Tom Poje

plus (if I remember correctly) it would eliminate any possibility of the shift of deferrals, but he does have a valid point administrators should possibly consider for plans with a discretionary match.

ERISA Outline Book (2003 edititon, 7.220) gives some examples. Adapting to your case: 1.ee could rollover 80,000, no withholding 2.ee could rollover 76,000 and then have 4,000 as a distribution he would never see because it would go for withholding. either method is fine. In your example you indicated the participant wanted 80,000 rolloed over so method 1 would have been the better method

1.414(q)-1T, A-3(b) any rounding method permitted, as long as you ar consistent (Minimum participation you have to round up, nhce concentration % you have to round down, but there is no restriction on HCE top paid group determination) of course, if you are an actuary, my understanding the answer is simply "What do you want it to be"

or simply define your classes as: 1. owners not by attribution 2. owners by attribution 3. all others

I remember that article well. Harwood is correct- the article did indeed appear in the Journal of Pension Benefits of Autumn 1999. I almost wrote one 'why I disagree with most of the points' About the only point I could truly agree with is that if plan uses a discretionary you are almost stuck with using current year, especially if client doesn't make a match one year. On the other hand, I have had very few plans (if any) where this has occurred. Ditto for QNECs. I think I have only one client that ever wanted to use the QNEC option. generally they are just too expensive. Example, from the article: point #1 said "The current year method almost always produces better results than the prior year method (meaning that the ADP for the NHCEs is usually greater under the current year method than under the prior year method...)" I question that statement. I have never seen any proof for that. My experience is that it simply varies from year to year, maybe around a happy medium. But to say the NHCE avg is almost always higher each year???? By that argument, if the avg is only .5 higher than the prior year, a plan that had an avg of 2.00 10 years ago, should be at 7.00 right now. Perhaps I am in the minority, but I still prefer prior method. At the busy crunch time in the early part of the year, I only have to get the comp and deferrals on the HCEs and I am done. I don't have to worry about the NHCE data at that moment in time. And therefore I can run the test a lot easier before 2 1/2 months are up. Perhaps the smartest client I had deferred up to the 402(g) limit each year, and simply accepted whatever refund he received. End result: that meant he deferred the maximum he could into the plan each year. In that case it wouldn't matter what method was used.

you might try reading some of the old discussions found here - Q and A 38 - 42 deal with improper return of excess contributions. There is some 'dust' on these, some of the discussion pertains to the self correction program when it first came out, but my guess is that most of the arguments are still valid. in particular see #41 - not recommended to make plan whole by making an additional contribution. In effect, this would give the HCE a 'free' nonelective contribution. *cough* gasp, I churned up plenty of dust just looking through my notes where to find the stuff http://benefitslink.com/qa_columns/plan_defects/archive/

well, I suppose if your document said 414s comp was any definition that satisfied 414s you could try testing comp-deferral, or if you used comp from date of entry maybe use year to date comp and see if you can make the test fail closer to the $400 refund. Hey, there is nothing in the regs that says you have to use the testing method which produces the best results - e.g. def of comp, testing otherwise excludbales separately, etc) bizarre idea, huh? go for it! if total amount of overpayment to a participant is less than $100, then rev proc 2003-44 (Self Correction) part III section 6.02c implies you don't have to request a refund The 2003 version is different from the 2002-47 in that an additional sentence is added - that you notify the participant that the refund is not available for rollover. Ha. this paragraph is primarily dealing with a regular distribution overpayment, but the distribution you are talking about is still an overpayment.

When formatting a number field, you have the option of enabling the currency field. I suppose the usual would simply be $ or %, but you could actually write more than that. For example, if you had a grand total of deferrals on your report, you could make the currency symbol "Wow. The total deferrals this year was and the report will print that expression in addition to the total. just make sure to include an extra space at the end of your so called currency symbol. (Or at the beginning if you set the format to print the currency after the field. Granted, you could accomplish the same thing with a formula, or use two fields - one being a text object and the other being your total field. But then, this tip was for the not-so adept. something easy. a simple modification to the report. shoot, this one is so easy and simple, yet it didn't even make the Crystal Reports for Dummies book. Date fields have the same option, in fact, you can add a prefix or a suffix or both. This all came about because we wanted to use the long form of date (e.g. December 31, 2003) A typical formula is written as "ADP/ACP Nondiscrimination Test for Plan Year End "+ToText({PLANEEKMTEST.YRENDDATE},"MM/dd/yyyy") however, the date will print as 12/31/2003 By using prefix you can accomplish the same thing without even having a formula. hey, I said this was for the not-so adept.

there have been lots of articles about this recently. these are just a few. everything is proposed at the moment, but will probably be finalized shortly. all you have to do is a google search on requirement for automatic rollovers and you will find a lot more http://www.mellon.com/hris/pdf/fyi_03_19_04.pdf http://www.benefitscounsel.com/erisaarchiv..._rollovers.html http://www.seyfarth.com/db30/cgi-bin/pubs/032204.pdf

yes, I meant to say deferrals are excluded. guess I can't type and think at the same time. dang, I can't think or type period. I will go back and edit that paragraph

thats true, the rev proc says amend plan to allow the ineligible and all other similar ees into the plan. some people argue this is simply one option and that the rev proc also states the plan should be restored to a position as if the error had not occurred - hence refund deferrals plus gains. its one of those tough calls. prior to the rev proc that allowed amending the plan (the ability to retro active amend is fairly recent), the general procedure was to refund the $

the match is a good thought. that would fix the problem going forward

you might try looking at #3 at http://www.abanet.org/jceb/2001/qa01irs.html again, any time you see a q and a with a response from the IRS, it represents an opinion of the IRS agent and not necessarily the stance of the Treasury. in other words, it is an unofficial view, but based on their unsterdanding of the regs.

how is an age weighted plan 'safe-harbor'? A safe harbor plan is one that does not have to perform a(4) testing. Either an even % is allocated to all participants (including possible permitted disparity) or the same $ amount allocated to each paricipant. There is also a non designed based safe haror - a points plan - but even that one requires some tesing. Age weighted requires testing, generally the E-Bars are the same, and perhaps this gives it the appearence of being safe harbor.

ok lets look at it this way you have a plan with 1 hce and 10 nhce 4 of the nhce receive 3% and everyone else receives 10%. Plus it is 401k so there are deferrals and matches. there are no hours restrictions or last day provision on any contribution. for simplicity purposes testing is done on an allocation basis. for 410(b), you have 3 tests: 401k 401m nonelective we made this easy, no restrictions, so everyone benefitted for each of the tests. the 5500 is easy, check box 3d. however, since people received different rates, nondiscrim testing must be done. again there are 3 tests ADP - the deferrals - must pass the 2 test (or in rare cases if the NHCEs defer over 8% on the avg, then the 1.25 test. I have never seen this happen) ACP test - the match - basically same as above nonelective - the profit sharing piece. Unlike ADP and ACP there is no '2 test'. And unlike those tests each HCE must pass on a stand alone basis, rather than by the average of the HCEs. We are considering the simple case with only one HCE. This HCE must have at least 70% of the NHCE in his rate group. (NHCE received an allocation rate or an accrual rate greater than the HCE) Only NONELECTIVES are looked at in this test. It excludes deferrals. In this example only 6 of 10 received the same rate, so testing failed. Proceed to the ABT. There are 2 parts, and you must pass both parts. There is the avg ben % test. this test includes ALL contributions, including deferrals and match. this test passes with a 70% rate. Again, all contribution are included. everything. Then there is part 2 of the ABT. This is the nondiscrimination classification. ONLY NONELECTIVES (and forfeitures since they are nonelectives are included) to pass this test, you take the ratio % test from above, and it must be greater than the 'midpoint' between the safe harbor and unsafe harbor %. this is derived from the NHCE concentration in the 'plan' being looked at (the nonelective portion). so when someone says you include all contributions in the ABT, yes, it is true, but only in one portion of that test, but not in the other. and you have to pass both to pass the ABT.

solution: without getting overcomplicated on creating reports modify summary account to print soc sec number and account data for money purcahse plan, and save as csv when running report writer. If you are not aware of it, any report run in report writer can be saved as an excel file or csv file or word or rich text file. import this data into census user fields in the 401k plan add a line to indicate these user fields on the certificate currently being used, and sum them into grand totals. ............ ok, I am a fool. I needed a break from my other work, so I designed the report for a reasonable fee. (Or at least the parties involved came to an agreement) I really fooled him. I designed it so it fit inside his bordered paper. I think my boss calls it Rent-A-Tom (RAT for short)

but only for the AB%T, not for the rate group group test. or perhaps a better way of saying it, yes, you include deferrals in the ABT, but only in the AB%T portion of that test, not the nondiscrimination classification portion of the test.

the right to make each rate of elective contributions - 1.401(a)(4)-4(e)(3)(iii)(D) its a BRF testing issue. If there is no cap on NHCEs then no problem, except perhaps making some HCEs disgruntled that they can't get as much as another HCE. but heck, they are only HCEs

410b is coverage - how many nhces get something compared to how much hces get something. the amount is not important a QNEC is a nonelective, and so is included in the coverage for nonelctives if you are actually talking about nondiscrimination [a(4) testing] the plan must pass a(4) both with and without QNECs

the actuaries already lost their days - back in 1582. in fact 10 of them. I vote their day can be the one that occurs every 2500 years. .............................. Pope Gregory XIII reformed the old Julian Calendar in A.D. 1582. His "Inter Gravissimas" papal bull described changes intended to return the annual cycle of Easter and other festivals to what was considered to be their proper calendrical positions. 1 To accomplish the desired goals he removed ten days from October of 1582, specified an intricate method of fixing Easter's date and modified the leap year rule. For the latter, Gregory decreed that February would continue to have an extra day every four years—except for centennial years, in which that month would have a 29th day only if the A.D. number was evenly divisible by 400. This was an improvement over the old Julian rule that designated every fourth year, without exception, to have an intercalary day. But our calendar could get at least one day further out of phase with solar years every 2500 years or so if we continue to follow Pope Gregory's specification. 2 Big deal? You bet it is! I don't think my software can handled the 2500 year probablem.

The way I understand it is as follows. You have a common ordinary 401k document. you provide notice saying 'wait and see' one month before plan year end (or earlier) you provide new notice saying 'we are going safe harbor' and at that point you amend plan to contain safe harbor language. what happens if you don't provide notice? Well, the document contains no language for safe harbor, I suppose you could argue by the mere silence and no plan amendment that things are still status quo. If the plan contains safe harbor language and you are using the wait and see approach, then the reverse is true. you must amend plan within one month of plan year end to remove safe harbor language. If you provide no notice, then I believe you are stuck. you have to provide the safe harbor because of document language, but you also have to test since you provided no notice. Without looking up the exact wording, I think the proposed regs really clear things up in the fact they say you can't contain language to do ADP testing if the plan is safe harbor - something like that.

see, I have to give this talk on cross testing at the fall ASPA conference, and its questions like that which I have to work into the discussion. in your particular case NHCE has SSRA of 67, so imputing disparity will add .65 to the EBAR the normal adjustment at SSRA 66 is .7 and at SSRA 65 is .75 however because HCEs have soc sec benefits taken out only up to the TWB, an adjustment is made to these numbers to take this into account. Thus the HCE adjustment will probably only be .4 to the EBAR in other words, the NHCE EBAR is increased by more than what the HCE EBAR is increased, thus helping your test.

well that depends. I have modified statements so the wording says Prior Money Purchase rather than Match, and simply coded the money purchase as a match account but of course that is a plan with no match, the money purchase is no longer in existance, and I wasn't showing investments, just totals by 'source' so to speak. If I remember correctly, even if money purchase is on going, you could still run it as an employer account and then switch to 'match' before printing. the typical 'language in the cert by source is something like If ({RPTPLANACCT.TYPECD} = 'A' or {RPTPLANACCT.TYPECD}='B' or {RPTPLANACCT.TYPECD}='M' or {RPTPLANACCT.TYPECD}='L') Then BegBal4:={RPTEEACCT.BEGBALAMT} so if your profit sharing account was #100 and the money purchase was #150 you could simply add that to the expression If ({RPTPLANACCT.TYPECD} = 'A' or {RPTPLANACCT.TYPECD}='B' or {RPTPLANACCT.TYPECD}='M' or {RPTPLANACCT.TYPECD}='L') AND acct number = 100 (that is not the exact expression, but it would be similar) then beg bal44:={RPTEEACCT.BEGBALAMT}

well, thats better than having software that simply allows you to test using SSRA! otherwise you might have simply ran it that way and had been satisfied. sounds like you only have a 10 year difference in age. (1.085^10 = 2.26) 9% * 2.26 = 20.34% if owner makes 200,000, then 20.34% of that is the 40,000 that assumes of course owner makes that much. imputing disparity helps a little.

well, you can of course do whatever you want , but I don't think that is what you are really asking. Is it permitted under the regs? yes, but... 1. when testing a plan you are supposed to use normal retirement age under the plan, so you can't simply decide to use that as your testing age. (see 1.401(a)(4)-12 defintion of testing age) 2. My understanding is that you could write a plan to contain language that for a(4) testing it would be SSRA - at that point it is also my understanding you would now have to check BRF , e.g. ees at SSRA 65, ees at SSRA 66, ees at SSRA67. chances are the plan would fail at that point. if your HCE is born in the first half of the plan year, using age definition nearest sometimes helps, especially if the NHCEs were born in the second half of the year.