Tom Poje

yep, just estimating. I think one of the ideas of this spreadsheet originated with the concept of guestimating the values as early as June for studies. It hasn't failed yet, though who knows what Septembers value will be with the way things are going. but it would have to drop dramatically for the limits to change what they show on the spraedsheet. Good grief. someone actually went out to the web site and looked up the values! 10 points! estimate for the taxable wage base is 106,000 though I am not quite sure how that one is rounded.

This is the file that I have used the last few years. the current input values are from June-July-Aug and not July-Aug-Sept. I believe the CPI-U table is updated every 3rd Wed (that is, it will be the 15th at the earliest of any given month that the new monthly value will be known) By inputting the monthly value into the spreadsheet you generally know ahead of time what the COLA value will be. The integration level is a different calculation so is not on the spreadsheet. Over the years, I have done my best to maintain the spreadsheet, most recently adding the catch-up level as well. good grief. A little history on this spreadsheet. BAck in 1997 I had written a brief Benefits Link article on how the indexed limits were calculated. Someone had seen the article and based on the notes created this spreadsheet and graciously sent me the spreadsheet. In 2002 we had the major regulation change pertaining to the limits (e.g. deferrals were set at 15,000, then 16,000, then 17,000, etc. )All the other limits were also redone, so I tinkered with the spreadsheet the best I could. The first page of the spreadsheet indicates the 'projected' values, the second page provides the spot to input the values for the new year (along with the link to the website to obtain the CPI-U values. There may be a slight rounding issue (e.g. I don't recall if things are supposed to be done to 2 or 3 decimal places), so the unrounded values may be off $1 or 2$ - but I never worried about that. As one of the earlier notes indicated, some of us knew a few months ago what the new limits would probably be, and this spreadsheet is one way of getting that idea.

couldn't tell you Mike. I simply copied and pasted the blurb. I didn't sign up for the free trial or anything, merely clicked on the link to the article. Now, I see, just a few days later, clicking on the link goes directly to the sign up options instead of the article.

but logic would say if the IRS feels an M & P would need such language to be definitely determinable, then it would only be logically for other plans to contain similar language...oh, we are talking logic and IRS. maybe that doesn't follow.

buckaroo, thanks for finding the source. I couldn't think of what it was and I was too lazy to look it up. for those who are interested, the LRM states If a sponsor wishes to use other gateways, it is important to ensure that the benefits provided under the plan remain definitely determinable. In order for plan benefits to remain definitely determinable, the plan document should specify which gateway is used. The plan document could allow adopting employers to elect between different gateways, but in order to provide definitely determinable benefits it is not sufficient for the plan document merely to specify that one of the gateway requirements will be satisfied. lets not confuse failsafe language with gateway language. failsafe language ensures a plan will pass coverage, basically it says keep bringing in bodies until you pass the ratio % test for coverage. (That is why you cant use the average benefits test for coverage when you have fail safe language) gateway language is for nondiscrim testing, and yes, if such language is not in the document (e.g. you haven't restated yet), then you use an -11g amendment to satisfy the requirements.

yes and no. as for 1) the owner received 10%. the gateway is the lesser of 5% or 1/3 the rate of the HCE, which will only be 3.334%. Thus, if you had someone who received the 3% safe harbor but failed last day or hours requirement (if they exist for the profit sharing) the individual would not be bumped up to 5% for gateway purposes. they would however have to receive the 3.334% since you would be cross testing. so, yes 5% is sufficient to satisfy the gateway, but there may be other issues involved. as for 2), if by the statement you mean there is one rate group test and it includes the safe harbor contribution as well as the profit sharing, then yes, that is true. as to whether that will actually satisfy the test, that depends on the ages of the NHCE. if the NHCEs are all older than the HCE, then no, as the statement is phrased, the 3% SH and 2% SH will not satisfy the rate group test.

the ADP test is the exclusive nondiscrimination test for deferrals the ACP test is the exclusive nondiscrimination test for matching contributions E-BARs show up in the avg ben pct test (which includes all contributions including deferrals, match but no catch ups) this test may be helpful to pass coverage. no gateway is needed. you are not testing for nondiscrimination, you are testing coverage. E-BARs also show up if the rate group test. strictly based on nonelective contributions. and you have to pass gateway before running the gateway test. the plan document does not have to allow for testing however there are some considerations: you can't simply give someone the gateway minimum. that is document driven. and my understanding is that the document must specify which of the gateways you will use. if testing for coverage AND the document contains failsafe language, then you can not use the average benefits percentage test.

BG5150 oddly enough, for whatever reason, the one time irrevocable election is only found under the 401k regs as far as i know- but look at the examples under 1.401(k)-1(a)(3)(vii) in particular example 5 which is a money purchase plan!

the answer to this depends on whether you are Catholic or Protestant, and what country you reside in. so just wait long enough and they will adjsut things to match your watch. they did it before, they can do it again. dang, I feel for the people who lost their vacation time! Historical fact: The Gregorian Calendar During the Middle Ages, astronomers and mathematicians observed that the calendar year was not completely accurate with matching solar years. Errors in the Julian calendar were noted by church officials and scholars because church holidays did not occur in their appropriate seasons. In 1582, Pope Gregory XIII (1502–85), who was pope from 1572 to 1585, and his astronomer and mathematician created a new, reformed calendar known as the Gregorian or New Style (N.S.) calendar. It was adopted first in Roman Catholic countries. Protestant countries adopted the calendar during the eighteenth century. In order to make the calendar adjustment in 1582, ten days were eliminated from October. Thus 4 October 1582 was followed by 15 October 1582. England and its American colonies did not adopt the reformed Gregorian calendar until 1752. Scotland adopted it earlier, celebrating the New Year on 1 January 1600 and subsequently on January 1st of each year. Interestingly, Alaska did not change from the Julian calendar to the New Style Gregorian calendar until 1867 because, up to that point, it was part of Russia. In order to make the calendar adjustment, eleven days were dropped from the month of September 1752. An eleven-day adjustment in 1752 was needed because one more day had been lost since the calendar was changed in 1582. The year 1751 began on 25 March and ended on 31 December 1751. The first day of the year was now January 1st and the last day was December 31st—the calendar we use today. Thus, 2 September 1752 was followed by 14 September 1752. In this way, the Julian calendar added one day between 1582 and 1752.

my understanding is that Sieve is correct, the regs quite cleary say "to the plan and any other plan of the employer (including plans not yet established)" plus, it can only be done when the ee is first eligible, all this according to 1.401(k)-1-a(3)(iv)

at the meeting I had a few weeks ago with the IRS they frowned upon such practice of changing a match during the year - regardless of whether the match was safe harbor or discretionary. if match is done on a payroll basis (as opposed to match which is deposited throughout the year, but with a true up at the end), then the rate of match becomes a problem. even if it was a match with a true up, and the match was increased you can still have problems. arguments were omething like this: someone who received a match earlier in the year and terminated might have deferred more if he knew the match was going to be greater. in a similar vein, at this late date, someone might have deferred more if they knew the match was going to be greater, but can't take full advantage as of this late date. etc., etc. I believe this will be discussed in greater detail at the ASPPA conference, but you never know how much time they have to discuss things at the Q and A session.

the one exception: we are talking 'nondiscrim testing' and in particular as it applies to non-elective contributions. the non-discrim testing for deferrals and match are the ADP and ACP tests. in those tests, a person must be actually 'eligible' to defer or receive a match. Therefore, if you had a last day provision, someone who quit would not be on the ACP test with a big fat ZERO. They simply wouldn't be on the ACP test. Yet, if the person worked over 500 hours, they would show up in coverage as includable and not benefiting.

this is based on the following rather 'startling' e-mail I received the other day (via The Defined Contribution and savings plan alert daily) ................... ASPPA Set To Merge With COPA - 09/30/2008 The American Society of Pension Professionals and Actuaries is about to be known by a new moniker—ACOPA. The newly-formed group comes as the result of ASPPA’s merger with the College of Pension Actuaries. “I am really excited by this combination,” Mike Preston, president of the new ACOPA, said. “I think it portends well for the actuarial profession.” The merger announced by both groups this past week is something of a reunion, as the founding members of COPA splintered off from the small plan trade group in 2005. It was then ASPPA most recently changed its name. ................ If I understand things correctly, it's not ASPPA changing their name, but rather there will be a subgroup within ASPPA - known as ACOPA. It is COPA that will be known by a new moniker. hence my comment about whose fault it was who wrote this thing. (OR maybe it really was a hostle takeover of ASPPA by the actuaries and I haven't heard about it)

was that simply discussed, or was an attempt made to test it in actual practice?

close, you have the correct reg cite, but the wrong definition look up 'non-excludable employee ' in Treas. Reg. Section 1.401(a)(4)-12. also recall that the gateway minimum need only be provided to those who are benefiting, by implication you have others that are not benefiting in the test. looks to me like Sieve is getting ready for another hockey season up in Detroit

ACOPA? 'Mea culpa' is a Latin phrase that translates into English as "my fault"

the term E-BAR, stands for Equivalent Benefit Accrual Rate. so you are simply converting a contribution to an equivalent benefit at retirement and dividing by the compensation to determine the rate. There is nothing to prevent or stop you from doing this anytime (unless I guess you are talking about an ESOP which is forbidden - well, you can still do it, but it doesn't help in this business). For 'business' purposes there wouldn't be much reason for calculating E-BARs unless you are failing nondiscrimination testing - otherwise you are spending your time doing something which isn't necessary. unless I am missing something in the question you are asking.

since it is permissible to use 3% for the look back year for the initial 401k plan year, why not go that route (giving the HCEs 5% deferral rate) and then become safe harbor next year?

This is probably an apt description (not mine, but from another source): Under the Three-Year Testing Cycle Guideline, Employers are not required to perform certain nondiscrimination tests more than once every three years provided that the Employer’s retirement plan meets certain requirements. Specifically, if there is no significant change in the make-up of the Employer’s workforce, no major change to salary structure, no amendment to the plan during a Plan Year, and the plan’s most recent test results indicate the plan satisfies the nondiscrimination requirements with a comfortable pass margin, an Employer may rely for two years on the results of a test which demonstrated compliance with the applicable nondiscrimination requirement now, in a small plan this is not likely to be the case, a change in one body or two amongst the NHCEs can make a big difference. if its a cross tested plan, then you wouldn't want to change the allocation % because that too would be a significant change. most plans I have seen that rely on the non-discrim test do not have what I would call a "comfortable pass margin", well at least the doctors and lawyers and such I have to run tests for want the max and provide the minimum. so I'm not a big fan of the concept of relying on 3 year testing to save a few bucks.

this sounds like one of those unintended consequences of the law. New Company (and it is late in the year) thus no one has a year of service. but under the regs, you can treat the HCE as having met the 1 year wait. so you compare the HCE to prior year NHCEs. since there are none, there are only HCEs and so the plan passes. in 2009, assuming the new company began after 7/1/08 no one will have 1 year of service, so the same thing happens. it appears the HCE stands alone again.

Kevin C: I agree with you. I see nothing referring to 'borrowing' in the regs. of course the term 'shifting' isn't used either, it is simply the term coined to referring to elective contributions used in the ACP test. The definition of eligible employee under 1.401(m)-5 indicates these are employees who are directly or indirectly eligible to receive a matching contribution or make employee contributions (after tax contributions). I see nothing in this definition that says "or have elective contributions used in the ACP test". the ACP for a group of eligible employees... 1.401(m)-2(a)(2) the determination of the ACP percentage is "eligible employee...sum of employee contributions and matching contributions...and QNECS and elective contributions..." 1.401(m)-2(a)(3) Thus, my reading of things would say you can't simply 'shift' numbers without considering other facts and circumstances. e.g. I don't see how you can shift deferrals for someone who is not eligble for the ACP in the first place. on the other hand, since it appears you can shift any amounts of elective deferrals it probably doesn't matter in most cases. as for whether you can shift after making corrections, it is unclear. my personal leanings are toward saying yes, since at that point in time you have a test that is passing, and since you don't rerun your test after making corrections, I don't see why shifting would make a difference (but I can understand if the IRS made it clear you couldn't)

this one developed into a lively discussion (we had the IRS Q and A pre-discussion meeting this past Friday, with no real answer being provided, but a promise from the IRS personel to look into the matter and get back to us. (the leanings seemed to be in the direction of including the indiviudal as a 0, but...) one point that was raised was what if the individual had indicated the desire to defer 25% on any pay that was not commission and this was simply a year he had no other pay. then what % would you use. ha, a nasty vicious circle.

ha, good comeback. I needed a good chuckle.

Sounds like a line from a song: where have you gone Po-je- magio my wild guess is there is some type of account data involved (e.g. you have a number of sources and investments) and therefore the HCE pops up printing each time it hits another account.) I've managed to accomplish that feat either with a bad sort routine or worse in the table links which I am never sure how to correct. If you recently added a table that wouldn't surprise me a bit.

I'd agree with Sieve. includable and not benefiting for coverage. not on the ADP test