Take NoteTake Note

Pre-age 59 1/2 series may use annual recalculation.

One of the 12 exceptions to the 10 percent penalty on plan distributions prior to age 59 1/2 (also called “premature distributions”) is for a payment that is “part of a series of substantially equal periodic payments” (SOSEPP) made for the life (or life expectancy) of the employee or the joint lives (or joint life expectancies) of such employee and his designated beneficiary.¹ In Notice 89-25,² the Internal Revenue Service provided three permissible methods for calculating the payments in a SOSEPP: (1) the minimum distributions method (comparable to computing lifetime minimum required distributions, or MRDs); (2) the amortization method (comparable to a level-payment self-amortizing mortgage); (3) and the annuitization method (comparable to converting the account to a fixed-payment annuity for the applicable period).

Although the amortization and annuitization methods both call for fixed payments for the life of the series, the IRS, in several private letter rulings, approved variations of these two methods in which the account balance, life expectancy and/or interest rate were redetermined annually, with the effect that the payments under the series varied from year to year.³

Then, in Revenue Ruling 2002-62,⁴ the IRS updated the three permissible methods and promulgated new rules for SOSEPPs, superceding the Notice 89-25 rules. Revenue Ruling 2002-62 stated that, “under the amortization methods, the account balance,…[the other factors] and the resulting annual payment are determined once for the first distribution year and the annual payment is the same amount in each succeeding year.” As such, the ruling appeared to preclude the annual revaluations the IRS had permitted in earlier letter rulings.

Now, in three 2004 PLRs issued to four individuals, the IRS has made clear that annual recalculation is alive and well, even after Revenue Ruling 2002-62.

In PLR 2004-32021, the IRS approved two SOSEPPs using the amortization method — one for a husband; one for his wife. (They obtained a joint ruling, thereby saving one “user fee”). Each spouse's SOSEPP used that spouse's life expectancy (based on the IRS single life table), an interest rate of 120 percent of the federal mid-term rate (the highest rate allowed under Revenue Ruling 2002-62), and the prior year-end account balance. The spouses proposed to recalculate their “fixed” amortization payments annually, using each year's then-current life expectancy, prior year-end account balance, and 120 percent-of-federal-mid-term-rate interest rate. An identical SOSEPP design was approved for another taxpayer in PLR 2004-32024.

In PLR 2004-32023, the IRS approved an annuitization-method SOSEPP that provided for annual recalculation of the payments in the same manner. Each year's payment would be calculated using the account balance and federal mid-term rate as of Dec. 31 of the year prior to the distribution year. The annuitization factor would be based on the participant's age in the distribution year.

The key to the IRS's approval in these rulings is that, even though the sizes of the payments will vary, the payment is determined exactly the same way each year. Each year's payment size is determined using the previous year's year-end balance, an interest rate of 120 percent of the federal mid-term rate as of Dec. 31 of the prior year, and the life expectancy of the participant in that distribution year.

The flexibility offered by annual recalculation is a welcome addition to the SOSEPP menu. As a reminder, to qualify for the SOSEPP exception, all elements of the SOSEPP design must be fixed at the outset, then followed religiously. Thus, someone who has already launched a fixed-payment SOSEPP using the amortization or annuitization method cannot change over to the annual recalculation method of determining his payments. Annual recalculation is available only if it is part of the SOSEPP design from the outset.

Should people use annual-recalculation SOSEPPs modeled on these three PLRs without getting their own PLRs? That's not clear. The IRS has stated as a matter of general information that any variation not explicitly blessed in Revenue Ruling 2002-62 would require an advance ruling. On the other hand, it is hard to see how the IRS could argue that a SOSEPP that follows a method explicitly blessed in one of these PLRs does not qualify for the penalty exception.

The IRS rejected another taxpayer's proposed SOSEPP design. Using his current account balance and his life expectancy from the applicable IRS table, the participant in PLR 2004-37038 proposed to determine the amount of his series payments by pretending that he was buying zero-coupon U.S. Treasury securities (“zeroes”) maturing at regular intervals over his life expectancy. Based on then-current market values, his account balance was just sufficient to purchase sequentially maturing zeroes that would provide him equal payments of $59,500 a year. The IRS denied approval on the grounds that the interest rate used to determined the series payment amounts (that is to say, the yield-to-maturity of the hypothetical zeroes) would be set by the bond market, and “may very well exceed 120 percent of the Federal mid-term rate,” which is the maximum rate permitted under Revenue Ruling 2002-62.

What if this fellow had already started taking “series” payments under his clever SOSEPP design before the IRS rejected it? No problem; he should seek IRS permission to roll those payments back into an IRA tax free, even if the 60-day deadline has passed. In PLR 2004-42033, the IRS granted a waiver of the 60-day rollover deadline to a taxpayer who had commenced a SOSEPP using an unapproved method, so he could unwind the defective SOSEPP, and put the payments he had already taken back into his IRA, late and tax free.

The next question is whether someone will now seek a PLR approving a SOSEPP with an annual cost-of-living adjustment, similar to those approved prior to Revenue Ruling 2002-62 in PLRs 9816028, 9747045, 9723035 and 9536031.

See-through trusts: The IRS uses the “snapshot” theory.

Generally, the slowest rate at which retirement benefits may be distributed, upon the death of the plan owner or “participant,” is the annual installments over the life expectancy of the individual who is named as beneficiary of the plan. With life expectancies under the IRS Single Life Table ranging as high as 80 years, this “life expectancy of the beneficiary” payout scheme can offer considerable continued in-come tax deferral after the death of the participant. Though this favorable treatment is generally available only to individual beneficiaries, a trust named as beneficiary also can qualify for it (based on the life expectancy of the oldest trust beneficiary), if the trust meets the requirements for a “see-through trust” under the IRS minimum distribution regulations.⁵ One of these requirements is that all trust beneficiaries must be individuals.

A question that has bedeviled practitioners through all three versions of the IRS's minimum distribution regulations is: Which trust beneficiaries, if any, can be ignored in applying the test that “all trust beneficiaries must be individuals” (and the related test that the oldest trust beneficiary's life expectancy becomes the applicable distribution period)? The regulations are, to put it mildly, unclear on this point.

Various alternative approaches have been suggested, one of which has been nicknamed the “snapshot” approach. Under the snapshot approach, the trust would be tested by assuming that all current temporary trust beneficiaries (for example, a life income beneficiary) died immediately after the trust was established, so that the “trust beneficiaries” (for purposes of applying the see-through tests) would be (1) the life beneficiary and (2) any remainder beneficiary(ies) who would take the trust outright if the life beneficiary died immediately after the participant. If the remainder beneficiaries were to take the trust outright upon the death of the life beneficiary, it would not be necessary (under this snapshot approach) to investigate further (for example, by counting beneficiaries who would take the trust property if the currently living remainder beneficiaries-apparent predeceased the life beneficiary).

In PLR 2004-38044, the IRS for the first time explicitly blessed such a “snapshot” approach in approving a trust as a see-through. In this PLR, Taxpayer A died, leaving his IRA payable to a trust. The trust benefited the participant's spouse, Taxpayer B, for her life. Upon her death the principal would be divided among the participant's “lineal descendants then living, per stirpes,” with each descendant's share held in trust for him until he had attained age 30.⁶ At the time of the participant's death, he had three living children, Taxpayers C through E, and apparently no deceased children. The three children had already attained age 30 at the time of the participant's death. Thus, if the spouse, Taxpayer B, had died immediately after the trust's establishment, the three children, Taxpayers C through E, would have taken the trust principal (including the remaining retirement benefits) outright.

The ruling found that, as Taxpayer B's interest in the trust was “not unlimited” (she was limited to a life income interest, plus principal in the trustee's discretion), it was “necessary to determine which other beneficiaries of Trust Y must be considered in determining who, if anyone, may be treated as Taxpayer A's designated beneficiary Y.” This statement illustrates that, if a trust beneficiary is not entitled to outright distribution of the entire trust, we must keep looking; we also must count as beneficiaries (for purposes of applying the tests in the IRS's minimum distribution trust rules) the beneficiary(ies) who will take the trust when this beneficiary's interest ends.

The ruling goes on to say, however, that we can stop our search once we reach the children who are the apparent remainder beneficiaries. Because they will take their shares outright when the prior beneficiary's interest ends, we do not need to go further and find out who would take the benefits if any of these three children predecease the surviving spouse. From the ruling: “Since the right of each child to his/her remainder interest in the… Y [trust] was unrestricted at the death of Taxpayer A, it is necessary to consider only Taxpayers B through E (that is to say, the spouse and the three children) to determine which of them shall be treated as the designated beneficiary of Taxpayer A's interest in” the IRA.⁷

Although a private letter ruling cannot be relied upon as precedent, the principle enunciated here will be of great help in drafting see-through trusts.

Assigning an IRA out of an estate or a trust.

PLRs 2004-32027B2004-32029 dealt with an IRA that was payable to a trust that, by its terms, was to terminate immediately upon the participant's death, with its assets being distributed in equal outright shares to his three children. The children sought and received rulings that, when an IRA is payable to a trust, and the trust terminates, the IRA can, following such termination, pay benefits directly to the children-beneficiaries. This ruling is no surprise, as it is supported by logic, good sense, the law and several prior, favorable rulings confirming this conclusion not only for a terminating trust, but also for a terminating estate.