TBA Law Blog

Posted by: Dan Holbrook on Aug 13, 2008

Journal Issue Date: Aug 2008

Journal Name: August 2008 - Vol. 44, No. 8

It's nice when the IRS concedes a position it really couldn't have won anyway, but could have used as a threat. It's nicer still when its concession encourages the use of a powerful estate planning tool too little used. Specifically, in a recent revenue ruling,[1] the IRS confirmed that when a grantor of a trust retains a Code  §675(4)(C) power to substitute assets of equal value, thus making the trust a "grantor trust" for income tax purposes, the power is benign for estate tax purposes. Why is this important? First, some background.

For most clients with big estates, a huge goal is avoidance of gift and estate taxes on the transfer of wealth to descendants. The first line of defense is usually the $12,000 per donee annual gift tax exclusion,[2] or $24,000 per donee on gifts from a married couple.[3] Thus, a couple with five descendants can easily transfer $120,000 per year from their estates to lower generations. Over 20 years, that's $2,400,000. Since the combined Federal and Tennessee death taxes have an effective top bracket over 50 percent,[4] such gifts can ultimately save more than $1,200,000 in death taxes.

Of course, not only is $2,400,000 of principal removed from the clients' estates, but so is all the net income such principal would have produced. Since we're looking at trusts, assume all the gifts have been made to one trust. Assume further that each year the trust earns 8 percent gross, pays 35 percent income tax,[5] and thus nets 5.2 percent. After 20 years, the trust will have grown to about $4,264,000. Such gifts to trust can ultimately save (at 50 percent rate) over $2,132,000 in death taxes.

Ah, but there's more, and here's where the use of a grantor trust can work its magic. A "grantor trust" refers to a trust that is treated for income tax purposes as if it didn't exist, so that income is taxed to the grantor rather than to the trust.[6] Since the grantor becomes legally liable for the trust's income taxes, the grantor can pay all such taxes, effectively making additional "gifts" to the trust that are not subject to gift tax.[7] The planning trick is to create an irrevocable trust that is complete for all gift and estate tax purposes, but incomplete and still taxed to the grantor for income tax purposes. Planners refer to these as "intentionally defective" grantor trusts ("defective" because they are designed to "flunk" the grantor trust rules).[8] Suppose the $2,400,000 of gifts are made to a grantor trust. If the trust earns 8 percent gross per year (tax-free because the grantor pays all the taxes), the trust will grow over 20 years to about $5,931,000, ultimately saving (at 50 percent rate) over $2,965,000 in death taxes. This is an increase in death tax savings of $833,000, or 39 percent,[9] simply because the gifts are made to a grantor trust.[10]

So how does one create an intentionally defective grantor trust? There are several ways under the Code,[11] but the most popular has been to include in the trust document a retained power by the grantor under Code  §675(4)(C) "to reacquire the trust corpus by substituting other property of an equivalent value," i.e., the power to purchase or swap assets of the trust at fair market value.[12]


   Total to Five Descendants Over 20 Years Accumulated Net of Income Taxes Accumulated Using Grantor Trust
Gifts $2,400,000 $4,264,000 $5,931,000
Death Tax Savings @ 50% $1,200,000 $2,132,000 $2,965,000


But wouldn't such a retained power by the grantor have genuine value? If you created a trust, wouldn't you appreciate the right to pick and choose specific assets of the trust you would like to be able to reacquire, even if you had to pay cash or swap assets at their fair market value to do so? And if the power has value, wouldn't it be a prohibited retained power that causes the entire trust to be includible in the grantor's estate under Code  §2036 or  §2038? Many planners, fearing the IRS might threaten to cause estate tax inclusion under such circumstances, have drafted trusts with the redundancy of at least one other power to make the trust a grantor trust. Unfortunately, every other such power has its own quirks and uncertainties, and the substitution power has always seemed the cleanest. Only the estate tax uncertainty clouded its use. With the publication of this recent revenue ruling freeing the power of substitution from estate tax concern, grantor trusts are easier to draft and more certain of result.

Perhaps our paradigm should be to make every gift trust a grantor trust unless there are reasons not to, rather than the other way around.

- - -


  1. Revenue Ruling 2008-22, 2008-1 CB __.
  2. Code  §2503(b) grants an annual exclusion of $10,000, indexed for inflation. The exclusion is currently $12,000.
  3. Each spouse can make the annual exclusion gift without having to file a gift tax return. Code  §2513 also permits one spouse to treat his or her gift as made one-half by him or her and one-half by his or her spouse, so long as the spouse consents, which requires filing at least short form gift tax returns to show the consent.
  4. In 2008 and 2009, the effective top combined federal and Tennessee death tax bracket is 50.225 percent. Since the top federal rate is 45 percent, the top Tennessee rate is 9.5 percent, and the Tennessee tax is deductible from the federal, then the top combined rate is 45 percent + 9.5 percent * (1 - 0.45) = 50.225 percent.
  5. 35 percent is the top bracket in 2008 for all individual, estate, and trust taxpayers. Trusts reach the top bracket of 35 percent on taxable income over $10,700 in 2008, so for large trusts the tax rate really is almost exactly 35 percent.
  6. More specifically, if the grantor is treated as owner of the trust, then items of income, deduction, or credit attributable to the trust will be included in calculating the grantor's taxable income. Codes  § §671 to 677 have existed in one form or another since the mid-1920s, to punish taxpayers who might try to "shift" income from themselves to a trust with a lower income tax rate. Since 1985, estate planners have instead seen the grantor trust rules as a blessing rather than a curse, using them to enhance the benefits of some estate planning techniques. This began with Revenue Ruling 85-13, 1985-1 CB 184, in which the IRS held that a transaction between a grantor and his grantor trust was not a sale for federal income tax purposes, so there was no capital gain consequence nor any effect on cost basis.
  7. Revenue Ruling 2004-64, 2004-2 CB 7.
  8. These trusts are also "Crummey" trusts for gift tax purposes, following the technique blessed in Crummey v. Commissioner, 397 F.2d 82 (9th Cir, 1968) to convert a gift to a trust to a "present interest," as required by Code  §2503 to qualify for the gift tax exclusion. So the ideal gift trust may be a "defective Crummey trust." Sometimes estate planning terminology is wonderfully colorful.
  9. These are only my numbers, calculated from a simple Excel spreadsheet, but experts' models confirm this analysis. The investment firm of Bernstein Global Wealth Management is a national leader in running sophisticated modeling programs on estate planning techniques with an enormous number of variables, to determine the probabilities of various outcomes. Its median result on annual exclusion gifts for 20 years, inflation-adjusted, based on globally diversified equities, shows almost exactly the same end result as my calculations, i.e., an increase in tax savings by using a grantor trust of about 36 percent. See Weinreb and Litman, "The Three Gs: Gifts, GRATs and Grantor Trusts May Be the Only Strategies Needed to Minimize Taxes on Even the Largest Estates," Trusts and Estates (March 2008).
  10. The projected tax benefits are dramatically greater if the number of years is greater, or if the trust is more "front-loaded" by holding more wealth earlier in the life of the trust, because in both circumstances the higher after-tax earnings have more opportunity to compound. For example, if the grantor in our illustration made an additional gift of $1,000,000 upon creation of the trust, paying Tennessee gift tax of $83,400 and using the $1,000,000 federal gift tax exemption to pay no federal gift tax, a non-grantor trust will accumulate after 20 years to $7,020,000, and a grantor trust will accumulate after 20 years to $10,592,000. After adjusting for the time value of use of exemptions and gift taxes paid, the death tax savings with a non-grantor trust are about $3,010,000, and with a grantor trust are about $4,712,000. That would be a 57-percent benefit in death tax savings by using a grantor trust over a non-grantor trust.
  11. More aggressive planning allows for the retained power that triggers grantor trust status to be turned off or on from year to year as the best interests of the various parties might dictate, a technique known as "toggling."
  12. There are some cautions and requirements in using a Code  §675(4)(B) power of substitution, to be sure. Under the statute, the power must also be "in a nonfiduciary capacity" and "without the approval or consent of any person in a fiduciary capacity." Treas. Reg.  §1.675-1(b)(4) provides that "the determination of whether the power [of substitution] is exercisable in a fiduciary or nonfiduciary capacity depends on all the terms of the trust and the circumstances surrounding its creation and administration." Revenue Ruling 2008-22 further makes clear that to escape any adverse estate tax consequence: (a) the trustee must have a fiduciary obligation (under local law or the trust instrument) to ensure the grantor's compliance with the terms of this power by satisfying itself that the properties acquired and substituted by the grantor are in fact of equivalent value; and (b) the substitution power cannot be exercised in a manner that can shift benefits among the trust beneficiaries. The ruling provides two safe harbors that satisfy (b), namely either (1) the trustee has the power (under local law or the trust instrument) to reinvest the trust corpus and a duty of impartiality with respect to the trust beneficiaries, or (2) the nature of the trust's investments or the level of income produced by any or all of the trust's investments does not impact the respective interests of the beneficiaries, such as when the trust is administered as a unitrust (under local law or the trust instrument) or when distributions from the trust are limited to discretionary distributions of principal and income. These requirements should almost always be easily met, although "equivalent value" under the (a) test is always a question of fact. The Tennessee Uniform Trust Code imposes a fiduciary duty to deal with any non-beneficiary third party (including the grantor) at arm's length, grants a Trustee power to reinvest trust corpus, and creates a duty of impartiality among beneficiaries. It would seem that a trust draftsman would have to add some provision negating part of the Tennessee Uniform Trust Code to run afoul of the requirements of the revenue ruling.

Dan Holbrook DAN W. HOLBROOK practices estate law with Holbrook, Peterson & Smith PLLC in Knoxville. He is certified as an estate planning specialist by the Tennessee Commission on Continuing Legal Education and Specialization, is a fellow of the American College of Trust and Estate Counsel, and serves on the TBA Probate Study Group reviewing and recommending legislation involving trusts and estates in Tennessee. He can be reached at dholbrook@hpestatelaw.com.