if you choose to disaggregate the plans. (The exception being if you choose to disaggregate otherwise excludables) if they aggregated the plans, then any nondiscrim would be also aggegated, and, assuming these are also 401k plans, that means combined ADP tests, and something in my head says that didn't happen, so we will assume disaggregated plans. for ADP and ACP tests you only include people actually eligible, so you never see the people from the other plans. for nonelective testing, my understanding is everyone else does indeed show up as 0 if you disaggregate. If plan A allocated 5% and plan B allocated 8% and each passes coverage on its own, then the results should be the same when testing on an allocation basis, which is what, I believe ERISA Toolkit is getting at.

I looked it up (just to make sure I wasn't imagining things. the document could read something like: Excess Elective Deferrals and Excess Contributions Not Required to Be Matched. Notwithstanding the above, to the extent Non-Safe Harbor Matching Contributions (including Qualified Matching Contributions) are contributed on an annual basis, no Non-Safe Harbor Matching Contribution (including Qualified Matching Contributions) will be required with respect to that portion of an Elective Deferral which for that Plan Year is determined to be either an Excess Elective Deferral (unless the Excess Elective Deferral is for a Non-Highly Compensated Employee) or an Excess Contribution. ............ of course in my opinion, this only makes sense if the match would be forfeited because it was related to the deferral. (And as a previous post pointed out, if you match catch ups then its not forfeited) In addition, you can always run the ACP test first, so if the test fails you can have a distribution.and if the HCE receives the distribution, then it's only to his benefit to receive the addition $100 or so anyway

check plan document. I have seen some that actually contain language that allow the match to be limited if it would cause ACP failure. (or at least I think I recall reading that in a document)

no clues, have never heard the issue addressed. obviously you do a 'good faith effort' in absense of any direction from the IRS in the regs. I'd think as of the date of the merger company A ceases to exist, so the safe harbor ceased at that point and was satisfied.

does the document require the match to be used in the ADP test? I have seen documents written as such. however, the regs only say 1.401(m)-2(a)(5)(iii) QMAcs used to satisfy the ADP test - QMACs that are taken into account for the ADP test are not taken into account in determining the ACP. so again, unless the document is specific, you can use the match in any way shape and form you want - part in adp and part in acp test

the regs say yes, plan is not considered to be top heavy. at some conferences they have indicated such. still, some, but not all, IRS reps have expressed an opinion that if it is entry date compensation there may be issues, but again that was only an opinion. if the plan provides that otherwise excludables receives no safe harbor then the regs are clear the top heavy free ride is lost, but there is no such provision for comp from date of entry plans. I think that means means some IRS officials have misgivings about the comp from date of entry issue, but there is little they can do about it.

from 2010 ASPPA Conference Q and A (1st page) (if it helps) DC plan is top heavy and has a plan year ending 12/31. The plan terminates on September 15, 2010. Normally, TH minimums are provided only if the employee is employed on the last day of the plan year. (Assume that there are salary deferrals during the year so that, if a top heavy minimum is required, it needs to be made.) Questions: (1) For the 2010 plan year, is 9/15/2010 treated as if it were the last day of the plan year, so that only non-key employees who are employed on that date are entitled to a TH minimum? (2) If (1) is Yes, is the 3% minimum calculated for compensation from 1/1/2010-9/15/2010? (3) If (1) is No, is there NO top heavy minimum for the 2010 plan year because the plan terminates before the end of the year (similar to the concept that there is no money purchase plan funding if the plan terminates before the end of the year and there is a last day employment requirement), or does the plan have to wait to see who is employed on 12/31/2010 to determine who is entitled to the TH minimum, even though the plan has terminated before that date? (4) Is the answer to any of the above affected by whether the employer continues in existence through the end of 2010? response (1) Of course, if there is no employer contribution, there would not be an obligation to provide top heavy minimum contribution. But, if there were contributions to keys during the year, including elective deferrals, there is a top heavy minimum based on compensation and employment through 9/15/10. Plan must liquidate within a reasonable time under Rev. Rul. 89-87 or else 9/15 date may not be reasonable. There is effectively a short plan year for top heavy purposes. (2) yes (3) n/a (4) no change

ha, I still have 2 years to report any 'significant' changes. My quick look at the section under the 'changes' or 'modifications' didn't find anything special. oh well.

guess it's a New Years gift to us. just so you don't have to hunt it down EPCRS_revproc2013_12.pdf

if everyone has the same retirement age and you do not impute disparity, then it makes no difference which mortality table you use - everyone will have the same APR so it becomes a constant. if you impute disparity then in a DC plan using 1983 IAF could make just enough of a difference. also, in your case, bevause the retiremnt ages are different then choice of mortality will make a difference because the people will have different APRs

there are 2 basic tests every plan must satisfy 1. coverage - and one way of passing coverage is by the average benefits test. there is nothing under 410(b) requiring the gateway. the gateway is under 401(a)(4). thus one can pass the avg ben test without worrying about the gateway minimum. The IRS confirmed this in at least one Q and A session, but sorry, at the moment I am too lazy to look it up. 2. 1.401(a)(4)-9©(3)(ii) restructuring not available for certain testing purposes ......in addition the minimum alloaction gateway cannot be satisfied on the basis of component plans. so if you are going to test at least one HCE on an accrual basis, then before you even get to that point, each NHCE who has received a nonelective must receive a gateway. then you are able to use whatever means at your disposal to pass nondiscrim.

again, since this keeps getting asked 1.401(k)-3(d)(3)Timing requirement (ii) the timing is deemed satisfied if at least 30 days before beginning of each plan year. then go back and look at (i) the determination of whether the notice has satisfied the timing requirement is based on facts and circumstances. so, can you give less notice? apparently less, it is just you have no guarantee if it is less. what is the purpose of the notice? so someone can make an informed decision as to wether to defer or not. so if the safe harbor is a match and you are now 'removing' it, then it is probably more of a problem than a safe harbor nonelective if it's a mom and pop shop and you only have to tell 3 people of the change, its probably less of a problem. if its a larger company (and adding the fact a number of people might be off for Christmas vacation) its probably more of a facts and circumstance problem.

the IRS has comments on this in 2 different spots, though they don't address the issue of withholding. neither of the comments says anything about 'unless the document says otherwise' http://www.irs.gov/Retirement-Plans/Plan-P...ibutions-(RMDs) The minimum amount you must withdraw from your account each year is called your required minimum distribution. • You can withdraw more than the minimum required amount. • Your withdrawals will be included in your taxable income except for any part that was taxed before (your basis) or that can be received tax-free (such as qualified distributions from designated Roth accounts). http://www.irs.gov/Retirement-Plans/Retire...Distributions#7 Can an account owner withdraw more than the RMD? Yes. .............. thus it would seem to say the regs say 1.You have to take a distribution at age xx. 2. It has to be at least $yyy. (but there is nothing to stop it from being more ) 3. taking more does not reduce the required amount in a futre year (except for the side effect of a smaller account balance) I'd assume this implies you could not roll over the 'excess' portion of the min distribution (else that portion wouldn't be treated as a minimum distribution) and if you tried, hopefully the document has an in service distribution feature that would permit you treat it this way - and then in that case I'd say you would have 2 withholding rates. but then, my theories are often off the mark.

if the person receives no ps or forfeitures then the general rule is you do not included the compensation. if the person received a match because they deferred then the compensation is included. if the person had deferrals only (or for that matter, could have deferred but didn't) it is a debatable issue. I've heard arguments either side. years ago, deferrals counted toward the deduction limit, so at that point you would include someone who could have deferred but didn't because they were considered benefitting. now that deferrals are no longer included toward the deduction limit the answer is a bit clouded. (e.g. is it a natural consequence that logically you should not include someone who receives no other contributions (besides the possibility of deferrals) since the deduction rules changed)

I am assuming you mean for 2013. well, if a notice is provided at least 30 days before plan year begins you have 'deemed satisfication of the timing requirement' 1.401(k)-3(d)(ii) if it is less than that, then it is really a facts and circumstances issue. will all employees have a chance to make an 'informed' decision before the end of the year? especially with Christmas holidays and the like? I think there is a good argument the answer is yes in a small office situation, arguably the more people involved and less 'contact' if that is a good term, then the answer gets like my hair - grayer.

well, never sure if my brain is functioning at 100%, but as I recall... assuming the plan is not pooled, because the rules are slightly different as I recall.... any one who has terminated has a distributable event, so loans are treated just like any other distribution. ha, watch you have a document that makes you wait until after the end of the year before you can take a distribution! and we will assume the people are beyond and grace period on the loan. so, anyone who quit has a distributable event and gets 1099 R with whatever distribution code. it is not a deemed distribution, the person had a triggering event so an actual distrubution, even though you don't see it at this time. anyone who is still active, has no distributable event, so now they have a 'deemed' distribution. this also requires a 1099 R but you include L in the distribution code and report it on a different line on the 5500. you also keep track on paper of the loan incurring interest, and if the person ever ever ever tries to take another loan it counts as an outstanding loan balance. my brain doesn't recall what happens if the plan only allows one loan at a time! well, that is what my brain remembers. or you could wait and see if those claiming the Mayan calendar end of the world is true and then the problem will go away shortly.

you are correct worst case: plan is top heavy so they are due top heavy minimums. testing: most likely no problem since you could test otherwise excludables separately, and since it's unlikely there is an HCE in the group its not an issue. even worst case: if the plan has 70 people in it an additional 49 (assuming all still active) would bring the count to 119, which is getting near the form 5500 audit level (assuming of course the count has never been over 100 in the first place.)

but again, for those individuals born 1936 or earlier, the definition of lump sum, according to the IRS (it's in publication 575) the entire balance distributed in a single taxable year (I guess as opposed to one single sum) in fact, if the company had 2 profit sharing plans the requirement is still the "entire balance", thus you have to take the distribution from both plans, and I doubt you would only get 'one' distribution check.

Then the Mayan calendar would run out and the world would end...no wait I almost forgot, Today is Popeye's perfect woman day - a '36' - because Olive Oyl is 12-12-12 ....................................... Many many many moons ago, lump sums qualified for 5 year averaging (pre 1999) As I recall, 10 year averaging still applies but only if you were born before 1936. so you would be 'dumb' to take 50% in Dec 2012 and 50% in Jan 2013 because it blows the lump sum out of the water.

Had to look it up, just to see if I remembered correctly from an old ASPPA test from the IRS website: A lump-sum distribution is the distribution or payment, within a single tax year, of a plan participant's entire balance from all of the employer's qualified pension, profit-sharing, or stock bonus plans. All the participant's accounts under the employer's qualified pension, profit-sharing, or stock bonus plans must be distributed in order to be a lump-sum distribution. so I always took that to mean that Gil T. S. Charged could take 50% of his balance in January 2012 and the remainder in December 2012 and it would still be considered a lump sum distribution.

and actually, once you file the short form you can file the short form every year until you go above 120 (not 100)

sorry, I don't think there is a way to code that formula. - in addition to a % based on years you also have a different cap for each grouap, and I don't think that is available. (unless something like plan specs coded 25% for 2 yrs svc, limit match to $400. run the transaction the plan specs coded 0% for 2 yrs svc, 50% next 3 yr svc, limit match to $800 run the transaction for match. then run 0% for 2 yrs, 0% for next 3 years and whatever. for the next x yrs. you still can't get around only having 3 groups, and it would take running three separate transactions. Are you to calculate the match, or is the payroll calculating and you are verifying things. you could of course calculate outside of Relius and then import into census. you will also have to perform a benefits rights features test. (if you are any good at Crystal, I have a report that checks BRF based on years of svc -it is plan specific based on the years of svc in the document but that could be modified. and another report that does an end of year check yo see if people received what they should have - again, would have to be modified for whatever formula is being used,

remember, you have 3 coverage tests, not one. 401k, 401m, and nonelective. and each of these coverage tests is a separate animal from the other, so you could test one using the otherwise excludable option, while not doing do in the other cases. if its a SHNEC then it is treated as a nonelective if its a safe harbor match it 401m.