Tom Poje

I would certainly agree you shouldn't use just comp from date of entry. I would hold that you should calculate an EBAR for the 401k based on total comp and an EBAR for the profit sharing based on comp from date of entry, and then aggrgate the two numbers. e.g. testing on an allocation basis if ee deferred 5% and nothing else his EBAR would have been 5% if ee received a 4% ps contribution from date of entry, his EBAR would have been 4% it seems reasonable to conclude if you had two separate plans the EBAR would be 9% if aggregated

yes it is possible to have each individual named as their own group yes it is possible to have language that simply states a gateway minimum will be provided. (In fact, you have to have such language to provide the minimum gateway) if each person is in their own group, then coverage testing may be affected if an individual does not receive an allocation. you fail reasonable classification test. however, chances are, such an individual would probably be HCE, so you would not need to rely on the Avg ben test to pass coverage anyway.

well, I suppose you could set up a DER to export comp and hours. e.g. plan has been on system since 1997 then import comp into census user field alpha 1 hours into user field numeric 1 populate user date field 1 as 12/31/97 (duplicate using DER?) then export for 1998 and import into user fields 2 etc..... once you have the basic set up then I guess each year you simply populate the next user field with the current year info then simply tell your report to print the user fields. I'm not quite sure what happens if a user field is left 'blank'. I've had people not print at all, but I could have done something wrong.

I must have missed this one under the pile of work I have! never use SSRA - regs require you to use NRA as testing age, currently SSRA is not an option for NRA as far as I know for document purposes. it is possible to write (at least I understand) to write it into the document that SSRA will be the testing age. At that point it is also my understanding that you would have to do a BRF test on each group - age 65, age 66, and age 67. it is doubtful you would pass that. rate banding - use only very cautiously, never the default, which simply takes the HCE and outs them at the high point and then determines the midpoint and the lo point. this is clearly in violation of the reg that says you can rate band as long as the rates of HCEs within the band are not significantly higher than those of the NHCEs within the range. impute disparity - no reason not to use this option. it doesn't matter if the plan is integrated or not. In fact, my challenge to people has always been try testing on an allocation basis on a plan integarted at 100% of the taxable wage base, using 5.7% permitted disparity. your e-bars will be the same for everyone. it is a way of checking your work. interest rate - 8.5% pre interest is best in almost all cases, I suppose if you had some owners kids getting a minimum contribution it might be best to use 7.5% Mortality - 1983 IAF, 7.5% will do slightly better than 8.5%. (Pre and post interest can be different) age definition - nearest, if HCEs born in the first half of the plan year (not calendar year)

for those interested see: 1.401(a)(4)-9 is the regs on DB/DC testing in particular -9(b)(2)(ii) speaks of testing the DB as an allocation rate. there is no special requirement for a gateway under this section -9(b)(2)(v) speaks of 'eligibility for teating on a benefits basis, which says you can't test this way unless you meet the aggregate gateway.

why would it not seem right? The rules only say what was the person's comp in the lookback year (unless the person is also a 5% owner) it does not matter if they were an HCE or not in prior year. and just because they only worked part of the year you do not annualize comp.

what would happen if you tested on an allocation basis? 1.401(a)(4)-8©. I've wondered about this - there is no 'minimum gateway', those rules are in -8(b) for when you convert a contribution to a benefit, but I don't see a corresponding rule - maybe it is because nobody will be in the HCE rate group, but then it would seem you could do an -11(g) to put some people in the rate group. oh, just ignore me and my twisted ideas. it was only a thought.

yes. I dont have a site per se but if you look at 8.82 ERISA Outline Book it would certainly imply that. (2003 edition)

on the other hand, if they were to eliminate the taxable wage base for purposes of soc security, then that would eliminate integrated plans, which might make cross tested more attractive. of course, imputing disparity would disappear as well...

you are correct, step 1, mandatorilly disagregate the 'plan' into 3 'plans' 1. 401 k 2. 401 m 3. nonelective now you begin testing you have your choice you can permissively aggregate the two 401ks for coverage - but if you do you run 1 ADP test withh all employees. there are some special rules if an HCE particpates in both plans. there is nothing to prevent you from permissively aggregating the 401k feature and not permissively aggregating the other features (or vice versa) but certaing nondiscrim options get tossed out (e.g. shifting deferrals to match) since you can only 'compare apples to apples'

the formula you cite sounds like one for failure of coverage rather than ADP. that is if you fail coverage you bring in the NHCE with the XXX hours or comp or whatever. that is known as fail safe language. as for if if it is possible to have a plan that only provides QNEC for those statutory includables, that should be possible, I would expect language similar to that used in safe harbor 401k plans, though I suspect it may take special language to do so

I guess if the document says 'all NHCEs receive' then I don't think you have a choice - regardless of testing method used.

I have a report that might work, but for whatever reason you can no longer post an .rpt on the board. if you give me your email (you dont have to post it on the board, but rather email me directly) no guarantees on any of my reports, they seem to work but I never know for sure. I think at some level of relius 10 there will be a report, but it will be awhile before I get converted

which is why along with the notice I would have employees in such a case signsomething indicating they realize and understand they would get a nice match if they deferred.

gap earnings must be calculated for plan years beginning after 12/31/05 however you are permitted to operate the plan under the new regs if you want (you just have to follow all the terms of the new regs you can't pick and choose) chances are you probably do'nt or won't and so there would be no gap income. The new regs describe how to calculate gap income, Treas Reg 1.401(k)-2(b)(2)(iv)(D) See also example 4 of treas reg 1.401(k)-2(b)(2)(viii) Treas Reg 1.401(k)-2(b)(2)(iv)(E) one of the examples has an error in the calculation (whoever said they can do simple math), but I don't recall which one.

easiest suggestion is to get someone on Benefits Link to do it for you the regs say any reasonable method. one such method would be Income allocable through the end of the plan year is calculated as follows: Corrective distribution * gains to the deferral account (Beginning balance of deferral account + total deferrals for the year) note: if you desire to operate the plans under the terms of the new regs, then you would also have to determine 'gap' income.

I am not sure if the amount is definitely determinable unless it is specified how testing will be done. for instance, what if interest rate assumption of 7.5% is used rather than 8.5%. or test using accrued to date, or test using allocation method, or even testing on a compnent plan basis, using age nearest or age last....

A couple of years ago at an ASPPA conference a speaker held the same position you count all, someone questioned that but it went no further than that. gee, no one carries around a copy of the regs to point out cites and argue the point, upon leaving the room I know at least 2 or 3 others that said you can't count all. so looks like we are all in a group wishing it was a lot clearer. At least the ERISA Outline Book, at least in some way, matched your quote from the regs about 'excludables'- or at least one way of looking at it.

I've tried looking for a clear black and white example, can't really find one. however ERISA Outline Book says (9.125, 2003 edition) Applying rate group test to disaggregated plans ...In applying coverage test to a rate group, the ratio test or the nondiscriminatory classification test, whichever is applied, generally would treat employees in disaggregated groups as excludable employees... emhasis from the book. Logically I still don't see how you can say I perform ratio percentage (Carving out those excludables) I now compare this to the safe harbor/unsafe harbor or even midpoint determined by the group of employees as a whole(not carved out). thats not comparing apples to apples. but then there is no logic to the regs at certain points. The problem is that the avg benefits test is really two parts. one part(avg ben %) you include all employees, the other part you dont. so for the non highly concentration % how do you calculate. I would hold be consistent, for the avg ben% test count all, but that is moot since you dont use the non highly concentration % for anything here. I suppose it gets down to the same argument some use in regardless to testing 'otherwise excludables'. There are no clear examples in the regs. some hold for the avg ben % test you cant test separately. You test 'All'. ERISA Outline Book gets into this discussion. see 8.93

perhaps the ERISA outline book sums it up fairly well 8.79 2003 edition "If 2 or more plans (or two or more portions of a single plan) must be tested seperately under the disaggregation rules, they cannot be aggregated to perform the ratio test or the nondoscriminatory classification test." or perhaps another way of looking at it I am testing coverage in a 401(k) plan for the 401(k) portion I include everybody to determine the nonhighly concentration % since everybody can defer. Usually this test passes at 100% so no one ever thinks of performing the nonhighly concentration %, but conceivably it is possible, especially if testing a controlled group separately -you might fail the ratio % test for coverage. but for the nonelective portion, since terminees are excluded from the contribution, a special rule allows me to exclude them entirely. I dont include the 401k terminees because the rule is I have to disaggregate bodies from the 401k bodies. it is only the avg ben % test where everything gets combined. now, the problem with the software is that it is hardcoded to excluded all terms with less than 500 hours. (Though I think if you enter a 0 for hours in plan specs after running everthing the system would actually include all these terminees.) if you have 1 or more HCEs who terminated with less than 500 hours it might be beneficial to include all terminees. the terminee < 500 hour rule only apples to 'participants'. so in the case of a controlled group, if you were looking at Company A separately, you could not exclude a terminee from company B because he was not a participant in Plan A. oh nuts, all that was probably as clear as mud.

nondiscrim classification test is found under 1.410(b)-4 It only refers to 'the plan'. 1.410(b)-7 defines plan and requires each part of a 401(k) to be disaggregated. 1.410(b)-5 refers to the avergae benefits percentage test. this test requires you to aggregate all 'plans' this is the exception to the rule. therefore, I hold the belief that when determining the nonhighly concentration % test you would only count those bodies in the actual 'plan' [disaggreagted] therefore, the terms with less than 500 hours would be xcluded if they didn't benefit. This is an option of course, you could count them, possibly you would want to if there was an hce in the bunch. If two or more controlled groups are involved and you are not aggregating for testing, then you would have to include the terminees from the other company not included.

is this one too aggressive or does it fall under the rules? HCE 5 4 NHCE 3 1 Plan passes ADP. now, I shift 1% of NHCE to ACP and arrive at HCE 5 4 NHCE 2 2 plan now passes ACP, but one of the rules for shift is plan must pass ADP afterwards. It doesnt. so can I now treat 1% of HCE as catch up?

Top heavy rules are slightly different, the code simply says ignore ees who have not performed 'any service' in the last year. it doesnt say 'took no comp' It is certainly possible the ee is performing service and not getting comp. I am assuming the ee inquestion was at one tome collecting comp, was in the plan, etc, but now for whatever reason is not taking comp.

Merlin: is the document in question new, and has it an approval letter? If so, I am not sure how a standardized plan can fail. Katherine may be right that this is an issue the IRS is not allowing any more, but I haven't seen anything on it, so I couldn't say for sure.

Instead of 'heard' or 'read' why not go the regs 1.406-1 t-24 ....(b) adjust for contributions due as of the determination date the gray area comes in when it says 'don't include if plan not subject to minimum funding requirements', though I have only herad that used with not including profit sharing contributions - though at an ASPPA conference a few years ago the IRS said you include those as well. I would hold things like deferrals are due as of the determination date, they have already been withheld, but simply haven't been deposited