Luke Bailey

I think: 1) The 5-year holding period for the Roth IRA begins with the first day of the calendar year in which the rollover amount is deposited into the Roth IRA, e.g., if this year, 1/1/2018. There is no tacking for the portion of the 5-year period already satisfied in the qualified plan. 2) Same answer. You did not ask about the taxation of any distributions from the Roth IRA that occur before the 5-year qualification period is up, but I will offer the following: * In case #1, the taxpayer's basis in the designated Roth account (i.e., the original 2014 conversion amount, but not any subsequent earnings in the qualified plan after 2014) is nontaxable basis in the Roth IRA following the rollover, but the portion of any distribution allocable to the non-basis amount (i.e., any earnings amount above the 2014 original conversion amount) is taxable until the five years has run for the Roth IRA. * In case #2, the entire amount rolled over, i.e. the 2014 conversion amount and all earnings up to the date of the rollover in 2018 is basis. Earnings on that amount within the Roth IRA (i.e., the post-Roth IRA rollover earnings) are taxable when distributed until the new 5-year period for the rollover Roth IRA has run.

I just think it's not completely clear under the statute and regs. All I'm saying. I can understand why the plan administrator and the IRA custodian are taking different positions as described in the original question. This is the IRC. It doesn't necessarily answer all questions or always make sense. It is still early enough in the year to fix. They just need to decide how important it is to them, pick a path, and go with it. If they can't agree, then perhaps someday the IRS will get involved. Most likely, the IRS's position will be the more conservative, i.e. that part of the distribution was 2018 RMD. Doesn't mean that is right or that IRS would win if litigated, although it might. Is this worth litigating? Of course not.

Thanks, Kevin C. Figured that some plans would have language like that, though not required. From a design perspective, such a provision has pros and cons, I think.

Could they have made this more complicated? On closer examination, I think that those of you who think that the first dollars out were RMD and not ERD may be right, but it's not completely clear to me. As for bringing 2017 into the picture, when I think we are all in agreement it is not in the picture, the original question of 52626 asserted that he thought the first dollars out in 2018 could not be rolled over because they were the 2017 RMD. We also all agree, I think, that this person's RBD is 4/1/2019, and, contrary to what I said yesterday, I agree that the third sentence of Treas. reg. 1.401(a)(9)-5(b)(1)(b) tells us that 2018 is a "distribution calendar year" for the individual, his first. Beyond the above, there may still be some ambiguity about what Congress and the IRS have said about this individual's situation. Treas. reg. 1.402(c)-2, Q&A-7(a) says that "if a distribution is required" for a calendar year, the first dollars out are "treated as RMD's under section 401(a)(9), to the extent that the total RMD for the calendar year has not been satisfied." Here, no distribution is required for this individual in 2018, because his 2018 RMD can be delayed until 2019. So reading only so far, one could argue pretty strongly, I think, that all of what he received in 2018 is an ERD. However, 1.401(a)(9)-5, Q&A-1(b) equates a "distribution calendar year" with "a calendar year for which a distribution is required," so one could read that equation, plus the fact that 2018 is a "distribution calendar year" for this individual, into the rule of Treas. reg. 1.402(c)-2, Q&A-7(a). So maybe a distribution IS required for 2018 for this individual. Moreover, the rule of Treas. reg. 1.402(c)-2, Q&A-7(a)(b), which says that any amount paid before January 1 of the year in which an individual turns 70-1/2 is not an RMD, implies that any amount paid in the year the person turns 70-1/2 (presumably, if the person retired in a prior year or is a 5% owner) is RMD, even if paid before the individual attained 70-1/2, and that in turn could be read as implying that the same rule (i.e., any amount paid on or after January 1 of the year in which an individual who is not a 5% owner retires) is RMD. However, 1.402(c)-2, Q&A-7(a) doesn't actually say that, and certainly that rule is highly impractical, since retirement, unlike attainment of age 70-1/2 or dying, is not necessarily a predictable event, and in the case of a plan that permits in-service distribution to someone of any advanced age (e.g., 59-1/2 or NRA), it would be impossible to administer in the case of a distribution requested in the year of retirement, but before retirement. Maybe there is other guidance on this that someone can find, but for me at this point, it's not really clear what Congress and the IRS are saying, here, thus making the position of the plan administrator, as described by 52626, not baseless, to say the least.

I agree with jpdrews. Per IRC sec. 401(a)(9)(C)(i)(II), if not a 5% owner and did not retire until 1/2/2018, his RBD is 4/1/2019. And 2018 is not a "distribution calendar year" for this guy. See Treas. reg. 1.401(a)(9)-5, Q&A-1(b). In this case, the year preceding the RBD year is not a distribution calendar year.

Good point about having to treat all QB's the same way, but in my experience most COBRA outsourcing companies do the bare minimum, i.e., nothing more than what the law requires. Usually, the contract between the employer and the COBRA outsourcing company compensates the COBRA outsourcing company based on the size of the active plan, not the number of COBRA continuees, so there is no incentive to maximize the number of COBRA continuees or the duration of their coverage. In a self-insured plan, at least, the employer may have some discretion, although limited, to modify the outcome in favor of the QB if the facts and circumstances warrant, such as a justifiable mistake of fact. Even in a self-insured plan, however, I would not exercise discretion without first getting approval from stop loss carrier.

Your initial COBRA notice must explain the premium payment requirements. For a missed payment there is a 30-day grace period, but no notice requirement to alert the COBRA continue that you did not get their payment timely within that period. The above is, you know, just "generally speaking."

ESOP Guy, I think that to actually get rid of the dead loan you need more than just a separation from service. The individual has to either request a distribution of there has to be some rule that allows you to distribute without consent, e.g. under $5,000, 401(a)(9), etc. The following sentence is from Treas. reg. 1.72(p)-1, Q&A-13(a)(2): "A distribution of a plan loan offset amount could occur in a variety of circumstances, such as where the terms governing the plan loan require that, in the event of the participant's request for a distribution, a loan be repaid immediately or treated as in default." Obviously, there is an ordering issue, here, e.g. if the dead loan is $10,000 and the individual has a balance of $100,000 and requests a distribution of $20,000, can you say the first $10,000 is the dead loan? Probably, but you would have to have that in your plan or perhaps loan document, or have the flexibility under your plan document to apply that as a uniform rule. Q&A-21 of the same reg actually deals with the issue of the participant's basis where he/she repays a previously deemed distributed loan. While the period in the example in Q&A-21 is much shorter than the three-year period in Andy Daniels' original question, I think the same principle would apply.

These strings get so long that sometimes we get so far into the weeds on the legal and tax issues that we forget the detailed facts of original question. Unleveraged trading of publicly traded stocks and bonds is not going to produce UBIT.

Larry's advice is of course good. When we merge plans and are in control (e.g., represent acquirer), we always prepare a "plan of merger" document that is adopted by sponsor of each plan. Will typically say that the beneficiary designations survive the merger. If you don't do this and someone dies before you get a new beneficiary designation, you will have a slight quandary. Because of 414(l), as you cite, I think that if there was a fight among potential beneficiaries (e.g., participant with no spouse named friend, but following death kids want to say that the beneficiary designation did not survive merger), the argument that the beneficiary designation survived would be stronger in the absence of any other fact or circumstance. But the plan fiduciaries, now that this issue has been identified, owe the participants a clarification one way or the other pending receipt of the new beneficiary designations.

You only count the W-2 from the S corp, since that is the only compensation for services. What comes out on the K-1 from the S corp is his share of the S corp's income (or, in this case, apparently a loss) after comp has been paid.

Google "Peek IRA case."

ESOP Guy, it makes sense if the loan was only deemed distributed (i.e., participant did not take a distribution of his/her account in connection with separation from service) and not offset.

Any pass through entity has the same problem, partnership, S corp, LLC taxable as either of those. The basic holding is from a Rev. Rul. from the 70's. I can find it if you want.

jashendorf, why do you say (c)(1)(D) instead of (c)(1)(E)?

Note that to the extent the problem is IRC 4975(c)(1)(E) (i.e., acting for the "plan" under a conflict of interest), the problem would not go away just because of the plan and individual would not be investors in the same entity. However, separate but simultaneous investment could avoid some "down the road" PT issues, probably.

Ma'am, you need a lawyer. If (as I infer) this is a state or local government plan, you need to find one in the jurisdiction of the plan.

Having worked in the area extensively, and as the previous astute (but opposing) comments demonstrate, the issue is not subject to resolution with the typical degree of certainty that we expect for most federal income tax issues dealing with retirement savings. This is a plan for only a husband and wife, so likely not subject to ERISA, making Elizabeth G's link to an article regarding IRA investment highly relevant. If there is a problem, it will be in IRC sec. 4975 or the UBIT rules of IRC secs. 511 et seq. Mere coninvestment, in and of itself, is not a transaction between the plan trustee/participant, as an individual, and the plan, and does not appear to transfer assets of the plan to the trustee/participant as an individual, so, depending on facts and circumstances, you may be able to check off, as inapplicable, IRC secs. 4975(c)(1)(A) through (D). (Note, however, that the potential for future transactions such as recapitalizations, capital calls, etc., or decisions relating to tax issues of the investment, may mean that down the road there will be indirect transactions between the plan and the trustee/participant as an individual, resulting in an unavoidable prohibited transaction at that time.) But what about IRC sec. 4975(c)(1)(E)? The rules tell us that even though not subject to DOL fiduciary rules, the individual in control of the investment (whether of a one participant plan or an IRA) is to be treated as a fiduciary for purpose of the prohibited transaction rules of IRC sec. 4975. See IRC sec. 4975(e)(3). How far do we take that and how do we apply it where the issue is conflict of interest (IRC sec. 4975(c)(1)(E))? Do we say, "Well, IF this were a plan subject to ERISA and IF the person controlling the investment decision was a real fiduciary and the plan contained other people's money (OPM), he or she certainly wouldn't be able to coinvest personally with the plan that he or she is a fiduciary of, because this would present a conflict of interest with respect to the OPM. Thus, the coninvestment is a prohibited transaction." Or, alternatively, do we say, "What are you talking about? It's this person's own money. As long as he or she honestly believes the purpose of the transaction is to benefit the plan or IRA at least as much as him- or herself individually, it's OK." To the best of my knowledge, this issue has not been settled by IRS, DOL, or the courts, other than in dicta in a footnote in the Ellis case. Because of the ERISA reorganization plan of 1978, the DOL, not the IRS, has all of the regulatory/guidance authority for IRC sec. 4975. But DOL does not have enforcement authority for IRC sec. 4975. Moreover, because all of the plans that are subject to ERISA (basically), and certainly all of the plans that the DOL is interested in, have OPM, the DOL has no interest in clearing up this issue of how closely to pus the "fiduciary" analogy in a case where OPM is not involved. The IRS, on the other hand has no regulatory authority for IRC sec. 4975, although they do occasionally make law as part of their enforcement jurisdiction by bringing cases such as Ellis. So I think in the end there is not going to be a certain answer to your question, Dalai Pookah, even assuming no complicating facts and circumstances, so you need to proceed carefully and with appropriate disclosure to the decisionmaker.

It seems to me that with what you have described you're satisfying all of the requirements of 1.401(m)-3(c). Optics might be better if you didn't call the HCE match discretionary, but say all participants get 4%, but subject to vesting schedule for HCEs. Form a policy perspective you are providing the NHCEs with the required incentive as much as if you said HCEs got the same fully vested match dollar for dollar on first 4%.

Thanks, Linda. That could be very useful if it comes to that. Hoping our facts are good enough will not have to challenge IRS's position.

Peter, you are spot on.

I agree with the preceding posts to some extent. There will be provisions in the partnership agreement that will determine whether the person is still a partner at the end of the year. The individual will likely be listed as a partner, or not, in a schedule, and his/her removal as a partner will be effective on a specific date. But I'm not sure if that's the question, since the word "employed" is not only ambiguous, but downright inappropriate, as applied to a partner. To prevent this sort of ambiguity, I usually include, in plans for law firms or similar entities, that a partner must still be actively providing services to or in connection with the partnership, or available to do so, in order to be considered eligible for an allocation as of a particular allocation date. This avoids a potential nondiscrimination issue where you would effectively have different standards for partners than for employees with respect to allocation eligibility, since a partner who had effectively stopped working for the partnership might still be a partner for some purposes, e.g. for receiving a share of the firm's receivables for a trailing period, or sharing in the appreciation of some of its assets if his/her capital account had not yet been paid to him/her.

If there is a basis for taking the position that the plan was adopted by the company effective 1/1/2018 (e.g., was approved by board of directors), you should be fine. Moreover, the DOL doesn't care (they just want to make sure employees get there benefits) and the IRS does not seem to be as punctilious with regard to welfare documents as they are with retirement plan documents. Even then, a retirement plan can generally be adopted as late as the last day of plan year and be retroactively effective as of first day of year, except that the IRS's position is that the plan must be adopted before you have elective deferrals (again, I would argue that adoption of an unsigned plan document is adoption). If you have employee pre-tax contributions for this health plan through a cafeteria plan, then you may have a greater audit exposure, so see if you can establish that there was other corporate or shareholder action (e.g., board minutes or action of the sole shareholder) showing legal adoption before the plan document was executed.

Thanks for your experience on this, Bob the Swimmer and Pension Panda.

beartd, that's not what Notice 2015-59 is saying. No change contemplated to Q&A-13(a) permitting a change on plan termination, only clarify that Q&A-14 for increases in benefit cannot be used to permit lump sum windows for individuals in pay status for a non-terminating plan. Go back and read carefully the second to the last sentence of fifth paragraph of Section II of notice and compare that with first paragraph of Section III.