In Focardi v. Commissioner,1 the Tax Court has once again shown a taxpayer the door in the case of a grantor retained annuity trust (GRAT) with a revocable annuity payable to the grantor and the grantor's spouse2 (a two-life annuity). The Tax Court's opinion makes it clear that the taxpayer lost; what's less clear is why. A look at the Tax Court's previous cases on two-life annuities helps to explain the confusion. And a look at the legislative history of Internal Revenue Code Section 2702 shows why the U.S. Court of Appeals for the Eleventh Circuit should reverse the Tax Court in Focardi.
The Tax Court started veering off track in its July 2000 opinion in Cook v. Comm'r.3 In this case, the court addressed a two-life annuity similar to the annuity in Focardi. Each trust in Cook provided that its grantor was to receive an annuity for a specified term of years or until the grantor's earlier death and, if the grantor died before the end of the term and the spouse survived, the spouse would receive the annuity payments that would have been paid to the grantor if the grantor had lived until the end of the term. In each case, the trust's grantor retained the right to revoke the interest payable to the spouse. At the end of the specified term, the remaining trust property was to be paid to a trust for the grantor's son.
At the time of the creation of the trusts involved in Cook, the Treasury Regulations provided that the “retention of a power to revoke a qualified annuity interest (or unitrust interest) of the transferor's spouse is treated as the retention of a qualified annuity interest (or unitrust interest).”4 As a result, if the interest payable to the spouse in Cook was a qualified interest, it would be treated as having been retained by the grantor and the value of that interest could be used to reduce the value of each transferor's taxable gift.5
The Tax Court in Cook held that the interest payable to the spouse was not a qualified interest. The court rested its decision on two alternative grounds. First, citing Treasury Regulations Section 25.2702-3(d)(3)6 as authority, the court found that the spousal interest was not qualified because it was contingent upon the spouse surviving the grantor.7 The court analogized the Cook spousal interest to the interest described in Treas. Regs. Section 25.2702-3(e), Example 5, because the spouse in Cook would not receive an annuity payment unless the grantor died before the end of the specified term of the trust and the spouse survived the grantor. Treas. Regs. Section 25.2702-3(e), Example 5, took the position that annuity payments to be made after the holder of the annuity dies were not included as part of the grantor's retained qualified interest, presumably because the holder would have to die before the end of the term for his estate to have the right to receive annuity payments.
Second, the court concluded that the spousal interest created the possibility that the retained annuity could extend beyond the grantor's life,8 a result the court believed was precluded by Treas. Regs. Section 25.2702-3(d)(3). The court distinguished the Cook spousal interest from the spousal interest described in former Treas. Regs. Section 25.2702-2(d)(1), Example 7, because the spouse in Example 7 would receive the annuity if he were living at the end of the term regardless of the timing of the grantor's death.
Five months after deciding Cook, the Tax Court handed down Walton v. Comm'r, in which it invalidated Treas. Regs. Section 25.2702-3(e), Example 59 and read Section 25.2702-3(d)(3) to include as a qualified interest an annuity payable over a term of years — even though that term might extend beyond the grantor's lifetime. Thus, after Walton, the Tax Court's decision in Cook rested upon that court's analysis of Example 7. And that's where the trouble began.
Schott v. Comm'r came next. The Tax Court issued Schott almost five months after Walton. By that time, it was bound by the precedents it had set in Walton and Cook, which likely gave Judge Mary Ann Cohen some reason for pause in deciding Schott. The issue in Schott had been addressed by the Tax Court in Cook, but the court's opinion in Walton had invalidated a regulation that served as the stronger of the two alternative grounds supporting the Cook opinion. Thus, Judge Cohen was left in the unenviable position of explaining how the spouse's interest in Example 7 is for a fixed term of 10 years.
To reach that conclusion, Judge Cohen found the regulation's “if living” language to mean that “if the spouse is living at the end of the grantor's 10-year term, annuity payments shall be payable to the spouse, but, if the spouse is not living at the end of the grantor's 10-year term, the spouse's 10-year term interest is payable to the estate of the spouse.”12 The court thus held that the spouse's interest in Example 7 was a fixed term annuity for 10 years. In effect, the “if living” language was read out of the text.
In reversing the Tax Court's decision in Schott, the U.S. Court of Appeals for the Ninth Circuit didn't bother addressing every conceivable reading of the regulations. Instead, the appeals court got to the point: As applied to the two-life annuity in Schott, the commissioner's reading of the regulations was invalid.13
Now in Focardi, the Tax Court has concluded once again — and despite the Ninth Circuit's decision to the contrary in Schott — that spousal interests are not qualified interests if “the interests by their terms are payable only if the grantor predeceases the spouse during the applicable term.” The court went on to mention vigor, force, substantial deference, valuation abuse, and a “marital contingency”14 — but none of these phrases help to clarify the issue.
IRC Section 2702 tries to measure what will be paid out of a GRAT. The Tax Court's fixation on what will be paid to the spouse is misplaced. If the spouse in Focardi is paid no portion of the unrevoked15 annuity, one of two things has happened: (1) The grantor and the spouse have died during the term, the probability of which is exactly what the two-life annuity tables are designed to calculate in measuring the retained interest and thus the gift; or (2) the grantor has survived the term, which means that the grantor's taxable gift resulting from the creation of the trust was in fact too large, because the two-life annuity calculation reduced the value of the annuity by the likelihood of death during the GRAT term (and that risk never materialized).
Congress defined a qualified annuity interest as “any interest which consists of the right to receive fixed amounts payable not less frequently than annually.”16 Congress did not use descriptions such as fixed-term, one-life or two-life annuity. The legislative history suggests, however, that Congress did intend that the qualified interests under Section 2702 be “similar to those permitted in charitable split interest trusts under Section 664.”17 Under IRC Section 664, the full values of fixed-term annuities, one-life annuities and two-life annuities are all respected.
So how did the Tax Court address this legislative history in its GRAT cases? In Walton, the court found the legislative reference to Section 664 persuasive. In Cook and Focardi, there was no mention of Section 664. In Schott, the court acknowledges that qualified interests under Section 2702 are similar to those permitted under Section 664 and that Section 664 allows a two-life annuity; it then states that if Congress intended to include a two-life annuity under Section 2702, Congress “could have” included an express reference to a two-life annuity in that particular section. That view overlooks the obvious: The language of Section 2702 does not refer to any particular type of annuity, so the failure to refer to a two-life annuity means nothing.
The brevity of the statutory standards of Section 2702 has led to extensive regulations, but Congress' reference to Section 664 in the legislative history must be taken into account in deciding how regulatory authority can be used to fill the gaps. The legislative history states how the statute is to be interpreted when applied to retained annuities. The requirements for qualified interests under Section 2702 are intended to be similar to the retained interests permitted under Section 664. The only logical conclusion from this congressional direction is that, in the absence of a specific and compelling ground for an exception to this parallel treatment, two-life annuities should be respected under Section 2702 just as they are under Section 664.
In February 2005, Treasury published T.D. 918118 and amended the final regulations to address a spousal interest like the one in Focardi. The final regulations add a new paragraph (2) to Treas. Regs. Section 25.2702-3(d). It provides that an interest will not be a qualified interest if it's subject to any contingency other than “the survival of the holder until the commencement, or throughout the term, of that holder's interest” or, in the case of a spousal interest, the possibility of a revocation by the grantor. As a result, an annuity that will be paid only if the grantor dies before the end of a GRAT's original term will not be a qualified interest. Treas. Regs. Section 25.2702-2(d)(1), Example 9, illustrates the new rule.
But the new rule and its example are inconsistent with the legislative history and are arguably invalid to the extent they conclude that a Focardi-type spousal interest is not a qualified interest. The Treasury's regulatory scheme should not be upheld to the extent it contradicts Congress' direction that qualified interests “are similar to those permitted in charitable split interest trusts under section 664.”
The preamble to the amended final regulations provides: “Just as the contingent reversion in Example 1 would increase the value of the retained interest without any certainty that the reversion would ever be paid, a revocable spousal interest of the kind at issue in Schott would similarly increase the value of the retained interest without any certainty that the spouse's survivorship interest would ever be paid.” This quote illustrates the source of confusion on the spousal annuity issue.
When it enacted IRC Section 2702, Congress prohibited the common practice of creating retained interests that were greater than a fixed-term annuity. Prior to the statute's enactment, taxpayers could accomplish this result by retaining a contingent interest most often referred to as a contingent reversion.19 For taxpayers, the attraction of a retained contingent reversion in estate planning was that it had actuarial value when the trust was created, and thus reduced the size of the gift, but that value would decrease over time. As a result, without any payment from the trust, the value of the contingent reversion would evaporate altogether if the grantor survived the annuity term,20 because the value of the contingent reversion was dependent on the grantor dying during the term.21
Section 2702 was intended to prevent taxpayers from reducing the value of gifts in trust by interests that would not be paid unless the grantor died during a specified term. Such interests have economic value when a trust is created because there is some probability that the grantor will die during the specified term, thereby triggering the payment. Yet as time passes, the probability that the grantor will die during the term decreases. As the probability of the grantor's death during the term approaches zero, so does the value of the retained interest since the event necessary to trigger the payment, the grantor's death during the term, becomes less likely. Eventually, once the grantor survives the term, the likelihood drops to zero. This type of interest was not intended to be a qualified interest under Section 2702, because its value disappears over time.22
In contrast, an annuity that is payable for the shorter of a specified term of years or until the prior deaths of both the grantor and the spouse is a qualified interest because it represents value that is being paid out of the GRAT. As each year passes during the term of years, the value of the “right to receive” annuity payments gets smaller — but only because the annuity payments actually are being paid out of the GRAT, either to the grantor or the spouse. Again, if the spouse is paid no portion of an unrevoked annuity, one of two things has happened: (1) The grantor and the spouse have died during the term, the probability of which is exactly what the two-life annuity tables are designed to calculate; or (2) the grantor has survived the term. In the latter case, the grantor's taxable gift caused by the creation of the GRAT was in fact too large, because the two-life annuity calculation reduced the value of the annuity by the likelihood of death during the GRAT term (and that risk never materialized). The commissioner has improperly characterized the spousal interest in a two-life annuity as similar to the contingent reversion.
Use of a retained two-life annuity to increase the value of a retained interest (as compared to one-life annuity) has little appeal after the Walton decision. Walton's treating a term-of-years annuity as a qualified interest in its entirety, despite the possibility that some payments could be made after the death of the grantor, allows taxpayers to retain an interest with a value even greater than the value of a two-life annuity. This is little consolation, however, for the grantors in Focardi who created GRATs with an annuity payable to the grantor and the grantor's spouse.23 The U.S. Court of Appeals for the Eleventh Circuit should follow the Ninth Circuit's lead and reverse the Tax Court in Focardi.
- T.C. Memo. 2006-56.
- In each of the cases outlined in this article, the spousal interest was revocable by the grantor.
- Cook v. Commissioner, 115 T.C. 15 (2000), aff'd, 269 F.3d 854 (7th Cir. 2001).
- Former Treasury Regulations Section 25.2702-2(a)(5).
- Internal Revenue Code Section 2702 requires, with few exceptions, that the value of a retained interest in a trust is to be treated as having a zero value for purposes of determining the amount of a grantor's transfer to a trust if the trust is for the benefit of certain family members unless the retained interest is in the form of a qualified interest as described in Section 2702(b).
- In the current version of the Treasury regulations, this regulation appears as Treas. Reg. Section 25.2702-3(d)(4).
- Cook, 115 T.C. at 23-24.
- Cook, 115 T.C. at 24-25.
- Walton v. Comm'r, 115 T.C. 589 (2000) at 604.
- See generally, Gerald A. Kafka and Rita A. Cavanagh, Litigation of Federal Civil Tax Controversies, 2d ed., at par. 2.06 (Warren, Gorham & Lamont, Boston, 1997).
- T.C. Memo. 2001-110.
- Schott, T.C. Memo. 2001-110, Slip op. at 13.
- Schott v. Comm'r, 319 F.3d 1203 (9th Cir. 2003).
- The Tax Court construed the terms of the trusts in Focardi to require not only that the grantor had to die during the term survived by his or her spouse in order for the spouse to receive annuity payments from the trust, but also that he or she be married to the grantor at the time his or her interest commenced. This is a contingency not contemplated by the Code or the regulations, and the existence of such a contingency would make it impossible to place a value on the spousal interest. This line of reasoning arguably provided an alternate basis for the Tax Court's decision.
- Revocation would be treated as a taxable gift. Treas. Regs. Section 25.2702-3(e), Example 8.
- IRC Section 2702(b)(1).
- 136 Congressional Record 30,540, n.30; see also Informal Senate Report on S.3209, 136 Congressional Record S15629, at S15682, n.30 (Oct. 18, 1990) (“These interests are similar to those permitted in charitable split interest trusts under section 664.”)
- 70 Federal Register 9222 (Feb. 25, 2005).
- Usually, the term “reversion” is used to refer to a future interest retained by a grantor in a trust and the term “remainder” is used to refer to a future interest created for some other person. Black's Law Dictionary 1292, 1320 (6th ed. 1990). The statute and regulations use the terms interchangeably. 26 U.S.C. 2702(b)(3); Treas. Regs. Section 25.2702-3(f)(2).
- See Mark R. Siegel, “Retained Possession And Enjoyment: Searching Out The Reality For Residential Transfers,” 24 Sw. U. L. Rev. 81, 112 n.244 (1994) (the value of a fixed-term interest could be increased with a contingent reversion). See also Mitchell M. Gans, “GRIT's, GRAT's AND GRUT's: Planning And Policy,” 11 Va. Tax Rev. 761, 803-804 (Spring 1992) (the contingent reversionary interest would “evaporate” in the event the grantor survived the annuity-retention period). Accord Department of Treasury, General Explanations of the Administration's Revenue Proposals, (“the Green Book“) at 129 (February 1998) (family limited partnerships and similar devices should not be permitted to make value disappear because “disappearing value is illusory”).
- For example, the right to collect $1 million if a specified person died within the next 10 days is a valuable right. Nonetheless, the value of that right goes to zero if the specified person survives the 10-day period.
- This principle justifies the core concept of the duration rule in Treas. Regs. Section 25.2702-3(d) that permits a term “for the life of the term holder, for a specified term of years, or for the shorter (but not the longer) of those periods” but it does not justify the conclusion in Focardi or the formulation in the new regulations whereby each spouse's interest in a two-life annuity must separately qualify under Section 2702 as if the two-life annuity was in fact two separate annuities.
- This is particularly true for the grantors in Focardi, because they acted in conformity with the commissioner's construction and application of the statute and regulations as articulated in a number of letter rulings issued prior to 1997. See Private Letter Rulings 9451056 (Dec. 23, 1994); 9449013 (Dec. 9, 1994); 9449012 (Dec. 9, 1994); and 9416009 (Dec. 30, 1993).
Captivating Contemporaries: Gerhard Richter's “Tante Marianne” is a self-portrait of the artist as a baby on the knees of his aunt. It sold at Sotheby's “Contemporary Art Auction” in London on June 21 for $3.942 million.