Quatloos! > Investment Fraud > Financial Planning > Private Annuity Exhaustion Test

Private Annuity Sales
and the Exhaustion Test

Notice Regarding Proposed Changes - Some Private Annuity Transactions Restricted, But Many Transactions Remain Advantageous Proposed Tax Changes regulations that provide guidance on the taxation of the exchange of property for an annuity contract. Stokes v. Commissioner In a case involving the National Association of Financial and Estate Planners, the court held that the private annuity trust in that particular case "constituted a sham trust that lacked economic substance". Not only did Mr. Stokes not receive the promised tax benefits, but he also got slammed with an accuracy-related penalty!

by Kevin McGrath, Esq.
The Saylor Law Firm of Atlanta, GA

Reprinted with Permission of the Authorh

[Editor’s Note: As one of the leading authorities on private annuities, Mr. McGrath will further discuss these issues at his upcoming presentation, “A Holistic Approach to Analysis of Private Annuity Trusts to Defer Capital Gains, Including Exit Strategies”, to be presented at 3:45p on Thursday, January 11, 2007, at the Heckerling Institute, http://www.law.miami.edu/heckerling/ which is the nation’s leading and most prestigious conference regarding estate planning issues. The following article was first printed by BNA after extensive review by their editorial staff for accuracy.]

INTRODUCTION

The exhaustion test presents a significant obstacle to successful implementation of a private annuity sale to an irrevocable trust. Ignoring the exhaustion test can result in gift and income tax, but meeting the exhaustion test is extremely difficult. Arguments against the validity of the exhaustion test may be difficult to make, so structuring transactions to avoid application of the exhaustion test may be the best option.

THE EXHAUSTION TEST

The regulations contain two limitations on the use of the §75201 annuity tables that may apply to private annuity sales to trusts.2 One is the so-called "exhaustion test," which states:

A standard section 7520 annuity factor may not be used to determine the present value of an annuity for a specified term of years or the life of one or more individuals unless the effect of the trust, will, or other governing instrument is to ensure that the annuity will be paid for the entire defined period. In the case of an annuity payable from a trust or other limited fund, the annuity is not considered payable for the entire defined period if, considering the applicable section 7520 interest rate on the valuation date of the transfer, the annuity is expected to exhaust the fund before the last possible annuity payment is made in full. For this purpose, it must be assumed that it is possible for each measuring life to survive until age 110. For example, for a fixed annuity payable annually at the end of each year, if the amount of the annuity payment (expressed as a percentage of the initial corpus) is less than or equal to the applicable section 7520 interest rate at the date of the transfer, the corpus is assumed to be sufficient to make all payments. If the percentage exceeds the applicable section 7520 interest rate and the annuity is for a definite term of years, multiply the annual annuity amount by the Table B term certain annuity factor, as described in Section 25.7520-1(c)(1), for the number of years of the defined period. If the percentage exceeds the applicable section 7520 interest rate and the annuity is payable for the life of one or more individuals, multiply the annual annuity amount by the Table B annuity factor for 110 years minus the age of the youngest individual. If the result exceeds the limited fund, the annuity may exhaust the fund, and it will be necessary to calculate a special section 7520 annuity factor that takes into account the exhaustion of the trust or fund. This computation would be modified, if appropriate, to take into account annuities with different payment terms.3

Regs. §25.7520-3(b)(2)(v), Ex. 5 ("Example 5"), explains how to calculate the special §7520 factor. If a trust will exhaust prior to the annuitant's attainment of the age of 110, then Example 5 requires that the annuity must be valued as an annuity payable for a term of years or until the prior death of the annuitant, with the term of years determined by when the annuity payments will exhaust the fund. Example 5 is reprinted in the Appendix.

The other limitation on the use of §7520 is contained in Regs. §25.7520-3(b)(1)(ii). It states that:

A restricted beneficial interest is an annuity, income, remainder, or reversionary interest that is subject to any contingency, power, or other restriction, whether the restriction is provided for by the terms of the trust, will, or other governing instrument or is caused by other circumstances. In general, a standard §7520 annuity, income, or remainder factor may not be used to value a restricted beneficial interest.

This is referred to herein as the "governing instrument" requirement. The governing instrument requirement is not analyzed herein.

What Is at Stake?

Failing the exhaustion test means that the annuity payable to the seller will be worth less than desired, potentially resulting in gift and income tax.

Example: A 65-year old desires to sell an asset worth $10,372,700 to an irrevocable trust4 in exchange for a private annuity at a time when the §7520 rate is 5.2%. Ignoring the exhaustion test, a $1 million life annuity will equal $10,372,700.5 Failing the exhaustion test means a gift will necessarily result. For example, assume that the irrevocable trust is funded with a seed gift of $1 million. Under Example 5, the trust will be assumed to run out of funds when the annuitant is 83.6

Additionally, if the trust is relying on Rev Rul 55-1197 to get basis in the purchased asset, then its basis in the asset will be only $9,276,307,8 so that upon its subsequent sale of the asset, some capital gain (likely short-term) will result.

Historical Background

Rev. Rul. 77-454

The starting point for the discussion of the current exhaustion test is Rev. Rul. 77-454.9 Rev. Rul. 77-454 interpreted former Regs. §25.2512-9, which laid out valuation tables that were then applicable to annuities, terms for years, remainders and reversions. In Rev. Rul. 77-454, the IRS claimed that annuity factors set forth in Regs. §25.2512-9 are based upon the assumption that payments can actually be received for each year even if the annuitant survives to age 109.

The IRS offered this example to support its point:

For example, an annuity payable only from a fund of $200 cannot be valued by using a factor from Table A(1) if an annuitant aged 51 years is to receive annual payments of $100 for life. In such a situation, the annuitant's right to payments is not worth $1,113 (the value produced through the use of a factor of $11.1308 from Table A(1)). On the contrary, it is worth something less than $200, due to the possibility that the annuitant may die before the fund is exhausted.

The IRS then concluded that an annuity from a fund that will exhaust prior to age 109 is to be valued as annuity for a term of years or the death of an individual, whichever occurs first. The term of years equals the number of years until the trust will exhaust.

From this revenue ruling comes the truism that an annuity cannot be worth more than the fund from which it will be paid. Economically, this is an irrefutable point. In other words, the IRS is correct that the annuity in the above example is not worth $1,113, because it is payable out of a $200 fund. The annuity cannot be worth more than $200. Second, the IRS claims that the annuity cannot be worth $200, because there is a risk that the 51 year old will die prior to receiving the entire corpus of the trust back. Again, economically, this is difficult to refute. At best, the annuitant will receive $200 (in today's value), and at worst, he will receive nothing. Therefore, the annuity has to be worth less than $200.

Third, the revenue ruling establishes a method for valuing the annuity. The IRS's method is to treat the annuity as a shorter of term or life annuity, with a term equal to the length of time until the trust runs out of money, determined by assuming that the trust grows at an assumed growth rate (then 6%).

Fourth, the IRS set forth the circumstances in which annuities must be valued under this method: if the trust does not have sufficient funding to pay the annuity through age 109, then this special valuation method must be used. This is probably the most controversial point in the ruling.

The assumption that the annuitant must receive the annuity through age 109 does not in any way follow from the $200 example. The IRS used an example in which the tables value an annuity at an amount well in excess of the fund from which the annuity would be paid to not only prove an easy point (that naked application of the tables is inappropriate in that circumstance), but to prove a different, unrelated point altogether: if the tables value an annuity at an amount LESS than the fund from which the annuity would be paid (so that the $200 example above is NOT on point), the tables are nevertheless inapplicable if there is a risk that the annuitant does not receive the benefit of the bargain inherent in an annuity contract (i.e., payments under the contract in the event the annuitant lives past his life expectancy).

While the age 109 assumption has little to do with the $200 example, it does in fact follow from the IRS's premise that the tables are based upon the assumption that payments can actually be received for each year even if the annuitant survives to age 109. Instead of illustrating the exhaustion test through the $200 limited fund example, the IRS should have made its point by illustrating that the value of an annuity is simply equal to the present value of each year's annuity payment multiplied by the chances of the annuitant living to receive each payment.10 Such an illustration would show that a portion of the assumed value of any annuity lies in the value of the potential payments beyond the annuitant's life expectancy.

This IRS's exhaustion argument was tested in Estate of Shapiro v. Comr.11

Estate of Shapiro

In Shapiro, the taxpayer argued that the tables under Regs. §20.2031-7(f) applied in valuing a $350,000 annuity payable to the 91-year old decedent out of a trust fund of $1,005,000. Application of the regulations resulted in the life annuity having a value of $943,425, and it was on this value that the decedent's estate calculated a credit for tax on prior transfers under §2013.

The IRS argued that, due to the trust's limited funding, the annuity should be valued as a term or life annuity. Although Rev. Rul. 77-454 was not cited by the Tax Court, the valuation method urged by the IRS in Shapiro was identical to the method prescribed by Rev. Rul. 77-454. Under the IRS's approach, the annuity had a value of only $680,896, thus reducing the tax on prior transfers credit. The IRS argued that the annuity should not be valued as a life annuity, but should be valued as an annuity for a term certain concurrent with one life because (1) the annuity was payable out of a depleting corpus, and (2) the trust was "underfunded" in that the corpus was not large enough to support the annuity obligation in case decedent had lived to be age 109.

The Tax Court addressed both arguments as if made in the alternative, but the second argument (that the corpus will fully deplete prior to age 109) is the real issue -- the first argument (that the annuity is payable from a depleting corpus) has no consequence unless the second argument (that the corpus will fully deplete prior to age 109) is true; and the second argument (that the corpus will fully deplete prior to age 109) cannot be true unless the first argument is also true (that the annuity is payable from a depleting corpus).12

The Tax Court sided in favor of the taxpayer, making several observations:

• "The actuarial tables, such as Table A of section 20.2031-7(f), Estate Tax Regs., are an administrative convenience in that they provide a `bright line' approach to valuation making it unnecessary to hypothesize as to the facts and circumstances surrounding each case."

• "It is well established that the actuarial tables generally are to be respected unless the established facts show that the result under the tables is unrealistic or unreasonable, or that the result ignores common sense."

• "Respondent confuses a TERM CERTAIN with a property interest that is TERMINABLE. In most cases, it is impossible to immediately ascertain the time at which a trust fund will exhaust with the certainty and specificity sufficient to equate the duration of the trust to a term certain. Although decedent's interest in the trust may have been TERMINABLE by virtue of the possibility of corpus depletion, the bequest of his interest was not definitively limited to a TERM CERTAIN. If respondent, in drafting this regulation, intended the meaning of term certain to comprise that period of time after which a property interest might be extinguished by the possible exhaustion of a trust fund, she should have set forth a special definition of `term certain' in the regulation. Because she did not, we are reluctant to transform `term certain' into a term of art in this case."

• "The fact that respondent, in section 20.2031-7, Estate Tax Regs., did not specify `terminable' interests, when she could have, further convinces us that a TERM CERTAIN is to be distinguished from an interest that is merely TERMINABLE DUE TO POSSIBLE DEPLETION OF THE CORPUS."

• With regard to the age 109 requirement, the Tax Court stated: "Taking respondent's argument to its theoretical conclusion, ANY TRUST created with corpus funds equivalent to the present value of a lifetime annuity obligation as computed under Table A would be deemed to be `underfunded' in that it would have insufficient funds to sustain the annual payments should the annuitant live beyond his or her average life expectancy. In this regard, respondent's argument contravenes the fundamental purposes and presumptions underlying the actuarial tables... The fair market value of an obligation to make future payments is deemed to be a sufficient amount to meet the entire obligation. Thus, by deduction, one of the fundamental presumptions underlying Table A is that the PRESENT VALUE of a lifetime annuity obligation is sufficient to sustain the stream of annuity payments as they come due over the course of an annuitant's expected life span."

• "A person, at any given age, is assumed to die within a time consistent with the average mortality rate for that age. Section 20.2031-7(f), Estate Tax Regs. Table A does not expect or presume that a 91-year-old person will live for 18 more years; to the contrary, the table implicitly recognizes the possibility of any person reaching 109 years of age is extremely remote."

• "In essence, respondent argues that unless an annuity is `guaranteed' throughout an annuitant's extreme life expectancy, just as a commercial annuity is guaranteed, the computation of the annuity's present value must be made on a case-by-case basis using a special actuarial factor supplied by the Internal Revenue Service; any computation of an unguaranteed, private annuity under Table A would be deemed invalid in respondent's view. Respondent's position, if it were correct, would vitiate the use of Table A as an administrative convenience and bright-line approach to valuation. Table A could not be used to determine the present values of all sorts of unguaranteed, privately funded annuities; they would all be subject to individual analysis under a facts-and-circumstances test."

The Tax Court in Shapiro not once, but twice, suggested that the Commissioner could promulgate its argument as a regulation. Not long thereafter, amendments to the §7520 regulations were proposed,13 codifying the exhaustion test of Rev. Rul. 77-454. In the preamble to the regulations, the IRS noted that it was rejecting the ruling of Shapiro.

Final Regulations

When Treasury finalized the exhaustion test regulations,14 it rejected comments suggesting that the exhaustion rule should be ignored in the case of less severe under-funding of trusts (apparently, there were no comments to the effect that the IRS had no authority to overrule Shapiro, or if there were, the IRS failed to address them). The IRS stated that the method described in the proposed regulations for determining the value of the annuity is consistent with fundamental principles for determining present value and long-standing IRS position. The IRS cited not only Rev. Rul. 77-454,15 but Rev. Rul. 70-452,16 Moffett v. Comr.,17 and U.S. v. Dean.18

These citations are somewhat disingenuous. While Rev. Rul. 77-454 was in fact the IRS's position, the Tax Court in Shapiro essentially shot it down. Rev. Rul. 70-452 and the Dean case are not remotely on point (they involve the deductibility for estate tax purposes of a contingent gift to charity). In Moffett, the government and the taxpayer stipulated on the valuation of the annuity (apparently employing an exhaustion-type test that benefited the taxpayer), but did so in a case the issue of which was the deductibility of a contingent bequest to charity. Nonetheless, the validity of the regulation is unlikely to hinge on the IRS's sloppy citations.

VALIDITY OF REGULATION

Section 7520(a) states that the value of any annuity shall be determined under tables prescribed by the Secretary. Section 7520(b) provides that §7520 shall not apply for purposes of any provision specified in the regulations. Section 7520(e) provides that the term "tables" includes formulas. In Chevron v. Natural Resources Defense Council19 the Supreme Court stated:

If Congress has explicitly left a gap for the agency to fill, there is an express delegation of authority to the agency to elucidate a specific provision of the statute by regulation. Such legislative regulations are given controlling weight unless they are arbitrary, capricious, or manifestly contrary to the statute.20

Because Congress expressly delegated the responsibility to enact tables to the Secretary, and because the statute has no meaning without the tables, the regulations under §7520 are legislative regulations, and should be upheld unless arbitrary or capricious.

Arguments against the validity of the exhaustion test typically take one or more of three forms: (1) the exhaustion test is not a "table," so it therefore goes beyond the scope of the legislative grant under §7520 (and accordingly should be deemed an interpretive regulation); (2) the regulation fails the arbitrary and capricious standard or the reasonableness standard, whatever is applicable, because the assumption that the annuitant will live to age 110 is arbitrary and capricious; and (3) Shapiro either overturned the regulation or severely undermines it.

The first point, that the exhaustion test is not a table, is a difficult argument. The exhaustion test explains how to apply the tables in certain circumstances (i.e., as a shorter of term or life annuity in the case of an exhausting fund). Surely Congress' delegation to enact tables includes a delegation to explain their use. Some commentators assume, like the Tax Court in Shapiro, that the exhaustion test mandates a facts and circumstances analysis, which is contrary to the intent of the statute -- to offer administrative convenience in the form of tables -- but Example 5 of Regs. §25.7520-3(b)(2)(v) clarifies that the tables are in fact used for an exhausting trust.

More importantly, however, Congress delegated to the Secretary the rule-making authority to except certain situations from the regulations. If Congress intended that the tables would carve out certain situations from their application, it is difficult to see how the middle ground approach taken by the Secretary -- not excluding use of the tables for exhausting funds, but rather applying the tables in a manner the Secretary feels is more akin to their substance -- it not within the delegation of authority under §7520.

Is the Age 110 Requirement Arbitrary and Capricious?

According to the Atlanta Journal-Constitution, there are 18 people in the United States who are age 110 or older.21 The same newspaper claims that an individual has a one in 5,000,000 chance of attaining such age.22 The tables relied upon by the IRS do not even recognize these long odds, as they assume that nobody will live past their 110th birthday.23 So is it arbitrary or capricious for the IRS to force us to assume that ALL taxpayers live that long?

Whether it is arbitrary or capricious probably does not matter: the assumption that an annuitant will attain the age of 110, while seemingly the heart of the exhaustion test, plays very little role in the negative consequences of failing the exhaustion test. In fact, the exhaustion test would be no less a nuisance than it currently is if the age requirement were dropped to 100, and it would be no more of a nuisance if the age requirement were increased to 210.

The reason is that while the exhaustion test must be carried out to the unrealistic age of 110, the consequences of failing the exhaustion test, i.e., the consequences of the trust running out of funds prior to age 110, are properly risk adjusted to take into account the unlikelihood of attaining each remote age.

For example, recall our 65 year old who sold a $10,372,700 asset for $1 million annuity for life. Table 1 illustrates the value of each $1 million annuity payment that the 65 year old may potentially receive. Column 4 represents the present value of each payment, assuming the annuitant lives to receive it. Column 6 represents the actuarial present value of each payment, which is the present value of each annuity payment multiplied by the probability that the annuitant will be alive to receive it.

TABLE 1

If application of the exhaustion test results in the trust running out of money at age 108, then the consequence is that we have to assume that the last two $1,000,000 annuity payments, at ages 109 and 110, will not be made. Column 4 illustrates that the present value of these two payments are low ($107,475 and $102,163 respectively). Column 5 illustrates how remote the chances of a 65-year old attaining the ages of 109 and 110 are (about .03% and .01%, respectively).24 Column 6 shows the value, for gift tax purposes, of failing to receive each of these payments (column 4 multiplied by column 5, or $34 for age 109 and $11 for age 110). Thus, the consequence of assuming that the seller does not receive the last two $1,000,000 annuity payments is only a $45 gift ($34 plus $11).

Column 7 illustrates the gift tax consequences of failing the exhaustion test by year, and shows that failing the exhaustion test on account of the trust running out of money at unrealistic ages has almost no gift tax consequence. For example, if instead of age 108, the trust will deplete at age 100, we must assume the seller will miss out on 10 $1 million payments. The seller's failure to receive the last $10 million results in a $6,540 gift, an amount that appropriately reflects the unlikelihood of attaining ages beyond 100, and amount which nobody will be fighting over. If the trust instead will deplete at age 95, the annuitant's assumed failure to receive $15 million in annuity payments still has a gift tax cost of only $48,307.

So if the consequences of failing the exhaustion test are so minimal, then why do we care? Because people do not just fail the exhaustion test, they fail it miserably. In a typically funded trust (either 7% or 10% seed gift), the trust will be scheduled to deplete just after life expectancy. In the example with the 65 year old, with a 10% seed gift the trust will deplete at age 83; the 65 year old's life expectancy is 82.25

If the trust will deplete at age 83, a huge gift results -- over $1 million. It is the failure to receive the bargained for annuity payments at ages 84, 85, 86, and so on, up to around age 95, that results in the bulk of the gift tax hit, because it is these ages that the grantor stands a reasonable chance of attaining. In fact, the exhaustion test could make us assume that the annuitant live to age 210, and the test would produce the exact same results: the consequences of failing to receive $1 million annuity payments from age 111 through age 210 will have zero gift tax cost; the burden associated with the exhaustion test would still be entirely associated with failing to receive annuity payments at the younger ages.26

Thus, if the exhaustion test is invalid on account of age 110 being "arbitrary and capricious," that is not a good answer because presumably some age (beyond the seller's life expectancy) will not be arbitrary and capricious. For example, if the exhaustion test were changed to an age 95 requirement instead of age 110, the results are nearly as disastrous (the gift would still be close to $1 million in the above example because the most of the gift tax hit (all but $48,307) occurs on account of the failure to receive payments before age 95, not after). While practitioners may view age 110 as being an absurd and onerous requirement, this is not the reason that the exhaustion test yields such terrible results, and arguing for a lower age will not be particularly helpful.

Did Shapiro Strike Down the Exhaustion Test Before It Was Even Enacted?

Many commentators have expressed the view that Shapiro struck down the exhaustion test before it was even enacted. Shapiro was an unpublished opinion, and is accordingly not binding on the Tax Court (and of course, any other court). Further, Shapiro interpreted a regulation (Regs. §20.2031-7(f)) in the face of an IRS argument to ignore the regulation in favor of the exhaustion test (and the court upheld the regulation), whereas today, a court would be interpreting a regulation (Regs. §25.7520-3(b)(2)) in the face of a taxpayer's argument to ignore the regulation. Thus, Shapiro does not constitute precedent of any sort, and even if it did, it is not on point. (And Congress did not ratify Shapiro in enacting §7520 in 1988, as §7520 predated Shapiro, decided in 1993, by five years).

Nevertheless, would a court, if faced with Regs. §25.7520-3(b)(2) today, make the same observations as the Tax Court in Shapiro, and if so, how would those observations hold up?

• Shapiro Court: The actuarial tables, such as Table A of Regs. §20.2031-7(f), are an administrative convenience in that they provide a "bright line" approach to valuation making it unnecessary to hypothesize as to the facts and circumstances surrounding each case.

The problem with this argument is that the exhaustion test isn't really a deviation from the tables at all. Rather, it explains how to apply the tables: an annuity from a fund which will exhaust at a certain time is valued as what it is in substance -- an annuity for a shorter of life or term of years. The term or life annuity is then valued -- in accordance with the "bright line" approach of the tables. Thus, the Shapiro court's fear that a facts and circumstances analysis would undercut the purpose of the tables - administrative convenience -- is not present. Thus, this argument should not hold up today, but then again, it should not have in 1993 either, as the valuation method proposed by the IRS was substantially the same, not the "facts and circumstances" approach feared by the Shapiro court.

• Shapiro Court: It is well established that the actuarial tables generally are to be respected unless the established facts show that the result under the tables is unrealistic or unreasonable, or that the result ignores common sense.

In 1993, the exhaustion test represented a departure from the tables. There was, at the time, no regulatory basis for it and it merely represented an argument by the IRS. Today, the exhaustion test is part of the regulations, and its application actually employs, rather than departs from, the tables. Thus, today the argument that the exhaustion test should not apply would represent the departure from the regulations and actuarial tables.

In contrast, if the governing instrument requirement, noted above, is failed, then a departure from the table is authorized. The Shapiro court's admonition against departing from the tables is still meaningful in the context of the governing instrument requirement (but in this regard, Shapiro has been somewhat weakened by the lottery cases, discussed below).

• Shapiro Court: Respondent confuses a TERM CERTAIN with a property interest that is TERMINABLE. ... If respondent, in drafting this regulation, intended the meaning of term certain to comprise that period of time after which a property interest might be extinguished by the possible exhaustion of a trust fund, she should have set forth a special definition of "term certain" in the regulation. Because she did not, we are reluctant to transform "term certain" into a term of art in this case.

And:

• The fact that respondent, in Regs. §20.2031-7, did not specify "terminable" interests, when she could have, further convinces us that a TERM CERTAIN is to be distinguished from an interest that is merely TERMINABLE DUE TO POSSIBLE DEPLETION OF THE CORPUS. (Bold and italics added, capital letters from original.)

In these statements, the Tax Court rejects a re-characterization of a terminable annuity (terminable due to trust exhaustion) into a term certain annuity. In the first quote above, the Tax Court invites the Commissioner to promulgate a regulation to this effect. In the second quote above, the Tax Court notes that the Commissioner could promulgate such a regulation. The implication of "when she could have" is that the Shapiro court believes that such a regulation would be valid, not invalid.

• Shapiro Court: Taking respondent's argument to its theoretical conclusion, ANY TRUST created with corpus funds equivalent to the present value of a lifetime annuity obligation as computed under Table A would be deemed to be "underfunded" in that it would have insufficient funds to sustain the annual payments should the annuitant live beyond his or her average life expectancy. In this regard, respondent's argument contravenes the fundamental purposes and presumptions underlying the actuarial tables... The fair market value of an obligation to make future payments is deemed to be a sufficient amount to meet the entire obligation. Thus, by deduction, one of the fundamental presumptions underlying Table A is that the PRESENT VALUE of a lifetime annuity obligation is sufficient to sustain the stream of annuity payments as they come due over the course of an annuitant's expected life span.

This logical deduction made by the Shapiro court is difficult to understand. If Table A says that a $1 million annuity for life is worth $10,372,700, one could just as easily state that a "fundamental presumption" underlying Table A is that the annuity will actually be paid. In fact, the IRS stated as much in Rev. Rul. 77-454. Such a presumption would validate the exhaustion test.

But how does one deduce, as the Tax Court did, that $10,372,700 is sufficient to pay the annuity, just because that is what the annuity is worth? If the fair market value of an annuity were sufficient to pay an annuity, why would one ever buy an annuity? One could merely self-fund his own annuity. Calling this a "fundamental presumption" underlying the Table drastically overstates the case. More accurately, the notion that $10,372,700 is sufficient to fund an annuity worth the same is a consequence of the old tables (and their then lack of an exhaustion test), not a "fundamental presumption" of them.

In any event, now that the exhaustion test is part of the regulations, the presumption that the annuity will actually be paid (no matter how long the annuitant's life expectancy) now appears to be a "fundamental presumption" underlying the tables. These are tables promulgated by the Secretary -- legislative regulations no less -- accordingly, the Secretary seems the logical party to decide what the fundamental presumptions of the tables are.

• Shapiro Court: A person, at any given age, is assumed to die within a time consistent with the average mortality rate for that age. Regs. §20.2031-7(f). Table A does not expect or presume that a 91-year-old person will live for 18 more years; to the contrary, the table implicitly recognizes the possibility of any person reaching 109 years of age is extremely remote.

To put this argument in its context, the Tax Court was responding to the IRS's argument that the trust in Shapiro was insufficiently funded to make annuity payments past the annuitant's 95th birthday. It is true that the regulations assume that an individual will die within a time consistent with the average mortality rate for his age, but this assumption is for purposes of valuing an annuity. The assumption has nothing to do with funding the obligor, or for how long the annuity will actually be paid. Because there was no exhaustion rule in the regulations, there was no need for any assumption built into the tables as to how long an annuity would be paid. Such an assumption would have been inconsequential, as it would not have factored into any calculation under the tables. The tables were concerned only with the value of the annuity, and the value of the annuity is determined by reference to how long the annuitant is expected to live.

It also is true that the tables implicitly recognize the possibility of any person reaching 109 years of age is extremely remote. Application of the exhaustion test reflects this fact, as the consequence of failing the exhaustion test appropriately reflects the unlikelihood of attaining unrealistic ages. It is not clear why this is an argument against the exhaustion test.

OTHER EXHAUSTION TEST OBSERVATIONS

Potential Conversion of a Private Annuity to a SCIN

One potential consequence of failing the exhaustion test is that arguably a private annuity thereby is converted to an installment note. The exhaustion test, when applicable, effectively converts a life annuity to a shorter of term or life annuity. It does so on substance over form grounds -- although the form of the annuity is a life annuity, in substance, the fund from which the annuity is payable is only sufficient to pay the annuity for a term of years.

The shorter of term or life annuity will most likely be for a term that is similar in length to the grantor's life expectancy. If the annuity is a transaction that is recognized for income tax purposes (i.e., the annuity transaction is not with a grantor trust), then the annuity could be re-characterized as a debt instrument, and therefore as an installment sale. There is little guidance as to how to characterize an annuity for term or life (whether it is characterized as an installment sale or private annuity).

One argument is that §1275, which defines debt instruments, governs the characterization because annuity contracts and debt are mutually exclusive instruments,27 Thus, if §1275 treats the contract as debt, then arguably it cannot be an annuity. Under Regs. §1.1275-1(j), an annuity contract will be excepted from the definition of a debt instrument only if the contract does not contain any terms or provisions that can significantly reduce the probability that total distributions under the contract will increase commensurately with the longevity of the annuitant.28 Regs. §1.1275-1(j)(6) provides that a contract with a maximum payout provision can significantly reduce the probability that total distributions under the contract will increase commensurately with the longevity of the annuitant, unless the period of time from the annuity starting date to the termination of the annuity is at least twice as long as the period of time from the annuity starting date to the expected date of the annuitant's death. Regs. §1.1275-1(j)(7) provides that if the amount of distributions during any contract year (other than the last year during which distributions are made) may be less than the amount of distributions during the preceding year, this possibility can reduce significantly the probability that total distributions under the contract will increase commensurately with the longevity of the annuitant. An exhausting trust could, under substance over form principles, be deemed a maximum payout provision or a decreasing payout provision.

Another argument is that GCM 3950329 governs the issue. GCM 39503 characterizes an annuity for a term of life expectancy or less as a debt instrument, and an annuity for a term of longer than life expectancy as an annuity. However, GCMs carry very little authoritative value (the same weight as a PLR), and when older than 10 years, and more particularly, as here, when intervening guidance is inconsistent (Regs. §1.1275-1(j)(6), which was added after GCM 39503 came out), they are given very little weight.30

Unfortunately, even if GCM 39503 can be relied on in this context, the news may not be good. The GCM requires the use of the life expectancy tables under §72, which have longer life expectancies than does Table 90CM used by §7520. A minimally funded trust will exhaust under §7520 at or around the time the grantor reaches his §7520 life expectancy. The Table 90CM life expectancy will be attained prior to the annuitant's §72 life expectancy. As a result, the shorter of term or life annuity, even if outside the annuitant's Table 90CM life expectancy, will often be WITHIN the annuitant's §72 life expectancy, and accordingly could be characterized as an installment sale even under GCM 39503.

Lottery Cases Irrelevant

The recent lottery cases31 under §7520 are worth mentioning, only because they have been cited by promoters of the so-called private annuity trust as cases that invalidate, or undermine, the exhaustion test. In the lottery cases, the estates of decedents who owned annuities won in state lotteries argued that use of the annuity tables created an unreasonable and unrealistic result because the valuation formula in §7520 does not take into account the lack of marketability of the lottery payments. In Cook and Gribauskas, the Tax Court refused to allow deviation from the tables on these grounds. In Shackleford, the Ninth Circuit did allow deviation from the tables on these grounds. The Fifth Circuit upheld the Tax Court in Cook, while the Second Circuit overturned the Tax Court in Gribauskas.

While these cases shed light on when taxpayers may avoid the tabular valuations of annuities (where the result produced by the tables is unreasonable), none of these cases concern the exhaustion test. The issue in each case is whether the taxpayer could ignore the regulations when the regulations produced an unreasonable result. Application of the exhaustion test is an application of the regulations, not a deviation from the regulations. If Cook controls, then it would appear to validate using the regulations, and therefore the exhaustion test. If Gribauskas and Shackleford control, then taxpayers could argue that the regulations, including the exhaustion test, produce an unreasonable result, but as discussed below, the result may not be to the taxpayer's liking.

Using the Exhaustion Test as a Sword: Exchanging Underwater Annuities

The exhaustion test does provide a private annuitant with one tactical estate planning advantage. Similar to a GRAT, if trust assets drop in value, the annuitant's estate can capture the "benefit" of trust depreciation, making private annuities, like GRATs a best of both worlds estate plan: it accomplishes an estate freeze if trust assets appreciate, but if trust assets depreciate, so could the estate.

Depreciation in the trust assets can be captured for the annuitant's estate tax benefit by exchanging the private annuity contract for another contract with lower annuity payments. In other words, private annuities can be refinanced, and the gift tax costs are relatively insubstantial. Some commentators and internet promoters of the so-called private annuity trust assume that annuities cannot be refinanced, at least without income tax consequence, because §1035, which governs exchanges of annuity contracts, requires the annuity to be issued by a commercial annuity company. However, §1035 is only needed as a non-recognition provision if §1001 would otherwise tax an exchange. In the private annuity context, Rev. Rul. 69-7432 protects against recognition of income. Rev. Rul. 69-74 provided for nonrecognition in the original transfer (the one giving rise to the original private annuity contract), and Rev. Rul. 69-74 should likewise apply in the subsequent transfer (the exchange of one private annuity contract for another).

From a gift tax perspective, the value of an annuity is determined under §7520. Unlike notes, which are presumed to be worth their face value, annuities have floating values. Their values change monthly, both as the §7520 rate changes and, in the case of "underwater" annuities, by changes in the value of the trust corpus. Changes in the value of the trust corpus affect the value of annuities because the value of the trust corpus determines when a trust will exhaust. As trust corpus drops, so does the value of the annuity owned by the annuitant, because under §7520, an annuity can never be worth more than the funds from which the annuity will be paid, and the annuity must always be worth less than the funds from which the annuity will be paid (due to the risk of death).

Example: A 65-year old sells an asset worth $10,372,700 to an irrevocable trust in exchange for a $1 million annuity at a time when the §7520 rate is 5.2%. Five years (and five annuity payments) later, the §7520 rate remains 5.2%, and the trust assets have underperformed, and have remaining corpus of $5 million. At such time, applying the exhaustion test, the now 70-year old's annuity in the trust is worth only $4,580,000.

If the trust assets recover, very little of the recovery is likely to inure to the benefit of the annuitant's family due to the substantial $1 million annuity payments coming back to the annuitant. If income tax was not an issue, the annuitant could settle the annuity contract for $4,580,000 at no gift tax cost, thereby locking in $420,000 at the trust level. Or, the annuitant could refinance the annuity, exchanging the $1 million annuity for a $508,810 annuity,33 although if the new annuity fails the exhaustion test (e.g., it is not guaranteed or if no other mechanism is employed to satisfy the test), then a $580,000 gift will result. This gift, however, represents a small cost, because it will almost be entirely recouped in one year, as the trust, due to the lower annuity payments, will have $491,190 (the difference between what the trust would have paid absent the refinancing, $1 million, and $508,810) in additional assets each year.

AVOIDING THE EXHAUSTION TEST

Does the Exhaustion Test Produce an Unreasonable Result, So That a Taxpayer Could Justify Avoiding the Regulations, as in Gribauskas and Shackleford?

What if we stipulated that the exhaustion test is valid, but rely on Gribauskas and Shackleford to claim that the regulations (including the exhaustion test) produce an unreasonable result? This is unlikely to help, because the tables tend to overvalue annuities. This is why taxpayers in estate tax cases (see the lottery cases, discussed above) seek to avoid the tables. Because the tables do not take into account the unsecured nature of the annuities and the source from which the annuities will be paid (except with respect to the exhaustion test), the tables are likely to overstate the value of annuities, and accordingly overstate estate tax.

In contrast, in the gift tax context, taxpayers want annuities to be overvalued (that way, more property can be given up, and removed from the estate, in exchange for the annuity). Because private annuities are unsecured, and payable from a limited fund, they likely have a much lower value applying traditional fair market value principles than they would in the §7520 context, where the unsecured nature of the annuity and the private nature of the annuity do not factor into valuation. Accordingly, applying §7520 usually will produce the desired higher value for the annuity. The exhaustion test negates much of the benefit of applying §7520 (by producing a lower value for the annuity), but traditional valuation principles (i.e., valuation outside the §7520 context) are likely to reduce the value of the annuity issued by an underfunded trust in the same way that the exhaustion test does. So, Gribauskas and Shackelford may provide sufficient authority for taxpayers, at least those in the Second and Ninth Circuits, to avoid application of the tables if they produce unreasonable results, but proving the unreasonable result may be difficult -- even with the exhaustion test, the result may be unreasonable in the other direction.

Meeting the Test with Guarantees

If individuals guarantee an annuity issued by a trust, then the exhaustion test most likely does not apply. The annuity is no longer payable from a limited fund. Although literally the exhaustion test applies to an annuity payable from a trust or other limited fund, suggesting that all trusts are subject to the exhaustion test (in addition to other entities with limited funds), the better interpretation is that the Secretary was using the term "trust" to describe entities that typically have limited funding. In any event, the rule looks to when the "fund" will exhaust, and nothing in the rule specifically limits the "fund" to trust corpus, and it would make little sense to ignore, for purposes of determining exhaustion, assets available to a trust to satisfy its obligation.

However, the use of guarantees is a short-term, or maybe a shortsighted, solution. Depending on who issues the guarantee, the guarantee may be a complete or incomplete gift, or if not a gift, the guarantee may give rise to income to the party for whose benefit the guarantee was issued, with any offsetting deduction potentially a below-the-line, §212 type (i.e., difficult to use). The gift and income tax issues associated with guarantees are beyond the scope of this article.

In addition to gift and income tax issues associated with the guarantees, if the guarantees actually are needed (i.e., the trust can no longer make its annuity payments), then reverse estate planning results, with substantial assets coming into, rather than going out of, the annuitant's estate each year. And the big problem with guarantees on private annuities is that guarantees stand an excellent chance of being called upon if there is any volatility in the performance of the trust assets. For example, recall the 65-year old example. Assuming a 10% seed gift, Monte Carlo simulation34 reveals that there is a 50% likelihood that the trust will run out of funds within 20 years. On the other hand, for annuitant's with shortened life expectancies, guarantees should provide sufficient cushion against application of the exhaustion test with little risk that the guarantees will actually be called upon, provided the gift and income tax issues are appropriately resolved.

Meeting the Test by Increasing Annuity Payments

One criticism of the exhaustion test is the supposed anomaly that if you increase the retained annuity, you fail the exhaustion test even worse, resulting in a larger gift. How can your gift go up by taking back a greater interest? If you have retained too much, then how can it logically follow that you've made a gift? This is actually a misconception. The exhaustion test actually works in the opposite way: as the retained annuity goes up, the resulting gift goes down.

Example: A 65-year old sells $10,372,700 asset to trust for $1 million annuity. The trust is funded with a $1 million seed gift. Application of the exhaustion test results in an assumption that the trust will run out of funds when the annuitant is 83, and an approximate $1.1 million gift. If the annuity were increased $1.5 million, then the exhaustion test will result in an assumption that the trust will run out of funds when the annuitant is only 75. However, the resulting gift drops to $735,000. If the annuity is increased to $1.75 million, the gift goes away altogether.

Then is this the answer to avoiding the exhaustion test? Maybe, but the increased annuity drastically increases the chances of wasting the seed gift. The reason that the $1.75 million annuity results in no gift is that the tables assume that the annuitant will receive his seed gift back in year 10 (i.e., the value of the right to receive the seed gift back in year 10 is equivalent in value to the potential loss of annuity payments due to early death). On the other hand, if the annuitant has a shortened life expectancy, increasing the annuity payment may be the best way to avoid the exhaustion test, as it is more feasible than meeting the test (see below) and the risk of returning the seed gift back to the annuitant is reduced.

Of course, if the annuity payment is increased to avoid the exhaustion test, then the annuity should, in form, be a term of years or life annuity, with a term equal to the term assumed by the exhaustion test. There is no sense in taking the risk that the funds of the trust will be sufficient to continue making annuity payments if the annuitant lives past the term assumed by the exhaustion test.

Increasing annuity payments in this manner is only recommended for estate planning private annuities (i.e., annuities entered into with grantor trusts). For private annuities entered into for income tax deferral, the increased annuity payments will not only decrease the period of deferral but most likely cause re-characterization of the annuity as an installment note, as explained above.

Meeting the Exhaustion Test through Large Seed Gift

Meeting the exhaustion test through a seed gift is extremely difficult, and probably unwise in any event. The 65-year old selling a $10,372,700 asset to a trust in exchange for a $1 million annuity would need to fund the trust with about a $7 million seed gift to avoid the exhaustion test altogether.35 A $4.5 million seed gift will allow the trust to have funding through age 95, resulting in an insignificant $48,307 gift. Thus, the last $2.5 million of funding only avoids a $48,307 gift.

In either case, however, meeting (or partially meeting) the exhaustion test through a seed gift shares the same problem as meeting the exhaustion test through guarantees. Unless the annuitant's life expectancy is shortened, there is a significant risk of "reverse estate planning" -- the return to the annuitant in the form of annuity payments of his large seed gift.

Structure the Sale as a SCIN

An interest only SCIN with a term equal to double the seller's life expectancy or less would likely achieve economic equivalence with a private annuity, as both structures provide a premium to compensate the seller for the risk of early death. The interest only SCIN, however, largely backloads the premium (the seller gets the principal balance in full or not at all), whereas the private annuity spreads the premium out over the annuitant's lifetime.

As discussed above, such a note should be characterized as a debt instrument under §1275 (which should have no consequence if the trust is a grantor trust) and the note should not constitute an annuity for purposes of §7520. However, if the interest payments are fixed, it is possible that the stream of interest payments would be valued under §7520, but in such event, it is far less likely to be impacted by the exhaustion test than a private annuity would be.36

CONCLUSION

The exhaustion test represents a substantial hurdle in any private annuity sale to a trust. The test can be mitigated through guarantees, seed gifts or larger annuity payments, but these solutions are recommended only where life expectancy is decreased. Where life expectancy is not diminished, an interest only SCIN with a payoff date that is prior to double the seller's life expectancy may achieve economic equivalence to a private annuity but with minimal exhaustion test concerns.

Appendix

Regs. §25.7520-3(b)(2)(v), Example 5

Eroding corpus in an annuity trust.

(i) The donor, who is age 60 and in normal health, transfers property worth $1,000,000 to a trust. The trust will pay a 10 percent ($100,000 per year) annuity to a charitable organization for the life of the donor, payable annually at the end of each period, and the remainder will be distributed to the donor's child. The section 7520 rate for the month of the transfer is 6.8 percent. First, it is necessary to determine whether the annuity may exhaust the corpus before all annuity payments are made. Because it is assumed that any measuring life may survive until age 110, any life annuity could require payments until the measuring life reaches age 110. Based on a section 7520 interest rate of 6.8 percent, the determination of whether the annuity may exhaust the corpus before the annuity payments are made is computed as follows:

Age to which life annuity may continue

110

less: Age of measuring life at date of transfer

60

- - -

Number of years annuity may continue

50

Annual annuity payment

$100,000.00

times: Annuity factor for 50 years

derived from Table B

14.1577

- - - - - - - - -

Present value of term certain annuity

$1,415,770.00

(ii) Since the present value of an annuity for a term of 50 years exceeds the corpus, the annuity may exhaust the trust before all payments are made. Consequently, the annuity must be valued as an annuity payable for a term of years or until the prior death of the annuitant, with the term of years determined by when the fund will be exhausted by the annuity payments.

(iii) Using factors based on Table 90CM at 6.8 percent (see §20.2031-7T(d)(7) of this chapter), it is determined that the fund will be sufficient to make 17 annual payments, but not to make the entire 18th payment. Specifically, the initial corpus will be able to make payments of $67,287.26 per year for 17 years plus payments of $32,712.74 per year for 18 years. The annuity is valued by adding the value of the two separate temporary annuities.

(iv) Based on Table H of Publication 1457 (a copy of this publication may be purchased from the Superintendent of Documents, United States Government Printing Office, Washington, DC 20402), the present value of an annuity of $67,287.26 per year payable for 17 years or until the prior death of a person aged 60 is $588,016.64 ($67,287.26 × 8.7389). The present value of an annuity of $32,712.74 per year payable for 18 years or until the prior death of a person aged 60 is $292,196.74 ($32,712.74 × 8.9322). Thus, the present value of the charitable annuity interest is $880,213.38 ($588,016.64 + $292,196.74).

Footnotes

1. Unless otherwise note, all section references are to the Internal Revenue Code of 1986, as amended, and the regulations promulgated thereunder.

2. The regulations contain other limitations (e.g., the lives of terminally ill individuals may not be used as a measuring life). This paper focuses on the limitations that may be applicable to an otherwise healthy annuitant.

3. Regs. §25.7520-3(b)(2)(i). There are identical regulations for income and estate tax purposes. See Regs. §§1.7520-3(b)(2)(i) and 20.7520-3(b)(2)(i). For simplicity, this paper will only cite the gift tax regulation.

4. Some private annuity sales are to non-grantor trusts. Such sales are intended to achieve income tax deferral. Other sales are to grantor trusts, intended to achieve estate tax benefits. This paper assumes that the trust will be a grantor trust, except where otherwise noted.

5. Calculations in this paper were made using Microsoft Excel or NumberCruncher. Excel is used to determine when, consistent with Example 5, a trust will exhaust. NumberCruncher is used to value annuities for a term certain or life. This method is able to replicate the numbers in Example 5. Calculations assumed a 5.2% §7520 rate, and disregarded income tax consequences of private annuities.

6. Example 5 uses a similar method to Rev. Rul. 77-454, 1977-2 C.B. 351, to determine when the trust will deplete. The principal is assumed to grow at the applicable rate (here, the §7520 rate of 5.2%) and then the annuity payment is made. The remaining principal again is assumed to grow at exactly the applicable rate, and the second annuity payment is made. This process is repeated until the trust runs out of funds. In this example, the trust has enough funds to make 17 full annuity payments and a partial payment in the 18th year. The value of the annuity is then computed by adding the value of an 18-year annuity (using the payment amount in year 18 as the annual annuity amount) to the value of a 17-year annuity (using the remaining portion of the annuity payment as the annual annuity amount). The $1 million annuity for life is accordingly valued as if it were an annuity for a term of years or life. The value of the annuity under §7520 is approximately $9,276,307, resulting in a gift of approximately $1.1 million.

7. 1955-1 CB 352.

8. While no IRS pronouncement has ever indicated that §7520 applies for purposes of determining the value of an annuity obligation for purposes of Rev. Rul. 55-119, §7520 does apply for income tax purposes (at least where not otherwise excepted) and Regs. §1.7520-3(b)(2)(i) does apply the exhaustion test for income tax purposes.

9. Rev. Rul. 77-454, 1977-2 C.B. 351.

10. This approach is illustrated in Table 1, infra.

11. T.C. Memo 1993-483.

12. The IRS had a third argument, that the annuity was not actually an annuity, an argument the Tax Court rejected. This aspect of the opinion is not pertinent to the issues herein.

13. 1994-2 C.B. 899.

14. T.D. 8630, 60 Fed. Reg. 63913 (12/13/95).

15. 1977-2 C.B. 351.

16. 1970-2 C.B. 199.

17. 269 F.2d 738 (4th Cir. 1959).

18. 224 F.2d 26 (1st Cir. 1955).

19. 467 U.S. 837 (1984).

20. Id.

21. White, "Healthy Living: 110 plus; GSU researcher tracks elite pack of supercentenarians for clues on longevity," Atlanta Journal Constitution, Jan. 31, 2006, at 1E.

22. Id.

23. See Table 90 CM, IRS Pub. 1457, Actuarial Values, Book Aleph (7-1999).

24. According to Table 90CM, about 1 in 4,678 sixty-five year olds will attain the age of 109. None will attain the age of 110. The probability of attaining the age of 110 is the average of the probability of attaining the ages of 109 and 110.

25. Incidentally, with no seed gift, contrary to the Shapiro court's assumption, the annuity will exhaust the trust PRIOR to life expectancy. As long as the average life expectancy is less than the median life expectancy, larger annuity payments are needed to compensate for the relative greater risk of early death.

26. This is so because the tables assume that the chances of living past age 110 are zero.

27. Dix v. Comr., 392 F.2d 313 (4th Cir. 1968); Rye v. U.S., 25 Cl. Ct. 592, 92-1 USTC ¶50,186.

28. Regs. §1.1275-1(j)(2)(i)(B).

29. GCM 39503 (5/7/86).

30. Regs. §1.6662-4(d)(3)(ii).

31. Estate of Gladys J. Cook v. Comr., T.C. Memo 2001-170, aff'd, 349 F.3d 850 (5th Cir. 2003); Estate of Paul C. Gribauskas v. Comr., 116 T.C. 142 (2001) rev'd, 342 F.3d 85 (2d Cir. 2003); Shackleford v. U.S., 262 F.3d 1028 (9th Cir. 2001), aff'g 84 AFTR 2d 5902, 99-2 USTC ¶60,356 (E.D. Cal. 1999); Estate of Donovan v. U.S., 2005-1 USTC ¶50,322 (D. Mass. 2005); Davis v. U.S., 2006-1 USTC ¶60,514 (D. N.H. 2005).

32. Rev. Rul. 69-74, 1969-1 C.B. 43.

33. A $508,810 annuity for a 70-year old would have a value of $4,580,000.

34. Monte Carlo simulation assumed that assets grow at an expected rate of 9% annually, with a volatility of 18%. This means that the average simulation will yield a 9% growth rate with a standard deviation of 18%. Five hundred trials were run, and 248 trials showed zero trust balance by the twentieth year.

35. The total at the bottom of column 4 of Table 1 represents the amount the trust would need to fund the annuity through age 110.

36. The term "annuity" is not defined in §7520. However, the Tax Court has held that "an annuity is commonly defined as a right to receive fixed periodic payments, either for life or for a term of years." Cook v. Comr., T.C. Memo 2001-170, aff'd, 349 F.3d 850 (5th Cir. 2003). If the SCIN involves fixed interest payments, the IRS could assert that the stream of interest payments is to be valued under §7520. However, this stream of income is unlikely to fail the exhaustion test because it represents a smaller annual payment than does the annual private annuity payment and it is limited to a term certain. There would be no exhaustion test for the principal balance of the note.