INTRODUCTION AND PURPOSE OF MEMORANDUM

Transcription

1 DRAFT MEMORANDUM REQUESTING THAT THE MRD RULES APPLICABLE TO DISTRIBUTIONS TO TRUSTS BE CLARIFIED NOTE THAT THIS HAS NOT BEEN SUBMITTED TO THE COMMITTEE FOR CHANGES YET SECTION 1.1 ARTICLE 1 INTRODUCTION AND PURPOSE OF MEMORANDUM Principal Drafters. This memorandum is respectfully submitted on behalf of the American College of Trust and Estate Counsel (ACTEC). The principal drafters of this memorandum are Jonathan G. Blattmachr, Natalie B. Choate, Virginia Coleman and Noel Ice. Note that they have not signed off on this yet. SECTION 1.2 Purpose. The purpose of this memorandum is to request that the Internal Revenue Service or the Treasury Department issue guidance that is sufficiently clear and unambiguous to allow the minimum required distributions (MRDs) of death benefits that are subject to the provisions of 401(a)(9) of the IRC 1 to be computed with relative certainty in those cases where such benefits are payable to a trust. SECTION 1.3 Where Guidance is Needed and Where it is Not. Trusts as Beneficiaries. The MRD to be paid to a Participant during the Participant s life is easily determined under the final regulations. The MRD to be paid to a Participant s Death Beneficiary is, however, anything but straight forward. If the beneficiary is a trust, the MRD may be impossible to determine with certitude by sole reference to the guidance existing today. This memorandum is concerned with this last mentioned problem. The problem is quite severe. People use trusts for all kinds of reasons, one of the most important of which is to provide for minors, incapacitated persons, and other persons for whom outright distribution could be harmful. This is as true for persons of modest means as it is for the wealthy; perhaps more so, for obvious reasons. Persons of modest means may need access to trusts even more than the wealthy, since the former have the most to lose relatively when faced with a choice between (1) not being able to use a trust for their children or other loved ones who need a trust s protection, or (2) running the risk of paying penalties, interest and additional taxes because they used a trust, without knowing what the rules are. If the child of a person of modest means is forced by the existing rules to receive a distribution outright, the child may end up with nothing by the time the child reaches maturity. This is not so likely a problem for the children of the wealthy, even if there is more to protect. The members of this committee are not generally considered among the most humble when it comes to their knowledge of this area of the law; but even though we have been studying this issue closely since 1986, almost a quarter century ago, none of us is willing to say with confidence that he or she knows what the MRD rules are as applied to trusts. Considering the source, this is a confession not lightly made, and one which should, we hope, move the Service to recognize that there really is a problem. 1 All references herein to the IRC are to the Internal Revenue Code of 1986, as amended, unless otherwise indicated. Page 1 of 26

2 ARTICLE 2 NOMENCLATURE. SECTION 2.1 Plan, Participant, Beneficiary and Designated Beneficiary. In this memorandum all arrangements subject to 401(a)(9) will be referred to generically as Plans and singularly as a Plan, including but not limited to IRAs described in IRC 408 and qualified plans under 401(a). In this memorandum, the term Participant includes but is not limited to an owner of an IRC 408 IRA and an employee under an IRC 401(a) qualified plan. Most Plans provide benefits to beneficiaries at or following the Participant s death, which herein will be referred to as Death Beneficiaries. The term Designated Beneficiary is a term of art, and it means a Death Beneficiary who is a human being. It is defined in IRC 401(a)(9)(E) as an individual [i.e., a human being] designated as a beneficiary by the employee. 2 SECTION 2.2 Minimum Required Distributions (MRDs). Treas. Reg (a)(9)-1 et seq. contain rules setting forth the minimum annual distributions (commonly called minimum required distributions or MRDs ) that a Plan must make to avoid a penalty being imposed under IRC SECTION 2.3 Required Beginning Date (RBD). The term Required Beginning Date (RBD) generally means April 1 of the calendar year following the later of the calendar year in which the employee attains age 70½ or the calendar year in which the employee retires from employment with the employer maintaining the plan. 3 SECTION 2.4 The Multiple Beneficiary Rule. When we refer to the Multiple Beneficiary Rule we mean that (a) if any beneficiary is not an individual there is no Designated Beneficiary, 4 and (b) if all beneficiaries are Designated Beneficiaries, the oldest Beneficiary will serve as the measuring life. 5 SECTION 2.5 The Beneficiary Determination Date. The Beneficiary Determination Date (BDD) that date that Death Beneficiaries are determined under Treas. Reg (a)(9)-4, A-4, for purposes of applying the MRD rules, September 30 of the calendar year following the calendar year of the employee's death. 2 See also Treas. Reg (A)(9)-4 Q&A 1. 3 Treas. Reg (a)(9)-2 Q&A 2. 4 See Dash 4, A-3. 5 See Dash 5, A-7. Page 2 of 26

3 SECTION 2.6 Subsection 2.6(a) Ignored. Mere Potential Successor Beneficiaries. Mere Potential Successor Beneficiaries Can be The regulations make it clear that for purposes of the Multiple Beneficiary Rule, mere potential successor beneficiaries can be ignored. 6 The meaning of this phrase is the source of the problem. The phrase contingent beneficiary is bandied about in the regulations and the preamble, but it is not helpful, and is, indeed, misleading, since the term has a meaning under the common law that is not the meaning being used by the IRS. Rather, the only useful approach is to figure out who is NOT a mere successor beneficiary. Subsection 2.6(b) First, Second and Termination Beneficiaries. Everyone Else is a Mere Potential Successor Beneficiary. If the Service might look at the life expectancy of every person alive on the BDD for whom an accumulation distribution will or might be made out of a trust. That would include the whole world in the case of a dynasty trust. Paragraph 2.6(b)(1) First Beneficiaries. An individual alive on the BDD for whom distributions will or might be made out of a trust without having to survive anyone is herein called a First Beneficiary. Paragraph 2.6(b)(2) Second Beneficiaries. A person alive on the BDD for whom distributions will or might be made out of a trust prior to its termination, if that person survives a term of years or survives a First Beneficiary is herein called a Second Beneficiary. A person alive on the BDD for whom an accumulation distribution will or might be made out of a trust prior to its termination, if that person survives a term of years or survives any other beneficiary living on the BDD is herein called a Second Beneficiary. Paragraph 2.6(b)(3) Termination Beneficiaries. An individual alive on the BDD for whom a termination distribution will be made if all of the First and Second Beneficiaries were to immediately die after the BDD are herein called Termination Beneficiaries. Paragraph 2.6(b)(4) Countable Beneficiaries. First, Second and Termination Beneficiaries are herein called Countable Beneficiaries. Subparagraph 2.6(b)(4)(A) Mere Potential Successor Beneficiaries. As used herein, a Mere Potential Successor Beneficiary is anyone who is not a Countable Beneficiary. Subsection 2.6(c) Example. In the example, income to A for 5-years (or sooner death), then to B for life, remainder outright to C or C s estate, A would be a First Beneficiary, B would be a Second Beneficiary, C would be a Termination Beneficiary, where A, B and C are all living on the BDD. A, B and C are Countable Beneficiaries. The beneficiaries of C s estate would be a Mere Potential Successor Beneficiary. 6 Treas. Reg (a)(9)-5, A-7(c)(1), first and second sentence. Page 3 of 26

4 SECTION 2.7 A Rule Consistent With the PLRs. A rule that is consistent with the PLRs, although otherwise odd, appears to be that Countable Beneficiaries (as above defined) are considered, and everyone else is a mere potential successor and therefore ignored for MRD purposes. ARTICLE 3 THE MRD RULES IN BRIEF SECTION 3.1 Exceptions and Exceptions to the Exceptions and the Exceptions to the Exceptions to the Exceptions. In order that this memo does not turn into a treatise, the following background rules are stated only generally, omitting some of the exceptions, the exceptions to the exceptions, and the exceptions to the exceptions to the exceptions, of which there are a multitude, except where necessary to mention them, which it sometimes will be. SECTION 3.2 MRDs During the Participant s Life. During the lifetime of the Participant, the MRD is straightforwardly determined each year by reference to the Uniform Lifetime Table found in Treas. Reg (A)(9)-9 A2, taking into account only the Participant s age, usually without regard to the age or identity of the Death Beneficiary. (Of course, there are exceptions, participants with young spouses as beneficiaries being one.) SECTION 3.3 MRDs Following the Participant s Death. Determining the MRDs following death is a complicated matter, and depends on who the Death Beneficiary is, the age of an individual Death Beneficiary (the age of a Designated Beneficiary), the number of Death Beneficiaries, the identity of the Death Beneficiary (whether or not the Death Beneficiary is an individual and, if an individual, whether the individual was the Participant s spouse), and whether or not the Participant died before or after the RBD, etc. If the beneficiary is a trust, the variables just mentioned all apply, but with a vengeance, such that, as previously admitted, not even the members of this committee are sure what the rules are. Subsection 3.3(a) Death Before or After the RBD, Where the Death Beneficiary is an Individual. If the Death Beneficiary is an individual (a Designated Beneficiary), or a Qualifying Trust (where the beneficiaries of the trust are treated as if they were Designated Beneficiaries), and the Participant dies (i) after the RBD, or (ii) before the RBD and the 5-year rule is not elected or mandated, then death benefits must be paid out based on the beneficiary s age each year using the factor set forth in Table V of IRC , also called the Single Life Table of Treas. Reg (a)(9)-9, A-1, with an exception if the Participant dies before the RBD and the beneficiary is the Participant s spouse. This is sometimes referred to as the Life Expectancy Method. The distribution period begins by determining the beneficiary s age in the year after the Participant s death. If the Death Beneficiary is not the spouse of the Participant, the beneficiary consults Table V (the Single Life Table ) only once. After first determining the life expectancy factor in the year following the Participant s death, the nonspouse Death Beneficiary subtracts one for each year thereafter. If the Participant s Death Beneficiary is the Participant s spouse, the Participant s spouse can rollover and defer taking benefits until the Participant s spouse turns 70½; or not rollover, and defer until the Participant would have been 70½. Also, a Participant s spouse determines the life expectancy factor by Page 4 of 26

5 referring to the Table V each year, afresh; in effect recalculating life expectancy each year, if there is no rollover. (If there is a rollover, it is as if the IRA were the Participant s spouse s all along, and therefore the MRDs are determined under The Uniform Lifetime Table of Treas. Reg (A)(9)-9 A2.) Subsection 3.3(b) Where the Death Beneficiary is Not an Individual. If a beneficiary is not an individual, the life expectancy tables obviously do not apply. In this case it makes a big difference whether death is before, of after, the RBD. Paragraph 3.3(b)(1) Death Before the RBD Without a Designated Beneficiary- The Five Year Rule. If a beneficiary is not a human being (hence not a Designated Beneficiary ), and the Participant dies before the RBD, the entire account in the Plan for the Participant must be distributed by the end of the fifth calendar year following the calendar year in which the plan Participant died. This is called the 5- year rule. Moreover, even if the Death Beneficiary is an individual, if and if the Participant dies before the RBD, the 5-year rule can still apply, if the Plan allows for it and the Death Beneficiary elects for it to apply, or if the Plan requires that it apply, etc.. 7 Paragraph 3.3(b)(2) Death After the RBD Without a Designated Beneficiary. If the Participant dies after the RBD without a Designated Beneficiary (i.e. where the Death Beneficiary is not a human being), then the death benefits must be distributed over the remaining life expectancy of the Participant (even if the beneficiary is a Non-Qualifying trust), under the same Uniform Lifetime Table of Treas. Reg (A)(9)-9 A2, the same Table that applied while the Participant was alive, except that the table is referred to only once, in the year following the year of death, and the life expectancy factor is simply reduced by one for each year thereafter. 7 Can the 5-Year Rule Apply if the Participant Dies After the RBD? The five year rule should have no application if the Participant dies after his or her RBD. This rule could be stated unequivocally were it not for one, important, controversial PLR, found in none of the published literature because no one believed there was such an exception until PLR was issued. The PLR applied 401(a)(9)(B)(iv)(II) and Treas. Reg (a)(9)-3, Q&A-5 to in a case where the Participant died after his RBD. 401(a)(9)(B)(iv)(II). The working common sense and correct assumption was and is that Treas. Reg (a)(9)-3, Q&A-5 only applies where the Participant dies before the RBD. This exception to the general rule, if it is indeed the law, is important enough to note. In that ruling the Participant named his wife as his sole beneficiary of his IRA. The Participant died after his RBD, survived by his wife. His wife died a few months after her husband before reaching her RBD, and without having taken any action with respect to her husband s IRA. (The IRA was the husband s; the wife was a mere beneficiary, not an owner.) Again, there was no rollover and the Participant dies after his RBD. The wife had not named a beneficiary -hardly having had the time to do so, and, not being the IRA owner, probably lacking the power to name one. The five year rule applied at the wife s death, even though it was not the wife s IRA and even though the husband, the IRA owner, died after reaching his RBD. This unexpected and controversial ruling, which, if correct, is probably the only time that the 5- year rule is ever applied where death is after the Participant s RBD (in the absence of a rollover or its tax equivalent), and is mentioned here only because of its possible application if a spouse is a beneficiary of a trust. We do note, in this context, that we believe that the Service s position is probably that no special rules otherwise applicable to Participants spouses can ever apply if the spouse is a beneficiary of a trust that is the Death Beneficiary. We believe that the Service should state this point unequivocally, assuming it to be the case, in order to make this point more clearly. Page 5 of 26

6 SECTION 3.4 Problem Areas. Subsection 3.4(a) Multiple Beneficiaries. When we refer to the Multiple Beneficiary Rule we mean that (a) if any beneficiary is not an individual there is no Designated Beneficiary, 8 and (b) if all beneficiaries are Designated Beneficiaries, the oldest Beneficiary will serve as the measuring life. 9 Subsection 3.4(b) The Multiple Beneficiary Rule. At one point the preamble states the rule this way. If this was really the rule, there would be no controversy: [I]f a beneficiary (subsequent beneficiary) is entitled to any portion of an employee's benefit only if another beneficiary dies before the entire benefit to which that other beneficiary is entitled has been distributed by the plan, the subsequent beneficiary will not be considered a beneficiary [merely because the person could become the successor to the interest of one of the employee's beneficiaries after that beneficiary's death]. The regulations say much the same thing at Dash 5, A-7(c)(1), using the term successor beneficiary in place of subsequent beneficiary. However, the regulations themselves add the bracketed clause, which is what throws the whole area into turmoil because no one knows what it means. The Preamble treats the statement that the subsequent beneficiary will not be considered a beneficiary as an exception to the general rule that- [I]n determining an employee's beneficiaries for purposes of applying the multiple beneficiary rule or determining if the employee's spouse is the employee's sole beneficiary, all beneficiaries of the employee's interest in the plan, including contingent beneficiaries, are taken into account. This is the first time the term multiple beneficiary rule is mentioned. It is mentioned one more time later in the same paragraph of the Preamble and never again. Notwithstanding, we are treated as if this were a well-known defined term. The term is not used in either and Dash 4, A-3 or Dash 5, A-7; nevertheless we think that Dash 4, A-3 and Dash 5, A-7 describe it. So, contingent beneficiaries are taken into account but subsequent beneficiaries are not, a subsequent beneficiary being (apparently) one who takes only if another beneficiary dies. Here is the problem: having articulated a general rule with a very big exception, but one which has the advantage of being relatively clear, the Preamble goes on to cloud the question- These regulations clarify that the exception from the multiple beneficiary rules for death contingencies only applies to a person who could be entitled to a portion of the employee's benefit by becoming the successor to the interest of one of the employee's beneficiaries after that beneficiary's death. [So far so good. But then...] The regulations provide that this rule [the exception?] does not apply to a person who has any right (including a contingent right) to an employee's benefit beyond being a mere potential successor to the interest of one of the employee's beneficiaries upon that beneficiary's death. [Emphasis added.] 8 See Dash 4, A-3. 9 See Dash 5, A-7. Page 6 of 26

7 Put another way: even though a beneficiary may be entitled to [a] portion of an employee's benefit only if another beneficiary dies, and is thus not ordinarily considered under the Multiple Beneficiary Rule, the rule will nevertheless apply unless such person is a mere potential successor. Again quoting from the Preamble- Thus, for example, if one beneficiary has a right to any income on an employee's individual account during that beneficiary's life and another beneficiary has a right to the principal but only after the death of the income beneficiary (with any portion of the principal distributed during the life of the income beneficiary to be held in trust until that beneficiary's death) 10, both beneficiaries must be taken into account in determining the beneficiary with the shortest life expectancy and whether only individuals are beneficiaries. 11 [Emphasis added.] According to the regs themselves- (1) A person will not be considered a beneficiary for purposes of determining who is the beneficiary with the shortest life expectancy under paragraph (a) of this A-7, or whether a person who is not an individual is a beneficiary, merely because the person could become the successor to the interest of one of the employee's beneficiaries after that beneficiary's death. 12 This sentence is followed by an important exception- However, the preceding sentence does not apply to a person who has any right (including a contingent right) to an employee's benefit beyond being a mere potential successor to the interest of one of the employee's beneficiaries upon that beneficiary's death. 13 If there is more than one beneficiary, then all of them must be human beings, or there is no Designated Beneficiary. 14 Further, if all of the beneficiaries are human beings, the life expectancy of the oldest among them is the life expectancy to be used. 15 These are the Multiple Beneficiary Rules, and they pose an especial problem for trusts, since it is metaphysically impossible to guaranty that under no circumstances can the trust property escheat to the state, or pass to a beneficiary s estate, without absurd language that would probably violate the Ascertainable Beneficiary Rule, 16 such as resorting 10 Trust principal or Death Benefits? It must be the latter, even though principal is a term normally ascribed to trust assets that are not allocated to income, and income and principal are not words ordinarily used to describe the assets of a Plan. This parenthetical is very confusing. Probably the principal distributed means benefits paid by the Plan ; it must. 11 So we know that if all the income goes to A for life, and principal to B on A s death, A & B are both beneficiaries. 12 Treas. Reg (a)(9)-5, A-7(c)(1), first sentence. 13 Treas. Reg (a)(9)-5, A-7(c)(1), second sentence. 14 Treas. Reg (a)(9)-4 A-3, second sentence, expresses one part of the rule this way: If a person other than an individual is designated as a beneficiary of an employee's benefit, the employee will be treated as having no designated beneficiary for purposes of section 401(a)(9), even if there are also individuals designated as beneficiaries. However, see A-5 of this section for special rules that apply to trusts and A-2 and A-3 of 1.401(a)(9)-8 for rules that apply to separate accounts. 15 Treas. Reg (a)(9)-5 A-7(a)(1). 16 Treas. Reg (a)(9)-4, A-1 and A-5(b)(3) describe the Ascertainable Beneficiary Rule. Page 7 of 26

8 perhaps to the telephone book to prevent escheat. There is an exception to the Multiple Beneficiary Rule if each beneficiary has his or her or their own certain separate account. 17 The exception that is necessary to makes the rule work is the one that ignores a beneficiary [who] is entitled to any portion of an employee's benefit only if another beneficiary dies before the entire benefit to which that other beneficiary is entitled has been distributed by the plan, if, but only if, the beneficiary a mere potential successor. Subsection 3.4(c) Requirement that Beneficiary be Identifiable. Treas. Reg (a)(9)-4, A-5(b)(3) provides that for the life expectancy method to be used, the beneficiary must be identifiable within the meaning of Treas. Reg (a)(9)-5, A-1. This is sometimes referred to as the Ascertainable Beneficiary Rule. Treas. Reg (a)(9)-5, A-1 also provides, inter alia, that where the trust beneficiaries are described by a class (e.g., the descendants of the plan participant), individuals who may be added to the class must be identifiable so it is possible to identify the member of the class with the shortest life expectancy. It is not clear what the limits of identifiability are if an actuarially unusual order of deaths is postulated. SECTION 4.1 ARTICLE 4 TRUSTS AS BENEFICIARIES Qualifying Trusts and Non-Qualifying Trusts. There is an exception to the rule that in order for the Life Expectancy Method to apply the Death Beneficiary must be an individual. In the case of a Qualifying Trusts, the regulations apply the Multiple Beneficiary Rule to the beneficiaries of the trust for the limited purpose of assigning a measuring life, treating, to that extent, the beneficiaries of the trust as the Death Beneficiaries of the Plan. 18 All of the beneficiaries must be human beings and numerous other requirements must be met. Trusts that qualify for the exception are referred to in this memo as Qualifying Trusts, (or See-Through Trusts), and those that do not as Non-Qualifying Trusts. SECTION 4.2 The Operative Question. Which beneficiary of a trust is to be treated as the Death Beneficiary for the purpose of the Multiple Beneficiary Rule is the operative question for which we seek guidance, because the regulations as interpreted in the PLRs leave so much uncertainty that the exception for trusts cannot be relied upon with confidence at least in the case of trusts that can accumulate Plan distributions ( Accumulation Trusts ). The regulations attempt to mitigate the last two possibilities by providing a special rule for Successor Beneficiaries, 19 but the approach, as reflected in the PLRs is unsatisfactory, and in some cases simply unknowable. See expanded discussion below. 17 Treas. Reg (a)(9)-8 and 1.401(a)(9)-5, A-7(a)(2). 18 Treas. Reg, 1.401(a)(9)-4 Q&A Treas. Reg (a)(9)-5, A-7(c). Page 8 of 26

9 SECTION 4.3 Contingent Beneficiaries, Subsequent Beneficiaries, Successor Beneficiaries, and Mere Potential Successors. It should be self-evident that a beneficiary can die before a trust terminates. It is probably impossible to prevent this contingency from happening. If this happens (or could happen), the trust corpus must pass either to the beneficiary s estate, to the state, or to some other beneficiary. Therefore, we have to ask, Who is the measuring life under the life expectancy rule? Recognizing this unavoidable issue, Treas. Reg (a)(9)-5, A-7(c) provides rules for determining when a contingent beneficiary of a trust will be considered to be the Participant s beneficiary for purposes of applying the MRD rules. As indicated previously, more guidance is needed in this area. In fact, this is the most important and most uncertain issue facing the drafter of every trust which holds as an asset the right to receive distributions from a Plan. The following Article is devoted entirely to this subject. SECTION 5.1 ARTICLE 5 THE EXISTING GUIDANCE The Regulations. The following, plus two examples in Dash 5 A-7(c)(3), constitute all of the regulatory guidance we have on the subject under discussion. Subsection 5.1(a) Dash 5, A-7(c)(1). Treas. Reg (a)(9)-5, A-7(c). Successor Beneficiary. (1)A person will not be considered a beneficiary for purposes of determining who is the beneficiary with the shortest life expectancy under paragraph (a) of this A-7, or whether a person who is not an individual is a beneficiary, merely because the person could become the successor to the interest of one of the employee's beneficiaries after that beneficiary's death. However, the preceding sentence does not apply to a person who has any right (including a contingent right) to an employee's benefit beyond being a mere potential successor to the interest of one of the employee's beneficiaries upon that beneficiary's death. Thus, for example, if the first beneficiary has a right to all income with respect to an employee's individual account during that beneficiary's life and a second beneficiary has a right to the principal but only after the death of the first income beneficiary (any portion of the principal distributed during the life of the first income beneficiary to be held in trust until that first beneficiary's death), both beneficiaries must be taken into account in determining the beneficiary with the shortest life expectancy and whether only individuals are beneficiaries. 20 Subsection 5.1(b) Dash 5, A-7(c)(3), Example 1 An Acceptable Accumulation Trust. Example 1. (i) Employer M maintains a defined contribution plan, Plan X. Employee A, an employee of M, died in 2005 at the age of 55, survived by spouse, B, who was 50 years old. Prior to A's death, M 20 We think that that if, in the example give, principal could have been distributed to the first beneficiary, the trust would still qualify. These were after all the facts in Example 1. Page 9 of 26

10 had established an account balance for A in Plan X. A's account balance is invested only in productive assets. A named a testamentary trust (Trust P) established under A's will as the beneficiary of all amounts payable from A's account in Plan X after A's death. A copy of the Trust P and a list of the trust beneficiaries were provided to the plan administrator of Plan X by October 31 of the calendar year following the calendar year of A's death. As of the date of A's death, the Trust P was irrevocable and was a valid trust under the laws of the state of A's domicile. A's account balance in Plan X was includible in A's gross estate under (ii) Under the terms of Trust P, all trust income is payable annually to B, and no one has the power to appoint Trust P principal to any person other than B. A's children, who are all younger than B, are the sole remainder beneficiaries of the Trust P. No other person has a beneficial interest in Trust P. Under the terms of the Trust P, B has the power, exercisable annually, to compel the trustee to withdraw from A's account balance in Plan X an amount equal to the income earned on the assets held in A's account in Plan X during the calendar year and to distribute that amount through Trust P to B. Plan X contains no prohibition on withdrawal from A's account of amounts in excess of the annual required minimum distributions under section 401(a)(9). In accordance with the terms of Plan X, the trustee of Trust P elects, in order to satisfy section 401(a)(9), to receive annual required minimum distributions using the life expectancy rule in section 401(a)(9)(B)(iii) for distributions over a distribution period equal to B's life expectancy. If B exercises the withdrawal power, the trustee must withdraw from A's account under Plan X the greater of the amount of income earned in the account during the calendar year or the required minimum distribution. However, under the terms of Trust P, and applicable state law, only the portion of the Plan X distribution received by the trustee equal to the income earned by A's account in Plan X is required to be distributed to B (along with any other trust income.) (iii) Because some amounts distributed from A's account in Plan X to Trust P may be accumulated in Trust P during B's lifetime for the benefit of A's children, 21 as remaindermen beneficiaries of Trust P, even though access to those amounts are delayed until after B's death, A's children are beneficiaries of A's account in Plan X in addition to B and B is not the sole designated beneficiary of A's account. 22 Thus the designated beneficiary used to determine the distribution period from A's account in Plan X is the beneficiary with the shortest life expectancy. B's life expectancy is the shortest of all the potential beneficiaries of the testamentary trust's interest in A's account in Plan X (including remainder beneficiaries). Thus, the distribution period for purposes of section 401(a)(9)(B)(iii) is B's life expectancy. Because B is not the sole designated beneficiary of the testamentary trust's interest in A's account in Plan X, the special rule in 401(a)(9)(B)(iv) is not available and the annual required minimum distributions from the account to Trust M must begin no later than the end of the calendar year immediately following the calendar year of A's death. 21 This makes it clear that what is being approved here is NOT a Conduit Trust. Trusts that must distribute all Death Benefits upon receipt are called Conduit Trusts, and we understand them. The problem is that for a Conduit Trust is almost like having no trust at all. 22 The children are younger, so we don t particularly care. However, there are several PLRs involving trusts for children for a term of years, but if the children failed to survive, then to the spouse, an uncle etc. In Example 1, it is almost certain that the children will take. In the PLRs just referred to it virtually certain that the children alone will take, and the contingent beneficiary will take only if there is an actuarially unusual order of deaths. Nevertheless, the PLRs use the life expectancy of the oldest, contingent beneficiary. Page 10 of 26

11 Paragraph 5.1(b)(1) Example 1, Reduced to Its Essence. Reduced to its essence, Example 1 says that a trust to A s spouse B for life, remainder to B s children works. Paragraph 5.1(b)(2) Open Facts in Example 1. What don t we know about the Example 1 trust? We don t know whether the trust can continue for so long as the children are minors. We don t know what the trust said was to happen if a child predeceased the spouse. And we don t know whether these questions are even relevant. The regulations suggest by their silence that the questions are not relevant. Under the PLRs briefed below, these questions are very relevant. Paragraph 5.1(b)(3) What Facts in Example 1 Are Not Crucial. Example 1 trust was apparently a marital deduction trust that required all the income to be distributed, and provided further that principal might or might not be accumulated. Our guess is that the fact that all the income had to be distributed is not crucial or relevant. What was relevant was that not all Death Benefits received by the trust had to be distributed during the life time of the First Beneficiary or Current Beneficiaries, i.e. beneficiaries living on the BDD who do not have to outlive anyone else to be eligible to receive distributions Subsection 5.1(c) Example 2. Dash 5, A-7(c)(3), Example 2 Conduit Trusts. (i) The facts are the same as Example 1 except that the testamentary trust instrument provides that all amounts distributed from A's account in Plan X to the trustee while B is alive will be paid directly to B upon receipt by the trustee of Trust P. (ii) In this case, B is the sole designated beneficiary of A's account in Plan X for purposes of determining the designated beneficiary under section 401(a)(9)(B)(iii) and (iv). 23 No amounts distributed from A's account in Plan X to Trust P are accumulated in Trust P during B's lifetime (a)(9)(B)(iii) and (iv): (iii) (iv) Exception to 5-year rule for certain amounts payable over life of beneficiary. If (I) (II) (III) any portion of the employee's interest is payable to (or for the benefit of) a designated beneficiary, such portion will be distributed (in accordance with regulations) over the life of such designated beneficiary (or over a period not extending beyond the life expectancy of such beneficiary), and such distributions begin not later than 1 year after the date of the employee's death or such later date as the Secretary may by regulations prescribe, for purposes of clause (ii), the portion referred to in subclause (I) shall be treated as distributed on the date on which such distributions begin. Special rule for surviving spouse of employee. If the designated beneficiary referred to in clause (iii)(i) is the surviving spouse of the employee (I) (II) the date on which the distributions are required to begin under clause (iii)(iii) shall not be earlier than the date on which the employee would have attained age 70½, and if the surviving spouse dies before the distributions to such spouse begin, this subparagraph shall be applied as if the surviving spouse were the employee. Page 11 of 26

12 for the benefit of any other beneficiary. 24 Therefore, the residuary beneficiaries of Trust P are mere potential successors to B's interest in Plan X. Because B is the sole beneficiary of the testamentary trust's interest in A's account in Plan X, the annual required minimum distributions from A's account to Trust P must begin no later than the end of the calendar year in which A would have attained age 70½, rather than the calendar year immediately following the calendar year of A's death. Subsection 5.1(d) Comments on the Examples. As set forth above Treas. Reg (a)(9)-5, A-7(c)(3) sets forth two examples to illustrate the contingent beneficiary rule when a trust is named as the plan beneficiary. In Example 1, a testamentary trust is named as the Death Beneficiary upon the Participant's death. Under the terms of the trust, all trust income is payable to the Participant's surviving spouse for life and then the principal of the trust is payable to the Participant's children. The example provides that no one has the power to appoint the property to anyone (e.g., the surviving spouse had no power to appoint the principal upon his or her death). The example states No other person [that is, no one but the surviving spouse and children] has a beneficial interest in [the] Trust. The example also states that the surviving spouse has the shortest life expectancy (that is, the surviving spouse is the oldest trust beneficiary) and, therefore, the distribution period for making MRDs is based upon his or her life expectancy. Example 2 is the same as the first example except all of the retirement plan distributions (and not just the trust s income) made to the trust during the lifetime of the surviving spouse must be paid to the surviving spouse. This second example concludes that the surviving spouse is treated as the sole plan beneficiary for purposes of determining MRDs. Since it is obvious that even in the two examples given by the regulations the trust could either escheat to the state or pass to an older collateral relative or pass to someone s estate, and that it is metaphysically impossible to draft around the problem and still have the beneficiaries be considered relatively identifiable, 25 what are we to do? All the beneficiaries of the trust in the two examples could die in a car crash two minutes after the Participant s death. Thus Treas. Reg (a)(9)-5, A-7(c)(3) must be interpreted with a grain of common sense. This is where guidance is needed. SECTION 5.2 The PLRs. At the end of 2009, there were nine PLRs construing or using the phrase mere potential successor under the Dash-4 trust MRD regulations, eight issued in 2006 and one in The eight issued in 2006 are as follows: , , , , , , and is the one ruling in It is the last one to date. Subsection 5.2(a) What Do the PLRs Have in Common/ Do they Suggest that there is a Rule or a Policy that the IRS Has Formulated Internally. The PLRs do have some consistency, and they do suggest that there may be a rule. That rule is not in the regulations, however, so one relies on it at one s peril, which is why this committee would like some clarification. 24 Here we have clear authority that the so-called Conduit Trust works, but a Conduit Trust defeats the reason for setting up a trust in the first place, most of the time. If you want Conduit Trust treatment, it is probably just as easy to provide in the beneficiary designation that the beneficiary is only to receive MRDs, and on the beneficiary s death, the amounts remaining, if any, will pass to descendants per stirpes. 25 Attempting to do so leads to absurdities. Page 12 of 26

13 Paragraph 5.2(a)(1) A Rule? The PLRs suggest that the it is not enough to say to C (children) for life, remainder to C s estate, and stop there. You have to look at who C s heirs-at-law are. Thus if C has no children, you have to go up the line. See PLR Conclusion, vesting is not enough. If C has children living on the BDD, C will be the measuring life. This is good. If C does not have children living on the BDD, you have to see who is next. You probably get to stop there, but this is not clear. PLR is the tough one: To C1 until 35, but if C1 fails to reach age 35, then to C1 s presently unborn descendants if any, but if none, to C2 outright, but if C2 is not age 35 at the time then to C2 s presently unborn descendants if any, but if none then outright and free of trust to 90 year old great uncle. Great uncle is the measuring life in that PLR. How do we articulate a rule to square that result with Example 1. Any rule has to be able to fit this example. Paragraph 5.2(a)(2) The Rule? The PLRs indicate that the IRS will take into account every potential beneficiary of the trust living on the BDD (1) who could directly benefit from the trust prior to its termination and while it is ongoing, and (2) everyone else (again, taking into account only persons alive on the BDD) who would take upon the trust s termination if all the other beneficiaries were to die immediately after the BDD. See Article 2 definitions of First, Second and Termination Beneficiaries, who are all Countable Beneficiaries. Everyone else is a mere successor beneficiary at best. Not an elegant rule or one easy to articulate, but are there any PLRs outside it? Paragraph 5.2(a)(3) Some Examples Applying the Rule?. In the example given in the definition section, income to A for 5-years (or sooner death), then to B for life, remainder outright to C or C s estate, A would be a First Beneficiary, B would be a Second Beneficiary, C would be a Termination Beneficiary, where A, B and C are all living on the BDD. A, B and C are Countable Beneficiaries. The beneficiaries of C s estate would be a Mere Potential Successor Beneficiary. Consider this example: To P (parent) for life, remainder (outright?) to C (children then living). This surely passes muster. This is, in fact, Example 1 under the regulations, if one assumes that the gift to C is not to be held in further trust. If the trust could continue for C, but if C dies prior to its termination others could benefit, the PLRs, but not the regulations, suggest we have to look further for more potential beneficiaries living on the BDD. How about to C for 10-years and then to C outright, but if see does not survive the term, then to C s estate. C has no descendants. C is the First Beneficiary. Who is alive on the BDD if C does not survive 10- years? That person is a Termination Beneficiary and therefore a measuring life if older than C. See PLR In the PLRs we looked to see who the intestate heirs were. They counted. Presumably whether C has a will or not would make no difference. The trust could name the successor to C if C were to die immediately, and that person, if living on the BDD would be in the class of measuring lives. See PLR Thus, where the disposition was in trust to a class of two then living minor children until age 35, but if both of the children die without issue prior to either reaching age 35 then outright and free of trust to a 90 year old great-uncle if living, the measuring life is the 90 year old if living, even though his chance of taking is less than that of winning the $100 million power ball (with or without a lottery ticket, which at that point is statistically irrelevant). See PLR Good thing the trust didn t continue for the 90 year old or it could have gotten worse. Page 13 of 26

14 Subsection 5.2(b) PLR This PLR does not contain the phrase mere potential successor, but is relevant nonetheless. Prior to the advent of this PLR, commentators thought that the actuarial likelihood that a trust would in fact accumulate distributed Death Benefits for the use of secondary beneficiaries was a relevant factor. This PLR makes it clear that it is not. The ruling is interesting because this was not a conduit trust. It was a trust for health, maintenance and support, which terminated when the child reached age 30. This pretty standard plain vanilla: No powers of appointment, no dynasty provisions, etc.; just your basic support trust that terminates at a young age. Specifically, in PLR , a support trust for two minor children was established by a California settlor who died before her RBD. A trust was the beneficiary of the settlor s IRA. There were two beneficiaries, ages 5 and 6, we believe. The trust was to terminate for each child when the child reached age 30. If one of the children died before age 30, the child s share went to the child s issue. If there were none, it went to the other child. If both children died before age 30 without issue, the trust passed to a 67 year old great-uncle. Though we don t know, the result would probably have been the same under the California intestacy statute if no mention had been made about where the property would go in that extremely unlikely event. The PLR did not recite that the if the child died prior to age 30 that the child s children, if any, would have been beneficiaries, but it did recite that the trust would go to the other beneficiary before going to the great-uncle. Apparently the IRS thought that all that was relevant was that the great-uncle might take if both children died before age 30 without issue. The Service failed to consider that if the greatuncle failed to survive both children and their issue (which was so unlikely that it should have been immaterial), that for all we know the trust would have passed to the great-uncle s great-uncle, who could possibly be celebrating his 110th birthday as you read this. Escheat to the state was possible too, one presumes. The ruling was issued after the final MRD (minimum required distribution) regulations were issued, and the author of the ruling was well aware of their content. What if the trust said to Child A until age 30, but if child dies prior to age 30, then to a 25 year old cousin (instead of the uncle), but if the cousin is not then living to the uncle, but if the uncle... What the trust said was to Child A but if... then to Child B. How is that situation materially different from Ex. 1? The life expectancy of the 67 year old great-uncle was used as the measuring life, since there was an extremely remote possibility that he might inherit trust accumulations. Why, we ask, was a similar concern not expressed in example 1. Were A s children for some reason thought to be less mortal than the children in PLR ? After all, they could die without issue too, in which case the accumulations would have to go somewhere, very likely to someone older than B, if not escheating to the state. How can Treas. Reg (a)(9)-5, A-7 (c)(3) Ex. 1 and PLR both be correct? The truth is that they cannot. Which leaves us wondering, what is the law? What was the actuarial likelihood that the uncle in PLR would benefit from the trust? Was it less than 5%? Can the actuarial likelihood off the uncle benefitting be computed easily? What is most absurd is that if the trust had failed to mention the great-uncle, then we would have had the exact same fact pattern as described in subparagraph (ii) in Example 1 of Treas. Reg (a)(9)-5, A-7 (c)(3), quoted above: A's children, who are all younger than B, are the sole remainder beneficiaries of the Trust P. No other person has a beneficial interest in Trust P. But what if one of A s children predeceased B? In Ex. 1 an older generation was interposed before the younger, which made it highly likely that the younger generation would take. In PLR the facts and the likely outcome Page 14 of 26

15 were the reverse. But really, there were interposed potential beneficiaries living at the death of the Participant: Child 1 and Child 2. Only if both died early would the uncle take. In Example 1, when the spouse died, the trust would pass to the children. If the children predeceased the spouse the interest would have passed to... That is, we don t know where it would have passed, which strongly suggests that the question is irrelevant. If irrelevant, what was the basis for using the uncle as the measuring life in PLR This is not complicated. The issue is simply when can we stop looking for potential beneficiaries? If we have to look at what will happen if a 5 and 6 year old each die before 30 without issue, then we might as well wonder what the rule would be if we said that instead of the great-uncle, the trust will pass to the first person in the Dallas phone directory who is under age 21. How about saying that on the child s death before 30 the trust terminates and is distributed to the child s estate. Is that okay? Perhaps the only reasonable approach would be to treat PLR as an aberration; take Ex. 1 at its word. Provide that benefits go to child A until age 30, but if child A does not reach 30 then to child B. Assert that this is example 1 which would certainly be a reasonable assertion- and leave everyone to guess where the property will go if all the descendants die PLR to be an aberration. Unfortunately PLR is not a total aberration. Let us consider the nine PLRs that actually interpret the phrase mere potential successor. This PLR surprised virtually all of the commentators. Subsection 5.2(c) PLR Taxpayer A was survived by one child, Taxpayer B, and B s 10-year son, Taxpayer C. Taxpayer B was married to Taxpayer D, C s mother. A trust was the beneficiary of the Plan. The trust provided that on A s death half of the trust passed to B outright, and the remainder was held for the sole benefit of C until C reached age 25, at which age the trust would pass outright to C if living, and if not living then to his heirs-at-law, who were B and D (at the time of A s death). As among B, C and D, B was the oldest. In nine pages the Service concluded that the life of C s mother, B, was the measuring life. Is some version of the Multiple Beneficiary Rule at work here, such that B is treated as a theoretical beneficiary of C s share because C s share somehow failed the separate account rule? Apparently. In this case, IRA X passes through Trust W prior to being distributed to its beneficiaries. Thus, pursuant to Section 1.401(a)(9)-4 of the Final regulations, Q&A-5(c), the separate account rules under A-2 of section 1.401(a)(9)-8 are not available to beneficiaries of Trust W with respect to the trust's interest in the Taxpayer A's IRA X. Therefore, the oldest beneficiary among Taxpayers B, C, and D will be the designated beneficiary of both Taxpayer B's and Taxpayer C's interests in IRA X. Taxpayer B is the eldest beneficiary. But D was a virtual stranger to the trust and should on no account have been included in the mix, as one of several (multiple) beneficiaries or otherwise. 26 We would have predicted that under Ex. 1 D would have been a mere potential successor beneficiary. Example 1 did not say where the trust would go if the C in that case (who ironically was a child rather than a grandchild) was not living at the date of death of the surviving spouse, and did not explore who 26 B got his half of the IRA outright at his father s death, but that was not enough, apparently, for it to be a separate share. This may moot the whole issue of what would have happened if B had not been an outright beneficiary of half the IRA. One presumes that because C might die in the next 10 years without children a very remote contingency indeed, statistically- C s father was the measuring life under the Multiple Beneficiary Rule if his trust and its beneficiaries are aggregated with C s subtrust. Page 15 of 26