The House of Representatives Votes to Restrict GRATs

The imposition of a ten-year minimum term on GRATs takes a possible step toward enactment, promising to change, but not stop, the creative use of GRATs.

Dear Readers Who Follow Washington Developments:

On March 25, 2010, the House of Representatives passed the “Small Business and Infrastructure Jobs Tax Act of 2010” (H.R. 4849).  Title III of the bill – captioned “Revenue Provisions,” translated pay-for’s – includes section 307, which would impose restrictions on the use of a grantor retained annuity trust (GRAT), including a requirement that a GRAT have a minimum term of ten years.  This is the only transfer tax provision in H.R. 4849.

A Minimum Ten-Year Term for GRATs

The ten-year minimum term for GRATs is adapted from the recommendation on page 123 of the Treasury Department’s “General Explanations of the Administration’s Fiscal Year 2010 Revenue Proposals” (“2009 Greenbook”), published on May 11, 2009, and discussed in Capital Letter Number 17, and on page 126 of Treasury’s “General Explanations of the Administration’s Fiscal Year 2011 Revenue Proposals” (“2010 Greenbook”), published on February 1, 2010.  Reminiscent both of the Greenbooks’ explanations and of the 1990 legislative history of section 2702 itself, the House Ways and Means Committee offers the following “Reasons for Change”:

The valuation rates and tables prescribed by section 7520 often produce relative values of the annuity and remainder interests in a GRAT that are not consistent with actual returns on trust assets.  As a result, under present law, taxpayers can use GRATs to make gifts of property with little or no transfer tax consequences, so long as the investment return on assets in the trust is greater than the rate of return assumed under section 7520 for purposes of valuing the lead and remainder interests.  The Committee believes that such uses of GRATs for gift tax avoidance are inappropriate.

In some cases, for example, taxpayers “zero out” a GRAT by structuring the trust so that the assumed value of the annuity interest under the actuarial tables equals (or nearly equals) the entire value of the property transferred to the trust.  Under this strategy, the value of the remainder interest is deemed to be equal to or near zero, and little or no gift tax is paid.  In reality, however, a remainder interest in a GRAT often has real and substantial value, because taxpayers may achieve returns on trust assets substantially in excess of the returns assumed under section 7520.  Any such excess appreciation passes to the remainder beneficiaries without further transfer tax consequences.

In addition, grantors often structure GRATs with relatively short terms, such as two years, to minimize the risk that the grantor will die during the trust term, causing all or part of the trust assets to be included in the grantor’s estate for estate tax purposes.  Because GRATs carry little down-side risk, grantors frequently maintain multiple short-term, zeroed-out GRATs funded with different asset portfolios to improve the grantor’s odds that at least one trust will outperform significantly the section 7520 rate assumptions and thereby allow the grantor to achieve a transfer to the remainder beneficiaries at little or no gift tax cost.

The provision is designed to introduce additional downside risk to the use of GRATs by imposing a requirement that GRATs have a minimum term of 10 years.  Relative to shorter-term (e.g., two-year) GRATs, a GRAT with a 10-year term carries greater risk that the grantor will die during the trust term and that the trust assets will be included in the grantor’s estate for estate tax purposes.  The provision limits opportunities to inappropriately achieve gift tax-free transfers to family members in situations where gifts of remainder interests in fact have substantial value.

H.R. Rep. No. 111-447, 111th Cong., 2d Sess. 55-56. (2010) (footnote omitted).

A footnote in both the 2009 and 2010 Greenbooks compared the proposed ten-year minimum GRAT term to the minimum ten-year term of “Clifford trusts” under section 673 (before its amendment by the Tax Reform Act of 1986), a comparison, by the way, that was completely lost on Capital Letters.  The Ways and Means Committee report discusses the grantor trust rules, but does not mention the ten-year requirement for pre-1986 “Clifford trusts” and does not explicitly tie the ten-year requirement from GRATs to the grantor trust rules.

It is well known that the benefits of any GRAT are partly offset by the fact that the transferor must survive the GRAT term for those benefits to be realized.  As the Ways and Means Committee Report noted, many GRATs have a very short term, often just two years, although in recent years there has been more use of longer-term GRATs to lock in relatively low interest rates or values.  Obviously, a requirement that a GRAT have a term of at least ten years would increase the likelihood that the GRAT will “fail” and be subject to estate tax upon the transferor’s death.  The Committee Report explicitly states that this change is “designed to introduce additional downside risk to the use of GRATs.”

Accompanying Changes to the GRAT Rules

In the single substantive change to the transfer tax proposals from the 2009 Greenbook, the 2010 Greenbook added that “[t]he proposal would also include a requirement that the remainder interest have a value greater than zero and would prohibit any decrease in the annuity during the GRAT term.”  Embracing that proposal, H.R. 4849 would impose those two requirements.  In a footnote, the Ways and Means Committee states that “[t]hese requirements are designed to prohibit circumvention of the ten-year minimum term requirement of the proposal.”  H.R. Rep. No. 111-447 at 55 n173.  For example, if the GRAT annuity payment could be reduced from year to year, then there is no reason why a ten-year GRAT funded with, say, $1,000 could not pay $474 the first year, $568.80 (20% greater) the second year, and one dollar in each of the last eight years.  At today’s 7520 rate of 3.2%, the taxable gift would be about 10 cents.  Moreover, if the grantor dies after two years but within ten years, it appears that under Reg. § 20.2036-1(c)(2)(i), with a 7520 rate of 3.2%, the amount included in the grantor’s gross estate would be $31.25 ($1/0.032).  It is hardly a stretch to view that as a “circumvention” of the additional downside risk the minimum ten-year term is designed to introduce.

Despite the requirement that that the remainder interest have a value greater than zero, the 2010 Greenbook went on, like the 2009 Greenbook, to say that “a minimum term would not prevent ‘zeroing-out’ the gift tax value of the remainder interest.”  The Greenbook seemed to use the term “zeroing-out” the same way most of us do, meaning the selection of a term and payments to produce only a small taxable gift.  The Ways and Means Committee Report does not repeat the “zeroing-out” comment, but the statutory language in proposed new section 2702(b)(2)(C) requires only that “the remainder interest has a value greater than zero determined as of the time of the transfer.”  It is hard to see how such a requirement would add to or reinforce the downside risk of a GRAT.  Even a nominal taxable gift requires reporting on a gift tax return, but failure to report even when the gift truly is zero would still not prevent the imposition of estate tax if the grantor dies during the GRAT term, which arguably is the real source of the proposed increased downside risk.

On the other hand, it could be argued that the estate tax if the grantor dies during the GRAT term is not a “downside” of creating a GRAT at all, because the property placed in the GRAT would have been subject to estate tax upon the grantor’s death if the grantor had forgone the GRAT and just held the asset until death.  The only losses from creating a GRAT are (i) transaction costs (attorney’s fees, appraisal costs, return preparation expenses, and the like), (ii) the loss of a small amount of unified credit (or the payment of a small amount of gift tax), and (iii) opportunity costs – that is, the loss of the potential benefits of forgone techniques such as installment sales, SCINs, AFR loans, private annuities, charitable lead trusts, and for that matter outright gifts.  So viewed, one could conclude that the only way to “introduce additional downside risk” would be to increase the amount of the initial taxable gift by introducing a minimum remainder value of, say, 10%, which would be roughly analogous to the 10% minimum value for the residual (junior) interest in a capital freeze entity under section 2701(a)(4).  Indeed, before we saw the 2009 Greenbook, many observers expected a 10% minimum remainder to be the GRAT reform we would get if we got anything.  Because the vague “greater than zero” statutory language shouts the question “how much greater,” it bears watching to see if someone discovers any “pay-for” potential from adding such a minimum.

An Answer Found in the Mechanics of the Federal Budget

Warning:  This section will be very tedious to all but true lovers of esoteric conjecture.  The reader will be forgiven for skipping to the next section, and will sacrifice no continuity by doing so.

The answer may lie in the most obvious place that is often overlooked – the “scoring” of revenue gains and losses by the staffs of the Joint Committee on Taxation and the Congressional Budget Office.  Most revenue measures are scored over a ten-year budget window, currently running through fiscal 2020, which ends September 30, 2020 (generally corresponding to gifts made and estates of decedents dying in calendar 2019).  With respect to GRATs, the universe of taxpayers who can contribute to increased federal revenues over the next ten years are

(i)      payers of gift tax – grantors who create GRATs between now and 2019 and pay more gift tax than they would otherwise have paid (either on the creation of the GRAT or subsequently during the ten-year budget window because of a reduced gift tax unified credit available),

(ii)     estates with less unified credit – the estates of grantors who create GRATs, use more unified credit than they otherwise would have used, and die before 2020 with less estate tax unified credit available as a result, and

(iii)    estates that pay estate tax on GRAT assets – the estates of grantors who create GRATs and die before 2020 before their GRAT terms expire, thereby subjecting the GRAT assets to estate taxes.

It may reasonably be assumed that the first group will pay little gift tax during the ten-year budget window.  Many GRATs currently result only in the use of gift tax unified credit, not the payment of actual gift tax.  Grantors of GRATs who pay significant gift tax, or who would have to pay significant gift tax if the gift tax rules were changed, may reasonably be expected to consider appreciation-and-interest-leveraged techniques other than GRATs if those rules were changed, such as installment sales, SCINs, AFR loans, private annuities, and charitable lead trusts.  The result is that revenue estimators might not assign much revenue potential to the first group of grantors.

As to the second and third groups – why, they turn out to be the same!  Every grantor who dies before 2020 will necessarily die before the expiration of any ten-year GRAT created after the effective date of the legislation.  So in general there will be no additional estate taxes from the second group by reason of the loss of unified credit, because the unified credit will in effect be restored by the exclusion of the gift to the GRAT from adjusted taxable gifts under the last phrase of section 2001(b).  The big revenue-raiser turns out to be the third group, which pays additional estate tax attributed to the inclusion of the value of GRAT assets in the gross estate.  And that is exactly the group targeted by the ten-year minimum GRAT term.  Despite the focus in the Ways and Means Committee Report on the understatement of the gift, the surprising continued indulgence of nearly-zeroed-out GRATs may be rational after all.

There are imperfections in the foregoing analysis, such as in the case of spouses of grantors who gift-split and whose unified credit is not restored under section 2001(b).  Married grantors also have the ability to qualify the GRAT for the marital deduction, thereby increasing the likelihood of pushing any additional estate tax revenue beyond the ten-year budget window.  But it still is reasonable to view the estate tax, not the gift tax, as the main potential revenue-generator with regard to GRATs.

Revenue Estimates

As to be expected with revenue gained mostly from the estate tax on GRATs created after the effective date of the legislation, the gains increase over the ten-year budget window.  The revenue estimates by fiscal year are found on page 64 of the Ways and Means Committee Report.  Because each fiscal year begins October 1, nine months after the start of the calendar year, and estate taxes are due nine months after death, a fiscal year’s estate tax receipts roughly reflect the estates of decedents who died the preceding calendar year.  Likewise, the gift tax on gifts made in any calendar year are generally due the following April, again causing gift taxes to be paid in the fiscal year following the calendar year of the gift.  The following table displays not only the revenue estimates by fiscal year in the Ways and Means Committee Report, but also the corresponding calendar year of the gifts and deaths that generally produce that revenue:

It is curious that the estimates include revenue gains in fiscal 2011 and 2012 – that is, for gifts made and decedents dying in calendar 2010 and 2011.  If the legislation were effective today, and a grantor created a GRAT today, there presumably would be no additional gift tax paid in fiscal 2011 (namely April 15, 2011), because the gift tax value of the remainder would be calculated in the usual way.  There would be additional estate tax payable in fiscal 2011 or 2012 only if the grantor died on or before December 31, 2011, but in that case even the value of the assets in a two-year GRAT would have been included in the grantor’s gross estate.  Indeed, it might be true that if the assets increase substantially in value, more value would be included in the gross estate under Reg. § 20.2036-1(c)(2)(i) in the case of a relatively high-payout two-year GRAT than in the case of a lower-payout ten-year GRAT.  Unless it is assumed that some taxpayers faced with a ten-year minimum term for GRATs will avoid the mortality risk by making outright taxable gifts, not by considering installment sales, SCINs, AFR loans, private annuities, charitable lead trusts, and the like, the additional $16 million expected in fiscal 2011 and 2012 is a mystery.

Regarding charitable lead trusts, by the way, the opening paragraph of the discussion in the Ways and Means Committee Report states that “[g]rantor retained annuity trusts (GRATs) and charitable lead trusts (CLTs) are two vehicles, often structured as grantor-owned, that are used to make transfers of temporal interests in property.”  H.R. Rep. No. 111-447 at 53.  Nevertheless, the committee does not refer to CLTs again, and, indeed, CLTs, which often can be structured to avoid inclusion in the gross estate, would not necessarily be affected by this type of legislative approach in any event.

Prospects for the House-Passed Bill

The House vote on H.R. 4849 was partisan.  Only four Republicans voted for it and only seven Democrats voted against it.  It goes to the Senate now, where, as we have seen, it is hard to pass partisan legislation that often requires 60 votes to be considered.  Many give H.R. 4849 little chance in its current form.  But the Senate Finance Committee, to which it has been referred, could change the bill to give it more bipartisan appeal, or the Senate leadership could package the bill with other measures that make it harder to ignore.  It is also possible that while H.R. 4849 is pending the GRAT provision will be used as a revenue offset in other legislation, including the much-needed and anxiously-awaited legislation to provide clarity and stability for the estate tax itself.

Effective Date

The GRAT legislation contained in H.R. 4849, like the earlier recommendations in the 2009 and 2010 Greenbooks, states that it will apply to transfers made after the date of the enactment – that is, after the date the President signs it into law.  There is no known inclination to change that effective date at this time, but that does not guarantee that the effective date will not be changed.  For example, the Senate Finance Committee could report out H.R. 4849 and change the effective date to the date of its report, allowing no time to react.  In any event, if the Senate passes H.R. 4849, which is viewed by the Administration and congressional leadership as an important economic initiative, the President might sign it very quickly.  Even if the Senate makes changes to the House-passed bill, the House might promptly approve the changes and send the bill to the White House.

While such developments are difficult to predict, the fact remains that this action in the House takes a theoretical proposal to limit GRATs one significant step closer to enactment into law.

GRATs are not for everyone, and there is probably no reason why this pending legislation should encourage someone to create a GRAT who would not otherwise do so.  However, for those for whom a GRAT at this time makes sense or those who have a pattern of creating GRATs from time to time, it may now be important to consider acting sooner rather than later to finalize the creating and funding of a GRAT or GRATs.

The Grantor’s Exchange of GRAT Assets

Short-term GRATs are often preferred to reduce the obvious mortality risk, but they are sometimes also seen as the appropriate response to the volatility of the value of GRAT assets.  A GRAT “works” when the assets in the GRAT greatly increase in value, especially in the first twelve months before the first annuity payment is due.  The upward potential of such assets is sometimes accompanied by downward volatility.

Volatility of value can and should be monitored and managed, without regard to the length of the GRAT term.  A common technique for this purpose is the grantor’s exchange of all or some of the highly appreciated GRAT assets for assets of equivalent value, but not necessarily equivalent volatility or equivalent income tax basis.  The exchange is often accomplished pursuant to the grantor’s reserved power under section 675(4)(C) of the Internal Revenue Code.  Such an exchange can also permit the grantor, often as the GRAT nears its termination, to swap into a GRAT less rapidly appreciating but valuable assets the grantor really wants to transfer to the next generation, and in any event an exchange can make the more rapidly appreciating asset withdrawn from the GRAT available for the funding of a new GRAT or GRATs.  Such exchanges with an existing GRAT to emulate the creation of new GRATs will be even more important if the GRAT is required to have a minimum ten-year term.

There is no known prohibition on such exchanges, and the grantor trust status of the typical GRAT permits such exchanges without concern for unwelcome income tax consequences.  Such exchanges used to be popular near the end of the term of qualified personal residence trusts (QPRTs), often to swap high-basis assets into the QPRT to be passed to the grantor’s children while returning the residence to the grantor to own and occupy until death, when it would receive a stepped-up basis.  Under Reg. § 25.2702-5(c)(9), QPRT governing instruments have been required to prohibit such exchanges since May 1996.  Again, there is no known inclination at this time to impose, by regulation, a similar prohibition on GRATs.  But again there are no guarantees, and such a prohibition would not be an unusual development.  Therefore, while a client who is considering creating a short-term GRAT now is well advised to act sooner rather than later, even the client who for any reason might be considering a ten-year GRAT after such a minimum term is in place will probably also be well advised to act sooner rather than later.  Things will never be dull.

Ronald D. Aucutt  © 2010 by Ronald D. Aucutt. All rights reserved