For a period of time, Tax Court decisions regarding family limited partnerships (FLPs) had a familiar read to them. A typical FLP case involved an assertion by the Internal Revenue Service that the transferor retained rights to transferred property that were impermissible under Internal Revenue Code Section 2036(a)(1), and the defense would stand or fall on whether the transfer was a bona fide sale for full and adequate consideration. Case analysis largely focused on what worked and didn’t regarding the estate’s ability to demonstrate legitimate non-tax purposes for the transfer. In 2017, Estate of Powell v. Commissioner raised new issues.1 While variety can be the spice of life—this case certainly left a bad taste for two reasons: (1) the revitalization of the IRS argument that rights held by a limited partnership interest could constitute impermissible control under IRC Section 2036(a)(2); and (2) a new method was used to compute the amount included in the gross estate when an FLP is included under Section 2036. After Powell, practitioners were left wondering whether this Section 2036(a)(2) argument would be routinely applied and if this new computational method would be applied again despite years of practice and precedent to the contrary. The Tax Court’s memorandum decision in Estate of Howard V. Moore v. Commissioner provided some insight on these issues as well as an important reminder to practitioners about what information in their files will come into evidence.2
Another Bad Facts FLP Case
Occasionally, there’s a good facts case regarding an FLP that informs practitioners of what they should advise their clients to do.3 More often though, the FLP cases that work their way to a decision in Tax Court are bad facts cases—and Howard Moore’s estate was a very bad facts case. It does, however, make for a good story and provides some useful lessons. Howard was the “rags to riches” American dream. Born in 1916 in Texas, Howard’s family lost their home and farm when Howard was a young child. The family moved to Arizona where the family had very little, and Howard’s education ended after eighth grade. Despite the rough start, Howard worked his way to financial prosperity. He was a land leveler in Arizona, where he leveled the ground for farm owners to help with irrigation. In exchange for this hard work, he was often paid in land as his customers had little cash. It was from this bartering (leveling land for land) that he amassed more than 1,000 acres over time, which he turned into his own farming operation.
Howard’s story wasn’t all positive though, as he had his personal struggles despite his financial success. He did get married and had four children (three sons and a daughter), but due to his battle with alcoholism, he became separated from his wife in the 1970s. His relationship with his children was complicated. At trial, his children described him as “strong,” “manipulative,” “firm” and “tough.” One anecdotal story included in the case recalled an incident when one of his sons came home injured from a school fight, only to have Howard tell him to fight harder, and the next time, he “[didn’t] want to see no skin on them knuckles.” Howard would also pit his sons against each other, which ultimately resulted in a total breakdown of the family. For example, in 1987, one son borrowed another son’s tractor without telling the owner. The son who owned the tractor became irate and unloaded an entire clip from a semiautomatic rifle into his own tractor. The damage to the tractor was minor compared to the irreparable damage to the Moore family.
After this 1987 blow up of the family (and tractor), an estate planner might see a prime opportunity to do estate planning. Howard’s wealth was tied up in an illiquid asset (an operating farm) that required active management.4 Further, with a dysfunctional family, a governance structure was needed to keep order and have proper succession of an operating farm.5 But, Howard’s controlling ways wouldn’t allow for proper estate planning. Instead, he carried on controlling every aspect of his operation. At one point, he made an unreported gift of farm land to one of his sons, but the land would effectively be returned to Howard before he died. In 2004, with Howard nearly a nonagenarian, he turned his attention towards liquidating his holdings. Howard and one of his sons had negotiated the sale of part of the land and had begun negotiations with a potential buyer for the rest of it.
In December 2004, before the deals to sell the land were finalized, Howard had a heart attack and a stroke. He was given less than six months to live, and Howard made clear he wanted to spend that time getting his affairs in order. On his metaphorical deathbed, though with full capacity, Howard decided to do estate planning. He called an attorney who had done Howard’s deceased wife’s estate plan and told him he wanted to try to “save the millions of dollars of taxes…” Howard then indicated during the “design phase” of the estate plan that his primary goals were to: (1) eliminate any estate taxes; and (2) maintain control. Just four days after being discharged from the hospital (still in December 2004)—Howard set up a partnership, a charitable foundation and some trusts.
The FLP Howard established in December 2004 had an initial funding of $10,000—$100 from each of his four children and $9,500 from a revocable trust he just established. These contributions accounted for the 99% limited partnership interest. Another trust he set up (the Management Trust) also contributed $100 for the 1% general partnership interest. In February 2005, Howard contributed the farmland into his revocable trust, which in turn contributed most of the farmland into the FLP. Simultaneously with this contribution, and despite the ownership structure in which the Management Trust should control the FLP, at Howard’s direction, all of the farmland was sold in exchange for $16.512 million from the previously identified buyers. As was customary in the area, Howard was allowed as part of the deal to continue to live on the land he just sold for the rest of his life. Also as part of the deal, he continued his farming operation after the sale, which wasn’t the local custom.
After the sale, an irrevocable trust established by Howard was gifted $500,000 as seed funding. The irrevocable trust then purchased the revocable trust’s FLP interest using the $500,000 it just received from the revocable trust and a promissory note for the balance of the purchase price. In terms of the purchase price, the trustee of the revocable trust (the seller in the transaction) testified that she didn’t know where the price came from. At trial, it came out that the price was based on the net asset value with a 53% valuation discount applied. After the partnership liquidated all of its real estate holdings, the assets were invested by a professional financial advisor, and the family was passive in its management of the partnership.
There’s more to this story and the case. There were purported loans to family members with no evidence that there was intent to repay. The IRS challenged the allowance of the attorney’s fees for estate administration purposes as it wasn’t clear what services he provided. There was a charitable deduction formula that was intended to eliminate all taxes if the planning was challenged but was ultimately held to be invalid by the court. While interesting, we’re not going to focus on those issues in this article. Instead, we’ll focus on the Sections 2036 inclusion and 2043 computation issues in relation to the FLP.
(1) is Moore Welcome Than (2)
One doesn’t really need to read past the description of the facts to know that the value of the FLP assets was going to be included in the gross estate by virtue of Section 2036—the only question is whether that value would be included under Section 2036(a)(1) or Section 2036(a)(2). So, if the taxpayer is going to lose on account of Section 2036, does it matter under which specific subsection the decedent’s planning falls? “When the fall is all that is—it matters,”6 and here planners can take some comfort in the Moore ruling. Historically, FLPs were often included under Section 2036(a)(1) because the decedent maintained some possession or right to income of the property. There’s also been a concerning development in which certain rights held by the interest can be considered an inclusionary power under Section 2036(a)(2) as “the right, either alone or in conjunction with any person, to designate the persons who shall possess or enjoy the property or the income therefrom.”7 In Estate ofStrangi v. Comm’r, there was clear inclusion under Section 2036(a)(1) as partnership assets were used to pay personal expenses of the donor, and the donor continued to use partnership property as his own. Nevertheless, the Tax Court went on to find that because the limited partnership interest held by the donor could give him a vote on the liquidation of the partnership, that also triggered Section 2036(a)(2).8 In Estate ofTurner v. Comm’r, the Tax Court found that Section 2036(a)(2) inclusion applied when the donor (and donor’s spouse) retained the general partnership interest because the donor had the ability to amend the partnership agreement and make distributions without the consent of the limited partners.9 In Turner, the donor also retained a majority of the limited partnership interests, thus controlling those rights as well.
More recently came the Powell decision. Like Moore, the Powell case involved a bad facts deathbed estate plan. Multiple statutory provisions could have been cited by the Tax Court to include the underlying FLP assets in Nancy Powell’s gross estate.10 Notwithstanding, the Tax Court relied on Section 2036(a)(2) in including FLP assets in the gross estate, focusing on the decedent’s retained right to vote as a limited partner on the liquidation of the FLP as impermissible control. This transformation of the Section 2036(a)(2) dicta from Strangi to the holding in Powell has given planners much to consider in both existing and new FLPs.11
The IRS asserted its Section 2036(a)(2) inclusion argument in the Moore case. However, the court first looked at the Section 2036(a)(1) issue—and there was a lot to look at. First, the fact that Howard continued to live and work the land in the same manner as he did prior to the contribution to the FLP and even after the sale was a clear retention of the possession and enjoyment of the property. Second, in addition to using the land, he used the sale proceeds to pay his own personal expenses. Third, although he put the general partnership interest into a trust—there was a clear implied agreement with the trustees (two of his children). The court described the power of the trustees as only “nominal ‘power’” as they did whatever Howard asked, and he had indicated maintaining control was a primary goal of his estate plan. Given the IRS prevailed on its Section 2036(a)(1) argument, the court found that it “need not address the Commissioner’s alternative arguments that Moore’s estate plan triggered their inclusion under section 2036(a)(2).”12 The holding in Moore certainly doesn’t undo the troubling analysis in Strangi or Powell, but at least it didn’t reinforce or further develop the Section 2036(a)(2) line of attack by the IRS.13
Keep It Bona Fide
Regardless of which subparagraph of Section 2036 the IRS is asserting, Moore is an important reminder of how practitioners must maintain their files and document their interactions with clients. With very few exceptions, taxpayers have been unsuccessful in a Section 2036 defense if they can’t meet the “bona fide sale for an adequate and full consideration” exception. Whether a transfer meets this exception depends on two questions. First, was the transfer for “adequate and full consideration?” This is a question of value (that is, did the taxpayer receive a partnership interest equivalent in value to what was contributed to the partnership). That’s usually an easy question to answer and satisfy. The second question is tougher, which is whether the sale was “bona fide,” and that question “turns on motive.”14 To satisfy the bona fide test, “the objective evidence must establish that the nontax reason was a significant factor that motivated the partnership’s creation.”15
Motive can be difficult to definitively determine in estate tax cases because the individual whose motive matters is deceased. Without the decedent’s direct testimony, the objective evidence that can be put forward to establish motive for creating an FLP can be divided into two categories: (1) evidence that was contemporaneous with the creation of the FLP; and (2) evidence of how the partnership was operated to achieve the decedent’s goals. In Moore, the objective evidence was pretty damaging, but the case serves as an important reminder of what not to do. As noted above, the attorney’s own records and testimony showed that tax was the primary factor underlying the planning with the second place finisher (maintaining control) being just as damning. The case is certainly an extreme example of what happens when the decedent did no estate planning for his entire adult life and then in just a few months engaged in a great deal of planning leaving clear evidence that the planning was tax driven.
Such blatant cases can serve as an important warning for planners to remember that communications with their clients can become discoverable. Practitioners must be proactive about discussing and documenting the non-tax benefits of FLPs. Transfer tax consequences certainly may be discussed, but why not also explain the income tax consequences in the same manner as transfer taxes? This may show that the practitioner was explaining all aspects of the planning to a client—not just pushing estate tax savings. Moore also highlights the importance of client intake notes and questionnaires. Having an item that welcomes a client to tell a planner, “I’m coming to you just to save taxes,” cuts the legs right out from under any bona fide sale defense. Communications with elderly clients or clients in poor health must detail the non-tax considerations of the planning because the Tax Court has repeatedly held that deathbed planning suggests the primary purpose was estate tax reduction.16
The second bona fide sale lesson from Moore is that reality needs to match the claimed purposes of an FLP, as a court won’t be inclined to find generic reasons to set up an entity persuasive. In Howard’s case, his children asserted two non-tax reasons. The first “was to bring his family together so that they could learn how to manage the business without him…”17 Establishing efficient and effective succession of management is certainly a valid non-tax reason that’s been recognized before. Unfortunately, the evidence didn’t support this was a valid reason in the Moore case. The children blindly signed the partnership agreement; there was no negotiating or independent counsel sought—they did as their father told them. Then there was no business to run and not even a business to sell as the deal for the farm was already negotiated by the time the farmland was contributed to the partnership. The partners only had one initial meeting, and an investment advisor managed the liquid assets after the sale was done. It was therefore easy for the court to dismiss business succession/family reunification as a valid business reason for the partnership.
The other non-tax reason claimed by the family was creditor protection. However, the court didn’t find objective evidence supporting this. Howard’s children couldn’t identify any actual or potential creditors. One son claimed that there was a risk of a claim because farmers routinely incur liability for illegal use of pesticides, yet the court noted there was no evidence Howard ever used such pesticides. The children also claimed a concern about potential “bad marriages” they had, but again, the court found that there was nothing in the record that supported that any of the children had or were having marital problems.18 The takeaway point here is that how a partnership operates is just as important as its structure in estate tax audits.
Further Exploring the Lacuna
One of the most (if not the most) troubling parts of the Powell decision was an introduction of a new way to compute the amount included in a decedent’s gross estate when an FLP is included under Section 2036. Historically, if there was inclusion of an FLP interest, the value of the FLP itself would be disregarded, and the value included would be that of the underlying assets without the benefit of any valuation discounts. This would avoid double counting an asset by including the FLP interest under Section 2033 and the underlying assets under Section 2036. While the parties in Powell contested whether there was estate tax inclusion, neither party raised what computation would be used if there was inclusion because the historical approach had longstanding precedent and produced an equitable result.
Nevertheless, the majority in Powell found that the logic behind the avoidance of double inclusion had gone “unarticulated” and that the case presented an opportunity to “fill that lacuna and explain why a double inclusion in a decedent’s estate is not only illogical, [but] it is not allowed under IRC Section 2043(a).”19 Given that the Tax Court raised the issue on its own, it didn’t have the benefit of competing arguments that undoubtedly would have exposed the problems with a new analytical framework and the fact that the court had previously found that the application of Section 2043(a) was inappropriate.20 There was a concurring opinion in Powell that accepted the Section 2036(a)(2) inclusion but disagreed with creating a new computation of the inclusion as it was “a solution in search of a problem.”21 Ultimately in Powell, the new analytical framework had no impact in the case given its unique facts, but both the majority and concurring opinions noted that there could be duplicative tax results in future cases in which this framework is applied.22
In Moore, Judge Mark V. Holmes found that the Powell computation needed to be applied as the court was required to apply the IRC as interpreted by a full Tax Court opinion despite the potential for what he called “odd results.” While the decision to give further life to the Powell analysis is unwelcome, the Moore opinion is useful because it delves into the computational detail in a way Powell didn’t. The Moore opinion presents an algebraic formula that ties together the various IRC sections in play. At the end of the day, we’re looking for the total value included in an estate (Vincluded). Prior to Powell, this was simply the fair market value (FMV) of the underlying partnership assets as of the decedent’s date of death (FMVd), with this value being included because of Section 2036. Thus, the equation was simple: Vincluded = FMVd.
What Powell effectively did was add two additional variables. The first is the FMV of the partnership as of the date of death. It’s included in determining Vincluded because Section 2033 would include the decedent’s holding in the partnership interest itself (this is in addition to the underlying assets being included under Section 2036). The Moore opinion referred to this as “Cd” because it was for the consideration (partnership interest) the decedent received when he contributed assets to the partnership valued at the date of death. Including both the underlying assets (FMVd and Cd) creates a double inclusion issue. This is where Section 2043(a) comes in according to Powell. The second variable is a reduction for the value of the partnership interest as of the date of the transfer of assets to the partnership (Ct), which is meant to address double inclusion. The Moore opinion formulated the Powell inclusion computation as: Vincluded = Cd + FMVd - Ct. Though many of us became lawyers because we didn’t want to do math—certainly this succinct formula more accurately captures the Powell computation than prose.
While the formula as stated in Moore is helpful, the ordering of the variables doesn’t drive home the change to the historical approach, so let’s reorder them. Using algebra, what you can end up with is Vincluded = FMVd + (Cd-Ct). What this reordering highlights is that, just as had been the historical computation, the value included is the date-of-death value of the underlying assets, but is now adjusted by the change in the FMV of the consideration received between the date of the transfer and the date of death. To the extent the value of the partnership appreciated (meaning Cd will be greater than Ct), then this new computational method will result in a greater value being included in the estate. To the extent the value of the partnership depreciated (meaning Cd will be less than Ct), then this new computational method will result in a lower value being included in the estate because the resulting number in the parenthesis will be negative.
To the extent that the value of the consideration stayed flat, then the number in the parenthesis will be zero, which means the outcome under the Powell formula (that is, Vincluded = FMVd + (0)) will be the same as the historical approach Vincluded = FMVd. This was the case in Powell, in whichthe partnership interest values at the date of transfer and death were the same given the deathbed planning. As is apparent, the more time that passes, the greater the probability that the Cd and Ct values will be different. The Moore decision illustrated this with hypotheticals, in which in all three cases, a decedent started with contributing $1,000 in land and getting a partnership interest back for $500 (that is, a 50% valuation discount). In one scenario, the values remain stagnant, in another the value of the land and in turn the partnership interest double and finally there’s a scenario in which the property and in turn the partnership interest have their values halved. See “Powell Formula,” this page.
What can be seen from the Powell formula are results that don’t necessarily achieve the intended policy objectives of Section 2036, which the pre-Powell computation did handle appropriately. If the value of the partnership increases, double counting value will result, as the date-of-death value of the partnership assets plus post-contribution appreciation of the partnership interest will be included in the decedent’s estate. Section 2036 wasn’t meant to be draconian—like all of the string provisions, it’s simply meant to undo a structure in which the decedent retained impermissible control or access to the property. Further, if the value of the partnership decreases, then the decedent is in better shape from an estate tax perspective had he not entered into the transaction—the exact opposite outcome Section 2036 seeks to achieve.
As is likely apparent, the more time that passes between the date of contribution to the partnership and the date of death, the greater the probability that Cd won’t equal Ct, especially in an active partnership (particularly one that makes distributions). Using an example given in Moore, a father contributed $1,000 of property to a partnership for a limited partnership interest while his son holds the general partnership interest, which creates a 25% valuation discount. The FLP then sells the land for $1,000 (so appreciation/depreciation doesn’t impact the hypothetical) and distributes $400 to the father, which leaves $600 of value in the FLP interest owned by the father. Before the Powell computational method, $1,000 would be included in the estate—$400 of the cash that was distributed out and the remaining $600 in the FLP with no valuation discount applied to it. Thus, the policy objective of Section 2036 is achieved when the father is in no better or worse a position after the application of Section 2036. The Powell method produces a different outcome though because Cd will be greater than Ct because the $400 of cash distributed out will be part of Cd with no discount applied, while that same property included in Ct has 25% discount applied. The result under the Powell formula is that Vincluded is $1,100. This occurs because FMVd is $1000, Cd is $850 ($400 of the cash distributed and the remaining $600 of underlying assets being reduced by the 25% valuation discount), and Ct is $750 (the $1,000 of assets prior to the cash distribution with a 25% valuation discount). It’s the $100 difference between Cd and Ct that creates the extra $100 of value when compared to if the father never entered into the FLP structure or did so but didn’t take a distribution.
While Moore touches on the impact of discounts in the context of whether FLP assets are distributed or not, there’s another issue with discounts not discussed. The examples in Moore all assumed the discount rates would remain the same even as the value of property fluctuated or the nature of the assets of the partnership changed (for example, a real estate holding company becoming liquid). Increasing and decreasing discount rates that occur between the date of the transfer and date of death could create similar fluctuations as changing values. If discount rates increase from the transfer to death, then Cd would become lower than Ct, which would reduce the amount included. Conversely, decreasing discount rates would result in Ct being higher relative to Cd.
Planning With Section 2043
After the decision in Moore, it appears that the Tax Court will routinely implement the new Powell formula. This means that for FLPs that are successful and appreciate, the total cost of Section 2036 inclusion has gone up so that a client will be worse off compared to the result if he hadn’t done the planning. Planners should reconsider the arithmetic before trying hastily implemented FLP planning when the client is on his deathbed, as an estate-planning Hail Mary (as was the case in Powell and Moore) because the old thinking that one is no worse off for doing the planning is no longer accurate. This should also give planners concerns about existing partnerships that have appreciated over time and whether that planning should be revisited, potentially ending the FLP, to avoid the draconian result.
There are also short-term possibilities in the current environment given the impact of COVID-19. The impact of the pandemic has driven asset values down relative to their worth before this crises. Correspondingly, discounts for lack of marketability and control are higher because the factors and benchmarks used to determine those rates have been impacted by COVID-19. This could result in scenarios in which Ct is higher relative to Cd, which as discussed above can result in the estate paying less in taxes had the planning never been undertaken. It’s certainly possible that an estate may wish to assert Section 2036 against itself for an FLP the decedent contributed to because the math could work out under the Powell computational method that would reduce the overall tax burden being lowered when compared to the FLP interest being outside of the estate. So much for going into law to avoid math.
—The authors would like to thank Jeremy Lent, a consultant at Withers Bergman LLP in New Haven, Conn., for his assistance with this article.
Endnotes
1. Estate of Powell v. Commissioner, 148 T.C. 392 (2017).
2. Estate of Moore v. Comm’r, T.C. Memo. 2020-40.
3. See, e.g., Estate of Purdue v. Comm’r, T.C. Memo. 2015-249.
4. See, e.g., Church v. United States, 85 A.F.T.R.2d 804 (W.D. Tex. 2000), aff’d 268 F.3d 1063 (5th Cir. 2001) (holding that operating a family ranch was a valid business purpose).
5. See, e.g., Estate of Stone v. Comm’r, T.C. Memo. 2003-309 (holding that using a family limited partnership (FLP) to settle disputes among quarrelsome siblings was a valid business purpose).
6. “The Lion in the Winter,” AVCO Embassy Pictures and Haworth Productions (1968).
7. For a discussion of potentially problematic powers in an FLP from an Internal Revenue Code Section 2036 perspective, see Chanie S. Fortgang and Christine R.W. Quigley, “Help for Control Freaks,” Trusts & Estates (November 2014).
8. Estate of Strangi v. Comm’r, T.C. Memo. 2003-145, aff’d 417 F.3d 468 (5th Cir. 2005).
9. Estate of Turner v. Comm’r, T.C. Memo. 2011-209. For more analysis of the Turner decision, see Stephanie Loomis-Price and N. Todd Angkatavanich, “Turn(er)ing the Tables on Taxpayers,” Trusts & Estates (July 2012).
10. Internal Revenue Code Section 2033 (if transfer of FLP interest was void under an invalid exercise of a power of attorney); IRC Section 2038(a) (if transfer of FLP interest was voidable under an invalid exercise of a power of attorney); IRC Section 2035(a) (as estate conceded the bona fide sale exception would apply and retained interests were terminated less than three years from death); and Section 2036(a)(1) (because of an implied agreement).
11. For more on this issue, see N. Todd Angkatavanich, James I. Dougherty and Eric Fischer, “Estate of Powell: Stranger than Strangi and Partially Fiction,” Trusts & Estates (September 2017).
12. Moore, supra note 2, at n. 17.
13. Note that this expansion of Section 2036(a)(2) isn’t limited to FLPs, as the Internal Revenue Service has successfully asserted it in other situations that are commonly used in estate planning. See Estate of Cahill v. Comm’r, T.C. Memo. 2018-84.
14. Moore citing Estate of Bongard, 124 T.C. 8 (2005).
15. Moore quoting Estate of Turner v. Comm’r, T.C. Memo. 2011-209.
16. Estate of Rector v. Comm’r, T.C. Memo. 2007-367; Estate of Erickson v. Comm’r, T.C. Memo. 2007-107; Estate of Rosen v. Comm’r, T.C. Memo. 2006-115.
17. Moore, supra note 2, at p. 2.
18. Query whether the only marital problems resulted from any of the children saying under oath they had marital issues.
19. Powell, supra note 1, at p. 410.
20. See Estate of Harper, T.C. Memo. 2002-121.
21. Powell, supra note 1, at p. 424.
22. For more on this issue with Powell, see supra note 11.