The rules governing distributions from qualified plans and IRAs are exceedingly and unnecessarily complex. Please read IRC §§72, 401(a)(9), 402 and 408(d) (and the regulations thereunder) very quickly at this point in the outline if the reader has the least doubt on this score. Quite frankly, there is no need for this sort of nonsense in our system of taxation. There may be areas of the law where complexity is necessary and unavoidable, but here there is no excuse, other than that the drafters are incompetent, lazy and unconcerned.
As if to make clear to all who doubt the wisdom of Congress, and in order to once again reaffirm that the taxation committees are entirely lacking in judgment and common sense, Congress as part of its yearly ritual will next year almost certainly make numerous changes in the already prolix and overburdened language which sets forth the rules governing distributions from qualified plans and IRAs.
In the light of these most recent changes it is time for each of us to ask, and to perhaps ask loudly: Is all of this complexity really necessary in order to raise revenue, achieve fairness and implement social policy? Is it really necessary to the achievement of these purposes that the rules be abstruse, disjointed and continually changed? Is it really impossible to achieve the common objectives in this area by the use of rules that are lucid, succinct and of long duration?
In advising clients with regard to distributions and contributions it is important to bear in mind the various penalty taxes on distributions and contributions of the wrong amounts or at the wrong times. Such penalties include the following, at least:
1. The recipient is taxed if he gets a distribution too soon. IRC §72(t).
2. The payee is taxed if the distribution is made too late. IRC §§401(a)(9) and 4974.
3. The payee is taxed if the distribution is too small. IRC §§401(a)(9) and 4974.
4. The employer is taxed if it makes a contribution that is too small. IRC §4971.
5. The employer is taxed if it makes a contribution that is too large. IRC §4972.
6. The employer is taxed if the nonhighly compensated employees make a contribution that is too small. IRC §§402(g), 401(k)(8), 401(m) and 4979.
7. The employer is taxed if the highly compensated employees make a contribution that is too large. IRC §§402(g), 401(k)(8), 401(m) and 4979.
8. The employer may be subject to liability to the PBGC if there are insufficient assets at termination. ERISA Title IV.
9. The recipient is taxed if he makes a contribution to an IRA that is too large. IRC §4973.
Almost everyone is now subject to the 10% tax on premature distributions, not just key employees or 5% owners. The tax applies to distributions from qualified plans (§401(a)), IRAs, and §403(a) and (b) plans.[1] The tax does not apply to IRC §457 plans,[2] because it is outside of the definition of “qualified retirement plan.”[3]
For many years, only “owner-employees” were subject to the 10% premature distribution tax. [4]TEFRA, P.L. 97-248, dispensed with the term “owner-employee” and applied the tax to any person during such time as the person was a key employee in a top-heavy plan, effective for plan years beginning after 1983. The 1984 Tax Reform Act, repealed the reference to key employees, and imposed the tax for any distribution attributable to years after 1984 in which the person was a 5% owner. Various problems with the effective dates resulted in the passage of a technical correction to the 1984 TRA amending §72(m)(5)(A) to make the tax applicable for benefits accrued before 1985, in the case of 5% owners. But then the 1986 TRA simply made the tax apply to virtually everyone, for taxable years beginning after 1984, and eliminated 72(m)(5) for taxable years beginning after 1986.[5]
A premature distribution is generally a distribution from a "qualified retirement plan" that is includible in the gross income of a distributee and which is made prior to age 591/2. For this purpose a "qualified retirement plan" is a plan described in IRC §4974(c), i.e., a qualified plan, IRA, SEP, or a 403(a) or 403(b) annuity.
Even an involuntary distribution, such as a distribution to a trustee in bankruptcy, can invoke the tax.[6]
There are a number of important exceptions to the application of the §72(t) 10% penalty tax.
The penalty does not apply if the distribution is made after the employee has separated from service if the separation from service occurred during or after the calendar year in which the employee attained age 55.[7] The plan apparently need not contain any special language dealing with early retirement in order for the age 55 exception to apply.[8] Note that this exception is not available if the distribution is from an IRA (including a SEP-IRA).[9]
The penalty does not apply unless the distribution is includible in gross income. Therefore, the return of nondeductible employee contributions and amounts rolled over are exempt by definition from the tax.
§72(t)(2)(A)(ii) provides that the tax does not apply after death. This exception is important to note. A distribution on account of the death of the participant can be made without fear of incurring the premature distribution tax, no matter what the age of the beneficiary,[10] unless, perhaps, the distribution is to a spouse who has rolled over the IRA. See below.
Query, does the exception for distributions on account of death[11] continue to apply after a spousal rollover made pursuant to IRC §§402(c)(9), 403(b)(8)(B) or 408(d)(3)(C)? (As previously set forth, it is now the rule that a participant’s spouse is the only beneficiary eligible to rollover a death benefit distribution of the participant’s interest in a qualified plan or IRA.)
As previously indicated §72(t)(2)(A)(ii) provides that the tax does not apply after death, and thus, the tax clearly does not apply to an inherited IRA. IRC §408(d)(3)(C)(ii)(I) expressly describes an inherited IRA as one “acquired . . . by reason of the death of another individual.”[12] But what if a spouse is the beneficiary of an inherited IRA, and the spouse either makes a rollover or elects to treate the IRA as his or her own. The IRS definitely is of the opinion that the death exception does not survive a rollover.[13] However, I don’t think that the answer to this question is as clear as the IRS would have you believe.
It is very important to note that the tax does not apply if the distribution is part of a series of substantially equal periodic payments made for the life or life expectancy of the participant or the joint lives of the participant and his beneficiary.[14] However, in the case of a qualified plan, but not an IRA, this exception does not apply unless the participant has separated from service.[15]
In Notice 89-25, I.R.B. 1989-12, 68, Q&A 12, the IRS described three generally favorable methods of satisfying the substantially equal periodic payment requirement. First of all, any method acceptable for the purpose of calculating minimum distributions under §401(a)(9) may be used here. Second, the account may be amortized using an interest rate that does not exceed a reasonable interest rate on the date payments commence. (What about an unreasonably low rate?) In this case, life expectancy is to be determined in accordance with Prop. Treas. Reg. §1.401(a)(9)-1. Finally, annuity factors can be employed by using a reasonable mortality table with a reasonable rate of interest.[16]
The three methods described in Notice 89-25 are not the only methods available.[17]
IRC §408(d)(2) explicitly states that “for purposes of applying section 72 to any amount described in paragraph (1) [which provides that any amount distributed out of an IRA shall be included in gross income under §72] (A) all individual retirement plans shall be treated as 1 contract, [and] (B) all distributions shall be treated as 1 distribution.”
However, it appears that this rule is limited to income taxation only, because a series of private letter rulings[18] have held that §72(t)(2)(A)(iv) does not require that IRAs ‑‑or qualified plans[19], for that matter‑‑ be aggregated in order to calculate a series of substantially equal periodic payments. These letter rulings stand for the proposition that distributions from a rollover IRA (or plans) can be calculated regardless of whether similar distributions are made from other IRAs (or plans), and that the account balances of the other IRAs need not be considered when calculating the amounts of the annual distributions necessary under §72(t)(2)(A)(iv).[20]
It also appears that the substantially equal periodic payments can be computed with respect to more than one, but less than all, of the taxpayer’s IRAs, and that, further, they may be satisfied by taking distributions out of one or the others, or from all[21]; provided, however, that the distribution requirement will not be satisfied by a distribution from an IRA that was not used to compute the aggregate payment amount.
This means that an individual may contrive to split one IRA into two IRAs, and to take periodic distributions from one but not the other, computing the necessary periodic distributions solely by reference to the size of the IRA from which the distributions will be taken, ignoring the other.[22]
If the payment method is modified for reasons other than death or disability before the end of the 5 year period beginning with the date of the first payment, or age 591/2 if later, then the IRC §72(t) premature distribution tax is to be increased in the year of modification by an amount equal to the tax that would have been imposed in the absence of the exception, plus interest.[23]
Can a participant who has computed the minimums with reference to one particular IRA, ignoring all others (as is apparently permitted), take distributions in varying amounts and at various times from other IRAs? Can a participant set up multiple installment distribution arrangements from multiple IRAs? Can a participant set up multiple IRAs through transfers or rollovers, solely for the purpose of attaining these objectives? PLR 8946045 and PLR 9050030 offer some support to an affirmative answer to each of these questions, but the answers falls short of definitive.[24]
The Tax Court believes that “subsequently modified” applies to an ad hoc distribution that exceeds the amounts required under the installment method selected. [25] In Arnold v. Commissioner the Tax Court held that where additional amounts, in excess of what was required, were taken after age 59&1/2 but before the end of the 5-year period, all installments prior to 59&1/2 were disqualified.[26]
In December 1989 when petitioner was 55 years old,fn he began receiving annual distributions from his IRA. The distributions from petitioner's IRA were as follows:
December 1989 $44,000
January 1990 44,000
January 1991 44,000
January 1992 44,000
January 1993 44,000
November 1993 6,776
* * * *
The sole issue for decision is whether the November 1993 distribution from petitioner's IRA impermissibly modified a series of substantially equal periodic payments so as to trigger the imposition of the 10-percent recapture tax under section 72(t)(4).
* * * *
Section 72(t)(4) fn dictates, however, that if the series of substantially equal periodic payments (which otherwise is excepted from the 10-percent tax) is subsequently modified (other than by reason of death or disability) within a 5-year period beginning on the date of the first distribution, then the 10-percent tax under section 72(t)(1) will be imposed retroactively on prior distributions made before the taxpayer attains age 59-1/2, plus interest. This retroactive application of the 10-percent tax under section 72(t)(4) is known generally as a recapture tax. See infra.
Petitioners contend that the November 1993 distribution of $6,776 did not impermissibly modify a series of substantially equal periodic payments. Petitioners make two principal arguments in support of this claim.
First, petitioners contend that the November 1993 distribution occurred after the series of substantially equal periodic payments was completed in January 1993, and thus no modification occurred. Respondent asserts that petitioners' contention contradicts the plain language of section 72(t)(4) which requires no modifications within a 5-year period. Respondent notes that in this case the 5-year period beginning with the date of the first distribution ran from 1989 through 1994. Thus, respondent argues, the November 1993 distribution was premature and hence impermissibly modified the series of substantially equal periodic payments.
* * * *
It is evident that the 5-year period in section 72(t)(4) closes at the end of 5 years from the date of the first distribution; it does not end on the date of the fifth annual distribution pursuant to a series of substantially equal periodic annual payments.
In the case herein, petitioner received the fifth distribution from his IRA in January 1993, slightly more than 3 years from the date of the first distribution. Under section 72(t)(4), petitioner was required to wait until sometime in December 1994 before he could receive additional distributions that would avoid modifying the prior series of substantially equal periodic payments. He did not meet the required waiting period. Instead, petitioner received his distribution in November 1993 [after age 59&1/2], prior to the close of the 5-year period as provided in section 72(t)(4).[27]
Similarly, the IRS issued a PLR holding that a taxpayer could begin a new schedule, increasing the installment payouts to reflect certain investment gains, but the price was that the recapture tax applied to withdrawals made under the original method.[28]
(This may be a tax planning opportunity.) The premature distribution tax does not apply in the case of a distribution from a qualified plan to an alternate payee under a QDRO.
Note, however, that IRAs (including SEP-IRAs) are not subject to QDROs.[29] Pursuant to IRC §408(d)(6) an IRA may be transferred to a spouse without recognition of income tax, but the order effecting the transfer is not a QDRO, and therefore, if the proceeds are withdrawn before the spouse is 591/2, the early withdrawal tax will ordinarily apply.
One of the more esoteric exceptions applies (if no other exception applies) to the extent the amounts distributed during the taxable year of the participant do not exceed the medical expense deduction allowable under IRC §213[30] (whether or not the employee itemizes deductions that year). Note that this exception is not available if the distribution is from an IRA (including a SEP-IRA).[31]
The tax does not apply to benefits payable pursuant to a written
designation made before
“(4) Transition Rule. – The amendments made by this section shall not apply with respect to any benefits with respect to which, a designation is in effect under section 242(b)(2) of the Tax Equity and Fiscal Responsibility Act of 1982.”[34]
In the case of a qualified plan, the premature distribution rules do not apply if as of March 1, 1986 the employee separated from service with the employer and as of this same date the accrued benefit was in pay status pursuant to a written election providing a specific schedule for the distribution of the entire benefit, and the distribution is in fact made pursuant to the written direction.[35]
The tax due on premature distributions is reported by filing IRS Form 5329 (see Part I). This same form is used for reporting (a) the excess contributions tax for IRAs, due under IRC §4973, (b) the tax on early distributions, due under IRC §72(t) and (c) the tax on excess accumulations (for failing to make the IRC §401(a)(9) minimum distributions), due under IRC §4974.
The distribution exception for death under §72(t)(2)(A)(i) does not strictly apply to distributions made to a predeceased spouse’s successors in interest under the community property laws, because the death referred to under the exception from the tax is that of the employee. However, the tax does not apply unless the distribution is includible in gross income.[36] Further, the tax is on amounts received by the taxpayer. Therefore, if the tax applies at all, it is arguable that it is owed by the beneficiary of the nonparticipant spouse’s interest.
Repealed by the 1997 Taxpayer’s Relief Act, Pub. L. 105-34.
Under prior law, if all but a de minimus amount of a taxpayer’s IRA and retirement plan death benefits passed to the taxpayer’s spouse, the excess accumulation’s tax could be postponed by making an election under §4980A(d)(5) to have the excess accumulation and excess distribution tax apply as if the spouse were the participant. Oftentimes this election was not made because if one benefit in paying the accumulations tax on the death of the participant is that neither the excess accumulations or the excess distributions tax would apply in the future. However, since §4980A has been repealed, neither the excess accumulations nor the excess distributions tax will apply in the future in any event. If the three/two year period for filing a claim for a refund under IRC §6511(a) has not expired, the estate may still make the §4980A(d)(5) election and file a claim for refund.[37]
The Deficit Reduction Act of 1984 (DEFRA[38]) placed substantial restrictions on the ability of a Participant who was a 5% owner of business that sponsored a qualified plan to avoid the recognition of income by deferring indefinitely the receipt of distributions from qualified plans, IRAs, etc. This was accomplished in large measure by amendments to IRC §401(a)(9), which now specifies when distributions must begin and in what amounts. IRC §4974 imposes the penalty for failure to comply with §401(a)(9). TRA ’86 expanded these rules to include everyone else.
It may be helpful in approaching this subject to separate the post death from the pre-death rules. Another helpful distinction can be made on the basis of whether the death of the participant occurs before or after the “required beginning date” (the RBD). But first, let us examine the statute.
|
§401 (a) Requirements for qualification. A trust created or organized in
the * * * * (9) Required distributions.-- (A) IN GENERAL. -- A trust shall not constitute a qualified trust under this subsection unless the plan provides that the entire interest of each employee -- (i) will be distributed to such employee not later than the required beginning date, or (ii) will be distributed, beginning not later than the required beginning date, in accordance with regulations, over the life of such employee or over the lives of such employee and a designated beneficiary (or over a period not extending beyond the life expectancy of such employee or the life expectancy of such employee and a designated beneficiary). (B) REQUIRED DISTRIBUTION WHERE EMPLOYEE DIES BEFORE ENTIRE INTEREST IS DISTRIBUTED. -- (i) WHERE DISTRIBUTIONS HAVE BEGUN UNDER SUBPARAGRAPH (A)(ii). -- A trust shall not constitute a qualified trust under this section unless the plan provides that if -- (I) the distribution of the employee's interest has begun in accordance with subparagraph (A)(ii), and |
|
(II) the
employee dies before his entire interest has been distributed to him, the remaining portion of such interest
will be distributed at least as rapidly as under the method of distributions
being used under subparagraph (A)(ii) as of the date of his death. (ii) 5-YEAR RULE FOR OTHER CASES. --A trust shall not constitute a qualified trust under this section unless the plan provides that, if an employee dies before the distribution of the employee's interest has begun in accordance with subparagraph (A)(ii), the entire interest of the employee will be distributed within 5 years after the death of such employee. [But see the exception.] (iii) EXCEPTION TO 5-YEAR RULE FOR CERTAIN AMOUNTS PAYABLE OVER LIFE OF BENEFICIARY. -- If -- (I) any portion of the employee's interest is payable to (or for the benefit of) a designated beneficiary, (II) 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 (III) 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. (iv) 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) 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 701/2, and |
|
(II) 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. (C) REQUIRED BEGINNING DATE.[39] For purposes of this paragraph -- (i) In general. The term "required beginning date" means April 1 of the calendar year following the later of -- (I) the calendar year in which the employee attains age 70 1/2, or (II) the calendar year in which the employee retires. (ii) Exception. Subclause (II) of clause (i) shall not apply -- (I) except as provided in section 409(d), in the case of an employee who is a 5-percent owner (as defined in section 416) with respect to the plan year ending in the calendar year in which the employee attains age 70 1/2, or (II) for purposes of section 408(a)(6) or (b)(3). (iii) Actuarial adjustment. In the case of an employee to whom clause (i)(II) applies who retires in a calendar year after the calendar year in which the employee attains age 70 1/2, the employee's accrued benefit shall be actuarially increased to take into account the period after age 70 1/2 in which the employee was not receiving any benefits under the plan. (iv) Exception for governmental and church plans. Clauses (ii) and (iii) shall not apply in the case of a governmental plan or church plan. For purposes of this clause, the term "church plan" means a plan maintained by a church for church employees, and the term "church" means any church (as defined in section 3121(w)(3)(A)) or qualified church-controlled organization (as defined in section 3121(w)(3)(B)). (D) LIFE EXPECTANCY. --For purposes of this paragraph, the life expectancy of an employee and the employee's spouse (other than in the case of a life annuity) may be redetermined but not more frequently than annually. (E) DESIGNATED BENEFICIARY. -- For purposes of this paragraph, the term "designated beneficiary" means any individual designated as a beneficiary by the employee. |
|
(F) TREATMENT OF PAYMENTS TO CHILDREN. --Under regulations prescribed by the Secretary, for purposes of this paragraph, any amount paid to a child shall be treated as if it had been paid to the surviving spouse if such amount will become payable to the surviving spouse upon such child reaching majority (or other designated event permitted under regulations). (G) TREATMENT OF INCIDENTAL DEATH BENEFIT DISTRIBUTIONS. -- For purposes of this title, any distribution required under the incidental death benefit requirements of this subsection shall be treated as a distribution required under this paragraph. [Emphasis added.] |
For the last 15 years or so the minimum required distribution rules have been set forth in proposed regulations only. It is therefore worth considering what legal effect a proposed regulation has. It is clear that a taxpayer may rely upon a proposed regulation under certain circumstances, but the IRS cannot use a proposed regulation as a sword. A proposed regulation has no more weight in court than that of a brief filed by the Commissioner:
First, we note that
although final regulations command our respect (Commissioner v.
The 401(a)(9) rules cannot be comprehended without constant reference to the “required beginning date” (the RBD).
“For an employee who attains age 701/2 after December 31, 1987 (i.e., age 70 after June 30, 1987), the term ‘required beginning date’ means April 1 of the calendar year following the calendar year in which the employee attains age 701/2.”[41]
If an employee attained age 70 prior to
In the case of a governmental or church plan,[43] the required beginning date means the later of (1) April 1 of the calendar year following the calendar year in which the employee attains age 701/2 or (2) April 1 of the calendar year following the calendar year in which the employee retires.[44]
Under §1404 of the SBJPA, IRC §401(a)(9)(C) was amended, effective beginning in 1997, to provide that a person who is not a 5% owner (as defined in IRC §416) is not required to begin receiving distributions until the later of retirement or the person’s RBD. Note that this rule will not apply to a participant’s interest in an IRA.
A plan would not have to implement this new provision, but the Committee Reports indicate that the plan may be amended so that minimum distributions to less than 5% owners who are still employed are no longer required, even if such payments have already begun. The amendment requires that when distributions begin, “the employee’s accrued benefit shall be actuarially increased to take into account the period after age 701/2 in which the employee was not receiving any benefits under the plan.” The Conference Agreement to the Committee Reports confirm that the actuarial adjustment does not apply to defined contribution plans.[45]
TRA’86 §1121(a) imposes a penalty tax on each payee who fails to receive a minimum required distribution. The penalty tax is 50% of the amount by which the distribution made, if any, is less than that required under the MRD Rules.
§1121(b) defines the term minimum required distribution (MRD) as the amount required under 401(a)(9), 403(b)(10), 408(a)(6), 408(b)(3), or 457(d)(2), as determined under regulations.
§1121(d)(4)(A) provides that neither §1121(a) nor (b) will apply with respect to benefits with respect to which a designation is in effect under §242(b)(2) of TEFRA.
§1121(d)(4)(B) provides that §1121(b) does not apply to someone who was 70½ before January 1, 1988, unless the person “is a 5-percent owner (as defined in 416(i) of the Internal Revenue Cod e of 1986), at any time during – (I) the plan year ending with or within the calendar year in which such owner attains age 66½, and (II) any subsequent plan year.”[46]
The minimum distribution rules may not apply if the participant
made a TEFRA §242(b)(2) election.[47]
In order for this exception to apply, the benefits must have been made payable
pursuant to a written designation made before
Further, TEFRA §242(b)(2) protection will be lost if the form or timing of the designated benefit is changed.[50] One way this could happen would be if a spouse refused to consent to a waiver of the joint and survivor annuity rules. Consent to the waiver does not invalidate the election.[51]
If the §242(b)(2) shelter is lost, the IRC §401(a)(9) rules could apply with a vengeance, since the proposed regulations take the position that in the year of revocation, the trust is required to distribute all of the benefits that would have been required to have been distributed in previous years if §401(a)(9) had been applicable all along.[52] This could easily result in the imposition of an excess distribution tax.
TRA §1121(b), which defines MRDs, does not apply to individuals who attained age 701/2 prior to January 1, 1988 (i.e., persons who were born prior to July 1, 1917), unless such individual was a 5% owner at any time during the plan year ending with or within the calendar year in which such owner attained age 661/2 and any subsequent year.[53] (The old law required distributions to more than 5% owners. The exemption does not protect them.)
Interestingly there is no exclusion from the application of TRA §1121(a). What does this mean? TRA §1121(a) imposes the 50% excise tax, and TRA §1121(b) defines the term MRD. My conclusion is that the tax applies, but the minimums are computed under the pre-TRA ’86 Rules, whatever they were!
It is probable that this
rule only applies to qualified plan distributions.
The provisions generally apply to years beginning after
A reference in the regulations to the “first distribution calendar year” is a reference to the year in which the participant or beneficiary attains age 701/2. The first distribution calendar year is the year immediately preceding the RBD.
It is worth noting here that if one waits until the RBD to make the first minimum distribution, that the distribution will have to encompass two years minimums (the minimum for the first distribution calendar year, due by April 1, and the minimum for the year that includes the RBD, due by December 31,[55] except in the case of certain annuity distributions from defined benefit plans.[56]
If the participant turned 701/2 in 1991, a distribution would have to be
made no later than
The new minimum distribution rules apply to qualified plans, SEPs, IRAs, tax sheltered annuities (403(b) plans),[59] and even to unfunded deferred compensation plans under §457.[60]
In the case of 403(b) plans the minimum distribution rules only
apply to amounts accruing after
A more complete treatment of this issue is found elsewhere, below, in this outline.
These general MRD rules are not generally effective until plan years beginning in 1989; however, many of the required distribution rules were already applicable to 5% owners under TEFRA and DEFRA.
The effective date for distributions under the IRC pre-TRA §401(a)(9), as it affected 5% owners, had been postponed by the Service in various Notices for the last several years, because of the absence of regulations. Under the proposed regulations, affected participants may have been required to receive up to four calendar years' worth of distributions by December 31, 1987, to make up for the distributions not made in earlier years while the plan was awaiting guidance as to the amount required to be distributed.
On
There is a penalty tax for failure to distribute the §401(a)(9) minimum distributions of 50% of the amount required to be distributed. Previously, the only penalty was plan disqualification.[62] IRC §4974(a) provides:
If the amount distributed during the taxable year of the payee under any qualified retirement plan or any eligible deferred compensation plan (as defined in section 457(b)) is less than the minimum required distribution for such taxable year, there is hereby imposed a tax equal to 50 percent of the amount by which such minimum required distribution exceeds the actual amount distributed during the taxable year. The tax imposed by this section shall be paid by the payee.[63]
I have heard it sanctimoniously repeated more times than I can count that the duty to determine the MRD is the IRA owner’s and it is not the custodian’s or trustee’s duty to make this determination. To the contrary, in the case of a qualified plan everyone admits that it is the duty of the plan administrator or trustee. However, in both cases, the person responsible for paying the tax is the payee. “The tax imposed by this section shall be paid by the payee.”[64]
The tax can be waived by the Service if it finds reasonable cause for the shortfall.[65] This provided in the IRC itself:
(d) Waiver of tax in certain cases. If the taxpayer establishes to the satisfaction of the Secretary that --
(1) the shortfall described in subsection (a) in the amount distributed during any taxable year was due to reasonable error, and
(2) reasonable steps are being taken to remedy the shortfall,
the Secretary may waive the tax imposed by subsection (a) for the taxable year.[66]
The regulations expand but slightly on the IRC:
Q-8. Are there any circumstances when the excise tax under section 4974 for a taxable year may be waived? The tax under section 4974(a) may be waived if the payee described in section 4974(a) establishes to the satisfaction of the Commissioner the following:
(a) The shortfall described in section 4974(a) in the amount distributed in any taxable year was due to reasonable error, and
(b) Reasonable steps are being taken to remedy the shortfall.[67]
Before the tax will be waived, it must first be paid, using IRS Form 5329 (see Part III), which is ordinarily filed with the Form 1040.
The instructions to the Form 5329 provide:
Note: The IRS may waive this tax on excess accumulations if you can show that any shortfall in the amount of withdrawals from your qualified retirement plan was due to reasonable error, and that you are taking appropriate steps to remedy the shortfall. If you believe you qualify for this relief, file Form 5329, pay this excise tax, and attach your letter of explanation If the IRS grants your request, we will send you a refund.[68]
A plan will be treated as failing the qualification requirements of IRC §401(a) for any plan year in which the §401(a)(9) minimum distribution rules are not satisfied[69]; however, a plan will not lose its qualification because of violations only in isolated instances.[70] A pattern or regular practice of failing to meet the minimum distribution rules, with respect even to one employee, “will not be considered an isolated instance even if each instance is de minimis.”[71]
As explained above in some detail, the tax due on failing to make minimum distributions is reported by filing IRS Form 5329 (see Part III). This same form is used for reporting (a) the excess contributions tax for IRAs, due under IRC §4973, (b) the tax on early distributions, due under IRC §72 and (c) the tax on excess accumulations in qualified retirement plans (including IRAs) for failing to make minimum distributions, due under IRC §4974.
If the participant’s entire benefit in all qualified plans, IRAs, §403(b) plans, and §457 plans is distributed by the RBD, this of course satisfies the 401(a)(9).[72] However, all that is required is that certain minimum benefit payments begin being made by this time.[73]
Specifically, the IRC gives a plan two alternatives, by providing that “[a] trust shall not constitute a qualified trust under this subsection unless the plan provides that the entire interest of each employee (i) will be distributed to such employee not later than the required beginning date, or (ii) will be distributed, beginning not later than the required beginning date, in accordance with regulations, over the life of such employee or over the lives of such employee and a designated beneficiary (or over a period not extending beyond the life expectancy of such employee or the life expectancy of such employee and a designated beneficiary).”[74]
Under the first alternative of IRC §401(a)(9)(A)(i), the plan or IRA must payout the entire benefit by the RBD. If the IRA or plan provides only the first alternative, then query whether compliance with the second alternative (whether or inadvertent) will avoid application of excise tax.
Recall that §4974(a) imposes an excise tax equal to “50 percent of the amount by which such minimum required distribution exceeds the actual amount distributed during the taxable year.”[75] The term “minimum required distribution” is defined for this purpose as “the minimum amount required to be distributed during a taxable year under section 401(a)(9), 403(b)(10), 408(a)(6), 408(b)(3), or 457(d)(2), as the case may be, as determined under regulations prescribed by the Secretary.”[76]
The phrase “as determined under regulations prescribed by the Secretary” appears to be a reference to alternative two, installment payouts under the qualifying plan language requirement of IRC §401(a)(9)(A)(ii). The “as the case may be” language implies that if the plan has the alternative one language, but not alternative two, then alternative two is not the case and is not applicable.
As discussed at length elsewhere, neither a participant nor the participant’s beneficiary spouse may rollover an amount that was required to be distributed under the minimum distribution rules.[77]
The determination of the size of the minimum distribution is made at the time of the distribution (or, more precisely, as of the valuation date in the immediately preceding calendar year), without reference to the past. Therefore, as a general rule, there is no “credit” for distributions made prior to the required beginning date or otherwise.[78] There is an exception to this rule which applies if payments have begun prior to the RBD under an irrevocable annuity,[79] and there is a further exception for amounts distributed in the first distribution calendar year (i.e., the year immediately preceding the required beginning date).[80]
Note that the distribution can be over life or life expectancy. Also note that unless there is a “designated beneficiary,” the participant’s life expectancy will be serve as the sole measuring life.[81]
The life expectancy tables are found in Treas. Reg. §1.72-9 and are reprinted in part below. See the attachment to the Model letters at the end of this outline.
Although the minimum distribution rules of IRC §401(a)(9) apply after age 701/2, this does not prevent a rollover of the amounts not required to be distributed.[82]
The effect of rollovers and transfers after the RBD is discussed
in Prop. Treas. Reg. §1.401(a)(9)-1 G, and Prop. Treas. Reg. §1.408-8, A-6,
(Proposed
Can an IRA owner rollover or even transfer of all of one IRA to another, after the RBD, with the MRD satisfied out of the transferee IRA, as otherwise permitted by Notice 88-38. Yes, according to a letter from the IRS to Universal Pensions. No, according to the second sentence of G-3.
Prop. Treas. Reg. 1.401(a)(9)-1, Q&A G-3 (Proposed
G-3. Q. In the case of a transfer of an amount of an employee's benefit from one plan (transferor plan) to another plan (transferee plan), are there any special rules for satisfying the minimum distribution requirement or determining the employee's benefit under the transferor plan
A. (a) In the case of a transfer of an amount of an employee's benefit from one plan to another, the transfer is not treated as a distribution by the transferor plan for purposes of section 401(a)(9). Instead, the benefit of the employee under the transferor plan is decreased by the amount transferred. However, if any portion of an employee's benefit is transferred in a distribution calendar year with respect to that employee, in order to satisfy section 401(a)(9), the transferor plan must determine the amount of the minimum distribution with respect to that employee for the calendar year of the transfer using the employee's benefit under the transferor plan before the transfer. Additionally, if any portion of an employee's benefit is transferred in the employee's second distribution calendar year but on or before the employee's required beginning date, in order to satisfy section 401(a)(9), the transferor plan must determine the amount of the minimum distribution requirement for the employee's first distribution calendar year based on the employee's benefit under the transferor plan before the transfer. The transferor plan may satisfy the minimum distribution requirement for the calendar year of the transfer (and the prior year if applicable) by segregating the amount which must be distributed from the employee's benefit and not transferring that amount. Such amount may be retained by the transferor plan and distributed on or before the date required or paid to an escrow account which in turn distributes such amount on or before the date required.
(b) For purposes of determining any minimum distribution for the calendar year immediately following the calendar year in which the transfer occurs, in the case of a transfer after the last valuation date for the calendar year of the transfer under the transferor plan, the benefit of the employee as of such valuation date, adjusted in accordance with F-5, will be decreased by the amount transferred valued as of the date transferred.[84]
The second sentence of G-3(a) clearly implies that even in the case of an IRA to IRA transfer (as opposed to a rollover), the MRD may not be transferred. This could come as a surprise to more than a few. There is a letter (not a PLR) from John G. Riddle, Jr., Acting Chief, Employee Plans Rulings Branch, to Bob Skomars of Universal Pensions, dated 6/21/94 stating that Notice 88-38 supercedes G-3. On the other hand, I had a friend ask Marjorie Hoffman, Senior Technician Reviewer, Employee Benefits and Exempt Organizations Branch - Internal Revenue Service, and principal author of the proposed regulations under §401(a)(9) what she thought of this letter, and I was told that Marjorie did not agree with it.
Before delving into the bowels of IRC §401(a)(9)(B) it should first be observed that all references to distributions under §401(a)(9)(A)(ii) are references to distributions occurring after the RBD, which can only occur if the participant should live so long. This is “Code.” Once this cryptogram has been deciphered, it follows (with one exception) that §401(a)(9)(B)(i) only applies if the participant dies after the RBD, and §401(a)(9)(B)(ii)-(iv) only apply if the participant dies before the RBD.
If distributions have begun under §401(a)(9)(A) and the participant dies before his entire interest has been distributed (i.e., the participant died after the RBD), the remaining interest must be distributed at least as rapidly as the method of distribution being used under §401(a)(9)(B)(i).[85]
This means that the maximum size of the minimum distributions will be more or less permanently fixed and determined for all time (pre-death and post-death) at the RBD, by the identity of the designated beneficiaries (if any) and the form of payment elected on the required beginning date.
If the participant does not have a designated beneficiary (meaning that on death the benefits are to be paid to the participant’s estate, to charity, or to a nonqualifying trust), then the payout will never need to change, unless the participant’s life expectancy is recalculated each year. Of course, the participant or beneficiary is ordinarily free to take ad hoc accelerated distributions. §401(a)(9) is only concerned with minimums.
If the participant has a designated beneficiary, and that beneficiary is changed, the minimum distribution period will decrease if the new beneficiary is older, but the period can never be lengthened even if the new beneficiary has a longer life expectancy than the person who was the beneficiary on the RBD.
If only the participant’s unrecalculated life expectancy is being used to calculate the §401(a)(9) minimums, it would appear that payments should be able to continue to the participant’s beneficiary after the participant’s death, on the same basis as during life, and this ought to be true whether or not the beneficiary is a “designated beneficiary,” i.e., payments under the original scheduled timetable ought to be able to continue being made even if the beneficiary is a trust or an estate! However, if life expectancy is being recalculated, the payout will be accelerated for obvious reasons.
Participant Rollovers are permissible after benefits have commenced under IRC §401(a)(9), i.e., after the participant’s RBD.[86]
The effect of a Participant rollover or transfer after the RBD is discussed in Prop. Treas. Reg. §1.401(a)(9)-1 G, and Prop. Treas. Reg. §1.408-8, A-6, (Proposed 7/27/87). Basically, what is required is that the rolled over amounts be separately accounted for, so that the minimum distribution rules continue to be satisfied with respect to the rolled over amounts on the same basis as if the rollover had not occurred. The minimum distribution will be increased, however, if the beneficiary under the receiving plan is older than the beneficiary of the distributing plan.[87]
It is possible to avoid the general rule that, after death and after the RBD, distributions must continue at least as rapidly as before death. A common and well-established post mortem technique to achieve this result is to have the spouse accelerate the distribution and roll it over, or, in the case of an IRA, to treat the IRA as his or her own under the spousal IRA rules. This is an extremely important planning opportunity and is key to most of the various post death distribution strategies.
One solution that has been suggested is that a rollover will usually be at least as rapid as what was being taken during life; however, I think that partial rollovers are allowable, and thus, a partial rollover could be in an amount less than required during the life of the participant. In that case, distributions would have to be made from the amount not rolled over in an amount that would satisfy the at least as rapidly rule for what remained of the participant’s original IRA. I almost said that distributions would have to be made from the amount not rolled over in an amount that, when added to what was rolled over, would satisfy the at least as rapidly rule for what remained of the participant’s original IRA. But I think that would be an error, perhaps fatal, because amounts required to be distributed under the MRD rules cannot be rolled over!
I have elsewhere expressed my concerned that when a rollover takes place in the year of the participant's death, that the MRD must be deducted from it first. I am not positive that this is required. In the case of an inherited IRA, one may not even know for a while whether it is to be treated as having been rolled over. The proposed ROTH IRA regulations jump on this principle as the basis for requiring a participant (not the spouse, and that may be a distinction, I just don't know) who turns 70&1/2 to take the MRD for the year into income in that year, before being entitled to make a Roth IRA rollover; and this is true even though in the case considered the RBD is not until the following year! The reason: the MRD is required for the first distribution calendar year, though it can be postponed until April 1 of the following year. If a rollover is to be made in the first distribution calendar year (or in any other year for that matter prior to December 31), the MRD for that year (even though not yet due) will have to be deducted first.
A spouse who is a beneficiary of a participant’s interest in a qualified plan, tax sheltered annuity or IRA, may rollover the inherited interest, or, in the case of an IRA, treat it has his own.[88] (However, the rollover cannot be to a qualified plan or 403(a) annuity.[89]) This rule is still applicable even after the RBD has been reached, as discussed more thoroughly elsewhere.[90] This means, in effect, that the spouse could (1) defer distribution until the April 1 of the calendar year following the calendar year in which the spouse attained age 701/2, and (2) make a new election with respect to recalculation of life expectancy, using a new beneficiary as another joint measuring life.[91]
Of course, the particular IRA involved would have to allow this. In most cases the IRA owner would have the power to extend or to deny this right to the surviving spouse, but this would be a matter that should not affect the initial election or the payout calculations, and therefore, the participant ought to be able to address the acceleration issue at any time before death.
In PLR 9005071, the IRS ruled:
Section 402(c)(9) of the Code provides, in general, that if a distribution attributable to an employee is paid to the spouse of the employee after the employee’s death, section 402(a)(5) shall apply to the distribution in the same manner as if the spouse were the employee.
* * * *
In general, a surviving spouse who, pursuant to section 402(c)(9) of the Code, receives a lump sum distribution from a plan by reason of her spouse’s death and transfers such amounts into an IRA within 60 days is considered the IRA holder. Distributions required by sections 402(a)(5)(G) and 401(a)(9) are not eligible for such rollover treatment. Furthermore, section 401(a)(9)(A), [distributions following the RBD while the participant is alive] rather than section 401(a)(9)(B), [distributions after the participant’s death] governs distributions from the surviving spouse’s rollover, and such distributions must begin with reference to the surviving spouse’s required beginning date.[92]
It is not entirely clear to me whether the spouse must deduct from the amount rolled over in the year of death of the participant the amount that would have been due had the participant lived until December 31. On balance, I believe that a deduction is required.
Treas. Reg. §1.402(c)-2 Q&A 7(a) provides in part:
A-7. (a) General rule. Except as provided in paragraphs (b) and (c) of this Q&A, if a minimum distribution is required for a calendar year, the amounts distributed during that calendar year are treated as required minimum distributions under section 401(a)(9), to the extent that the total required minimum distribution under section 401(a)(9) for the calendar year has not been satisfied. Accordingly, these amounts are not eligible rollover distributions. . . . [93] [Emphasis added.]
It is worth noting in this context that the Roth regulations
prevent a Roth rollover in the year prior to the RBD unless the participant
deducts from the rollover, and pays tax on, the MRD for the first distribution
calendar year (which isn’t even due to be paid until the following year!):[94]
Since the RBD has been reached, distributions under the rule are to continue “at least as rapidly as under the method of distributions being used under subparagraph (A)(ii)[95] as of the date of his death.”[96] This would ordinarily require a distribution no later than December 31 of the calendar year of death. If the participant dies on December 30, it is unlikely that anyone will be in the proper state of mind, or even have the authority, to make a timely distribution the next day. However, this could be true of any one of a number of other tax deadlines as well, so there is no presumptive reason to doubt the conclusion suggested by the regulation, even if the result may be draconian in operation.
But for the spousal rollover, a distribution would have been due by December 31 of the year of death. Does a spousal rollover in the year of death but prior to December 31 of the year of death have to take into account the distribution that would have been due on December 31? In other words, must the amount that would have been distributable on December 31 be deducted from the amount rolled over?
Treas. Reg. §1.402(c)-2 Q&A 7(a), quoted above, suggests to me that to the extent an amount is required to be distributed by December 31 of the year of the participant’s death, after the participant’s required beginning date (RBD), as a result of the minimum distribution rules, this amount cannot be rolled over by the surviving spouse, even if the rollover takes place prior to December 31 of the year of death, and even though the minimum distribution would not otherwise have been required prior to December 31, had the participant lived so long.
Whether Treas. Reg. §1.402(c)-2 Q&A 7(a) conflicts with the first sentence of Treas. Reg. §1.408-8, Q&A A-4(b) is unclear to me:
“(b) In the case of an individual dying after December 31, 1983, the only beneficiary of the individual who may elect to treat the beneficiary's entire interest in the trust (or the remaining part of such interest if distribution thereof has commenced to the beneficiary) as the beneficiary's own account is the individual's surviving spouse. . . . ”[97] [Emphasis added.]
It is important to note that (1) if the participant’s life is being recalculated, and (2) if the rollover takes place in the year after death, and (3) if a minimum required distribution (MRD) is required to be made based on the status of the account prior to the rollover, then a very large minimum distribution may be required before the account can be rolled over.
If the MRD is not based on the status of the account prior to the rollover, then there may be no MRD at all that year, even if the spouse is over 70&1/2, since the spouse’s RBD would be December 31 of the year following the rollover, if PLRs 9311037 and 9534027 are taken at face value. Under this analysis, there is no MRD in the rollover year. I like this result, but I am skeptical about the analysis, because the result is clearly anomalous. Requiring a MRD prior to the rollover appears to me to fit the letter and spirit of the law, even if the result is not taxpayer friendly.
The year that a taxpayer reaches age 70&1/2 is the “first distribution calendar year.” However, distributions need not begin until April 1 of the following calendar year (the RBD). The adjusted gross income of a taxpayer who is going to make a Roth IRA rollover cannot have income that exceeds $100,000. However, the taxpayer need not include in AGI for this purpose income recognized as a result of a Roth IRA rollover. If the taxpayer is going to make a Roth IRA rollover in the year prior to the RBD, must the taxpayer include the MRD for the first distrbution calendar year in income
Q- 6. Can an individual who has attained at least age 701/2 by the end of a calendar year convert an amount distributed from a traditional IRA during that year to a Roth IRA before receiving his or her required minimum distribution with respect to the traditional IRA for the year of the conversion?
A- 6. (a) No. In order to be eligible for a conversion, an amount first must be eligible to be rolled over. Section 408(d)(3) prohibits the rollover of a required minimum distribution. If a minimum distribution is required for a year with respect to an IRA, the first dollars distributed during that year are treated as consisting of the required minimum distribution until an amount equal to the required minimum distribution for that year has been distributed.
(b) As provided in
A-1(c) of this section, any amount converted is treated as a distribution from
a traditional IRA and a rollover contribution to a Roth IRA and not as a
trustee-to-trustee transfer for purposes of section 408 and section 408A. Thus, in a year for which a minimum
distribution is required (including the calendar year in which the individual
attains age 701/2 ), an individual may not convert the assets of an IRA (or any
portion of those assets) to a Roth IRA to the extent that the required minimum
distribution for the traditional IRA for the year has not been distributed.
(c) If a required
minimum distribution is contributed to a Roth IRA, it is treated as having been
distributed, subject to the normal rules under section 408(d)(1) and (2), and
then contributed as a regular contribution to a Roth IRA. The amount of the
required minimum distribution is not a conversion contribution.[98]
The Preamble to the Roth IRA proposed regulations states:
A required minimum distribution may not be converted to a Roth IRA because section 408(d)(3)(E) prohibits the rollover of any such distribution. Under the proposed regulations, if a non-Roth IRA owner has reached age 701/2, any amount distributed (or treated as distributed because of a conversion) from the IRA for that year consists of the required minimum distribution to the extent that an amount equal to the required minimum distribution for that year has not yet been distributed (or treated as distributed). Thus, if a taxpayer who is required to receive a minimum distribution of $10,000 from his or her non-Roth IRA for a taxable year attempts to convert $11,000 to a Roth IRA prior to receiving the required minimum distribution, $10,000 of the conversion amount would be treated as the required minimum distribution and would be ineligible for conversion. This result is not affected by the means through which the taxpayer effects the conversion or by whether an amount greater than or equal to $10,000 remains in the taxpayer's non-Roth IRA after the conversion.[99]
Note that until 2005 the MRD is required to be included in determining modified AGI for purposes of determining whether a taxpayer qualifies for a Roth IRA rollover:
Q- 6. Is a required minimum distribution from an IRA for a year included in income for purposes of determining modified AGI?
A- 6. (a) Yes. For taxable years beginning before January 1, 2005, any required minimum distribution from an IRA under section 408(a)(6) and (b)(3) (which generally incorporate the provisions of section 401(a)(9)) is included in income for purposes of determining modified AGI.[100]
What happens if the surviving spouse is already past his required beginning date when the participant dies?
If the surviving spouse is the beneficiary of the deceased spouse’s interest, can the spouse still rollover? Yes.
Can the spouse designate new beneficiaries for the first time and compute the minimums based upon their ages in the calendar year in which the spouse attains age 70 1/2? Yes.
When is the first distribution due? December 31 of the calendar year following the rollover.
When must the beneficiary’s be designated for purposes of determining the minimum distributions required? December 31 of the calendar year following the rollover.
(a) PLR 9311037.
In PLR 9311037,[101] both the participant and the spouse had reached the RBD at the participants death in 1991. The spouse established an IRA in the year of the participant’s death, and proposed to rollover the decedent’s entire IRA account balance the same year. (The facts also recite that the minimum distribution for 1991 had been made prior to death.[102]) Whether the spouse’s IRA was established before or after death is unclear, but what was clear is that the spouse’s IRA was established after the spouse’s RBD. The Service specifically ruled that the minimum distribution on the amount rolled over in 1991, as a result of the decedent’s death in 1991, was not required to be made to the spouse before December 31, 1992.
“In this case, you reached age 70-1/2 in ****. Your required beginning date had already passed when you established your IRA in 1991. You are not required to take a distribution until 1992, pursuant to section 1.408-8, Q&A A-6, because the amount distributed from Individual A's IRA in 1991 and rolled over to your IRA must be considered to be a part of your account balance as of December 31, 1991, for purposes of determining the required minimum distribution for 1992.” [103]
The spouse asked for a ruling on the question of whether she could treat her beneficiaries as designated beneficiaries of her rollover IRA for purposes of computing her required minimum distributions. The Service ruled that she could:
“While the proposed regulations do not specifically answer your second ruling request, in the absence of final regulations, issues may be resolved by a reasonable interpretation of the proposed regulations and statutory provisions. Accordingly, it is a reasonable interpretation of the minimum distribution requirements, with respect to ruling request two, that ***** may be treated as designated beneficiaries for purposes of section 408(a)(6) since they were designated before your first required distribution date of December 31, 1992.”[104] [Emphasis added.]
So, the date the minimum distribution rules would first be determined and applied was December 31 of the year following the rollover, and it would therefore be the beneficiary designation in effect on that day that would determine and limit the minimum distribution schedule applicable thereafter.
The question remains, how exactly are the minimums to be computed? Should the new calculation begin afresh —which would certainly be the most straight forward approach— or should the calculation be adjusted to take into account distributions that were not made during the intervening years between the spouse’s RBD and the participant’s death? Should unrecalculated life expectancies be calculated as of the spouse’s RBD or as of the first required distribution date, or as of the year before?
(b) PLR 9534027.
In PLR 9534027 the Service ruled that a spousal rollover was available even though both the participant and the spouse had passed their respective RBDs.[105]
In this case the decedent died in 1994, naming his spouse as his beneficiary. The deceased IRA owner reached age 701/2 in 1989. The spouse reached age 701/2 in 1990. Minimum distributions began in 1990 and continued to be made based upon the unrecalculated life expectancy of the decedent and his spouse through 1995.[106] The spouse rolled over her inherited interests (less, presumably the 1995 minimum required distribution determined as if there were no rollover) into three new IRAs in 1995, and named her three children as beneficiaries. The spouse proposed to make the first minimum distribution out of the new IRAs by December 31, 1996.
“Section 1.408-8 of the proposed regulations, Q&A A-4, provides that a surviving spouse is the only individual who may elect to treat a beneficiary interest in an IRA as the beneficiary's own account. Q&A A-4 further provides, in pertinent part, that an election will be considered to have been made by a surviving spouse if either of the following occurs: (1) any required amounts in the account (including any amounts that have been rolled over or transferred, in accordance with the requirements of section 408(d)(3)(A)(i), into an IRA for the benefit of such surviving spouse) have not been distributed within the appropriate time period applicable to the decedent under section 401(a)(9)(B), or (2) any additional amounts are contributed to the account (or to the account or annuity to which the surviving spouse has rolled such amounts over, as described in (1) above) which are subject, or deemed to be subject, to the distribution requirements of section 401(a)(9)(A). The result of such an election is that the surviving spouse shall then be considered the individual for whose benefit the trust is maintained.
“The above cited sections of the proposed regulations do not preclude a surviving spouse's receiving a distribution from a Code section 401(a) qualified retirement plan and rolling over, or transferring, said distribution into more than one IRA and treating each of the several IRAs as her own IRA.
“In this case, Taxpayer B will receive a distribution of the full amount due her from Plan X. She will then roll over, or transfer by wire, pursuant to sections 402(c)(9) and 401(a)(31) of the Code, the distributed amount, less any portion of the distribution required under Code section 401(a)(9), into three IRAs, IRA F, IRA G, and IRA H, set up and maintained on her behalf. Such action on the part of Taxpayer B constitutes an election on her part to treat the IRAs as her own IRAs and is consistent with the regulation cited above.
* * * *
“Section 1.401(a)(9)-1, Q&A F-1(b) and (c), of the proposed regulations states, in part, that the distribution required to be made on or before the employee's required beginning date shall be treated as the distribution required for the employee's first distribution calendar year. A calendar year for which a minimum distribution is required is a distribution calendar year. The first calendar year for which a distribution is required is an employee's first distribution calendar year. The distribution required for distribution calendar years (other than a distribution required to be made on or before the employee's required beginning date) must be made on or before December 31 of that distribution calendar year.
“Section 1.408-8, Q&A A-6, of the proposed regulations provides, in part, that in the case of a rollover to an IRA of an amount distributed by another IRA, the amounts must be separately accounted for and the minimum distribution with respect to such amounts must be separately determined as described in section 1.401(a)(9)-1, Q&A G-2. In the case of an IRA, the value of the account balance is determined as of December 31 of the year prior to the year for which the minimum distribution is being determined. Section 1.401(a)(9)-1, Q&A G-2, provides, in part, that if an amount is distributed by one plan (distributing plan) and is rolled over to another plan (receiving plan), the benefit of the employee under the receiving plan is increased by the amount rolled over. Generally, the distribution has no impact on the minimum distribution required to be made by the receiving plan for the calendar year in which the rollover is received. But, under the general rule, if a minimum distribution is required to be made by the receiving plan for the following calendar year, the rollover amount must be considered to be part of the employee's benefit under the receiving plan.
“Section 1.401(a)(9)-1, Q&A G-2, of the proposed regulations, further provides, however, that for purposes of calculating the benefit under the receiving plan, if the amount rolled over is received by the receiving plan in a different calendar year from the calendar year in which it is distributed by the distributing plan, the amount rolled over is deemed to have been received by the receiving plan in the calendar year in which it was distributed by the distributing plan.
“In this case, Taxpayer B will reach age 70-1/2 in 1995. Thus, her required beginning date, as provided in Code section 401(a)(9)(C), will be April 1, 1996. However, pursuant to section 1.401(a)(9)-1, Q&As A-6 and G-2, of the proposed regulations, Taxpayer B is not required to take a required distribution for 1995 from any of the three IRAs she intends to set up in her name because the amounts distributed from Plan X and rolled over into her three IRAs must be considered a part of her IRA account balances as of December 31, 1995. Thus, calendar year 1996 is Taxpayer B's first distribution calendar year, and no distribution is required from any of the IRAs for any calendar year prior to calendar year 1996.
“We concluded with respect to your second letter ruling request, that Taxpayer B's actions referenced above constitute an election on her part to treat the three IRAs referenced herein as her own IRAs. As the owner of said IRAs, she may name one of her children as the designated beneficiary for each of her IRAs for purposes of sections 401(a)(9) and 408(a)(6). Although the proposed regulations do not specifically answer Taxpayer B's fourth ruling request by indicating by which date she must designate her beneficiaries, in the absence of final regulations, issues may be resolved by a reasonable interpretation of the proposed regulations and statutory provisions. Accordingly, it is a reasonable interpretation of the minimum distribution requirements, with respect to ruling request four, that Taxpayer B's children may be treated as designated beneficiaries for purposes of section 408(a)(6) as long as she designates them as such prior to Taxpayer B's first required distribution date of December 31, 1996.
* * * *
“Thus, with respect to your third and fourth ruling requests, we conclude as follows:
“3. That Code section 401(a)(9) minimum required distributions need not be made for calendar year 1995 from any of the three IRAs referenced herein which Taxpayer B intends to establish; and
“4. that Taxpayer B may choose a distinct Code section 401(a)(9) minimum required distribution period for each of the three IRAs, IRA F, IRA G, and IRA H. Furthermore, as long as Taxpayer B designates her children, Taxpayers C, D, and E, as the beneficiaries of her IRAs no later than December 31, 1996, distributions from each of said IRAs may be made over the joint life expectancy of Taxpayer B and either Taxpayer C, Taxpayer D, or Taxpayer E. However, such distributions must conform to the MDIB rules found in section 1.401(a)(9)-2 of the proposed regulations.”[107]
Note that the “new” RBD for the spouse was said to be December 31 of the calendar year following the calendar year in which the rollover election is made. Any elections made by the original participant become irrelevant following the rollover.
Again, the Service ruled that since the Proposed Regulations do not address the method for calculating the minimum distributions in this case, a reasonable approach would be to treat 12/31/96 as the RBD, following the 1995 rollover. The year following the rollover is treated as the first distribution calendar year. The year of the rollover is the first valuation calendar year, and the minimum distribution in 1996 is determined by the value of the account balance as of December 31, 1995.
Although the ruling does not tell us, we might assume that since the RBD is 12/31/96, the spouse’s age determined as of her birthday in 1996 should determine the minimum distribution. If this is true, then, even if the spouse is not recalculating life expectancy, the initial life expectancy factor (from which “1” would be subtracted each subsequent year) is determined on the spouse’s first distribution calendar year, i.e., the rollover year, rather in the year the spouse turned 701/2.
(b) PLR 9711032.
The participant (P) died after his RBD, with his spouse (Sp) as the sole primary beneficiary of his IRA. The spouse, who was over 70&1/2 at P’s death, immediately rolled over her husband’s IRA. The IRS approved the rollover, holding further that the spouse’s RBD was December 31 of the year following the rollover year.
Under the facts of the ruling, P died in July of 1995 without having yet taken his minimum required distribution (MRD) for that year. Sp apparently rolled over the entire IRA in September, and took her husband’s MRDs in September and November of the same year. The IRS ruled that SP’s first MRD would be due no later than December 31 of 1996.
The observant reader will note that the IRS analysis, while favorable, is contrary to what I have elsewhere suggested. As otherwise discussed in this article, I would have thought that the spouse could not rollover in 1995 the MRD for that year because amounts required to be distributed under the minimum required distribution rules are not supposed to be rolled over. Since the tax result is the same —the withdrawal is made and made subject to tax in the right year by the right person, albeit out of the wrong account according to my analysis—, there is no tax policy reason for being punctilious in an area already overly technical. All sides ought to applaud the IRS for being willing to cut the taxpayer a little slack where no harm is done.
(c) PLR 9713005.
In this ruling, P died in 1995, after the RBD. P’s beneficiary was S, P’s spouse. P died before taking the MRD for the year, and P’s MRD for 1995 was in fact not taken after P’s death. This triggered the deemed rollover rule as of 12/31/95. The MRD, however, was paid to S in June, 1996. The relevance of this payment is not clear. On January 20, 1996, S made a formal rollover. (Is that relevant?) Consistent with PLRs 9311037 and 9534027, the IRS ruled that the RBD was 12/31/96,[108] and that 1996 was the first distribution calendar year. Further, for purposes of the MRD rules, S could use the joint life expectancy of S and S’s designated beneficiary C, who we are told could be designated as late as 12/31/96.
Again, 1995 was the rollover year under the deemed rollover rule, but the Service graciously treated 1996 as the RBD, consistent with PLR 9311037. This would not itself do away with the MRD of the participant for the year of death, at least not technically. Noted without further comment was that the MRD for 1995 was made in June of 1996, after the rollover. The excise tax applicable to P for 1995 was not put in issue; nor was the effect of the payment of P’s 1995 MRD in 1996 on the 1996 MRD of S discussed: “No opinion is expressed with respect to the minimum required distributions from Individual A's IRA and any applicable excise taxes for the 1995 calendar year.”
I wonder about three things: (1) Why is the year following the rollover treated as the RBD? (2) How is it that the spouse gets to rollover the participant’s RBD for the year? (3) Finally, of what significance is the fact that the 1995 MRD was taken out of the rollover account in June of 1996?
Technically, it seems clear to me that a rollover of the MRD is impermissible; which means, among other things, that there is no deemed rollover. If there is a rollover, I would think that the RBD would be in the year of the rollover, but that no MRD would be required because the there was a zero balance in the rollover account as of the last day of the preceding year (the valuation year). If there is a deemed rollover, and the rollover year is the RBD –the worst of all possible combinations‑ then it will be very likely that the spouse will not have designated a beneficiary by the end of the deemed rollover year. Consider what happens if the IRA owner dies on New Year’s eve on the way to the bank to withdraw the MRD for the year. I am therefore grateful for this rare show of liberality on behalf of the IRS, which gives much needed relief here, at the cost of some uncertainty and illogicality, however.
Treas. Reg. §1.402(c)-2 Q&A 7(a) provides in part:
A-7. (a) General rule. Except as provided in paragraphs (b) and (c) of this Q&A, if a minimum distribution is required for a calendar year, the amounts distributed during that calendar year are treated as required minimum distributions under section 401(a)(9), to the extent that the total required minimum distribution under section 401(a)(9) for the calendar year has not been satisfied. Accordingly, these amounts are not eligible rollover distributions. . . . [109] [Emphasis added.]
It is worth noting in this context that the Roth regulations prevent a Roth rollover in the year prior to the RBD unless the participant deducts from the rollover, and pays tax on, the MRD for the first distribution calendar year (which isn’t even due to be paid until the following year!).[110]
Treas. Reg. §1.402(c)-2 Q&A 7(a), quoted above, suggests to me that to the extent an amount is required to be distributed by December 31 of the year of the participant’s death, after the participant’s required beginning date (RBD), as a result of the minimum distribution rules, this amount cannot be rolled over by the surviving spouse, even if the rollover takes place prior to December 31 of the year of death, and even though the minimum distribution would not otherwise have been required prior to December 31, had the participant lived so long.
Arguably, Treas. Reg. §1.402(c)-2 Q&A 7(a) conflicts with the first sentence of Treas. Reg. §1.408-8, Q&A A-4(b):
(b) In the case of an individual dying after December 31, 1983, the only beneficiary of the individual who may elect to treat the beneficiary's entire interest in the trust (or the remaining part of such interest if distribution thereof has commenced to the beneficiary)[111] as the beneficiary's own account is the individual's surviving spouse. . . . [112] [Emphasis added.]
It is important to note that (1) if the participant’s life is being recalculated, and (2) if the rollover takes place in the year after death, and (3) if a minimum required distribution (MRD) is required to be made based on the status of the account prior to the rollover, then a very large minimum distribution may be required before the account can be rolled over.
If the spouse can rollover the entire account, there may be no MRD at all that year, even if both the participant and the spouse are in their 80s. The reason, if true, is that the spouse’s RBD with respect to the rollover account would be December 31 of the year following the rollover, if PLRs 9311037, 9534027 and 9713005 are taken at face value. Under this analysis, there is no MRD in the rollover year. I like this result, but I am skeptical about the analysis, because the result is clearly anomalous. Requiring a MRD prior to the rollover appears to me to fit the letter and spirit of the law, even if the result is not taxpayer friendly. However, the IRS position reflected in letter rulings and informal comments appears to be that since the legislative history favors spousal rollovers, the spouse can rollover the participant’s last MRD.
G-2 and G-3 create thorny issues in this context. PLR 9713005 discussed G-2 as follows:
Section 1.408-8, Q & A A-6 of the proposed regulations provides, in part, that in the case of a rollover to an IRA of an amount distributed by another IRA, the rules in section 1.401(a)(9)-1 of the proposed regulations will apply for purposes of determining the account balance for the receiving IRA and the minimum distribution from the receiving IRA. Thus, certain amounts rolled over to an IRA must be separately accounted for and the minimum distribution with respect to such amounts must be separately determined as described in section 1.401(a)(9)-1, Q & A G-2 of the proposed regulations. In the case of an IRA, the value of the account balance is determined as of December 31 of the year prior to the year for which the minimum distribution is being determined. Section 1.401(a)(9)-1, Q & A G-2 provides, in part, that if an amount is distributed by one plan (distributing plan) and is rolled over to another plan (receiving plan), the benefit of the employee under the receiving plan is increased by the amount rolled over. However, the distribution has no impact on the minimum distribution required to be made by the receiving plan for the calendar year in which the rollover is received. But, if a minimum distribution is required to be made by the receiving plan for the following calendar year, the rollover amount must be considered to be part of the employee's benefit under the receiving plan.[113]
Not mentioned is F-5(d), which reads:
(d) If an amount is distributed by one plan and rolled over to another plan (receiving plan), G-2 provides additional rules for determining the benefit and minimum distribution under the receiving plan. If an amount is transferred from one plan (transferor plan) to another plan (transferee plan), G-3 and G-4 provide additional rules for determining the minimum distribution and the benefit under both the transferor and transferee plans.
The Service’s reliance on G-2 may be somewhat misplaced. G-2 reads:
G-2. Q. If an amount is distributed by one plan (distributing plan) and is rolled over to another plan (receiving plan), how are the benefit and the minimum distribution under the receiving plan affected
A. (a) Except as otherwise provided in paragraph (b), if an amount is distributed by one plan (distributing plan) and is rolled over to another plan (receiving plan), the benefit of the employee under the receiving plan is increased by the amount rolled over. However, the distribution has no impact on the minimum distribution required to be made by the receiving plan for the calendar year in which the rollover is received. But, if a minimum distribution is required to be made by the receiving plan for the following calendar year, the rollover amount must be considered to be part of the employee's benefit under the receiving plan. Consequently, for purposes of determining any minimum distribution for the calendar year immediately following the calendar year in which the amount rolled over is received by the receiving plan, in the case in which the amount rolled over is received after the last valuation date in the calendar year under the receiving plan, the benefit of the employee as of such valuation date, adjusted in accordance with F-5, will be increased by the rollover amount valued as of the date of receipt. For purposes of calculating the benefit under the receiving plan pursuant to the preceding sentence, if the amount rolled over is received by the receiving plan in a different calendar year from the calendar year in which it is distributed by the distributing plan, the amount rolled over is deemed to have been received by the receiving plan in the calendar year in which it was distributed by the distributing plan.
(b) If an amount is distributed by the distributing plan after the employee's required beginning date under both the distributing plan and the receiving plan, and the designated beneficiary of the employee under the receiving plan is a designated beneficiary with a life expectancy that is longer than the life expectancy of the designated beneficiary under the distributing plan, the following rule will apply. In such case, the receiving plan must separately account for the amount rolled over and treat it as a separate benefit. It must then begin distribution of such separate benefit in the calendar year following the calendar year in which the amount rolled over was distributed by the distributing plan. The separate benefit attributable to the rollover amount must be distributed over a period not exceeding the period (including any adjustments for recalculation under section 401(a)(9)(D), if applicable) used by the distributing plan to determine the employee's minimum distribution with respect to the benefit attributable to the amount rolled over. For purposes of determining the life expectancies or lives used to determine the minimum distribution under the receiving plan, the designated beneficiary under the distributing plan will be the designated beneficiary under the receiving plan (with respect to the benefit attributable to the amount rolled over). If such beneficiary is changed under the receiving plan to a different beneficiary from the designated beneficiary under the distributing plan, or a beneficiary is added who was not a beneficiary under the distributing plan, the rules in E-5 applicable to changes in beneficiaries will be used to determine the period over which distributions must be made by the receiving plan.
G-2 is fairly clear to me. The MRD is to be made out of the distributing plan, not the receiving plan. So if the IRA owner cannot rollover (or even transfer, see below) an MRD during life, how is it that the spouse gets to do so. I think the answer is simply one of practicality, and perhaps an IRS recognition of the Congressional purpose of favoring spousal rollovers.
* * * *
What if the IRA owner is still living? Can an IRA owner rollover or even transfer of all of one IRA to another, after the RBD, with the MRD satisfied out of the transferee IRA, as otherwise permitted by Notice 88-38. Yes, according to a letter from the IRS to Universal Pensions. No, according to the second sentence of G-3.
Prop. Treas. Reg. 1.401(a)(9)-1, Q&A G-3 (Proposed 7/27/87) provides:
G-3. Q. In the case of a transfer of an amount of an employee's benefit from one plan (transferor plan) to another plan (transferee plan), are there any special rules for satisfying the minimum distribution requirement or determining the employee's benefit under the transferor plan
A. (a) In the case of a transfer of an amount of an employee's benefit from one plan to another, the transfer is not treated as a distribution by the transferor plan for purposes of section 401(a)(9). Instead, the benefit of the employee under the transferor plan is decreased by the amount transferred. However, if any portion of an employee's benefit is transferred in a distribution calendar year with respect to that employee, in order to satisfy section 401(a)(9), the transferor plan must determine the amount of the minimum distribution with respect to that employee for the calendar year of the transfer using the employee's benefit under the transferor plan before the transfer. Additionally, if any portion of an employee's benefit is transferred in the employee's second distribution calendar year but on or before the employee's required beginning date, in order to satisfy section 401(a)(9), the transferor plan must determine the amount of the minimum distribution requirement for the employee's first distribution calendar year based on the employee's benefit under the transferor plan before the transfer. The transferor plan may satisfy the minimum distribution requirement for the calendar year of the transfer (and the prior year if applicable) by segregating the amount which must be distributed from the employee's benefit and not transferring that amount. Such amount may be retained by the transferor plan and distributed on or before the date required or paid to an escrow account which in turn distributes such amount on or before the date required.
(b) For purposes of determining any minimum distribution for the calendar year immediately following the calendar year in which the transfer occurs, in the case of a transfer after the last valuation date for the calendar year of the transfer under the transferor plan, the benefit of the employee as of such valuation date, adjusted in accordance with F-5, will be decreased by the amount transferred valued as of the date transferred.
The second sentence of G-3(a) clearly implies that even in the case of an IRA to IRA transfer (as opposed to a rollover), the MRD may not be transferred. This could come as a surprise to more than a few. There is a letter (not a PLR) from John G. Riddle, Jr., Acting Chief, Employee Plans Rulings Branch, to Bob Skomars of Universal Pensions, dated 6/21/94 stating that Notice 88-38 supercedes G-3. On the other hand, I had a friend ask Marjorie Hoffman, Senior Technician Reviewer, Employee Benefits and Exempt Organizations Branch - Internal Revenue Service, and principal author of the proposed regulations under §401(a)(9) what she thought of this letter, and I was told that Marjorie did not agree with it.
This is obviously another area fraught with great confusion.
The proposed regulations point out that if a participant dies prior to the RBD, it makes no difference that the death was during or after the first distribution calendar year, or that the participant had already taken a distribution on account of the first distribution calendar year.[114] This rule is important to note in those cases where the participant has attained age 701/2 during a calendar year, but dies before April 1 of the year following that calendar year.
How long does a spouse have to make the inherited IRA election in those cases where the spouse is a beneficiary and all minimum distributions, if any are required, have been made, and no additional amounts have been contributed by the spouse?
After the decedent’s RBD, a minimum distribution would be required each year, if the spouse is a beneficiary rather than an owner. If such amounts are not distributed, the proposed regulations provide that an election by the spouse to treat the IRA as his or her own will be considered to have been made.[115] Except for this rule, it may be that there is no limit, under these facts.
§401(a)(9)(B)(ii)-(iv) is awkwardly organized, and as a result, its terminology is misleading. The distribution rule that generally applies in the event a participant dies before the RBD (distribution over life expectancy) is described as the exception, while the exception (the 5-year rule) is proffered as if it were the rule. Unhappily we will adhere to the “Code” terminology since it is our authority.
If distributions have not begun under §401(a)(9)(A) at the time of the participant’s death (i.e., the participant died before the RBD), the participant’s entire interest must be distributed by December 31 of the calendar year that contains the fifth anniversary of the participant’s death,[116] unless, as presaged above, the “exception” applies.
Actually, the IRC says that the entire interest must be distributed within 5 years after death.[117] The proposed regulations gratuitously extend the period to December 31.[118]
This means, for example, that if an employee dies before the RBD, but after post separation from service installment distributions have commenced, post death distributions are to be made under the §401(a)(9)(B)(ii)-(iv) rules (the 5-year rule and the exceptions to the 5-year rule) rather than under §401(a)(9)(B)(i) (which requires that distributions post death be at least as rapid as pre-death).
The proposed regulations point out that this is true even if the employee dies in the year he attained age 701/2 (i.e., the year before the RBD, A/K/A the “first distribution calendar year”), and even though the plan has distributed the minimum distribution for the first distribution calendar year before death.[119]
If any portion of the Participant's interest is payable to (or for the benefit of) a “designated beneficiary” (as it usually will be), that portion of the benefit may be paid over the life or life expectancy of the beneficiary,[120] provided distributions begin no later than December 31 of the calendar year immediately following the calendar year of the Participant's death.[121]
Actually, the IRC says that distributions must begin no later than 1 year after death “or such later date as the Secretary by regulations may prescribe.”[122] It is the regulations that gratuitously extend the date to December 31.[123]
The determination of life expectancy is made by reference to the latest date on which distributions must commence, i.e., the beneficiary’s life expectancy is determined by reference to the beneficiary’s birthday occurring in the calendar year following the year of the participant’s death.[124]
If the beneficiary is the participant’s spouse, life expectancy can be redetermined (recalculated) annually.[125]
Under the IRC, if death is prior to the RBD and the designated beneficiary of a participant’s interest in an IRA or qualified plan is the participant's spouse, then distributions need not begin any earlier than the end of the calendar year in which the participant would have attained 70&½[126] (or the end of the calendar year following the year of death, if later[127]).
(b) Spousal beneficiary. In order to satisfy the rule in section 401(a)(9)(B) (iii) [i.e., the exception to the 5-year rule, where death is prior to the RBD] and (iv) [i.e., the beneficiary is the spouse and death is prior to the RBD], if the designated beneficiary is the employee's surviving spouse, distributions must commence on or before the later of (1) December 31 of the calendar year immediately following the calendar year in which the employee died and (2) December 31 of the calendar year in which the employee would have attained age 70&½ .[128]
This proposed regulation requires careful reading, lest one interpret it as applying after the RBD. If it applied after the RBD, then §1.401(a)(9)-1, Q&A C-3(b)(1) would apply, because “December 31 of the calendar year immediately following the calendar year in which the employee died” would be later than “December 31 of the calendar year in which the employee would have attained age 70&½” and the MRD for the year of death (after the RBD) would be postponed until the following year, if the spouse were the beneficiary. Such a rule would be helpful where the participant, after the RBD, died late in the year, before taking the MRD for the year, and would also be relevant in the determination of whether the spouse has been deemed to have elected to have treated the participant’s IRA as the spouse’s own, under the deemed rollover rules. However, as indicated, a careful reading of the regulation indicates that it applies only where death is prior to the RBD. The reason for “the later of” language is that a participant could die in the year prior to the RBD (i.e., in the year the participant reached age 70&½), in which case “(1) December 31 of the calendar year immediately following the calendar year in which the employee died” would be later than “(2) December 31 of the calendar year in which the employee would have attained age 70&½.”
If the spouse dies before such distributions begin, the minimum distribution rules are applied as if the surviving spouse were the participant.[129] This rule cannot be reapplied to a new spouse of the surviving spouse.[130]
Note the rule under discussion could be beneficial where the surviving spouse is older than the predeceased spouse.
Once benefits commence to the spouse, are they distributed in accordance with the spouse’s life expectancy as first determined at that time, or as determined and adjusted with reference to the participant’s date of death? Presumably, the former.
As discussed in detail above, under IRC §§ 402(c)(9), 403(b)(8)(B) and 408(d)(3)(C), the spousal rollover/inherited IRA rules, a spouse who is a beneficiary of a participant’s interest in a qualified plan, tax sheltered annuity or IRA, may treat the interest as his or her own (in the case of an IRA), or roll it over, in all cases. (A rollover under these rule must be made to an IRA, i.e., it cannot be made to a qualified plan.[131])
After the rollover, §401(a)(9) will apply solely with reference to the surviving spouse and the surviving spouse’s beneficiary.[132] Among other things, this should permit the spouse to defer distribution until the April 1 of the calendar year following the calendar year in which the spouse attains age 701/2, if the spouse has not reached his or her RBD.[133] Further, 401(a)(9) will be applied to the to the surviving spouse and the surviving spouse’s beneficiary even if the spouse has already reached the RBD at the time of the rollover.[134]
In the case of an IRA (but not a qualified plan), an election will be considered to have been made to treat the remaining interest as the surviving spouse’s own if (1) the minimum distributions that would otherwise have been required (in the absence of an election) are not made or (2) if the spouse makes a contribution to the account that is subject to the minimum distribution rules.[135]
How long does a spouse have to make the inherited IRA election in those cases where no required amounts applicable if the spouse were a beneficiary have not been distributed, and no additional amounts have been contributed by the spouse? Prior to the RBD, a minimum distribution will never be required before at least five years have expired,[136] and, if the designated beneficiary is the Participant's spouse, distributions need not begin any earlier than the end of the calendar year in which the participant would have attained 701/2 (or the end of the calendar year following the year of death, if later).[137] If such amounts are not distributed, the proposed regulations provide that an election by the spouse to treat the IRA as his or her own will be deemed to have been made.[138] Except for this rule, it may be that there is no limit. There are two private letter rulings that hold that if a spouse who is under 591/2 takes a distribution from an IRA without paying the §72(t) 10% early withdrawal penalty tax (which would not be applicable if the spouse were a beneficiary), the spouse is deemed to have elected beneficiary status.[139]
The proposed regulations provide that if distributions irrevocably (except for acceleration) commence to an employee before the RBD over a period permitted for distributions permitted after the RBD and the distribution is in the form an annuity sanctioned in §F-3 and F-4 of Prop. Treas. Reg. §1.401(a)(9)-1 (Proposed 7/27/87), then distributions will be considered to have begun on the actual commencement date, even if the employee dies before the RBD.[140]
If the plan does not specify the method of distribution following the death of the participant prior to the RBD, then distributions must be made under the 5-year rule, by default, unless the beneficiary is the participant’s spouse.[141] Contrariwise, if the beneficiary is the participant’s spouse, then distributions must be made under the installment method.[142]
If the participant dies prior to the RBD, the plan may allow the participant or beneficiary to elect which rule to use. If the plan allows for it and if an election is to be made, it must be made, “in operation,” no later than the earlier of (1) December 31 of the calendar year in which distribution would be required to commence in order to satisfy the requirements for the exceptions to the 5-year rule, or (2) December 31 of the calendar year which contains the fifth anniversary of the date of death of the participant. As of this date the election must be irrevocable as to the beneficiary (and all subsequent beneficiaries) and must apply in all subsequent years.[143]
When would (2) above ever be earlier than (1)? The only case I can think of is under the exception provided by IRC §401(a)(9)(B)(iv) that allows a surviving spouse to defer distributions until the deceased spouse would have reached 70½.
The election must be made “in operation.” Does this mean that the beneficiary need not actually sign anything? The election must be irrevocable. Is this an operational matter as well? It appears that the answer to both questions is a qualified “yes.”
(Note that the following appears to be the result of the application of this rule: Unless the beneficiary is a surviving spouse, the election would have to be made by December 31 of the calendar year following the date of death. A spouse, however, would have until December 31 of the calendar year in which the employee would have attained age 701/2 in order to satisfy the requirements for the exceptions to the 5-year rule, or, if sooner, December 31 of the calendar year which contains the fifth anniversary of the date of death of the participant. Surely, it is not the intent of the rule to require the spouse to take a distribution earlier than otherwise required.)
(A) May Some Beneficiaries Choose the 5 Year Rule While Others
Chose the Life Expectancy Method?
If the separate account rule applies, some beneficiaries may elect the 5-year rule and others the life expectancy rule.[144]
(B) Who Decides Whether To Use The Life Expectancy Method Or
The 5-Year Rule Where the Separate Account Rule Does Not Apply?
What if the separate account rule does not apply? If the separate account rule does not apply, the life expectancy of the oldest beneficiary must be used if the life expectancy method is adopted; however, who decides whether to use the life expectancy method or the 5-year rule? The proposed regulations do not tell us.
(c) Plan Default Provision.
Again, the plan may specify whether and under what conditions either the installment method or the 5-year method will apply.[145] The plan may also provide a default provision that is only applicable if the participant or beneficiary fail to elect.[146] Therefore, it would appear that if the plan gives the beneficiary an option, but the default provision is suitable, the beneficiary may do nothing other than see to it that withdrawals meet the minimums required by the default method.
If the plan gives the participant or beneficiary an option, but does not have a default provision that operates if the participant or beneficiary fails to exercise it, then the general default provisions applicable where there is no plan provision at all operate.[147]
In addition to the other requirements of IRC §401(a)(9), the minimum distributions must also satisfy the minimum distribution incidental death benefit rule.[148] This rule only applies while the participant is still alive and where the designated beneficiary is someone other than the participant’s spouse.[149] In other words, the MDIB rule disappears, or goes away, at the death of the participant.
The incidental death benefit rule has long been a part of the qualified plan rules, even before being enshrined in §401(a)(9)(G).[150] The rule is derived from the general principle that a qualified plan is for retirement and is not to be established primarily to provide a benefit to others besides the employee.[151]
If the designated beneficiary is someone other than a spouse, the "incidental death benefit rule" forces the participant to use a joint life expectancy table that treats the beneficiary as if he or she were no more than approximately 10 years younger than the participant.[152] This means that if the beneficiary is not the participant’s spouse, the maximum joint life expectancy will be 25.3 for the first distribution calendar year if the participant reaches age 71 in that year, or 26.2 if the participant is turns 70 in that year.[153]
However, the MDIB rule no longer applies after the participant’s death, and a longer post death payout may therefore become available at that time![154]
For example, if the beneficiary were 50 years younger than the participant, the incidental death benefit rule would require joint life expectancy to be calculated as if the beneficiary were only 10 (rather than 50) years younger than the participant. In the year of the participant’s death, the divisor used to determine the minimum distribution fraction would be computed as if there were only a 10 year difference in age. However, in the year following the participant’s death, the beneficiary’s actual life expectancy (and the participant’s remaining life expectancy, assuming death is after the RBD and there is no recalculation) is used, without regard to the MDIB. This means that the divisor will be a much larger number than previously used.
In drafting language to be used in an IRA, a qualified plan, or in the beneficiary designation, to state the minimum required distribution period as being the lesser of the joint life factor or the MDIB maximum (if the latter is applicable). I have seen language similar to the statute, which says that at the death of the participant, distributions shall continue to be made at least as rapidly as during life. This would be a mistake, if taken literally, since during life, the distributions were subject to the MDIB rule, so in actuality, distributions can be made less rapidly after death —not that I am a nitpicker.
Consider the effect of recalculation here. There is none; or rather, there is very little. If the MDIB rule applies at all, the beneficiary is more than 10 years younger than the participant. During the participant’s life, it will matter not at all whether life expectancy is being recalculated, because the MBIB rule will override the normal calculation in any event (presumably). At death, it may make some difference whether or not recalculation was chosen, since if it had been, then only the beneficiary’s life expectancy will be relevant. If recalculation had not been chosen, then you will find that the single life expectancy of the beneficiary will be very close to the joint life expectancy, particularly if the difference is large, the larger the difference in ages, the less the difference between single and joint life expectancy. To illustrate the point, consider that the joint life expectancy of a 100 year old and a 1 year old is just about the same as the life expectancy of the child, as you would expect. Having said all that, it still would be foolish to have elected recalculation, since there will always be some benefit, however small, to being able to use joint life expectancy after the death of the participant, and I can think of no case where recalculation would help anyone subject to the MDIB rule during life!
Typically the beneficiary(ies) may consist of one or more individuals, or trusts, or even the participant’s probate estate. During the participant’s lifetime, only the participant can be the recipient of benefits.[155] The named beneficiary therefore, generally refers to the recipient of what remains undistributed of the participant’s benefit, following the participant’s death.
The beneficiary will generally be whomever the participant designated (on a beneficiary designation form) to receive his benefit, assuming the plan has a death benefit, and assuming it allows the participant to choose a death beneficiary (and most plans do, subject to the survivor death benefit rules of REA discussed above).[156]
What if the participant did not complete a beneficiary designation form? In virtually every case, the plan will designate a beneficiary if the participant did not affirmatively chose a beneficiary, and this will suffice under §401(a)(9).[157] There are many possibilities. Probably the plan will designate the surviving spouse as the beneficiary. Remember, the survivor benefits guaranteed by REA will come into play here. If there is no surviving spouse, the plan may designate the participant’s children or the participant’s estate as the beneficiary.
“D-2. Q. May an individual who is not designated as a beneficiary under the plan be considered a designated beneficiary for purposes of determining the minimum distribution required under section 401(a)(9)
“A. (a)(1) Except to the extent provided in E-5 with respect to former beneficiaries, designated beneficiaries are only individuals who are designated as beneficiaries under the plan. An individual may be designated as a beneficiary under the plan either by the terms of the plan or, if the plan provides, by an affirmative election by the employee (or the employee's surviving spouse) specifying the beneficiary. A beneficiary designated as such under the plan is an individual who is entitled to a portion of an employee's benefit, contingent on the employee's death or another specified event. For example, if a distribution is in the form of a joint and survivor annuity over the life of the employee and another individual, the plan does not satisfy section 401(a)(9) unless such other individual is a designated beneficiary under the plan. A designated beneficiary need not be specified by name in the plan or by the employee to the plan in order to be a designated beneficiary so long as the individual who is to be the beneficiary is identifiable under the plan as of the employee's required beginning date, or as of the date of the employee's death (in the case of distributions governed by section 401(a)(9)(B) (iii) and (iv)), and at all subsequent times. The members of a class of beneficiaries capable of expansion or contraction will be treated as being identifiable if it is possible at the applicable time to identify the class member with the shortest life expectancy. The fact that an employee's interest under the plan passes to a certain individual under applicable state law does not make such individual a designated beneficiary unless such individual is designated as a beneficiary under the plan.
“(2) This paragraph (a) is illustrated by the following example.
“Example. Employee X attains age 701/2 in calendar year 1990. As of April 1, 1991, X designates as his beneficiaries under the plan his spouse and his children. X does not specify them by name. Even though X did not specify his spouse and his children by name, they are identifiable based on their relationship to X as of his required beginning date. Further, it is irrelevant that additional children of X may be born after his required beginning date and thus that the class of beneficiaries is capable of expansion.
“(b) See E-5 for the rules which apply if there is a change in beneficiaries under the plan with respect to an employee.”[158]
Ordinarily the beneficiary may be changed either before or after the RBD. However, if the change is after the RBD, it may result in a shorter payout period, but it cannot result in a longer payout period.[159]
If the beneficiary whose life expectancy is being used to calculate the minimum distribution dies after the participant’s RBD, the fact that a new designated beneficiary is later appointed will not affect the payout period. This is true, even if the new designated beneficiary is older than the one who died. This means, for example, that the participant could name the estate as beneficiary, and still use the joint life expectancy method, because that method would have become fixed upon the death, after the RBD, of the prior beneficiary.
“If the designated beneficiary whose life expectancy is being used to calculate the distribution period dies on or after the applicable date, such beneficiary's remaining life expectancy will be used to determine the distribution period whether or not a [new] beneficiary with a shorter life expectancy receives the benefits.”[160] [Emphasis added.]
This rule obviously does not apply if the beneficiary is a spouse whose life expectancy was being recalculated.[161]
Does it make any difference whether the new beneficiary is appointed before or after the death of the original beneficiary? The proposed regulation just quoted does not appear to draw such a distinction.
Most plans are woefully inadequate in designating the takers in default if there is no surviving spouse. A distribution in equal shares to children is sometimes provided. What about predeceased children and orphaned grandchildren? Does the state’s anti-lapse statute apply? The resolution of these issues is beyond the scope of this outline, but it is worthwhile noting here that the default beneficiary designation scheme under a plan will, more likely than not, be ambiguous, inadequate and poorly drafted. The same may be true of a beneficiary designation that was not thoughtfully completed.
The term “designated” beneficiary is a term of art and has a special meaning under the minimum distribution rules. The IRC says that a "designated beneficiary" is an individual designated as a beneficiary by the employee.[162] However, the regulations go beyond simple definitions. Perhaps the term “designated beneficiary” is best defined in terms of what is not a designated beneficiary. Virtually all beneficiaries are designated beneficiaries. An exception would be the participant’s probate estate and certain trusts.[163] Trusts are a special case. A corporation or a charitable organization would not be a “designated beneficiary.”[164]
If the participant is living on the RBD, the beneficiary must have been determined by that time,[165] subject to the annuity distribution exception discussed above.[166]
If there is no designated beneficiary, and the participant dies before the RBD, then, as indicated above, the entire benefit must be distributed by December 31 of the fifth calendar year which contains the anniversary of the participant’s death.[167]
IRC §401(a)(9)(A)(ii) requires that distributions begin not later than the RBD “over the life of such employee or over the lives of such employee and a designated beneficiary . . . ”
This may appear to leave open the question of whether an election must be made which way to go. If the employee has a designated beneficiary, then joint life expectancy will always be longer than single life expectancy, and so one wonders why the option appears to be there. The answer, I think, is that the single life expectancy method applies if there is no designated beneficiary, and the joint life expectancy method applies if there is one, without need or even option to elect the appropriate treatment. Although it is not entirely clear under the proposed regulations, it is my opinion that if a participant having a beneficiary chooses to take distributions over single life expectancy, it merely means that the participant is taking distributions faster than is necessary, and is not bound to continue doing so.[168]
In the case of the decision as to whether or not to recalculate life expectancies, and whether to take distributions upon death prior to the RBD under the 5-year rule or the life expectancy method, the proposed regulations contain specific default rules that operate in the absence of an election and in the absence of a plan term solving the problem. But if the participant survives to the RBD, every indication is that the joint life expectancy method applies automatically, if the beneficiary of the employee’s death benefit is a human being.[169] In other words, the minimum required distribution determined at the RBD is not an “election” that must be made; rather, it is determined solely by reference to (i.e., it is a function of) the nature and age of the death beneficiary. The MRD and is not determined by the size of the actual distributions taken in fact by the participant. After all, the participant is perfectly entitled to take more than the MRD. This conclusion and analysis is confirmed by PLR 199951053.
It should be noted at the outset that a trust, no matter what its terms, can be made a beneficiary of an IRA or qualified plan. The only issue is the timing of the distributions. If a beneficiary of the trust is to be treated as a “designated beneficiary” the trust must meet certain requirements. If, however, trust beneficiary is not treated as a designated beneficiary, then the distribution timing rules operate differently, depending upon whether death occurs before or after the required beginning date (RBD).
If death occurs before the RBD, then, in the case of a trust the beneficiaries of which are not treated as designated beneficiaries, distributions will have to be made under the 5-year rule. However, if death occurs after the RBD, then, if the trust does not qualify, distributions will ordinarily be calculated with reference to the life expectancy of the participant alone; and if that life expectancy is not being recalculated, it would appear that the distributions should be able to continue to the trust (or to the estate for that matter) at a rate no less rapid than during the participant’s lifetime.
The question here is whether we need regulations to treat the beneficiaries of a trust as designated beneficiaries. Does a reasonable interpretation of the statute allow us to treat the beneficiaries of a trust as designated beneficiaries?
Under 401(a)(9)(B)(iii)(I), the exception to the five year rule is explicitly made applicable if “any portion of the employee's interest is payable . . . for the benefit of . . . a designated beneficiary.[170] So, it seems clear to me that at least in the case of distributions on account of death prior to the RBD, a distribution to a trust for the benefit of a human being ought to qualify, even in the absence of regulations. This “for the benefit of” language is not found in the IRC in the case of a distribution after the RBD; so perhaps in that case regulations might (arguably) be necessary.
In the case of distributions after the RBD, IRC §401(a)(9)(A)(ii) provides that a trust is not qualified unless the plan provides that the entire interest of the employee “will be distributed, beginning not later than the required beginning date, in accordance with regulations, over the life of such employee or over the lives of such employee and a designated beneficiary.”[171] (Emphasis added.)
As pointed out previously, it is clear that a taxpayer may rely upon a proposed regulation under certain circumstances, but the IRS cannot use a proposed regulation as a sword. A proposed regulation has no more weight in court than that of a brief filed by the Commissioner:
First, we note that
although final regulations command our respect (Commissioner v.
According to the Blue Book, the plan will satisfy the requirement that the distribution be made on behalf of the designated beneficiary, if payments are made to a trust which itself meets the requirements of 401(a)(9).[173] There remain many unanswered questions respecting the application of the rules to trusts.
Besides the statement in the Blue Book, all that we really have to go on at the present time are three Q&As found in the proposed regulations. These regulations, which were originally proposed in 1986, were re-proposed in amended form on December 30, 1997. Since this is such an important area, these three provisions are set forth in full immediately below. In fact, again because the issue is of critical importance to the all too common situation where a trust is a beneficiary, I am even reproducing the preamble to the re-proposed regulations.
[REG-209463-82]Required Distributions from Qualified Plans and Individual Retirement Plans.
Background. On July 27, 1987, Proposed Regulations (EE-113-82) under sections 401(a)(9), 403(b), 408, and 4974 of the Internal Revenue Code of 1986 were published in the Federal Register (52 FR 28070). Those proposed regulations provide guidance for complying with the rules relating to required distributions from qualified plans, individual retirement plans, and section 403(b) annuity contracts, custodial accounts, and retirement income accounts. This document contains amendments to proposed §1.401(a)(9)-1 (hereinafter referred to as the Existing Proposed Regulations) that was included in EE-113-82. Specifically this document contains amendments to Q&As D-5 and Q&A D-6 of the Existing Proposed Regulations which prescribe specific requirements that must be met when a trust is named as a beneficiary of an employee's benefit under a plan, and adds a new Q&A D-7 to the Existing Proposed Regulations. Proposed §§1.408-8 and 1.403(b)-2 (also included in EE-113-82) provide that the provisions of proposed §1.401(a)(9)- 1 generally apply to individual retirement plans, and section 403(b) annuity contracts, custodial accounts, and retirement income accounts. Accordingly, these amendments and additions also generally apply to such plans, contracts, and accounts.
The amendments and additions to the Existing Proposed Regulations in these proposed regulations are issued in response to comments and questions received regarding the Existing Proposed Regulations with respect to section 401(a)(9). Treasury and the IRS continue to welcome additional comments concerning the Existing Proposed Regulations and the other sections of EE-113-82.
As in the case of the Existing Proposed Regulations and the other sections of EE-113-82, taxpayers may rely on these proposed regulations for guidance pending the issuance of final regulations. If, and to the extent, future guidance is more restrictive than the guidance in these proposed regulations, the future guidance will be applied without retroactive effect.
Explanation of Provisions. Overview. Section 401(a)(9)(A) provides that, in order for a plan to be qualified under section 401(a), distributions of each employee's interest in the plan must commence no later than the “required beginning date” for the employee and must be distributed over a period not to exceed the joint lives or joint life expectancy of the employee and the employee's designated beneficiary. Section 401(a)(9)(B) provides that if distribution does not commence prior to death in accordance with section 401(a)(9)(A), distributions of the employee's interest must be made within 5 years of the employee's death or, generally, commence within one year of the employee's death and be made over the life or life expectancy of the designated beneficiary.
Section 401(a)(9)(E) defines the term “designated beneficiary” as an individual designated as a beneficiary by the employee. The Existing Proposed Regulations provide that, for purposes of section 401(a)(9), only individuals may be designated beneficiaries. A beneficiary who is not an individual, such as the employee's estate, may not be a designated beneficiary for purposes of determining the minimum required distribution, but nevertheless may be designated as the employee's beneficiary under the plan. If a beneficiary who is not an individual is designated to receive an employee's benefit after death, the employee is treated as having no designated beneficiary when determining the required minimum distribution. In that case, under section 401(a)(9), distributions commencing before death must be made over the employee's single life or life expectancy and distributions commencing after death must be made within 5 years of the employee's death.
However, the Existing Proposed Regulations provide that if a trust is named as a beneficiary of an employee's benefit under the plan, the underlying beneficiaries of the trust may be treated as designated beneficiaries for purposes of section 401(a)(9) if certain requirements are satisfied. In response to comments, these proposed regulations modify these trust beneficiary requirements as explained below by:
Permitting the designated beneficiary of a revocable trust to be treated as the designated beneficiary for purposes of determining the minimum distribution under section 401(a)(9), provided that the trust becomes irrevocable upon the death of the employee.
Providing relief from the requirement that the plan be provided with a copy of the trust document if certain certification requirements are met.
Irrevocability of Trust. The Existing Proposed Regulations generally provide that a trust must be irrevocable as of the employee's required beginning date in order for the beneficiaries of the trust to be treated as designated beneficiaries under the plan for purposes of determining the distribution period under section 401(a)(9)(A). Commentators have indicated that most trusts established for estate planning purposes and designated as the beneficiary of an employee's plan benefits are revocable instruments prior to the death of the employee. In response to those comments, these proposed regulations provide that a trust named as beneficiary of an employee's interest in a retirement plan be permitted to be revocable while the employee is alive, provided that it becomes irrevocable, by its terms, upon the death of the employee. The requirements in the Existing Proposed Regulations that the trust be valid under state law (or would be but for the fact that there is no corpus) and that the beneficiaries be identifiable from the trust instrument are retained.
Information to Plan Administrator. In order to permit the plan administrator to substantiate that the requirements for treating the beneficiaries of the trust as designated beneficiaries under the plan are satisfied, the Existing Proposed Regulations require that a copy of the trust instrument be provided to the plan administrator by the earlier of the required beginning date or the date of the employee's death. In response to comments, this proposed regulation permits an alternative method of substantiation.
As under the Existing Proposed Regulations, a copy of the trust instrument may be provided to the plan administrator. However, because the trust need not be irrevocable, under this method, the employee must also agree that if the trust instrument is amended at any time in the future, the employee will, within a reasonable time, provide a copy of each such amendment.
Alternatively, the employee may provide a list of all of the beneficiaries of the trust (including contingent beneficiaries) with a description of the portion to which they are entitled and any conditions on their entitlement, and certify that, to the best of the employee's knowledge, this list is correct and complete and that the other requirements for the beneficiaries of the trust to be treated as designated beneficiaries are satisfied. Under the second method, the employee must also agree to provide corrected certifications to the extent that the amendment changes the information previously certified. Finally, the employee must agree to provide a copy of the trust instrument to the plan administrator upon demand.
In addition, these proposed regulations provide that, if the minimum required distributions after death are determined by treating the beneficiaries of the trust as designated beneficiaries, a final certification as to the beneficiaries of the trust instrument must be provided to the plan administrator by the end of the ninth month after the death of the employee. This rule applies even if a copy of the trust instrument were provided to the plan administrator before the employee's death. Alternatively, an updated trust instrument may be provided.
The proposed regulations also provide that a plan will not fail to satisfy section 401(a)(9) merely because the terms of the actual trust instrument are inconsistent with the information in the certifications or trust instruments previously provided to the plan administrator if the plan administrator reasonably relies on the information provided in the certifications or trust instruments. However, the minimum required distributions for years after the year in which the discrepancy is discovered must be determined based on the actual terms of the trust instrument. For those years, the minimum required distribution will be determined by treating the beneficiaries of the employee as having been changed in the year in which the year the discrepancy was discovered to conform to the corrected information and by applying the change in beneficiary provisions found under the Existing Proposed Regulations. However, for purposes of determining the amount of the excise tax under section 4974 (including application of a waiver, if any, for reasonable error under section 4974), the minimum required distribution is determined for any year based on the actual terms of the trust in effect during the year.
Special Analyses. It has been determined that this notice of proposed rulemaking is not a significant regulatory action as defined in EO 12866. Therefore, a regulatory assessment is not required. It also has been determined that section 553(b) of the Administrative Procedure Act (5 U.S.C. chapter 5) does not apply to these regulations. Moreover, it is hereby certified that the regulations in this document will not have a significant economic impact on a substantial number of small entities. This certification is based on the fact that the reporting burden is primarily on the plan participant to supply the information rather than on the entity maintaining the retirement plan and the fact that the number of participants per plan to whom the burden applies is insignificant. Accordingly, a regulatory flexibility analysis under the Regulatory Flexibility Act (5 U.S.C. chapter 6) is not required. Pursuant to section 7805(f) of the Internal Revenue Code, this notice of proposed rulemaking will be submitted to the Chief Counsel for Advocacy of the Small Business Administration for comment on its impact on small business.
Comments and Requests for a Public Hearing. Before these proposed regulations are adopted as final regulations, consideration will be given to any written comments (preferably a signed original and eight (8) copies) or comments transmitted via Internet that are submitted timely to the IRS. All comments will be available for public inspection and copying. A public hearing may be scheduled if requested in writing by a person that timely submits written comments. If a public hearing is scheduled, notice of the date, time, and place for the hearing will be published in the Federal Register.
Drafting Information. The principal author of these regulations is Cheryl Press, Office of the Associate Chief Counsel (Employee Benefits and Exempt Organizations), IRS. However, other personnel from the IRS and Treasury Department participated in their development.
D-5. Q. If a trust is named as a beneficiary of an employee, will the beneficiaries of the trust with respect to the trust's interest in the employee's benefit be treated as having been designated as beneficiaries of the employee under the plan for purposes of determining the distribution period under section 401(a)(9)(A)(ii)?
A.(a) Pursuant to D-2A of this section, only
an individual may be a designated beneficiary for purposes of determining the
distribution period [on and after the RBD] under
section 401(a)(9)(A)(ii).[174]
Consequently, a trust itself may not be
the designated beneficiary even though the trust is named as a beneficiary. However, if the requirements of
paragraph (b) of this D-5A are met, distributions
made to the trust will be treated as paid to the beneficiaries of the trust
with respect to the trust's interest in the employee's benefit, and the
beneficiaries of the trust will be treated as having been designated as
beneficiaries of the employee under the plan for purposes of determining the
distribution period under section 401(a)(9)(A)(ii). If, as of any date on or after the employee's required beginning date,
a trust is named as a beneficiary of the employee and the requirements in
paragraph (b) of this D-5A are not met, the employee will be treated as not
having a designated beneficiary under
the plan for purposes of section 401(a)(9)(A)(ii). Consequently, for calendar
years beginning after that date, distribution must be made over the employee's
life (or over the period which would have been the employee's remaining life
expectancy determined as if no beneficiary had been designated as of the
employee's required beginning date).
(b) The requirements of this paragraph (b) are met if, as of the later of the date on which the trust is named as a beneficiary of the employee, or the employee's required beginning date, and as of all subsequent periods during which the trust is named as a beneficiary, the following requirements are met:
(1) The trust is a valid trust under state law, or would be but for the fact that there is no corpus.
(2) The trust is irrevocable or will, by its terms, become irrevocable upon the death of the employee.[175]
(3) The beneficiaries of the trust who are beneficiaries with respect to the trust's interest in the employee's benefit are identifiable from the trust instrument within the meaning of D-2 of this section.[176]
“(4) The documentation described in D-7 of this section has been provided to the plan administrator.[177]
(c) In the case of payments to a trust having more than one beneficiary, see E-5 of this section for the rules for determining the designated beneficiary whose life expectancy will be used to determine the distribution period. If the beneficiary of the trust named as beneficiary is another trust, the beneficiaries of the other trust will be treated as having been designated as beneficiaries of the employee under the plan for purposes of determining the distribution period under section 401(a)(9)(A)(ii), provided that the requirements of paragraph (b) of this D-5A are satisfied with respect to such other trust in addition to the trust named as beneficiary.
D-6.[178] Q. If a trust is named as a beneficiary of an employee, will the beneficiaries of the trust with respect to the trust's interest in the employee's benefit be treated as designated beneficiaries under the plan with respect to the employee for purposes of determining the [pre-RBD] distribution period under section 401(a)(9)(B)(iii) and (iv)[179]?
A.(a) If a trust is named as a beneficiary of
an employee and the requirements of paragraph (b) of D-5A of this section are satisfied as of the date of the employee's
death or, in the case of the documentation described in D-7 of this section, by the end of the ninth month beginning after the employee's date of
death, then distributions to the trust for purposes of section
401(a)(9) will be treated as being paid to the appropriate beneficiary of the
trust with respect to the trust's interest in the employee's benefit, and all
beneficiaries of the trust with respect to the trust's interest in the
employee's benefit will be treated as designated beneficiaries of the employee
under the plan for purposes of determining the distribution period under
section 401(a)(9)(B)(iii) and (iv). If the beneficiary of the trust named as
beneficiary is another trust, the beneficiaries of the other trust will be
treated as having been designated as beneficiaries of the employee under the
plan for purposes of determining the distribution period under section
401(a)(9)(B)(iii) and (iv), provided that the requirements of paragraph (b) of
D-5A of this section are satisfied with respect to such other trust in addition
to the trust named as beneficiary. If a
trust is named as a beneficiary of an employee and if the requirements of
paragraph (b) of D-5A of this section are not satisfied as of the dates
specified in the first sentence of this paragraph, the employee will be treated
as not having a designated beneficiary under the plan. Consequently,
distribution must be made in accordance with the five-year rule in section
401(a)(9)(B)(ii).
(b) The rules of D-5 of this section and this D-6 also apply for purposes of applying the provisions of section 401(a)(9)(B)(iv)(II) if a trust is named as a beneficiary of the employee's surviving spouse. In the case of payments to a trust having more than one beneficiary, see E-5 of this section for the rules for determining the designated beneficiary whose life expectancy will be used to determine the distribution period.
D-7.[180] Q. If a trust is named as a beneficiary of an employee, what documentation must be provided to the plan administrator so that the beneficiaries of the trust who are beneficiaries with respect to the trust's interest in the employee's benefit are identifiable to the plan administrator?
A. (a) Required distributions commencing before death. In order to satisfy the requirement of paragraph (b)(4) of D-5A of this section for [post RBD] distributions required under section 401(a)(9) to commence before the death of an employee, the employee must comply with either paragraph (a)(1) or (2) of this D-7A:
(1) The employee provides to the plan administrator a copy of the trust instrument and agrees that if the trust instrument is amended at any time in the future, the employee will, within a reasonable time, provide to the plan administrator a copy of each such amendment.
(2) The employee —
(i) Provides to the plan administrator a list of all of the beneficiaries of the trust (including contingent and remainderman beneficiaries with a description of the conditions on their entitlement);
(ii) Certifies that, to the best of the employee's knowledge, this list is correct and complete and that the requirements of paragraphs (b)(1), (2), and (3) of D-5A of this section are satisfied;
(iii) Agrees to provide corrected certifications to the extent that an amendment changes any information previously certified; and
(iv) Agrees to provide a copy of the trust instrument to the plan administrator upon demand.
(b) Required distributions after death [and after the RBD]. In order to satisfy the documentation requirement of this D-7 for required distributions after death, by the end of the ninth month beginning after the death of the employee, the trustee of the trust must either —
(1) Provide the plan administrator with a final list of all of the beneficiaries of the trust (including contingent and remainderman beneficiaries with a description of the conditions on their entitlement) as of the date of death; certify that, to the best of the trustee's knowledge, this list is correct and complete and that the requirements of paragraph (b)(1), (2), and (3) of D-5A of this section are satisfied as of the date of death; and agree to provide a copy of the trust instrument to the plan administrator upon demand; or
(2) Provide the plan administrator with a copy of the actual trust document for the trust that is named as a beneficiary of the employee under the plan as of the employee's date of death.
(c) Relief for
discrepancy between trust instrument and employee certifications or earlier
trust instruments.
(1) If required distributions are determined based on the information provided to the plan administrator in certifications or trust instruments described in paragraph (a)(1), (a)(2) or (b) of this D-7A, a plan will not fail to satisfy section 401(a)(9) merely because the actual terms of the trust instrument are inconsistent with the information in those certifications or trust instruments previously provided to the plan administrator, but only if the plan administrator reasonably relied on the information provided and the minimum required distributions for calendar years after the calendar year in which the discrepancy is discovered are determined based on the actual terms of the trust instrument. For purposes of determining whether the plan satisfies section 401(a)(9) for calendar years after the calendar year in which the discrepancy is discovered, if the actual beneficiaries under the trust instrument are different from the beneficiaries previously certified or listed in the trust instrument previously provided to the plan administrator, or the trust instrument specifying the actual beneficiaries does not satisfy the other requirements of paragraph (b) of D-5A of this section, the minimum required distribution will be determined by treating the beneficiaries of the employee as having been changed in the calendar year in which the discrepancy was discovered to conform to the corrected information and by applying the change in beneficiary provisions of E-5 of this section.
(2) For purposes of determining the amount of the excise tax under section 4974, the minimum required distribution is determined for any year based on the actual terms of the trust in effect during the year.[181] [Emphasis added.]
To paraphrase the proposed regulations: “[A] trust itself may not be the designated beneficiary even though the trust is named as a beneficiary.”[182] However, if the trust is named as the beneficiary, the beneficiaries of the trust will be treated as designated beneficiaries if:
“as of the later of the date on which the trust is named as a beneficiary of the employee, or the employee’s required beginning date, and as of all subsequent periods during which the trust is named as a beneficiary,”[183]
(1) The trust must be valid under state law, or would be but for the fact that there is no corpus.
(2) The trust must be irrevocable “or will, by its terms, become irrevocable upon the death of the employee.”
(3) The beneficiaries of the employee’s interest under the trust must be identifiable from the trust instrument.
(4) The
documentation described in D-7 must be provided to the plan administrator.[184]
(Query: “Who is the “Plan Administrator”
of an IRA?) The documentation requirements of D-7, depend upon whether
death is before or after the RBD.
(A) If death is before the RBD, the documentation requirements are met (under one alternative) if “by the end of the ninth month beginning after the death of the employee, the trustee of the trust . . . [provides] the plan administrator with a copy of the actual trust document for the trust that is named as a beneficiary of the employee under the plan as of the employee's date of death.”[185]
(B) If death is after the RBD, the requirements are met (under one alternative) if “[t]he employee provides to the plan administrator a copy of the trust instrument and agrees that if the trust instrument is amended at any time in the future, the employee will, within a reasonable time, provide to the plan administrator a copy of each such amendment.”[186]
Note that there is no requirement that the trust make pass through distributions to the beneficiary.[187]
The re-proposed regulations contain a few technical problems: (1) A testamentary trust is not a trust under state law until death, though perhaps it would be “but for the fact that there is no corpus.” (2) An IRA owner is never an “employee” though a qualified plan participant usually (but not always) will be. (3) An IRA has no plan administrator, so how can the delivery requirements be met. Assuming that the term “plan administrator” is cryptographic, does it stand for the IRA owner or the IRA institutional sponsor. Obviously, if the former, delivery is meaningless. These issues are discussed below in more detail.
A trust no longer need be irrevocable in order to qualify as a “designated beneficiary”, if it “will, by its terms, become irrevocable upon the death of the employee.”
This has never been much of an issue if death is prior to the RBD, because in that case the “applicable date” (i.e., the sooner of date of death or the RBD) is date of death, and the trust would almost certainly be irrevocable at that time.[188] Under the old proposed regulations, the irrevocability requirement could pose a problem
Obviously, a testamentary trust is irrevocable at death, and so it ought to qualify as a “designated beneficiary” if death occurs before the RBD.[189] This has always been true, and ought to be true under the re-proposed regulations as well.
Presumably a testamentary trust could never be a designated beneficiary after the RBD if the irrevocability requirement is relied upon, because it is revocable on the RBD. However, under the re-proposed regulations, a testamentary trust may qualify, since “by its terms, [it will] become irrevocable upon the death of the employee.”[190]
It is perhaps true, as a technical matter, that a testamentary
trust is not a trust under state law until death. Perhaps it would be “but for
the fact that there is no corpus,” but a testamentary trust cannot really have
corpus. Nevertheless, I feel very confident that the re-proposed regulations
were intended to cover testamentary trusts and will be construed accordingly.
It still would have been
The last of the four requirements listed in the D-5(b) Re-proposed Regulations is that “[t]he documentation described in D-7 must be provided to the plan administrator.”[191]
D-7 is new, and it is lengthy and detailed. The regulation was set forth in its entirety above. It is doubtful that many IRA owners will accidentally comply with this regulation. As a matter of public policy perspective, this is unfortunate. From the benefits lawyer’s perspective it may be a boon, but even in that case the ease of making an innocent mistake is too great for comfort.
As set forth below in some detail, the re-proposed regulations contain lengthy delivery, agreement and certification requirements. Why are the proposed regs so nit-picky? For one thing, compliance with the regulations allows the plan administrator to be able to compute the minimum distribution rules, but it would seem that an imaginative reg writer could have solved this problem with more parsimony.
The only other possible policy purpose this requirement could serve (other than to harass the taxpayers) would be to lessen the incentive for a beneficiary to commit forgery and possibly perjury by altering the instrument to change the beneficiary after the death of the participant. This is a somewhat cynical approach to the problem, assuming the likelihood of this happening to be significant, which I doubt.
Further, in the case of death after the RBD, any forgery would have to be consistent with the life expectancy elections that were in place at date of death, so what would be the point? (Again, it is perfectly permissible to change one’s beneficiary at any time, whether before or after the RBD, by substituting one irrevocable trust for another, so long as the shortest measuring life expectancy is used in the case of changes after the RBD.)
The re-proposed regulations state in the preamble:
Proposed §§1.408-8 and 1.403(b)-2 (also included in EE-113-82) provide that the provisions of proposed §1.401(a)(9)- 1 generally apply to individual retirement plans, and section 403(b) annuity contracts, custodial accounts, and retirement income accounts. Accordingly, these amendments and additions also generally apply to such plans, contracts, and accounts.[192]
Notwithstanding this general statement, the re-proposed regulations uniformly use the terms “employee” and “plan administrator” and never directly hint that the term “employee” includes an IRA owner (who is never an employee with respect to the IRA, and certainly need not be or ever have been an employee with respect to anyone). Further, the term “plan administrator” would appear to have no application to an IRA and even if it did, it would not be clear whether it refers to the IRA owner or to the IRA institutional sponsor, who may be either a custodian or a trustee.
D-7A(a)(1) and (2) require that if the employee is alive on the RBD, the employee must, presumably by the RBD or as of any date the beneficiary is subsequently changed, either—
(1) provide to the plan administrator a copy
of the trust instrument and agree “that if the trust instrument is amended at
any time in the future, the employee will, within a reasonable time, provide to
the plan administrator a copy of each
such amendment”[193];
or,
(2) Comply with all of the provisions of D-7A(a)(2), which are that the employee —
(i) Provides to the plan administrator a list of all of the beneficiaries of the trust (including contingent and remainderman beneficiaries with a description of the conditions on their entitlement);
(ii) Certifies that, to the best of the employee's knowledge, this list is correct and complete and that the requirements of paragraphs (b)(1), (2), and (3) of D-5A of this section are satisfied;
(iii) Agrees to provide corrected certifications to the extent that an amendment changes any information previously certified; and
(iv) Agrees to provide a copy of the trust instrument to the plan administrator upon demand.[194]
I submit that the alternative (1)[195] will be the easiest to comply with and the one most likely to result in certainty of compliance.
Obviously, the delivery and certification must be made prior to the RBD to be effective on the RBD. Delivery and certification could be made later, if there was a designated beneficiary on the RBD. Presumably in the latter case, the delivery and certification would have to be done simultaneously with the change in beneficiary designation.
Query whether it is necessary to agree “that if the trust instrument is amended at any time in the future, the employee will, within a reasonable time, provide to the plan administrator a copy of each such amendment”[196] if the trust is irrevocable! Literally, the regulation requires this.
In addition, the
certification, or, alternatively, an updated copy of the trust, must be given
all over again, after the date of death and before the end of the ninth month
beginning after the death of the employee.[197]
The re-proposed regulations provide that by the end of the ninth month beginning after the death of the employee, the trustee of the trust must either — :
(1) Provide the plan administrator with a final list of all of the beneficiaries of the trust (including contingent and remainderman beneficiaries with a description of the conditions on their entitlement) as of the date of death; certify that, to the best of the trustee's knowledge, this list is correct and complete and that the requirements of paragraph (b)(1), (2), and (3) of D-5A of this section are satisfied as of the date of death; and agree to provide a copy of the trust instrument to the plan administrator upon demand; or
(2) Provide the plan administrator with a copy of the actual trust document for the trust that is named as a beneficiary of the employee under the plan as of the employee's date of death.[198] [Emphasis added.]
Again, as was the case where the participant or IRA owner is living on the RBD, the recommended procedure is to simply deliver the document, rather than to try to make all of the required certifications, which could involve the certifier in the nuances of properly characterizing contingent beneficiaries under the law of estates and future interests.
The delivery must be made within nine months of death and it must be made by the trustee. If the delivery were by the “employee” during life, would this requirement be satisfied? Apparently not.
According to the Preamble of the re-proposed regulations, and as reflected in D-6(a) and D-7(b) of the reproposed regs.
In addition, these proposed regulations provide that, if the minimum required distributions after death are determined by treating the beneficiaries of the trust as designated beneficiaries, a final certification as to the beneficiaries of the trust instrument must be provided to the plan administrator by the end of the ninth month after the death of the employee. This rule applies even if a copy of the trust instrument were provided to the plan administrator before the employee's death. Alternatively, an updated trust instrument may be provided.[199] [Emphasis added.]
This requirement is found in D-6(a) where death is pre-RBD, and in D-7(b) where death is post RBD.
The life expectancy of the oldest beneficiary of the trust is generally used to determine the maximum permissible payment period.[200] However, see the separate account exception[201] discussed below.
If a trust does not qualify as a designated beneficiary, the benefit must be paid out no later than 5 years following death, if death occurs before the RBD.[202] After the RBD, a nonqualifying trust is simply ignored as a joint measuring life. Distributions can still be made over the life expectancy of the participant, and these distributions can continue after death if life expectancy was not being recalculated and if the participant did not outlive the expected period.
A thorny problem that does not lend itself to an easy solution is the common situation where the trustee of a testamentary trust or a living trust (that becomes irrevocable at death if not sooner) is named as the beneficiary of the qualified plan or IRA benefits.
It may be that the beneficiaries are ascertainable, but that the share of the IRA or qualified plan benefits that any particular beneficiary will receive is uncertain, because the fiduciary has a “pick and choose” power to allocate the assets to one trust or beneficiary or to another, as he or she sees fit. Is it sufficient that the beneficiaries are ascertainable, or is it implied that the shares to which the beneficiaries are entitled be also ascertainable on the applicable date? That is the real question.
What if one of the beneficiaries in not an individual. For example, what if the trust makes a number of small pecuniary bequests, one of which is a $1000 memorial to a church or synagogue. Under the multiple beneficiary rule,[203] subject to the separate account rule perhaps,[204] if one of the beneficiaries is not an individual, the life expectancy method will not be available. Will this be ignored is the gift to the non-individual is de minimis. Will the IRS take a wait and see approach, waiting to see just what it is that the trustee does, and then apply to the rules to facts as they turn out to be? We do not know.
Note that applying the minimum distribution rules here, probably requires the application of different principles than those employed in the spousal rollover context. In that context, the IRS routinely (and perhaps surprisingly) often does take a wait and see approach. For instance, if the beneficiary is the estate, the spouse is the executor, the will has a formula marital deduction clause, the executor/spouse elects to allocate the IRA proceeds to the marital deduction trust, and the marital deduction trust gives the spouse a lifetime general power of appointment, and the spouse proposes to exercise the power to take a distribution of the benefits from the trust and roll them over, the IRS has allowed the rollover![205] I do not think these rollover cases are necessarily relevant, though they might be useful to us by analogy, and if the same approach is applied in the 401(a)(9) area, would argue in favor of a wait and see approach.
The proposed regulations under the minimum distribution rules do not specifically require that the shares be ascertainable, and at this point, with what little there is to go on, I believe that a good faith interpretation of the proposed regulations and the statute is that the ultimate size of the shares need not be ascertainable. We don’t really need to know the shares to determine the minimum distribution required, unless we are going to attempt to apply the separate account rule[206] (discussed elsewhere), because the multiple beneficiary rule[207] (discussed elsewhere) requires, in any event, that we take the life expectancy of the oldest beneficiary of the trust as the measuring life. Despite my opinion on what a good faith reading would disclose, the safer course would be to try to draft in such a manner that these uncertainties are eliminated, if that is possible.
We know that an estate (or the beneficiaries of an estate) cannot be designated beneficiaries for purposes of avoiding the 5-year rule where death is prior to the RBD, or for purposes of using the joint life expectancy method at and after the RBD. Can the beneficiaries of a trust that functions for all practical purposes in the same manner as a decedent’s probate estate be treated as “designated beneficiaries,” and if so, why should this designation work, when a designation of the estate will not?[208] The answer, which admittedly begs the question, may simply be that a trust is okay in this circumstance, because the proposed regulations and legislative history say it is, but that both the proposed regulations and the legislative history are unequivocal that the life expectancy payout method is not to be available to a decedent’s estate or to the beneficiaries of a decedent’s estate.
To cite just one example, consider the common two trust estate plan, in which the trustee is directed to collect the proceeds as to which it is a beneficiary, and to allocate these proceeds between trust A (the marital deduction trust) and trust B (the bypass trust) in accordance with a formula under the will or trust? This achieves the desired flexibility necessary to fine tune the estate plan, but it engenders other problems.
The problem is that the amounts determined under credit shelter or marital deduction formula clauses are often not determinable until the estate is closed. For example, to the extent that administration expenses chargeable to principal are not deducted by the estate, the amount available for funding a credit shelter gift will be reduced. This is because the estate is only allowed a marital deduction for the net value of the property passing to the spouse (adjusted for gains and losses, perhaps). If the will employs a formula clause, the marital deduction claimed would have to be net of such expenses. Therefore, if the property passing to the spouse were subject to reduction for administrative expenses, this would have to be reflected in the size of the marital deduction claimed. In practice, this means that charges to principal that are not deducted on the 706 reduce the credit shelter gift. Since the credit shelter gift is, in effect, the dollar amount left over after the marital deduction has been computed, the credit shelter gift must be reduced.[209]
Generally, IRC §642(g) limits the taking of a double deduction for “swing items,” i.e., items which are deductible on the estate’s income tax return (Form 1041) and are also deductible on the estate tax return (Form 706). Administration expense is a common example.
In order for the trust to be a “designated beneficiary” so that the life expectancies of the beneficiaries can be used to determined the minimum required distributions, “the beneficiaries of the trust who are beneficiaries with respect to the trust's interest in the employee's benefit [must be] identifiable from the trust instrument, within the meaning of D-2.”[210] Again, this regulation does not state that the share to which the trust or the beneficiaries is entitled need be identifiable, only that the beneficiaries be identifiable.[211]
D-2 contains the following example, which is only marginally helpful here:
Employee X attains age 70 1/2 in calendar year 1990. As of
Can the requirement that the beneficiaries be identifiable be met if the trustee has the discretion to determine which trust, or in what proportions, or to whom, to give the benefit? We ought in that case to be able to use the beneficiary of either trust having the shortest life expectancy as the measuring life, at a minimum, but we just do not know if that will be good enough. But, if the beneficiaries of both trusts are the same, it is hard to see how identifying the beneficiaries could be the issue, and, again, the regulation says nothing about the share that the beneficiary is to receive being identifiable.
What if the trustee has the discretion to use plan benefits for other purposes, such as payment of taxes or debts or to satisfy pecuniary legacies? This is even more troubling.
And what about the irrevocability issue? Is the trust irrevocable at death, if the trustee has discretion to use the benefits to satisfy different gifts or different trusts with the benefits? Again, we do not know, and must await final or temporary regulations that we hope will expound upon, or better yet, change, the meager rules that we now have.
Consider the following suggested trust provision:
Retirement Plan Benefits. If the trustee is named as the beneficiary of retirement plan benefits that are subject to the minimum distribution rules of IRC §401(a)(9), and if under the circumstances existing at date of death the benefits or the right to receive the benefits may continue to be held in trust (and for this purpose the survival of the beneficiaries for any post death survivorship period shall be presumed), then if (a) the participant dies prior to his required beginning date (as that term is used in IRC §401(a)(9)), or (b) (i) the participant dies on or after the required beginning date and (ii) if this trust is a “designated beneficiary” so that the joint life expectancies of the beneficiaries and the participant are being used for purposes of the minimum distribution rules, then, in either case (a) or (b), the following rules apply:
Such retirement plan benefits shall not be used to pay debts or expenses or pecuniary bequests, if other assets are available for those purposes, notwithstanding the rules otherwise applicable to apportionment, abatement and the payment of debts and expenses. Further, the rules otherwise applicable to apportionment and abatement of death taxes are hereby expressly limited to provide that in no event shall retirement plan benefits be used to pay death taxes that are not directly attributable and proportionate to the estate tax value of the benefits.
If the trust estate is divided into fractional shares, the interests of the trustee (and the beneficiaries) of a particular trust in retirement plan benefits shall be likewise divided, proportionately [if residuary marital], except that the right to receive the retirement proceeds shall be allocated to the trustee of a trust that is not intended to qualify for the marital deduction before being allocated to a fractional share trust that is intended to qualify for the marital deduction, to the extent possible without accelerating income taxes or affecting the estate tax value of the marital deduction gift[endif].
It is the purpose of these rules to insure that the trust is considered irrevocable and that the beneficiaries of the trusts be identifiable, so that the life expectancies of the beneficiaries may be used to calculate the minimum distributions required by IRC §401(a)(9), and this paragraph shall be interpreted with this intent being paramount to any other direction in it.
These clauses are still evolving, and great care should be taken before using them. Exempting the benefits from being used to pay debts, taxes and expenses could place the remaining trust estate under considerable strain. On the other hand, using the assets for such purposes may not be physically compatible with a long term payout, even if permissible. However, because of the possibility that the estate tax burden could be severe, I believe it may be prudent to leave open the option to charge the benefits with its proportionate share of those taxes, as more particularly specified in the estate tax apportionment provisions of the instrument. We certainly hope that such a provision would not limit our ability to use one of the life expectancy methods.
The facts state that a wife inherited an IRA from her husband. Whether this was before or after the husband’s RBD is not clear, but the facts state that the wife rolled over the distribution to an IRA in her name and had attained age 701/2 before she died. The beneficiary of her IRA was a trust that became irrevocable at her death. The Service ruled that the minimum distribution rules could not be applied by reference to the husband’s life because of the spousal rollover, which has the effect of removing the husband from the picture for 401(a)(9) purposes from the husband’s death forward.
The inexplicable part of the ruling is that, based upon the taxpayer’s representation that the D-5 proposed regulation had been satisfied as of the wife’s death, the D-6 trust requirements (applicable where death is prior to the RBD) would be met. (The trust was irrevocable upon death of the wife.) Therefore, distributions from the IRA following the wife’s death were allowed to continue using the life expectancy method of IRC §401(a)(9)(B)(iii) (based upon the life expectancy of the oldest trust beneficiary. The life expectancy method of IRC §401(a)(9)(B)(iii) applies where death occurs prior to the RBD.
To account for this result, one must conclude that although the wife died after reaching 701/2, she must have died before her RBD (April 1, of the following year).
Incidentally, the separate account rule was not mentioned. The Service simply stated that since there was more than one beneficiary of the trust, the beneficiary with the shortest life expectancy would have to be used as the measuring life.
There is a strange rule found in IRC §401(a)(9)(F) to the effect that any amount paid to a minor child shall be treated as if paid to the surviving spouse if such amount will be payable to the surviving spouse upon such child reaching majority.
(F) Treatment of payments to children. Under regulations prescribed by the Secretary, for purposes of this paragraph [401(a)(9)], any amount paid to a child shall be treated as if it had been paid to the surviving spouse if such amount will become payable to the surviving spouse upon such child reaching majority (or other designated event permitted under regulations).[213] [Emphasis added.]
This rule appears to require regulations in order to be effective. It is clear that distributions will have to be payable to the spouse upon the child reaching majority, in the absence of regulations designating other permissible events; but it is not clear to me whether regulations are required in order for 401(a)(9)(F) to be generally effective. As of the time this is written, there are no regulations, and almost no commentary from third parties (your author excepted).
The purpose of this rule is not clear to me. It might be that, but for 401(a)(9)(F), if the beneficiary of an IRA or qualified plan were A for a period of years (in this case a minor child until 18), followed by distributions to B (in this case the spouse), neither would be a “designated beneficiary.”
I can think of several instances where the application of the rule might be helpful. Say that death is prior to the RBD and the beneficiary is the child until the child reaches 18, whereupon distributions are to me made to the spouse. If the beneficiary were the spouse all along, distributions would not be required until the year after the decedent would have reached 70&1/2, which may be after the child has reached 18. But then what would be the point of naming the child as beneficiary unless distributions will actually be made? Perhaps what the rule seeks to accomplish is to permit distributions to be made only if needed; otherwise distributions won’t be made — the latter option being one that is unavailable if someone other than the spouse is a beneficiary. But who makes that decision? Perhaps the IRA trustee, under special provisions in the beneficiary designation.
Can the IRA trustee be given discretion whether or not to make distributions, if distributions to a spouse would not be minimum required distributions (MRDs). Presumably the answer is yes, and that is the reason for the rule. But I am not sure about either conclusion. If the child didn’t need a distribution, and if distributions to the spouse would not be required, then distributions could be deferred under circumstances that would otherwise have required annual distributions but for the exception.
Maybe the income tax rate is a consideration.
Here is an important question, for more reasons than one: Is 401(a)(9)(F) necessary if a trust is named as beneficiary, where distributions are keyed to the needs of the child, but that after the child attains 18, distributions will be made solely to the spouse? Perhaps. Because in that case distributions will need to be made to the trust annually, absent 401(a)(9)(F), whether or not distributed to the child. I assume that the fact that the spouse is one of several beneficiaries will not by itself be sufficient to use the spousal deferral exception of §401(a)(9)(B)(iv)(I).
Incidentally, the 401(a)(9)(F) exception would not qualify the trust for QTIP treatment because the fiction is only “for purposes of this paragraph” and the paragraph referred to is not within §2056.
Under Prop. Treas. Reg. §1.401(a)(9)-1, Q&A E-5(a), if there is more than one death beneficiary, the beneficiary with the shortest life expectancy is generally used as the measuring life or joint measuring life with the participant. We will refer to this as the Multiple Beneficiary Rule.
“Except as otherwise provided in paragraph (f) [designations by beneficiaries], if more than one individual is designated as a beneficiary with respect to an employee as of the applicable date for determining the designated beneficiary, the designated beneficiary with the shortest life expectancy will be the designated beneficiary for purposes of determining the distribution period.”[214]
If one of the beneficiaries is not an individual (or a qualifying trust), then the participant will be treated as not having any designated beneficiaries. (Certain contingent beneficiaries are ignored for this purpose. It is an important exception, and I will discuss this later.)
“However, except as otherwise provided in D-5, D-6 [trusts as beneficiaries], and paragraph (e)(1) of this E-5 [death contingency], if a person other than an individual is designated as a beneficiary, the employee will be treated as not having any designated beneficiaries for purposes of section 401(a)(9) even if there are also individuals designated as beneficiaries.”[215] [Emphasis added.]
“If any designated beneficiary involved is the employee’s spouse and the life expectancy and the life expectancy of the spouse is being recalculated, the life expectancy of the spouse as recalculated will be compared in each calendar year to the remaining life expectancy of the other applicable designated beneficiary or beneficiaries, not recalculated, and the shortest life expectancy will be used for determining the minimum distribution required for that calendar year.”[216] [Emphasis added.]
It is important to note at the onset that the death of a beneficiary after the applicable date is generally irrelevant, unless the beneficiary is a spouse and recalculation was elected:
If the designated beneficiary whose life expectancy is being used to calculate the distribution period dies on or after the applicable date, such beneficiary's remaining life expectancy will be used to determine the distribution period whether or not a [new] beneficiary with a shorter life expectancy receives the benefits.[217]
This rule hints at an important theoretical principle.
I am suggesting in this subparagraph a possible articulation of the contingent beneficiary rule that focuses on the identification of the beneficiaries who are alive or in existence on the applicable date (i.e., either the RBD, or date of death if sooner). If, examining that point in time only, all of the beneficiaries are human beings, and if they would be the only persons capable of receiving benefits from the plan or IRA should they live long enough, then the inquiry ought perhaps to be over, with the conclusion being that all of the relevant beneficiaries are designated beneficiaries.
It occurs to me that one possible way of looking at the rule would be to ask the following question: “If all of the primary beneficiaries who are alive or in existence on the applicable date survive the complete distribution of the beneficial interest, is there any contingency under which anyone who is not an individual might receive a distribution (as a beneficiary) of the plan or IRA proceeds?” In the case of a trust, the question would be similarly posed, except that it would be expanded to include all of the contingent beneficiaries of the trust who are living (or in existence) on the applicable date, assuming of course that the contingent beneficiaries are all identifiable.
A person who is named as a beneficiary may either (a) fail to survive the participant or (b) having survived the participant, may fail to survive the complete distribution of the interest. The proposed regulations address the effect of both cases on the computation of the minimum distribution period.
In order to determine whether or not the joint life expectancy method is available, one looks solely to circumstances existing on the applicable date —i.e., either the RBD or date of death if sooner. (This is a point you emphasized in your article.) Therefore, in order to determine whether or not there are any beneficiaries who are not designated beneficiaries, one might ask the question posed above, another formulation of which might be: If the IRA or qualified plan interest were distributed in full during the lifetime of all of the primary beneficiaries who are alive or in existence on the applicable date, is there any contingency under which the benefit could pass to anyone who is not a human being or a qualified trust? If the answer is “no” then all of the beneficiaries are designated beneficiaries; otherwise, perhaps not.
If a trust is the beneficiary, then to determine whether or not it is qualified, one might ask a similar question: If the IRA or qualified plan interest were distributed in full to the trust during the lifetime of all of the beneficiaries (contingent or contingent)who are alive or in existence on the applicable date, and if the trust thereupon terminated, is there any contingency under which the benefit could pass to anyone (other than to the trust itself) who is not a human being?
Approaching the issue in the manner suggested by the above questions is consistent with the notion found in Prop. Treas. Reg. §1.401(a)(9)-1, Q&A E-5(e)(2) that we can ignore the death of an individual after the applicable date (other than for recalculation purposes). It is also consistent with PLR 9820021 and PLR 9846034. We begin, for testing purposes, with the assumption that if all of the beneficiaries live long enough, they will eventually receive all of the proceeds that were subject to 401(a)(9). We ignore —as §1.401(a)(9)-1, Q&A E-5(e)(2) implies we are entitled to do— the “contingency” that they might not live so long. I believe this approach to be consistent with Prop. Treas. Reg. §1.401(a)(9)-1, Q&A E-5(b) & E(5)(e)(1) as well, discussed below. The only limitation on the application of the rule as thus formulated, as applied to trusts, is that the human beings who are the trust beneficiaries must be more than nominal beneficiaries.[218]
Before addressing the issue posed in the heading, I ask the reader to first recall Prop. Treas. Reg. §1.401(a)(9)-1, Q&A E-5(a), the multiple beneficiary rule:
“E-5. Q. If an employee has more than one designated beneficiary or if a designated beneficiary is added or replaces another designated beneficiary after the date for determining the designated beneficiary [i.e., after the RBD or date of death if sooner], which designated beneficiary's life expectancy will be used to determine the distribution period?
“A. (a) General rule. (1) Except as otherwise provided in paragraph (f) [designations by beneficiaries], if more than one individual is designated as a beneficiary with respect to an employee as of the applicable date for determining the designated beneficiary, the designated beneficiary with the shortest life expectancy will be the designated beneficiary for purposes of determining the distribution period. However, except as otherwise provided in D-5, D-6 [trusts as beneficiaries], and paragraph (e)(1) of this E-5 [death contingency], if a person other than an individual is designated as a beneficiary, the employee will be treated as not having any designated beneficiaries for purposes of section 401(a)(9) even if there are also individuals designated as beneficiaries.”[219] [Emphasis added.]
(A-1) The E-5(b) Rule.
To this rule there is an exception provided in Prop. Treas. Reg. §1.401(a)(9)-1, Q&A E-5(b):
“(b) Contingent beneficiary. Except as otherwise provided in paragraph (e)(1), if a beneficiary’s entitlement to an employee’s benefit is contingent on an event other than the employee’s death (e.g., death of another beneficiary), such contingent beneficiary is considered to be a designated beneficiary for purposes of determining which designated beneficiary has the shortest life expectancy under paragraph (a).”[220] [Emphasis added.]
Both the E-5(b) and the E-5(a) regulations, quoted above, are subject to E-5(e)(1). The (e)(1) exception referred to is very important.
“(e) Death contingency. (1) If a beneficiary’s entitlement to an employee’s benefit is contingent on the death of a prior beneficiary, such contingent beneficiary will not be considered a beneficiary for purposes of determining who is the designated beneficiary with the shortest life expectancy under paragraph (a) or whether a beneficiary who is not an individual is a beneficiary. This rule does not apply if the death [of the beneficiary] occurs prior to the applicable date for determining the designated beneficiary.”[221] [Emphasis added.]
The applicable date is presumably either the RBD or the participant’s date of death if sooner.[222] We will refer to this exception to the Multiple Beneficiary Rule as the Contingent Beneficiary Rule.
|
E-5(b) |
E-5(e)(1) |
|
(b) Contingent beneficiary. Except as otherwise provided in paragraph (e)(1), if a beneficiary’s entitlement to an employee’s benefit is contingent on an event other than the employee’s death (e.g., death of another beneficiary), such contingent beneficiary is considered to be a designated beneficiary for purposes of determining which designated beneficiary has the shortest life expectancy under paragraph (a).”[223] [Emphasis added.] |
(e) Death contingency. (1) If a beneficiary’s entitlement to an employee’s benefit is contingent on the death of a prior beneficiary, such contingent beneficiary will not be considered a beneficiary [ at all, designated or otherwise] for purposes of determining who is the designated beneficiary with the shortest life expectancy under paragraph (a) or whether a beneficiary who is not an individual is a beneficiary. This rule does not apply if the death [of the beneficiary] occurs prior to the applicable date for determining the designated beneficiary.”[224] [Emphasis added.] |
Do yourself a favor and skip the rest of this paragraph. Under E-5(b), if it applies, a beneficiary is a designated beneficiary if the “beneficiary’s entitlement to an employee’s benefit is contingent on an event other than the employee’s death.” This language is somewhat misleading, because it implies that if, under the last sentence of E-5(e)(1), the death of the beneficiary is prior to the applicable date, that beneficiary is a designated beneficiary. Because this is absurd, we know that this interpretation is incorrect (perhaps because the predeceased beneficiary had only an expectancy rather than an entitlement to the benefits?). In any event, I feel confident that a beneficiary who dies prior to the applicable date is to be ignored, even though such beneficiary’s entitlement is contingent on surviving the applicable date, rather than surviving the employee, and even though such entitlement is not technically contingent on the employee’s death (which is a certainty, not a contingency). Phew! So much for that.
A more awkward cross-reference scheme is difficult to imagine. If the meaning is clear to you at first reading, your MENSA dues will be waived. Here is what I think the quoted regulations mean, but be advised that I am not smart enough to confidently say I am correctly interpreting them.
Reading E-5(b) and E-5(e)(1) Together. The E-5(e)(1) exception referred to within E-5(b) reads almost identically to E-5(b), except that the result is reversed. E-5(b) is expressly subject to the exception in E-5(e)(1). But E-5(e)(1) does not apply (i) if the death of the beneficiary is prior to the applicable date, or (ii) the beneficiary’s entitlement does not depend on the death of a prior beneficiary, which means (apparently) that E-5(b) applies without exception in those cases.[225]
If a beneficiary dies before the applicable date, then I believe the beneficiary is ignored, although you really cannot get there from reading these two regulations. E-5(b): However, if the beneficiary(ies) survives the applicable date, and entitlement to the benefit is contingent on some event other than the death of a prior beneficiary, then the E-5(b) regulation applies, and the beneficiary or beneficiaries who survived the applicable date are treated as designated beneficiary(ies). E-5(e)(1): If the beneficiary dies after the applicable date, and the beneficiary’s entitlement is contingent on the death of a prior beneficiary, then the E-5(e)(1) regulation applies and that beneficiary is NOT treated as a designated beneficiary.
What I think the E-5(b) regulation is merely saying, in the most abstruse manner imaginable, is that if participant P designates A as his beneficiary, but if A does not survive P then to B, then, if A is not living on P’s RBD or prior death, B will be the designated beneficiary, not A. The E-5(e)(1) regulation on the other hand would apply under the same circumstances except that A is living on the P’s RBD or prior death, in which case, A would be the designated beneficiary, not B. If however, B takes if A is unmarried at P’s death —to give but one example of a contingency other than (or more properly in addition to) the employee’s death— then both A and B are considered beneficiaries and subject to the multiple beneficiary rule.
Note that the phrase that ends the first sentence of E-5(e)(1) states that, if A survives P, it makes no difference that B is not an individual, because “such contingent beneficiary will not be considered a beneficiary for purposes of determining . . . whether a beneficiary who is not an individual is a beneficiary.” Thus, even if B is not an individual, the fact that B may succeed to A’s interest may be ignored, if B’s “entitlement to . . . [the] benefit is contingent on the death of [A].”
Although E-5(b) seems to be explicit that for purposes of applying the multiple beneficiary rule after the applicable date, we can ignore beneficiaries whose entitlement is “contingent on the death of a prior beneficiary” this begs the issue of who or what this means.
(C-1) An Example From
the Proposed Regulations.
The proposed regulations give the following example:
Example. The designated
beneficiary of an unmarried participant (X) as of X's required beginning date on
This example clearly implies that if A is living on X’s RBD or later death, but dies before withdrawing all of X’s benefits, B will take what is left.
What would be an example of a situation where B’s interest was not contingent on A’s death, so that B and A would both be considered to be beneficiaries under the multiple beneficiary rule? How about this: A is the death beneficiary unless A is still married to her worthless husband at X’s death, in which event, B takes.
There is some uncertainty about the application of the Contingent Beneficiary Rule.
Note that the only contingency recognized under E-5(e)(1) is the death of the primary beneficiary, assuming the primary beneficiary is living on the participant’s date of death or the RBD if earlier. Note further, that since the primary beneficiary will be an individual, the death of the primary beneficiary is not a contingency, it is a certainty. And finally, note that it is entitlement to the benefit, not enjoyment, that must be contingent. The focus, therefore, is certainly not on whether the primary beneficiary will die, but rather, on whether the secondary beneficiary’s entitlement to any interest remaining at that time is contingent on this death. It is mind boggling to view the contingency from this aspect, given that the death of the primary beneficiary is not itself contingent.
It may be that the way to answer the question of whether or not the secondary beneficiary is entitled to this benefit is to apply the law of future interests to determine whether the interest is vested or contingent. If the secondary beneficiary is an individual, survivorship will usually be a condition of vesting, and an interest in a remainderman that is expressly contingent on surviving the holder of a predecessor interest is normally considered contingent.[227] If this analysis is correct, then the Contingent Beneficiary Rule will apply if “a beneficiary’s entitlement to an employee’s benefit is contingent on [surviving] the death of a prior beneficiary.” This conclusion forms a basis for distinguishing between individual remaindermen and nonindividuals, a distinction we have reason to believe the Service wishes to make.
More regulatory guidance is definitely needed on the question of what is a death contingency.
I am assuming the same analysis that is generally applicable to determining whether a beneficiary is a contingent beneficiary for purposes of the Contingent Beneficiary Rule will apply regardless of whether the beneficiary is a beneficiary of a trust. That is to say, if the interest of a beneficiary of a trust is contingent on surviving a prior beneficiary, then the Contingent Beneficiary Rule should operate to allow us to ignore the life expectancy of the secondary beneficiary in determining the minimum payments required under IRC §401(a)(9), and that this is true even if the beneficiary is not an individual.
One reason for wondering whether a different analysis might be in order is that property received by the trust can and usually will be accumulated for distribution to the primary beneficiary later in life, or for distribution to the secondary and contingent beneficiaries, following the death of the primary beneficiaries. However, we know that distributions made to a trust under the minimum distribution rules are not required to be redistributed to the trust beneficiaries at the same time received.[228] If the rule were otherwise there would be little point in having a trust as a designated beneficiary! Further, unless the Contingent Beneficiary Rule is given literal application to trusts, insurmountable difficulties emerge at once.
All well drafted trusts are going to provide for the contingency that all of the beneficiaries might die prior to the complete distribution of the trust estate. At that point, the remaining trust estate could (a) be distributed to the beneficiaries’ probate estates, (b) be distributed as the beneficiaries might appoint, with gift over in default, or (c) be distributed to other individuals, such as heirs at law determined under the laws of intestate succession. None of these alternatives, with the possible exception of (b),[229] can be drafted to assure that no one other than an individual younger than the oldest primary beneficiary will ever receive the remainder of the trust estate. Moreover, the rule against perpetuities requires that a beneficiary’s interest in a trust must vest sooner or later, and, if the beneficiary has died, the only way it can vest is to be distributed to the beneficiary’s estate or distributed pursuant to a general power of appointment in the beneficiary. The fact that it is virtually impossible to draft a trust instrument that will assure that the ultimate beneficiary will under every contingency be an individual younger than the oldest primary beneficiary is a persuasive reason for believing that the Contingent Beneficiary Rule is equally applicable to accumulation trusts as to individual beneficiaries.
What if the remote beneficiary is not an individual? What was indicated above, is reiterated here: if the beneficiaries include someone who is not an individual, then the Multiple Beneficiary Rule will provide that no one can be treated as a designated beneficiary, unless the beneficiary is a contingent beneficiary.[230] A trust is an exception to this rule. However, if the beneficiaries of the trust include someone who is not an individual, then the Multiple Beneficiary Rule could apply to the trust, unless the beneficiary is a contingent beneficiary under the Contingent Beneficiary Rule.
If a nonindividual is entitled to receive distributions (if any remain to be made) after the death of a prior beneficiary, then perhaps there is no “designated beneficiary.” This may be true, whether or not an accumulation trust is a medium. The reason is that the interest of the nonindividual will likely be vested. For one thing, there is no survivorship contingency. Thus, the death of the prior beneficiary merely affects timing, not entitlement.
If the remainderman is an estate of the prior beneficiary, we have
a special case, since the estate of a beneficiary is an extension of the
beneficiary, and this is merely a common form of vesting the interest. Moreover,
the last clause of the first sentence under Prop. Treas. Reg. §1.401(a)(9)-1,
Q&A E-5(e)(1). (Proposed
The problem is more likely to be found where the secondary beneficiary is a charity, not because the charity is not an individual, but because the interest of the charity is (arguably) not contingent. (Never mind that the only reason it would not be contingent is because it is not an individual.)
In the case of a charitable remainderman, although receipt of the benefit will be contingent on the death of the primary beneficiary, the charity’s entitlement to the benefit will not be expressly conditioned on survival. Unlike individuals, charities can theoretically live forever and will be presumed to survive. Therefore, the charity is more likely to be treated as a vested remainderman in a situation where an individual might more properly be treated as a contingent remainderman. Or, using the language of the proposed regulations, entitlement to (as opposed to enjoyment of) the benefit is not contingent on the death of the prior beneficiary if the secondary beneficiary does not have a limited life.
To give a common example, if a QTIP trust were the beneficiary, and a charity was the remainderman, it might be the case that the distribution must be made at the end of five years (if the participant dies before the RBD), or at the end of the participant’s life expectancy (if the participant dies after the RBD and is not using the recalculation method).[231] The concern is that if the interest of the charity is treated as vested, then under the Multiple Beneficiary Rule the participant will be considered as having no designated beneficiaries. Whereas, if the interest is treated as “contingent on the death of a prior beneficiary,” it may be ignored under the Contingent Beneficiary Rule. The result of the charity being treated as a noncontingent beneficiary of a QTIP trust, in a case where the participant dies prior to her RBD, is that (a) not only must all of the income be distributed annually under the QTIP rules, but (b) the entire remaining interest must be distributed— and income taxes paid— at the end of five years, under the minimum distribution rules,[232] substantially diminishing both the amount passing to the spouse and the amount passing to charity.[233]
The same problem is present, even without a trust, if a charity is a remote beneficiary. Here, even if the prior beneficiary were vested (with a general power of appointment, power of acceleration, etc.), the interest of the charity might be considered as a vested remainder subject to divestment by exercise of the power of appointment in the prior beneficiary.[234] Again, if there was a survivorship requirement, the gift over would be in the nature of a contingent remainder which cannot vest until survivorship in the remote beneficiary is established.[235]
Much of the above analysis assumes that the law of future interest is either directly applicable or analogous, which may or may not turn out to be the case upon the promulgation of final regulations.
A recent PLR held the decedent’s wife was not a designated beneficiary under a trust where the spouse (apparently?) had a general testamentary power of appointment, in default of the exercise of which the remainder would pass to charity.[236] This ruling is shocking, to say the least.
Trust M provides, in part, that the income and principal of the Marital Funds will be distributed as follows:
(1) commencing at Individual A's death and during the life of Individual B, the trustee shall pay the income to Individual B in annual or more frequent installments as may be convenient to her;
(2) in addition the trustee from time to time may pay to Individual B such amounts from the principal as the trustee considers necessary for Individual B's health and medical needs, taking into consideration the income and cash resources known to the trustee to be available for such purposes from other sources; and
(3) upon Individual B's death, all accrued or undistributed income shall be distributed to or for the benefit of such appointees (including Individual B's estate) as Individual B may provide by will making specific reference to the power of appointment.
At the death of Individual B, any interest in Marital Fund II (which also receives amounts in Marital Fund I after Individual B's death), not effectively disposed of shall be divided into four equal parts, after the payment of certain debts and distribution of real estate interests, and distributed by the trustee as follows:
(1) two parts to University Q;
(2) one part to School R; and
(3) one part to Reservation S.
* * * *
Because additional amounts that are distributed from Plan X could remain in Trust M during Individual B's lifetime, three organizations (not individuals), University Q, School R and Reservation S, are entitled to benefits while Individual B is alive unless the trustee of Trust M considers the amounts necessary for Individual B's health and medical needs, even though access to these amounts may be delayed until after Individual B's death. Absent the occurrence of this contingency, the death of Individual B affects the timing rather than the availability of their benefits. Thus, the entitlement of University Q, School R and Reservation S is not contingent on the death of Individual B. As a result, these beneficiaries are designated beneficiaries for purposes of applying the rules in section 1.401(a)(9)-1, Q&A E-5 and Individual B is not treated as the sole beneficiary. Pursuant to section 1.401(a)(9)-1 Q&A D-2A(b) of the proposed regulations Individual A is treated as having no designated beneficiary for purposes of section 401(a)(9) of the Code, since persons other than individuals are designated as beneficiaries of Individual A's account.[237] [Emphasis added.]
There is an important exception to the rule that the beneficiary with the shortest life expectancy is used as the measuring life. The exception applies where the benefit is held in a separate account or segregated share.[238] In that case, the single beneficiary of each separate account or segregated share may be used as the measuring life with respect to that share. This avoids application of the otherwise applicable rule that the beneficiary with the shortest life expectancy must be used as the measuring life.[239]
The proposed regulations provide rules under which a separate account will qualify. Generally, an account under a defined contribution plan or IRA must bear its own prorata share of gains and losses and otherwise be separately accounted for,[240]