John Feldt ERPA CPC QPA

Bill, the 50% owner of the corporation, buys the other 50% from his sister Mary in mid-2009. He does not have enough cash to do this, so his sister holds a promissory note (more on this later). The company has a calendar year 401(k) plan. Mary is an employee and is still an HCE in 2010 due to her prior ownership. Mary did not retain any repurchase rights, but she held liens/collateral interest in her company shares, the building, some land, etc. as security for the promissory note from her brother since he did not have the cash to pay for her shares outright. Mary's wages have always been only about $50,000 per year so she is not an HCE due to wages. Is Mary an HCE in 2011? Is Mary an HCE in 2012?

Watch out for the SPD language if you have a system generated SPD. We had to do a lot of manual SPD modifications for a recent plan we did like this. Also, the checkboxes on any current 403(b) plan are not like the checkboxes for a prototype DC plan or a vol sub DC plan formatted with an adoption agreement. Instead, all 403(b) plans are currently individually designed (even if it uses an adoption agreement and a basic document). No IRS agent has provided approval for the 403(b) language you are using right now. So you might not have the kind of restrictions listed in your adoption agreement like you usually see for a pre-approved adoption agreement. FWIW.

Although somewhat rare, a 403(b) plan could be written such that it only has employer contributions - no deferrals. So, yes I think it is possible to have a January 1, 2013 effective date as long as your deferral effective date is current or prospective. Will your document allow for that?

The proposed regulations allow a fixed 5% rate. Since these are not final regulations, you might be able to use a higher rate and say you used good faith compliance that your rate did not exceed the market rate? Some say the final regulations might allow a fixed rate of 5.5% (the 415 interest rate). We'll see soon, hopefully.

Derrin Watson recently did an advanced topic in cross-testing session at the SunGard Advanced Pension Conference in Chicago where he had about 7 or 8 slides which included a good explanation and a good example. He credits Tom Finnegan of The Savitz Organization for his assistance. Not sure if this topic might be in any of the ERISA workshops SunGard is doing, but if you have any SunGard product, perhaps you could ask Derrin for a copy of those slides and see how it goes. Or maybe test using highest consecutive 3-year average compensation, or highest 4-year average, or 5-year, etc. Or comp from date of entry. Or if you need testing fixed for just a year (or a few years), perhaps accrued-to-date testing might help.

To get the maximuim, shERPA is right: 25% of $134,000 = $33,500 (maximum deductible employer contribution is 25% of eligible compensation). This amount is not subject to FICA tax. plus a $17,500 deferral. This is subject to FICA. This amount is not part of the 25% employer deductible contribution limit. The deferral is deducted by the employer as employee wages (regardless of whether the employee chose to defer as Roth or pre-tax). Total $51,000 If the employee was born on 12/31/1963 or earlier, they can also defer an additional $5,500 in 2013. The $5,500 is treated like a deferral above. Total $56,500 Assumes no other deferrals made that are subject to the 402(g) limitation, since the 402(g) limit is an overall individual limit, not just a single plan limit.

Okay, and the silly example should have used 60.01% as top heavy, not just 60%. The dividing zero by zero thing is more tongue-in-cheek for me, not for serious consideration - it's an obvious mathematical logic fallacy. Your mathematical proof is certainly more pursuasive: can $0.00 exceed 60% of any balance, including zero? Certainly not. Good thing the rule is not equal to or exceeds 60%.

Also, but probably not your concern, you may want to mention to your client that it's assumed the CPA has advised them regarding reasonable compensation. There are many tax cases out there where the IRS comes in and declares certain amounts really are wages. Since that could, in turn, affect the plan, it might be worth mentioning.

Not a QNEC. Under the 401(a)(4) regulations, "plan" has special meaning. See the definition in 1.401(a)(4)-12, which refers you to 1.410(b)-7. There you find that the portion of a plan that is a section 401(k) plan and the portion of a plan that is not a section 401(k) plan are treated as separate plans. Similarly for match (401(m) plans). So when 1.401(a)(4)-11(g)(3)(vii) applies and it used the term "401(k) plan" it means only the deferral portion of a 401(k) profit sharing plan. That subparagraph requires the correction for a 410(b) failure with respect to the plan's deferrals to be done using a QNEC. I see no other section in -11(g) that requires a QNEC. That being said, are you trying to provide additional "profit sharing" (nonelective) to pass testing but the folks getting the extra PS are terminated non-vested? If so, then you would have an issue with 1.401(a)(4)-11(g)(4) regarding the substance requirement. Is this why they are asking for a fully vest allocation?

No to the first part - the 5500 count is not determined by counting only those who defer and/or counting just those with accounts. I remember something about perhaps ASPPA having asked the DOL to consider changing the accountant's opinion requirement to be applied, at least for 403(b) plans, to only those plans with over 100 participant accounts? Maybe I was just dreaming that they asked the DOL about this? Anyway, we have seen this exact problem and the solution presented was to go back and get audit reports for the Form 5500 and file under DFVCP. It's just not the kind of advice that wins over a prospect.

Perhaps implementing a 500 hour rule for accrual would help the plan going forward? Unless, of course, that then prevents enough terminate participants from helping the plan pass 401(a)(26).

I'll make a general statement: you must consider all benefits of all the employer's "plans", and that includes each component of this plan, but you can ignore certain mandatorily disaggregated plans. So this employer probably just has the one "plan", right, a 401(k) profit sharing plan with match? You run one average benefits percentage test for that entire plan. You can run that on a benefits basis or on a contributions basis. If your document does not require otherwise, you test using average compensation, current compensation, compensation from date of entry, using any definition of compensation which passes 414(s). You can test with imputed disparity, you can average some employees' benefit rates, you can use various mortality and rates from 7.5% to 8.5% differing for pre/post as an assumption, you can use nearest age, last age, you can try accrued-to-date testing and rate banding, plus many other options until you can't stand testing anymore. You do not run an average benefits percentage test for component #1 and then another for component #2. Component testing does not in and of itself change the result of the average benefits percentage test (but your assumptions and methods certainly can and do). If component #1 tests the nonelective contribution as a benefit payable at the testing age (cross-tests), and if the HCEs rate groups do not each pass with a 70% pass rate, then your average benefits percentage test for the plan must be at least 70%. And yes, the average benefits percentage test must include the other "plans". Those "plans" include the deferral "plan" and matching "plan" within the overall "plan". I hope this helps?

If the plan only has HCEs, the usual profit sharing or 401(k) document with each participant in their own allocation class should be easy enough. No cross-testing (or any 401a4 testing) needed if the only eligible employees are HCEs. Assumes the plan passes coverage.

So this does not become a benefit right or feature that needs testing, when ten years later, the HCE is the only one who has this?

Perhaps this boils down to the definition of the "first plan year" under 416. If you argue that for purposes of 416, the 2011 year is truly the "first plan year", then the plan could be considered as not top heavy in 2012. On the other hand, if you argue that 2012 is truly the "first plan year" because no funds had been deposited to mark the beginning of the plan's trust yet for 2011, then the plan could be considered as top heavy in 2012. I don't see any guidance to this specific issue. Would I be able to stand in front of an auditor and say 2012 is not top heavy and keep a straight face? Maybe, but I would rather discuss the situation with the client and see if they think it may be worth getting a legal opinion, but definitely let them decide the route to take here. Either way chosen, it works out mathematically: 6 / 2 = 3, as proof of this: 2 x 3 = 6 0 / 0 = 59.9, because 59.9 x 0 = 0 (not top heavy) 0 / 0 = 60, because 60 x 0 = 0 (top heavy)

I do not disagree.

No later than the last age in the actuarial chart? Maybe FAPInJax meant there is no maximum age, but left out the "no"?

I doubt the IRS would approve a basic document to have language in it that allows an allocation condition to the gateway. I seriously doubt ASC would write it that way too. That does not mean you can allocate any extra PS above the gateway however, because that would be a profit sharing allocation, which you stated has a 1000 hour requirement. Put each person in their own class and remove the allocation conditions. When each person is in their own class, you could still choose to not allocate PS to those with under 1000 hours and/or are employed on the last day as long as all the nondiscrimination tests and coverage is met.

Very useful, yes - thanks!

Okay. I should also add that any steps taken should not further the plan's operational error. For example, a document I am looking at does not default the plan into "test mode" for ADP purposes if the employer merely fails to give the SH notice. The document indicates that the employer can utilize EPCRS, including SCP, VCP etc. to fix any errors. Thus, it could be another operational error if the plan forged ahead, ran an ADP test, and just made refunds to the HCEs. In this document, the only way that could be correct would be if the employer, decided that was how EPCRS would have it fixed for that year. The actions to fix must also align with the document. edit to say: Aw, PensionPro beat me to making that point.

Maybe the punishment is the mere payment to the IRS for the VCP application. Calendar year plan. Suppose the first payroll of the year is on January 31. The employer hands out the notice on the 2nd of January, just 29 days before that first payroll - yes it's too late anyway, regardless. What is the harm in asking the IRS under VCP to allow the plan to be considered as safe harbor for testing purposes for that year? Certainly they could say, "no way, Jose, we won't agree to the fix you presented here." But it seems reasonable to at least ask, and that certainly does not mean you must ADP test, perhaps they will agree with some other way to give the plan its fix for the year.

First, a minor point: the SH notice deadline is, according to the regulations, due within a reasonable period of time before the first day of the plan year. The IRS will not question a notice provided at least 30 days and no more than 90 days before the beginning of the plan year, but I have not personally seen them try to challenge a notice given 25 days before the beginning of the plan year. As long as the plan's written language itself did not require the 30 days, you have no operational error for giving the notice with less than 30 days to go, but you do fall outside the "we're definitely safe" zone, into the "I wonder what an IRS agent may think of this" zone. No big deal here. Now to the real question: What if you have the language in the document but you didn't give the notice? You have an operational error. What does the plan say in it to fix operational errors? I'd bet it might mention the EPCRS program. Anyway, the plan has failed to act in accordance with its terms. If you discover this problem just a few days after the beginning of the plan year, and if the employer is otherwise eligible for self-correction (has procedures and has reliance on the document language), then perhaps the employer could argue that this is self-correctable by merely giving the notice right away. If so, then document the error, document the steps taken to "fix" it and when it was fixed, and document the procedural change so it can be prevented from happening again. If the employer does not think SCP is the way to go here, then they could submit a VCP application with their mea culpa and offer to the IRS something they think the IRS might agree with, almost anything can be tried here, and you won't know what they will truly accept until you show them all of the facts and explain the circumstances to them. Maybe the worst option is for the employer to do nothing about it and play Russian roulette with the plan.

I would agree. If something happens that subjects the participant to require a gateway (such as a 3% SH allocation, or a top heavy minimum allocation, or a forfeiture allocation), then the gateway minimum for this participant should not be offset by the cash balance accruals that are credited to the other NHCEs.

There is no such thing as a safe harbor match that is contingent solely upon the issuance of a participant notice being provided each year. If a plan is going to provide a safe harbor matching contribution to avoid the ADP test and presumably the ACP test, the safe harbor match provisions must be in place in writing before the beginning of the plan year and a notice must be provided a reasonable amount of time prior to the beginning of that plan year. Thus, as Tom also stated, for the plan to be a safe harbor match plan, some written plan document language or plan amendment language must say it is. A mere statement saying the plan might provide a safe harbor match - indicating no SH match occurs when no notice is provided, and a SH match is contributed if the employer provided the notice, should not be sufficient to satisfy the regulations as it places the safe harbor provisions outside of the written terms of the plan and into certain actions that are discretionary by the employer.

Yes, although not all of them must benefit at 0.50%. Under 401(a)(26), at least 40% must have "meaningful" benefits, that's where the 0.50% comes in. For larger plans it's just 50 employees instead of 40% of the population. For a small employer, if the only 2 employees meet the age and service rules, then at least 2 employees have to have meaningful benefits. In contrast, under 401(a)(4), as discussed in the posts above, we just look for anyone who is merely "benefitting" - which is any accrual. For 401(a)(26) we must count the number of NHCEs whose benefitting amounts are "meaningful". The IRS generally believes this means the participants need to get an accrual in the form of an annuity payable at the participant's normal retirement age of at least 0.50% of pay in order to be "meaningful". This is not a regulatory minimum, it is from internal IRS documents on what they think "meaningful" should be. Again, as before, the 0.50% is an annuity (the accrued benefit) as converted from the cash balance credit, not the actual lump sum cash balance credit itself. So, in the example given where a recent design provided a 2.20% pay credit in the cash balance plan, it turns out that this provided a meaningful benefit to 50% of the employees. edit: typo