Tom Poje

so in answer to the question, if the HCE(s) are born in the first half of the year as compared to NHCEs then it is probably better to use age nearest. If the HCEs are born in the second half of the year, attained age is probably better. Suppose you have a number of HCEs, and one of them fails rate group testing, but barely. well, it might make a difference whether you used age nearest or age attained. this is an assumption and is not document driven (at least at this time)

my 'chart' says it would probably be 230,000 in comp and 46,000 for 415 limit, but no change to deferral or catch up limit

this is another one of those helpful 'tricks' in cross testing. but instead of directly answering your question, lets put it in the form of an exanmple and see if you can determine what the effect will be. Employee 1 Employee 2 Plan year end 12/31/2003 12/31/2003 DOB 5/31/1943 8/31/1943 Actual Age 60 yrs 7 mths 60 yrs 4 mths Yrs to age 65 4 yrs 5 mths 4 yrs 8 mths If expressed in whole years: 4 yrs 5 yrs Employee 1 has one less year to retirement! Now, for the quiz: (I think people learn more and remember it better by solving some of this stuff) 1. if your pre retirement interest rate assumption is 8.5%, what do you think this will do to the E -BARs? 2. if the HCE is born in the first half of the plan year (PLAN year, not calendar year) would his E-BAR be greater than, less than or equal to a similar employee born in the 2nd half of the plan year. 3. what is your conclusion about selection of nearest vs last.

ha. I mention this little bit into this in my ASPPA talk. (but probably most people fall asleep by the time I get to that point in the talk so...zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz) One of the steps in the calculation of an individuals E-BAR is to divide by the APR (that is, based on whatever mortality table you have chosen) now, if all employees have the same retirement age, then that would be a constant, and thus have no effect on anything. if someone has continued working past normal retirement (or retirement age is 65/5, those E-BARs are generally so small that the mortality table choice makes no difference. Now, all that being said, the choice of mortality table does make a difference even though it is a constant. If you impute disparity then it makes a difference, and enough of a difference to possibly pass or fail testing. 1983 IAF produces the largest APR and that is generally the one I would use.

the 90 day rules is listed under the heading "Special rules for certain withdrawals from EACA" and not under the section pertaining to QACA. (of course the rules don't take effect until the 1/1/08 plan years.)

since one or more of the HCEs ratio % is less than 70% (but at least greater than the midpoint) then you need to pass the average benefits % test as well.

no wonder imputing disparity doesn't help. the HCEs have comp less than the taxable wage base/and or covered comp limit

I am assuming your document permits the allocation of QNECs as described. the rule is as follows: A 5% (or less) QNEC is deemed not to be disproportianate, so it would appear either example you provide would be possible. you case #2 is a good example of what can happen. 8 NHCEs but 2 are terminated. NHCEs who are employed on the last: 6 so 50% of this group is 3 bodies. the lowest rate for any 3 bodies (D,E and F) is 2.51% so you could actually use twice that (5.02%) if you had allocated a larger QNEC to anyone. thus you could have given D=3 E=3 F=6 as well again, assuming the document permits such an allocation.

what happens if you test on an allocation basis (and impute disparity)?

you have 10 nhce and 3 hce, so nhce concentration is 10/13 = 76.xxxxxxx always round down - so you have 76%. this equals a midpoint of 33%. this means you need 4 NHCEs in the rate group (4/10 > 33%) you have some big problems with a 49 year old hce and the 4 closest nhces are 26, 44, 47 and 48. This means you only have a 1 year difference - maybe a 2 year difference in age (depending on age definition 'nearest' or 'last' between the 'oldest' of the NHCEs in the group. since the max interest rate is 8.5% and based on a 1 year age difference, at very best the HCE can receive 1.085 times more than the NHCE. (if testing on current accrual. imputing disparity will help some (2.564 * 1.085) = 2.782 2.782 + .65 (max disparity) = 3.432 hce is at 3.244, but the HCE will also get some disparity added, while most likely less than .65, without knowing the comp, would guess around .45. thus you will still fail, even if avg ben % test passes. if hce is born in the first half of the year and the NHCE is born in the second half of the year then it is possible the ages will be 47 and 49 instead of 48 and 49 - then you might have a chance of passing, but based on the info provided this plan will fail on current accrual testing.

puts my limited attempts at song writing to shame.

well, from my notes from prior conferences: only mention of pro rating 415 limit is with a change in limitation year. you don't have that. at 1996 asppa conference the response was 'there is no initial short limitation year. it is a full 12 month period working backward from end of the year. pro ration of comp limit: depends on the period used for determining allocations. and it is certainly possible to use a 12 month period ending with in the plan year, so that would be a document issue.

my response was to only answer the question that was posed as follows from above: 1) What if the employee does not defer anything in the second calendar year? Wont he then have contributions being recharacterized for 2006 when he has not physically made contributions to 2006? 2) Next is his W-2 will show 0 for the 2006 year. If the client gets audited and actually checks his W-2 for the deferral amounts, the ADP/ACP test, and the re-charaterized amounts, then, again, he'll have contributions being recharacterized for 2006 when he has not physically made contributions to 2006? the example from the regs is simply one year different from the question posed above. ee deferred 600 from 11/1/06 - 12/31/06. no mention is made of what took place in 2007. yet the 600 is still treated as catch up for the plan year ending 10/31/07. thus, even if the ee defers nothing in 2007 (the second calendar year), and thus his w-2 will show 0 deferrals, he still had something treated as 'catch-up' for the plan year ending in 2007 - that is - amounts that can not be used in the ADP test. (Of course, this 600 is not applied to the calendar year 2007)

enough already! read example 5 of the regs! There are 5 points to this. plan year is 11/1/05 - 10/31/06 (i). ee deferred 3200 from 11/1/05 - 12/31/05 and 16,000 from 1/1/06 - 10/31/06 so he has 19,200 for the PLAN YEAR (ii) oh no, he exceeded the 402(g) limit in 2006 of 15,000. thus 1000 is automatically counted as catch up. so for testing purposes only 18,200 is used in the ADP test (iii) oh no ADP test. he can only have 14,800 for the PLAN YEAR (iv) thus 18,200 - 14,800 = 3400 in catch up (v) THIS IS THE CRITICAL ONE. REPEAT. THIS IS THE CRITICAL ONE. ee can defer an additional 600 in calendar year 2006 because he has only used up 4400 in catch up. NOW READ THE LAST SENTENCE the 600 in catch will NOT (I repeat) will NOT be taken into account for the plan year ending 10/31/07. the emphasis might be mine, but the example is the regs. end discussion! conclusion: he deferred 600 in 2006 (not 2007) yet it is still treated as catch up in 2007.

How is the definition of limitation year defined in the document?

here is the link to the data for calculating the taxable wage base. although it is not quite official yet, it looks like it is pretty much decided - no matter which of the tables you use - they all have the same projected value for 2008. the wage base is calculated as follows: 1994 base * 1992 avg wage index / wage index for year (Y-2) [rounded to nearest multiple of $300] the 1994 base is 60,600 so now you know! e.g. for 2007 you divided by the 2005 avg wage index http://www.ssa.gov/OACT/TR/TR07/V_programatic.html#wp187848

so I think we agree that the plan in question might pass testing, but not necessarily for the reason cited by prior administrator (whose reason appears to be wrong) I am not sure on the comment you must pass 'reasonable classification' when splitting into component plans. 1.401(a)(4)-9©(2) says you can select the group of employees used for THIS purpose in ANY manner, and the composition of the group MAY be changed from plan year to plan year. In addition, 1.401(a)(4)-2©(3)(ii) says a plan passes nondicrimin classification test (including reasonable classification) if and only if you pass the.... midpoint or ratio % test. my understanding, at least from the talks I have attended, there really isn't a 'reasonable classifcation test' once you get to the a(4) portion of testing.

if I understand component plan testing correctly you have the following: step 1 would be to determine if the 'plan' as a whole passes coverage. as you indicated, if you tried to test the 'plans' as seperate, you would fail coverage. so you have aggregated the plans and all benefit (100%) so that is not a problem for coverage. based on what you indicated, plan will not pass if all employees are tested on either a contribution or accrual basis. so, now you want to do some component testing. there is no requirement you split people between A and B. you can put people into different groups, as long as each employee is in one and only one group. for component testing, each group must pass 410b (obviously wont work if you only use group B) and 401(a)(4) however, it might be possible to take 5 or 6 'old' hces from group b, and test them with group A instead. (e.g. a component plan) that would almost double your percentage of 14.58 because you cut the hce group almost in half. now you have a fighting chance if you can pass a(4) again, that is if I understand the component (or restructuring) rules properly.

buckaroo, lets really simplify the example. Mr. Big defers $10,000 in November 05 nothing in 06 through the end of the plan year. plan fails ADP test. by your 'logic' would you say he couldn't get any catch up because he has no deferrals in 06?

How can your brain freeze when there is no guidance in the regulations - see 1.401(k)-5! Its reserved for mergers/spinoffs etc.!! well, okay, my copy is still blank, maybe your copy of the regs has something written by someone. One argument might be using an example when an individual switches from union to nonunion or vice-versa - only use comp through the date of eligibility, which is what you suggest.

hmmm. hadn't thought about something like that, though the notes I have say that the amendments of this section shall apply to plan years beginning after 12/31/2007, and I read that to say the 90 day distributions rules don't apply to deferrals made in a plan year prior to that date.

I'm not sure what you mean. The first possible 'payroll deduction' for automatic enrollment would be Jan x. the person would have 90 days after that date to make an 'election' to take a distribution. In effect through the end of March.

Tom's quote on the 'first payroll' is not saying that is the only deferral available for distribution. the regs say that is the date when the 90 day clock starts ticking. the rest of 2( C) says ....and any succeeding payroll period beginning before the effective date of the election.... (but Tom has other work sitting on the desk so he can't type everything, he does expect someone to go back and look and see what the regs say - ha!) 3( C) of the same regs does say ...in the absence of an investment election by the participant, contributions are invested ['in a default fund' (my paraphrase0] so that would seem to say once you make an election to a fund all bets are off. since these plans don't satrt until the new year, I would expect a good chance of same clarification on some of the issues.

well, recall that you originally couldn't even have a safe harbor match on a payroll basis. the Q and A of Notice 2000-3 is what permitted it. under what other conditions would the last sentence of Q2 apply? If the match failed safe harbor to start with, then its a moot point. so why didn't they just say, yes, you can do the safe harbor match on a payroll basis. I suppose you could say its not Both 1 AND 2 but rather was 1 satisfied AND then was 2 satisfied but then you are left with saying that the regs say if you change a fixed match then you have to run an ACP test, but a discretionary you can do whatever you want every payroll period. (though of course if there is a loophole in the way things are written people will take advantage of them.) the 4% cap on comp for discretionary was effective 1/1/2000 so some people 'got away' with it in 1999 by providing a larger discretionary match. regardless, based on you comments, you said the employer 'notifies' people that the discretionary was being changed. I thought the safe harbor notice was something done 30 days prior to the beginning of the year.

perhaps Q-2 of Notice 2000-3 which says (last sentence) The payroll period method applies only for purposes of satisfying the ADP safe Harbor matching contributions of section 401(k)(12) AND the ACP safe harbor matching contributions of section 401(m)(11) I have a big problem with the word 'AND'. a discretionary match fails the first part (it does not satisfy ADP safe harbor.)