TIFD III-E INC. v. U.S.

MEMORANDUM OF DECISION

TIFD III-E Inc. ("TIFD III-E") has sued the United States ofAmerica to recover approximately $62 million that TIFD III-Edeposited with the Internal Revenue Service ("I.R.S.") insatisfaction of an alleged tax liability. That tax liabilityarose from the I.R.S.'s determination that TIFD III-E hadincorrectly calculated and reported the amount of income TIFDIII-E earned as a partner in Castle Harbour-I Limited-LiabilityCompany ("Castle Harbour"). The case was tried to the court overeight days. The court's findings of fact and conclusions of laware set forth below.

I. Procedural History

Castle Harbour was a Nevada limited liability company. In everyyear from 1993 to 1998, it filed a partnership tax return, alsoknown as a Form 1065. (J. Exs. 22, 40, 50, 56, 59, 70) TIFDIII-E, a Delaware corporation, was one of the owners of CastleHarbour. Because Castle Harbour was treated as a partnership fortax purposes, TIFD III-E paid United States tax on the incomeattributed to it on Castle Harbour's Form 1065.

In 2001, the I.R.S. issued two notices of Final PartnershipAdministrative Adjustments ("FPAAs") concerning Castle Harbour.(Pl.'s Exs. 377, 378) The FPAAs attributed approximately $310 million of additional income to TIFD-III E, resulting in anadditional tax liability of $62,212,010.

Pursuant to section 6226(a) of the Internal Revenue Code("I.R.C."), TIFD III-E filed a complaint against the UnitedStates challenging the FPAAs. Before filing, TIFD III-E depositedthe disputed sum of $62,212,010 with the I.R.S. as required byI.R.C. section 6226(e). Because it is the tax matters partner ofCastle Harbour, TIFD III-E is authorized to bring this suit.I.R.C. § 6226(a)

II. Findings of Fact

On the basis of the testimony, exhibits, and designateddepositions, I make the following findings of fact.

A. Background

TIFD III-E is a wholly owned subsidiary of the General ElectricCapital Corporation ("GECC"), a subsidiary of the GeneralElectric Company ("GE"). Among other things, GECC1 is inthe business of commercial aircraft leasing. (July 21, O'Reilly,48; July 21, Lewis, 134)2 Typically, airlines do not ownaircraft, principally because airlines do not ordinarily producesufficient income to take advantage of the tax depreciationdeductions generated by commercial aircraft. (July 21, Brickman,204-05) Instead, a company with greater taxable income, such asGECC, will buy the planes and lease them to airlines, therebygiving the airlines the use of the aircraft and the lessorcompany the tax deductions. (July 21, Brickman, 204-05) In the early 1990s, the airline industry experienced a numberof setbacks, including several bankruptcies. (July 21, O'Reilly,49; July 21, Lewis, 137; July 22, Dull, 299; July 22, Parke,385-86; July 23, Nayden, 451) These events caused GECC concernfor its own prospects in the aircraft leasing business. (July 21,O'Reilly, 66-72; July 21, Lewis, 141-42; July 22, Parke, 386) In1992, at least partially in response to this concern, GECCexecutives began looking for ways to reduce GECC's risk in theaircraft leasing business. (July 21, Lewis, 155-56) To do this,GECC initiated what it referred to as a "sell-down" effort — anattempt, among other things, to raise immediate cash againstGECC's aircraft assets. (July 21, O'Reilly, 77-78; July 21,Lewis, 158-59; July 26, Nayden, 455-56) In other words, ratherthen simply awaiting return in the form of the — now less certain— rental income, GECC wanted to lower its risk by raisingimmediate cash against that future stream of income.

Selling the aircraft or borrowing money against them, twostraightforward ways of raising capital, were not options. Salewas not a realistic option because, in general, the secondarymarket for aircraft was weak. (July 21, O'Reilly, 70) This wasparticularly true with respect to GECC's older aircraft — knownas "Stage II" aircraft — which did not meet certain regulatorystandards, including those for noise. (July 21, O'Reilly, 50;July 22, Dull, 308) Non-recourse debt was not an option for tworeasons. First, in order to maintain its AAA credit rating, GECChad an agreement with credit rating agencies that prohibited GECCfrom borrowing more than eight times its common equity. (July 21,O'Reilly, 95-96; July 22, Parke, 379) In 1993, GECC'sdebt-to-common-equity ratio was 7.96 to 1, giving it little roomto borrow. (July 22, Parke, 381) Second, a number of GECC'smedium-term and long-term debt instruments contained a "negativepledge" — a covenant prohibiting GECC from using its assets tosecure debt other than purchase money debt. (July 21, O'Reilly,96; July 22, Parke, 384)

With the two most obvious options for raising money off thetable, GECC sought outside advice concerning other possibilities.In May 1992, GECC submitted Requests for Proposal ("RFPs") toseven investment banks. (Pl.'s Exs. 141, 142, 143, 146; J. Exs.8, 10) The RFPs sought proposals that would, in essence, allowGECC to solicit outside investment in at least part of itsaircraft leasing business. All of the investment banks submittedproposals; none of them met all of GECC's objectives. (July 21,O'Reilly, 88-89) Nevertheless, in March 1993, after some back andforth, the investment bank Babcock & Brown submitted a revisedproposal that GECC found acceptable. (July 21, O'Reilly, 90)Babcock & Brown eventually received a $9 million fee for itswork. (July 21, Brickman, 225)

Babcock & Brown's final3 proposal called for thecreation of a separate entity to which GECC would contribute anumber of aircraft. (D.'s Ex. 22) Investors would then besolicited to purchase ownership shares in the new entity. (J. Ex.17) The result would be that GECC would trade some of the risksand returns of those aircraft to the outside investors inexchange for a cash contribution to the newly created entity. Theproposal also called for the investors to be foreign tax-neutralentities, an arrangement that would offer lucrative tax savingsto GECC.

After GECC senior management approved the proposal, it wasimplemented in two stages.

First, on July 26, 1993, GECC formed a Nevada limited liabilitycompany known as GE Capital Summer Street-I Limited LiabilityCompany ("Summer Street"), which was owned by three GECC subsidiaries: TIFD III-E, TIFD III-M, and General ElectricCapital AG. (Pl.'s Ex. 118; July 22, Dull, 315) Through thesesubsidiaries, GECC contributed to Summer Street: (a) 63 "StageII" aircraft worth $530 million, but subject to $258 million ofnon-recourse debt (a net value of $272 million);4 (b) $22million of rents receivable on the aircraft; (c) $296 million incash; and (d) all the stock of GECC subsidiary TIFD VI, which hada value of $0. (July 22, Dull, 314-15, 331)

Second, on October 6, 1993, the GECC subsidiaries sold $50million of their interest in Summer Street, which included all ofGE Capital AG's interest, to two Dutch banks, ING Bank N.V. andRabo Merchant Bank N.V. (collectively, the "Dutch Banks"). (July22, Dull, 322-23) The Dutch Banks also contributed an additional$67.5 million, bringing their total investment to $117.5 million.(July 22, Dull, 322-23) Summer Street then changed its name toCastle Harbour-I Limited Liability Company ("Castle Harbour")(Pl.'s Exs. 118, 119), and TIFD VI changed its name to CastleHarbour Leasing, Inc. ("CHLI") (July 22, Dull, 331)

B. Structure of Castle Harbour

When the final stage of the Castle Harbour transaction wascompleted on October 6, 1993, the parties entered into an Amendedand Restated Operating Agreement ("the Operating Agreement"). (J.Ex. 1) The Operating Agreement dictated the way gains and losseswere allocated between TIFD III-E, TIFD III-M (collectively the"GECC entities"), and the Dutch Banks. Consequently,understanding the terms of the Operating Agreement is essentialto understanding the tax consequences at issue in thislitigation. 1. Overview

Castle Harbour was a self-liquidating partnership. Through twoentities,5 GECC contributed to Castle Harbour a net $246million in cash6 and, more importantly, approximately$294 million worth of leased aircraft.7 The Dutch Bankscontributed approximately $117.5 million in cash. Each partnerreceived an allocation of the net income of the partnership. TheDutch Banks were referred to as Class A partners, the GECCentities as Class B partners. (J. Ex. 1 at Bates 101430-31) Overeight years, the Dutch Banks' ownership interest was to be almostentirely bought out with the income of the partnership. Thisself-liquidation mechanism can best be understood by consideringan actual year in the partnership.

Castle Harbour's first full year of operation was 1994. At thebeginning of 1994, each partner's ownership interest — recordedin a "capital account" entry on Castle Harbour's financialstatements (J. Ex. 1 at Bates 101403-31) — was approximately thesame as at the start of the partnership. GECC's capital accountwas approximately $540 million.8 (J. Ex. 24 at Bates8812) The Dutch Banks' combined capital accounts wereapproximately $112 million.9 (J. Ex. 24 at Bates 8812) In 1994, Castle Harbour received approximately $100 million ingross income — mostly rent from aircraft leases.10 Inaccordance with the Operating Agreement, the net income for 1994,approximately $9.8 million, was distributed $9.6 million to theDutch Banks and $0.2 million to GECC. (J. Ex. 37 at Bates 16344)(The mechanics of net income calculation and allocation arediscussed below.) Accordingly, the Dutch Banks' combined capitalaccounts increased to approximately $122 million, and the GECCentities' accounts increased negligibly, remaining at around $540million.

The Dutch Banks' capital accounts, however, did not actuallyincrease because part of the gross rent of $100 million was usedto "buy-out" approximately $40 million of the banks' combinedownership interest. (J. Ex. 37 at Bates 16344) Thus, at the endof 1994, the Dutch Banks' capital accounts, originally at $112million, had increased by $10 million (allocation) but decreasedby $40 million (buy-out), resulting in final combined capitalaccounts of approximately $82 million.11

This general pattern was to be repeated for eight years. Eachyear the Dutch Banks were to have their capital accounts debitedor credited, depending on whether the partnership had received again or suffered a loss, and each year the Dutch Banks were tohave a significant portion of their ownership interest bought outby the partnership. The amount of the annual buy-out payment wasset forth in the Operating Agreement at Exhibit E, giving rise tothe name "Exhibit E payments."12 (J. Ex. 1 at Bates 101662) The Exhibit E payments were scheduled to paycash annually in amounts that would provide the Dutch Banks withan internal rate of return13 of 9.03587% over eightyears. At the end of eight years, if the Dutch Banks' capitalaccounts had actually earned a rate of return 9.03587%, the DutchBanks' capital accounts, i.e., ownership interests, would bedecreased to near zero — in other words, Exhibit E payments wouldhave cancelled out the Dutch Banks' capital account increases andreturned the Dutch Banks' initial investment. Similarly, if theDutch Banks' capital accounts were credited with partnershipincome at a rate less than 9.03587%, the capital accounts wouldbe negative after eight years; if the capital accounts werecredited at a rate greater than 9.03587%, the capital accountswould be positive. Positive capital accounts would result inpayments to the banks when the partnership wound up; negativeaccounts would mean the banks owed money to the partnership. (Op.Agmt. § 12.2-12.3, J. Ex. 1 at Bates 101492-93). If the banks'interests were not liquidated after eight years, the banks wouldstill have their capital accounts credited or debited byallocations of income or loss in successive years.

This arrangement provided the Dutch Banks, in return for theirinitial cash investment, with (a) an ownership interest in CastleHarbour that was increased or diminished by allocations of incomeor loss and (b) a stream of payments over eight years that wouldrepay the Dutch Banks' initial investment at an internal rate ofreturn of 9.03587%. Any discrepancies between these two items —the Dutch Banks' ownership shares and the total payments made tothe banks over the eight years — would be reconciled upon liquidation of the banks' partnership interests.

The arrangement allowed GECC to reduce part of its interest infixed assets to cash, i.e., it "monetized" or "securitized" partof the assets. Before the formation of Castle Harbour, GECC —through its wholly owned subsidiaries — owned 100% of its $272million of aircraft (net), $22 million in associated rentreceivables, and $240 million in cash. After the formation ofCastle Harbour, GECC owned approximately 82% of a partnershipthat itself owned the aircraft, receivables, GECC's cash, and anadditional $117.5 million in cash. In other words, pre-CastleHarbour, GECC owned 100% of $272 million in aircraft, 100% of $22million in receivables, 100% of $240 million cash, and 0% of$117.5 million cash. Post-Castle Harbour, GECC owned 82% of eachasset, i.e., $223 million of aircraft, $18 million ofreceivables, and $290 million in cash.14 Over the firsteight years, however, while the Dutch Banks' ownership interestwas being bought out using Castle Harbour's income, GECC'sinterest was increasing inversely. Consequently, at the end ofthe eight years (had the partnership lasted that long), GECCwould have regained almost complete ownership of the aircraft.

In sum, through Castle Harbour, GECC received $117.5 million incash from the Dutch Banks, in return for which the banks receiveda self-liquidating (i.e., limited-term) ownership interest inaircraft and an allocation of the rental income from thoseaircraft. Put another way, GECC sold the Dutch Banks whatamounted to an eight-year investment in a commercial leasingpartnership. 2. Allocations

The general structure of Castle Harbour just described isfairly simple and perhaps not very different from that of a greatmany partnerships. The complexity of the transaction, and thesource of this litigation, comes principally from the way inwhich the partnership's income was allocated between the GECCentities and the Dutch Banks.

Crucial to its allocation scheme, the Operating Agreementdefined two categories of income (or loss): OperatingIncome15 and Disposition Gains/Losses.

a. Operating Income

Operating Income was comprised of income less expenses. Incomewas rent and interest on investments. Expenses consisted ofnormal administrative expenses, interest owed on aircraft debt,depreciation of the aircraft, and guaranteed payments to GECCentities, also known as Class B Guaranteed Payments. (J. Ex. 1 atBates 101425-26) Depreciation was calculated on a straight-linebasis, starting with the fair-market value of the aircraft at thetime of Castle Harbour's formation and assuming a useful life foreach aircraft of the greater of 7 years or 125% of the timeremaining on the then outstanding lease of the aircraft. (J. Ex.20 at Bates 8785) The Class B Guaranteed Payments were madeannually to the GECC entities in the amount of either $500,000 or$2 million and did not reduce the capital account of thereceiving GECC entity. (Op. Agmt. § 4.1, J. Ex. 1 at Bates101447) Notably, expenses did not include payment of principal onthe airplane debt or Exhibit E payments to the Dutch Banks. Once Operating Income had been calculated, it was allocated tothe capital accounts as follows. If Operating Income waspositive, i.e., an Operating Gain, it was allocated 98% to theDutch Banks and 2% to the GECC entities. (Op. Agmt. § 3.1, J. Ex.1 at Bates 101435) If Operating Income was negative, i.e., anOperating Loss, then it was (a) first allocated in an amountsufficient to offset the cumulative Disposition Gains allocatedto any of the partners in previous years, (b) the remainder wasthen allocated 98% to the Dutch Banks until they had beenallocated, cumulatively, $3,854,493 of Operating Losses, and (c)the remainder was allocated 99% to the GECC entities and 1% tothe Dutch Banks.16 (Op. Agmt. § 3.2, J. Ex. 1 at Bates101435-36)

Operating Gain allocation was straightforward. For example, in1997 Castle Harbour's net Operating Income was $2,304,000, whichwas allocated $2,258,000 to the Dutch Banks and $46,000 to theGECC entities — a simple 98/2 split. (J. Ex. 61 at Bates 18002)

Castle Harbour never experienced an Operating Loss, but ahypothetical situation will illustrate how one would have beenallocated. Assume that in 1994 Castle Harbour had an OperatingGain of $10 million and a Disposition Gain (described below) of$10 million. The Operating Gain would have been allocated $9.8million to the Dutch Banks and $200,000 to the GECC entities. TheDisposition Gain, assuming it was Castle Harbour's first ever,would have been allocated (for reasons explained later)approximately $3 million to the Dutch Banks and $7 million toGECC. If in 1995 Castle Harbour had a — very unlikely — OperatingLoss of $15 million, it would have been allocated as follows.First, $3 million would be allocated to the Dutch Banks and $7 millionto the GECC entities to offset their prior Disposition Gains.Second, the remaining $5 million would be allocated 90% to theDutch Banks and 10% to the GECC entities until the Dutch Bankshad been allocated $3,854,493 of cumulative Operating Incomelosses. Assuming the Dutch Banks in previous years had never beenallocated an Operating Loss, this second step would mean that$4,282,770 would be allocated 90/10, resulting in $3,854,493going to the Dutch Banks and $428,277 going to the GECC entities.Third, the remaining approximately $717,000 would be allocated99% to the GECC entities and 1% to the Dutch Banks.

b. Disposition Gain/Loss

A Disposition Gain or Loss was the result of the differencebetween the sale price of an asset, usually an aircraft, and itsbook value. (Op. Agmt. § 3.3(h), (j), J. Ex. 1 at Bates 101438-39)Alternatively, if an aircraft was distributed back to one of theGECC entities, the fair market value was deducted from thereceiver's capital account and the difference between theaircraft's fair market value at the time of distribution and itsbook value was treated as a Disposition Gain or Loss. (Op. Agmt.§ 10.8(a)(i)(B), J. Ex. 1 at Bates 101488) Similarly, if the GECCentities were to buy out entirely the Dutch Banks' interest inthe partnership, the difference between the fair market value ofall held assets and their book value was to be treated as aDisposition Gain or Loss. (Op. Agmt. § 10.8(b)(i)(B), J. Ex. 1 atBates 101488)

Disposition Gains and Losses were allocated much like OperatingLosses: (a) first, Disposition Gains were allocated to offsetprior Disposition Losses and prior Operating Losses;Disposition Losses offset prior Disposition Gains,17 (b) theremainder was then allocated 90% to the Dutch Banks until theyhad been allocated, $2,854,493 of either Disposition Gains orLosses, (c) the remainder was allocated 99% to the GECC entitiesand 1% to the Dutch Banks.18 (Op. Agmt. § 3.3(h), (j), J.Ex. 1 at Bates 101438-41)

For example, in 1995 Castle Harbour disposed of a number ofaircraft, some to TIFD III-E and some to third parties. Theaircraft distributed to TIFD III-E had a fair market value ofapproximately $27 million, and consequently TIFD III-E's capitalaccount was reduced by that amount. (J. Ex. 49 at Bates 8969) Theaircraft sold to third parties were sold for approximately $21million. (J. Ex. 49 at Bates 8969) The book value of all theseaircraft was approximately $74 million, causing Castle Harbour aDisposition Loss of approximately $26 million.19 (J. Ex.49 at Bates 8969) That loss was distributed as follows. First,because in prior years cumulative Disposition Gains ofapproximately $3 million had been allocated to the Dutch Banks,and cumulative Disposition Gains of approximately $1.5 millionhad been allocated to the GECC entities (J. Ex. 37 at Bates16347), those amounts of loss were allocated to each respectively.20Second, the Dutch Banks were allocated 90% of the remaininglosses and GECC, 10%, until the Dutch Banks had been allocated$2,854,493 in losses, and GECC had been allocated about $0.3million. Third, the remainder was allocated 99% to GECC(approximately $18 million) and 1% to the Dutch Banks(approximately $0.1 million). In total, therefore, the DutchBanks were allocated approximately $6 million in DispositionLosses,21 and the GECC entities, approximately $20million.22 (J. Ex. 49 at Bates 8970)

c. Operating Income vs. Actual Income

Operating Income, as defined by the Operating Agreement, mightat first glance look like a simple measure of the net cashreceived by Castle Harbour in its normal operations, i.e., grossnondisposition income less expenses. That was not the case.Operating Income in fact defines a nonobvious category of income,primarily because it includes as expenses items not clearlyconsidered expenses, e.g., Class B Guaranteed Payments, andexcludes items that appear to be expenses, e.g., debt paymentsand Exhibit E payments. It is worth pointing out some of theeffects of this definition.

As noted above, the Class B Payments guaranteed to the GECCentities were treated as expenses and did not reduce thereceiver's capital account. Consequently if one were to considersuch payments as allocations — which would make sense given thatthey represented income going directly to GECC entities23 — the income allocated to GECC would thenbe significantly more than 2%. For example, in 1997, as discussedabove, Operating Income was approximately $2.3 million, and wasallocated 98% to the Dutch Banks and 2% to GECC. If one considersthat GECC also received a Class B distribution of $2,000,000prior to any allocation and not deducted from its capitalaccount, it might make sense to consider its allocation to be$2,046,000, making the actual split closer to 50/50.

The treatment of aircraft depreciation as an expense, and itsconsequent deduction from Operating Income, also had someinteresting effects. The depreciation schedule for the aircraftwas fairly aggressive, usually coming out to between 60 to 70percent of the rental income for a given year. The effect of thisdepreciation was that a large portion of the cash that came intoCastle Harbour was not reflected in Operating Income.24Of course, aggressive depreciations meant Castle Harbour would bemore likely to realize a gain when the assets were sold (or theDutch Banks were bought out), but this gain would be aDisposition Gain and therefore allocated more favorably toGECC.

3. Other Provisions

The principal features of the Castle Harbour transaction havebeen described: it was a selfliquidating partnership with acomplex scheme for allocating gains and losses. There are threeother features of the partnership relevant to this case. a. CHLI

Most of the cash invested in Castle Harbour was not held byCastle Harbour. Instead it was transferred to Castle HarbourLeasing Inc. ("CHLI"), a domestic corporation and wholly ownedsubsidiary of Castle Harbour. (July 22, Dull, 322-23, 358; July22, Parke, 394) This arrangement allowed income from any asset(cash or aircraft) to be recognized only as a Disposition Gainrather than as Operating Income, simply by moving that asset toCHLI. For example, interest earned on the reinvestment of theapproximately $360 million cash initially invested by thepartners would have counted as Operating Income had the cash beenheld in Castle Harbour. Because the cash was moved to CHLI,interest accumulated there and was allocated to the partners as aDisposition Gain when the Dutch Banks were ultimately boughtout. Similarly, CHLI purchased several aircraft during CastleHarbour's existence. (July 22, Dull, 325) Rental income generatedby those aircraft did not count towards Operating Income, as itwould have if the aircraft were owned by Castle Harbour, and thatincome was not allocated to the partners until the buyout of theDutch Banks, when it was allocated as a Disposition Gain. (July22, Dull, 370)

b. Investment Accounts

Under the Operating Agreement, Castle Harbour was required tomaintain "Investment Accounts" for the Dutch Banks. (J. Ex. 1 atBates 101405-06) No cash was actually paid into these accounts;they merely kept track of a hypothetical balance. Id. Theopening balance of these accounts was the initial investment madeby the Dutch Banks. Id. That balance was to be recalculated atthe time the Dutch Banks exited the partnership as if every yearthe balance had been increased by a defined Applicable Rate butalso reduced by the Exhibit E payments. Id. The Applicable Ratewas either 9.03587%25 or 8.53587%, depending on the reasonfor the Dutch Banks' exit. Id.

If at the time of the Dutch Banks' exit from Castle Harbour theInvestment Account balance exceeded the algebraic sum of theDutch Banks' allocation of (a) Operating Gains, (b) OperatingLosses that had been allocated at the 98% rate, which could notexceed approximately $4 million, (c) Disposition Gains, and (d)Disposition Losses that had been allocated at the 90% rate, whichcould not exceed approximately $3 million, that amount would bepaid to the Dutch Banks, instead of the amount in their capitalaccounts. That payment, if made, was labeled a Class A GuaranteedPayment. Id.

c. Core Financial Assets

As discussed above, Castle Harbour put most of its cash in itssubsidiary, CHLI. CHLI was not free to dispose of that cash as itchose. Under the Operating Agreement, CHLI was required to keephigh-grade commercial paper or cash, referred to as "CoreFinancial Assets," in an amount equal to 110% of the currentvalue of the Investment Accounts. (J. Ex. 1 at Bates 101408; Op.Agmt. § 5.8(b), J. Ex. 1 at Bates 101471-72)

As it turns out, CHLI ended up investing most of its $360million in GECC commercial paper. (July 22, Dull, 324) Thisbenefitted GECC because, by having a GECC subsidiary — CastleHarbour — purchasing GECC commercial paper, GECC was buying backor "retiring" its debt, thereby decreasing its debt-to-equityratio and freeing GECC to borrow more money. (July 22, Parke,393) d. Management Rights

Castle Harbour was managed primarily by the GECC entities whoelected the partnership's managers. The Dutch Banks' role wasminimal. They did not vote for managers26 and none oftheir employees worked for Castle Harbour. They did participatein annual member meetings (July 22, Dull, 362), but, for the mostpart, the only control they had was negative (July 22, Dull,361).

The actual day-to-day operations of Castle Harbour, such asfinancing and accounting activities, were outsourced to other GEentities — first, GE Capital Advisory Services Ltd., then,General Electric Capital Aviation Services, Ltd. ("GECAS Ltd.").(July 22, Dull, 412-13; July 26, Tewell, 646-47)

4. Risks and Returns

Having explained Castle Harbour's structure, it is worthsummarizing how this complex structure allocated the risks andreturns of the Castle Harbour business.

The Dutch Banks received the lion's share of Operating Income,though this number was greatly reduced from gross rental incomebecause depreciation was treated as an expense. By contrast, theDutch Banks were not likely to receive much upside from thedisposition of assets. As a practical matter, their return onasset dispositions was capped at about $3 million. Although theyalso received 1% of Disposition Gains above $3 million, thatamount was insignificant compared to their overall investment.For example, even if Castle Harbour had sold all its assets atapproximately $303 million over book value, the Dutch Banks wouldonly have recognized an additional $3 million of return. By comparison, Castle Harbour actually disposed of the assets ata cumulative gain of $137 million, which netted the Dutch Banksonly $1.3 million above their initial $3 million allocation.

Similarly, under the Operating Agreement, the Dutch Banks wereexposed to little more than a $3 million risk of DispositionLosses and a $4 million risk of Operating Losses, i.e., a totalrisk of little more than $7 million. Again, that risk was cappedbecause, for amounts above that range, the Dutch Banks wereexposed to only 1% of the risk for each type of loss (Dispositionor Operating). For example, if the aircraft had simply been givenaway (i.e., a Disposition Loss of roughly $530 million) in 1994,the Dutch Banks would only have been exposed to approximately$5.27 million dollars of loss above their initial $3 millionallocation of Disposition Losses.

Accordingly, under the Operating Agreement, the Dutch Bankswere entitled to any Operating Income upside, but probably littlemore than approximately $3 million of any Disposition upside. Onthe other hand, they were not likely to be allocated much morethan $7 million in losses, $4 million from Operating Losses and$3 million from Disposition Losses.

The Dutch Banks were actually protected against the possibilityof even that $7 million in losses by their Investment Accounts.The Operating Agreement provided that, if the Dutch Banks'Investment Accounts exceeded the sum of their Operating Gains,Disposition Gains, Operating Losses up to $4 million, andDisposition Losses up to $3 million, the banks would be paid thedifference. Thus, even if the Dutch Banks were allocated up to $4million of Operating Losses and $3 million of Disposition Losses,they would still receive the full amount of their InvestmentAccounts. That is, the only negative effects of these losseswould be to limit the Dutch Banks' payout to the return rate oftheir Investment Account (either 9.03587% or 8.53587%). That is not to say that the Dutch Banks were guaranteed toreceive at least their Investment Account amounts. Those amountscould be reduced by the 1% allocation of Disposition Losses orOperating Losses, because such loss amounts were not included inthe determination whether a Class A Guaranteed Payment wasrequired.

To illustrate the risks to the Dutch Banks, it is useful tostart with what actually happened. When the Dutch Banks wereactually bought out, their capital accounts were worth $31.1million (J. Exs. 71, 72) and their Investment Accounts were worthapproximately $29 million (using the applicable interest rate of8.53587%) (July 22, Dull, 366). Because their allocated gains,$31.1 million, exceeded their Investment Accounts, no Class AGuaranteed Payment was made; in other words, the InvestmentAccounts were irrelevant. (July 22, Dull, 365)

The situation would have been different if Castle Harbour haddone worse. If, for example, Castle Harbour had a cumulativeOperating Loss of $1 million and no Disposition Gain or loss, theDutch Banks would have received a lower return. They would not,however, have received a negative return, that is, they would nothave been required to pay more money into Castle Harbour. CastleHarbour would have suffered a loss, and that loss would have beenallocated 98% to the Dutch Banks, bringing their capital accountto negative $980,000. The result would have been that theirInvestment Accounts ($29 million) would have exceeded the sum oftheir Disposition Gains and Losses ($0) and their OperatingLosses below $3 million ($980,000). Consequently, the Dutch Bankswould have been bought out for only $29 million. Thus, thoughthey would have received a lower return than if Castle Harbourhad done better, their return would still have been positive.

The Dutch Banks would only have received less than the amountin their Investment Accounts if Castle Harbour had done badly enough to cause losses to bedistributed to the Dutch Banks at the 1% allocation rate, i.e.,greater than approximately $4 million in Operating Losses or $3million in Disposition Losses. Even then the effect in mostscenarios would be minimal because the allocation would be only1%. For example, the Dutch Banks would only have received a zeroreturn, i.e., an allocation of $28.8 million in losses, if CastleHarbour experienced approximately $2.8 billion in Disposition orOperating Losses, an unlikely scenario given that the combinedassets of Castle Harbour never exceeded a value of $700million.27

In short, the Castle Harbour Operating Agreement shifted to theDutch Banks principally the upside of the aircraft rentals andthe risk that the upside would not exceed their InvestmentAccounts. The Dutch Banks also received a small portion of anyupside from disposition of assets. Theoretically, the Dutch Banksalso bore some risk from Disposition and Operating Losses, but,for the reasons just given, the risk was minimal.

5. Tax Consequences

The tax consequences of the Castle Harbour partnershipallocations were significant. They are also relatively simple tounderstand. As described above, 98% of Operating Income wasallocated to the Dutch Banks. Operating Income was reduced byexpenses, including asset depreciation, which in most yearsequaled close to 70% of gross rental income. For tax purposes,the same allocation was made; Operating Income was allocated 98% to the Dutch Banks.Again, Operating Income was reduced by expenses, includingdepreciation, but all the aircraft in Castle Harbour had alreadybeen fully depreciated for tax purposes. Consequently, althoughnominally depreciation was an expense for tax purposes, it didnot actually reduce taxable income. Accordingly, the taxableincome allocated to the Dutch Banks was greater than their bookallocation by the amount of book depreciation for that year. TheDutch Banks, however, did not pay United States incometaxes.28 Thus, by allocating 98% of the income from fullytax-depreciated aircraft to the Dutch Banks, GECC avoided anenormous tax burden, while shifting very little book income.

Put another way, by allocating income less depreciation totax-neutral parties, GECC was able to "re-depreciate" the assetsfor tax purposes. The tax-neutrals absorbed the tax consequencesof all the income allocated to them, but actually received onlythe income in excess of book depreciation. Thus, the full amountof book depreciation was available, pre-tax, to Castle Harbour touse.29

C. Operation of Castle Harbour

Having explained the structure and mechanics of Castle Harbour,I will now summarize the actual results of its five years ofoperation.

Castle Harbour operated from its formation on October 6, 1993until December 31, 1998 when GECC liquidated the Dutch Banks' interest.30 Itsprincipal place of business was in Bermuda. For most of itsperiod of operation, Castle Harbour's lease portfolio was managedby GECAS under the terms of an administrative services agreement.(J. Exs. 26, 31, 32; July 22, Dull, 412-13; July 26, Tewell,646-47) GECAS also helped Castle Harbour place aircraft whentheir existing leases expired. (J. Ex. 32; July 26, Hyde, 534-35)

From 1993 to 1998, Castle Harbour's cumulative Operating Incomewas approximately $28.6 million. (J. Exs. 20, 24, 37, 49, 53, 61,68) Approximately $28 million of that income was allocated to theDutch Banks. Id. During its period of operation, Castle Harbourdisposed of a number of aircraft at a cumulative loss of about$24 million. Id. When the Dutch Banks were bought out, thevalue of all aircraft and of CHLI exceeded their respective bookvalues by approximately $161 million. (J. Ex. 71 at Bates 24679)Consequently, Castle Harbour had a cumulative Disposition Gain ofapproximately $137 million. Approximately $4 million of that wasallocated to the Dutch Banks. Exhibit E payments were madethrough 1997 in an amount totaling approximately $118.5 million.At the time of the buyout, the Dutch Banks had a positive balancein their capital accounts of approximately $31 million. They werebought out at that price.31 In total, the Dutch receivednearly $150 million ($118.5 million Exhibit E payments, plus $31million buyout) over five years for their $117.5 millioninvestment. Stated more usefully, they received an internal rateof return of approximately 9.1%. (July 27, Myers, 817)

The GECC entities were allocated approximately $600,000 of the$28.6 million of Operating Income, and approximately $133 millionof the $137 million Disposition Gains. (J. Exs. 20, 24, 37, 49,53, 61, 68) Over the period of Castle Harbour's operation, GECCreceived $6 million in Class B payments, approximately $20million in distributions from its capital accounts, anddistributions of aircraft worth about $41 million. Id. In 1998,after GECC bought out the Dutch Banks for approximately $31million, it became the sole owner of the assets of CastleHarbour, worth approximately $692 million. In total, GECCreceived nearly $728 million ($6 million Class B payments, plus$20 million distributions, plus $41 million aircraft, plus $692assets, minus $31 million buyout) over five years for its initialinvestment of around $591 million. These payments gave GECC apre-tax internal rate of return of approximately 5.5%. (July 27,Myers, 762)

For the reasons given above, the Dutch Banks were allocatedmuch more taxable income than book income. Specifically, theDutch Banks were allocated approximately $310 million in taxableincome. Had this income been allocated to GECC, GECC would havebeen required to pay approximately $62 million in taxes on thatincome.

III. Conclusions of Law

The government argues three alternative theories in defense ofits reallocation of Castle Harbour's income. First, thegovernment argues that Castle Harbour was formed with no non-taxpurpose, making its formation a "sham" transaction to bedisregarded when calculating taxes. Second, the government arguesthat, even if the arrangement had a business purpose, the DutchBanks were, for tax purposes, only lenders to Castle Harbour, notpartners, and could not, therefore, be allocated any partnership income. Third, the government argues that, even ifCastle Harbour should be treated as a partnership for taxpurposes, the way it allocated income violated the "overall taxeffect" rule of section 704(b) of the Internal Revenue Code.

A. Standard of Review

The parties do not dispute that I am to review TIFD III-E's taxliability de novo, ignoring the factual findings and legalanalysis of the Commissioner of Internal Revenue ("theCommissioner"). R.E. Dietz Corp. v. United States, 939 F.2d 1,4 (2d Cir. 2001). The ultimate determination of the Commissioner,however, is presumptively correct. Therefore the taxpayer, TIFDIII-E, bears the burden of persuading me that the determinationis incorrect. Id.

B. Was Castle Harbour's Formation a "Sham" Transaction?

Regardless of its literal compliance with the tax code, atransaction will be deemed a "sham" and disregarded whencalculating taxes if it has no business purpose or economiceffect other than the creation of tax benefits. Jacobson v.Commissioner, 915 F.2d 832, 837 (2d Cir. 1990). There is nodispute that the Castle Harbour transaction created significanttax savings for GECC. The critical question, however, is whetherthe transaction had sufficient economic substance to justifyrecognizing it for tax purposes. Newman v. Commissioner,902 F.2d 159, 163 (2d Cir. 1990).

To determine whether a transaction has economic substance oris, instead, a "sham," a court must examine both the subjectivebusiness purpose of the taxpayer for engaging in the transactionand the objective economic effect of the transaction. Gilman v.Commissioner, 933 F.2d 143, 148 (2d Cir. 1991). TIFD III-E takesthis to mean that if I find either a subjective businesspurpose or objective economic effect, the transaction is not asham. The government adopts another reading of the law, arguing that I should apply a more flexible standard thatconsiders both factors but makes neither dispositive.

The decisions in this circuit are not perfectly explicit on thesubject. Recently, for example, Judge Arterton adopted the moreflexible standard, but acknowledged some potentially contrary, orat least ambiguous, language in Gilman. Long Term CapitalHoldings v. United States, 2004 WL 1924931, *39 n. 68 (D. Conn.Aug. 27, 2004). That ambiguity, however, does not affect thedecision of this case. As I will explain, under either reading Iwould conclude that the Castle Harbour transaction was not a"sham." The transaction had both a non-tax economic effect and anon-tax business motivation, satisfying both tests and requiringthat it be given effect under any reading of the law.

1. Economic Effect

In light of my findings of fact, I have little troubleconcluding that the Castle Harbour transaction had a real,non-tax economic effect. In return for a significant portion ofCastle Harbour's Operating Income,32 the Dutch Bankscontributed approximately $117 million dollars, which was used byCastle Harbour's subsidiary CHLI either to purchase aircraft orto retire GECC debt.33 The economic reality of such atransaction is hard to dispute. The Dutch Banks gave up $117million and received part of Castle Harbour's Operating Income inreturn. Castle Harbour received the Dutch Banks' $117 million,assigned part of its Operating Income in return, and put the $117million to use in an ongoing, substantial business.

The government does not dispute that, if this is what occurred,the transaction would have economic effect. Instead, thegovernment argues that various provisions of the OperatingAgreement stripped these apparent realities of any substance.Specifically, the government argues that: (1) the Dutch Banks didnot really give up their $117 million, because the combination ofInvestment Accounts and Exhibit E payments guaranteed return ofthat money; and (2) Castle Harbour did not really gain the use ofthe $117 million because the "Core Financial Assets" provision ofthe Operating Agreement essentially "froze" that money.

As explained in my findings, the Investment Accounts andExhibit E payments served different purposes. The Exhibit Epayments provided the Dutch Banks with a guarantee that theywould receive fixed payments over eight years, resulting in theultimate buyout of most of their ownership interest. The netresult of the Investment Accounts, on the other hand, was toprovide the Dutch Banks with some amount of security for theirinvestment, namely, that they would almost certainly receive noless than an 8.5% return. In other words, Exhibit E paymentsguaranteed the manner of payment, whereas the InvestmentAccounts guaranteed the amount.

It is hard to see how an assurance about the manner in whichreturns would be paid could undercut the economic reality of aninitial investment. In year one, Castle Harbour received, andused, $117 million from the Dutch Banks. Over subsequent years,that amount was paid back in installments out of the gross rentalproceeds of the aircraft leasing business. The latter fact doesnot change the former — Castle Harbour received an economicallyreal, up-front payment of $117 million. In truth, I am not surethat the government takes issue with this point. Rather, thegovernment appears to believe that Exhibit E payments somehow guaranteed the Dutch Banks anamount of return. They did not. The Exhibit E payments reducedthe banks' capital accounts, while Operating Income increasedthem. If, when the Dutch Banks were bought out, the Exhibit Epayments had "overpaid" the banks' accounts they would owe money(and conversely, if the banks were underpaid they would be owedmoney). In other words, it was quite possible for the DutchBanks' allocation to fall short of the sum of Exhibit E paymentsand, in that case, the banks would be required to make a paymentto Castle Harbour.

The Investment Accounts, by contrast, did provide the DutchBanks with some guarantee of return. As explained earlier, theDutch Banks were almost entirely certain of at least an 8.5%internal rate of return on their investment. Latching on to thisfact, the government asserts that, because of this guarantee,there was no risk to the Dutch Banks and, therefore, no economicreality to their investment. I do not agree.

First, a lack of risk is not enough to make a transactioneconomically meaningless. Even with an 8.5% guaranteed return,the Dutch Banks still participated in the — economically real —upside of the leasing business. The better the leasing businessdid, the more money the Dutch Banks made. In fact, they made areturn of approximately 9.1% on their investment (i.e., an amountgreater than 8.5%), and, had things gone better, they would havemade even more. Participating in upside potential, even with someguarantee against loss, is economically substantial.34Second, the government's premise, that a guarantee of a positive return indicates no risk, is simplistic.Whether an investment is "risky" to the investor depends on anumber of factors, including the investor's cost of capital andopportunity costs. Had the Dutch Banks, for example, borrowed the$117 million at 8.6% or foregone an opportunity to lend thatmoney at 8.6%, the chance that they would only earn an 8.5%internal rate of return would have unquestionably represented areal risk. In short, entirely open-ended risk is not the onlyeconomically real risk.

The government also argues that Castle Harbour did not reallyraise any money, because the Operating Agreement required CHLI tomaintain 110% of the Dutch Banks' Investment Accounts in "CoreFinancial Assets," i.e., high-grade commercial paper or cash.Thus, says the government, the $117 million was effectively"frozen," not available for use. This argument is missing a step.It is true that the Operating Agreement contained the provisionin question, but that provision only restricted the money's use;it did not forbid it. At the very least, CHLI was permitted toinvest in GECC commercial paper, which it did. That investmentwas not meaningless, because it allowed GECC to retire thatamount of commercial paper, thereby improving its debt-to-equityratio and creating borrowing capacity under GECC's agreement withthe credit rating agencies.

2. Business Purpose

TIFD III-E contends that GECC's non-tax purpose in enteringinto the Castle Harbour transaction was to raise capital and,more importantly, to demonstrate to investors, rating agencies,and GECC senior management, that it could raise capital on itsfleet of aging Stage II aircraft. The most direct evidence insupport of this contention was the testimony of five GECCexecutives, who all swore that "demonstrating liquidity" and"monetizing" Stage II aircraft were important motivations. In evaluating the economic substance of a transaction, courtsare cautioned to give more weight to objective facts thanself-serving testimony. Lee v. Commissioner, 155 F.3d 584, 586(2d Cir. 1998). Were the executives' testimony the only evidencebefore me, I am not sure how persuaded I would be of GECC'smotives. As things stand, however, against the backdrop of theobjective economic reality of the Castle Harbour transaction —i.e, that GECC did raise $117 million and increase itsliquidity by retiring debt — I find the testimony of GECC'sexecutives persuasive.

Consequently, I find that GECC was subjectively motivated toenter into the Castle Harbour transaction, at least in part, by adesire to raise capital and a desire to demonstrate its abilityto do so.

C. Were the Dutch Banks Partners in Castle Harbour?

The government takes the alternative position that, even if theCastle Harbour transaction as a whole had economic substance, fortax purposes the Dutch Banks were not partners of the GECCentities but rather were their creditors. Though neither partyhas put it exactly this way, there are two circumstances underwhich the Dutch Banks would not be considered partners: (1) ifthere was no economic reality to the label "partner;" and (2) if,regardless of the economics of the situation, the tax code simplyclassifies them as something else. In either case, the practicaleffect of declaring that the Dutch Banks were not partners wouldbe to reassign all of Castle Harbour's income to GECC.

In the first circumstance, the analysis is essentially the sameas the "economic substance" analysis just undertaken, but, ratherthan examining the substance of the entire transaction, I wouldonly address the narrow question whether there was any economicreality to the choice of the partnership form. Indeed, there is anarrow subset of case law, derived in part from the generaleconomic substance doctrines and in part from the Supreme Court'smore specific partnership analysis in Commissioner v. Culbertson, 337 U.S. 733, 742 (1949), thataddress just this point.

In the second circumstance, the question is not whether theform chosen has economic reality, but whether tax law requires adifferent choice of form, i.e., a different classification. Ihave not seen a case in which a court has undertaken this type of"classification" analysis of a partnership, and I am skepticalthat under the current tax code there is much chance that aneconomically substantial partnership would ever be classified assomething else. Nevertheless, the government urges me to examinethe Dutch Banks' partnership interest using the same standardused in determining whether a commercial instrument is classifiedas "debt" or "equity" for the purpose of taxing distributions.For reasons I will explain below, I do not think the debt/equitytest is relevant to classifying a partnership — the Tax Code'sdefinition of a partnership is extremely broad and easily met inthis case. Moreover, even applying the test urged by thegovernment, there is no question in my mind that the Dutch Banks,for tax purposes, held equity in Castle Harbour.

1. Economic Substance of the Dutch Banks' Partnership Interest

The decision to form a partnership may be economicallyinsubstantial, even though the partnership undertakes alegitimate business. In other words, a transaction may haveeconomic effect, yet there may be no non-tax reason to jointogether with a third party to engage in that transaction.

This situation has received particular attention in a recentline of D.C. Circuit decisions starting with ASA InvesteringsPartnership v. Commissioner, 201 F.3d 505 (D.C. Cir. 2000). InASA Investerings, the D.C. Circuit examined a complexpartnership and concluded that, even if the taxpayer had alegitimate non-tax business purpose for engaging in thetransaction in question, the formation of a partnership withvarious foreign entities to accomplish that goal served nonon-tax purpose. Specifically the court asked the question (derived from the Supreme Court'sanalysis in Culbertson),35 "whether, all factsconsidered, the parties intended to join together as partners toconduct business activity for a purpose other than taxavoidance." ASA Investerings, 201 F.3d at 513. The D.C. Circuitanswered the question in the negative, noting that "[t]here is noreason to believe that AlliedSignal[, the taxpayer,] could nothave realized Matthews's interest rate play[, one of thetransaction's benefits,] without the partnership at far, farlower transaction costs." Id. at 516. The court also concludedthat the foreign investor's interest did not look much like apartner's interest because the investor was guaranteed an exactreturn — "A partner whose risks are all insured at the expense ofanother partner hardly fits within the traditional notion ofpartnership." Id. at 515.

In a subsequent case, involving a nearly identical transaction,the D.C. Circuit was even more explicit that a partnership wouldnot be recognized if its formation served no business purpose."The only logical explanation then, for the partnership'sformation was the exploitation of Temp. Treas. Reg. §15A.453-1(c)(3)(1). . . . The absence of a non-tax businesspurpose is fatal to the recognition of the entity for taxpurposes." Boca Investerings Partnership v. United States,314 F.3d 625, 632 (D.C. Cir. 2003).

Neither ASA Investerings nor Boca Investerings enunciate anew standard of review. Both cases are only applications of thegeneral economic substance or "sham transaction" doctrine. TheASA Investerings court said as much, noting, "[courts] treat`sham entity' cases the same way the law treats `sham transaction' cases." ASA Investerings,201 F.3d at 512. In fact, the D.C. Circuit directly rejected the argumentthat different standards might apply: "Although the Tax Courtsaid it would not consider whether the transactions at issuelacked `economic substance,' its decision rejecting the bonafides of the partnership was the equivalent of a finding thatit was, for tax purposes, a `sham.'" Id. (emphasis inoriginal).

I have already explained why the Castle Harbour transaction didnot lack economic substance. Nevertheless, ASA Investeringsraises the possibility that, though the transaction as a wholehas economic substance, the formation of a partnership toaccomplish the transaction might, absent tax considerations, beeconomically meaningless. On the facts of this case, however,such a holding is not possible.

If I had concluded, for example, that Castle Harbour was not asham because it was engaged in the business activity of leasingaircraft,36 or if I concluded that it was not a shambecause it earned GECC a pre-tax profit,37 then I mightvery well need to consider separately the question whether theuse of a partnership with foreign banks to conduct such activitywas a sham, serving no non-tax purpose. In a transaction where apart of an ongoing business is spun off into a separatepartnership, the fact that the underlying business has economicsubstance does not necessarily preclude a finding that the creation of the spin-off was a sham transaction.38 Inother words, when two parties form a new entity, the fact thatthe newly created entity has a business purpose does not alwaysmean that the act of creating the entity was not a sham.

Here, however, I did not conclude that the Castle Harbourtransaction had economic substance because Castle Harbour engagedin a legitimate business. Rather, I concluded that thetransaction that created Castle Harbour was not a sham. In otherwords, I concluded there was valid business purpose and economicreality in the arrangement by which the GECC entities and theDutch Banks came together to form Castle Harbour, i.e., there waseconomic substance in not only the actions, but also theformation, of the partnership.

In this respect, the Castle Harbour transaction differssignificantly from the transaction at issue in ASAInvesterings. In that case the court was principally concernedthat (1) the outside "investors" appeared to have absolutely nostake in the partnership and (2) there seemed to be no reason toform a separate entity to engage in the underlying transactions,other than to avoid taxes. Neither situation is present here.

The Dutch Banks had a very real stake in the transactionbecause their return was tied directly to the performance of theaircraft leasing business. If the business did better, theirreturn was greater; if the business did worse, their return wasless. By contrast, the foreign banks in ASA Investerings were guaranteed an exact amount of return regardless of thebusiness's performance. It is true that the Dutch Banks' risk wasmitigated by the existence of the Investment Accounts, whichguaranteed them a minimum return, but in ASA Investerings thecourt was careful to distinguish the case where an investor'sexact return is guaranteed, making the investor entirelyindifferent to the partnership's activities, from the case wheredownside is limited but not upside, giving the investor anobvious interest in the performance of the partnership. Theformer situation indicates a sham, the latter — the situationpresent here — is not. See ASA Investerings, 201 F.3d at 514;see also Hunt v. Commissioner, 59 T.C.M. (CCH) 635 (1990)(holding valid partnership existed even though one partnerguaranteed a minimum return of 18%).

With respect to the reason for forming a separate entity, notonly was there a legitimate non-tax reason to create a separateentity — so that the Dutch Banks and GECC could share aninvestment in a specific business — but it is actually hard toimagine an alternative to creating a separate entity. As alreadydiscussed, a non-recourse loan on the aircraft was not possiblebecause of the "negative pledge" and eight-to-one debt/equitycovenant. Accordingly, given that GECC wanted to raise moneyagainst its aircraft, and given that it could not borrow againstthem, it is difficult to see what else it could have done otherthan create a separate entity and seek investments in thatentity. Of course, even if there had been another way ofachieving this financing, it would not change my analysis; thecreation of a partnership was one — even if not the only —legitimate way of achieving the non-tax purpose of raisingcapital against some of GECC's Stage II aircraft. That is all theeconomic substance test requires.39 2. Characterization of the Dutch Banks' Interest in Castle Harbour

The government argues that, even if the Dutch Banks had aneconomically substantial interest in Castle Harbour, theirinterest should, as a matter of tax law, be characterized as acreditor's interest, not a partner's. In support of thisargument, the government cites exclusively to the line of casesstemming from the Supreme Court's holding in John Kelley Co. v.Commissioner, 326 U.S. 521 (1946). Those cases discuss whethervarious commercial instruments are considered debt or equity fortax purposes — usually in order to know whether to treat returnon those instruments as deductible interest or non-deductibledividends.

Debt/equity analysis differs fundamentally from "sham"transaction analysis. John Kelley, 326 U.S. at 523 ("There isnot present in either situation the wholly useless temporarycompliance with statutory literalness which this Court condemnedas futile, as a matter of law, in Gregory v. Helvering."). Infact, the two analyses are exclusive. "Classification" analysisasks the question what formal classification is appropriate fortax purposes, whereas "sham transaction" analysis asks whether anotherwise appropriate formal classification should bedisregarded.

Accordingly, rather than asking me to apply the reasoning ofASA Investerings, as it did in its previous argument, thegovernment here asks me to question what ASA Investeringspresupposed, namely, that the entity in question was, as a formal matter,correctly classified as a partnership. See ASA Investerings,201 F.3d at 511 (noting that despite compliance with the taxcode's formal definition of a partnership, the question was"whether the formal partnership had substance").

Given that the question is how to classify the Dutch Banks'interest in Castle Harbour, I am more than a little puzzled bythe cases cited. If the question is whether Castle Harbour wasformally a partnership, it makes sense to look not to the caselaw distinguishing between debt and equity for the purpose ofdetermining appropriate deductions but to the tax code'sdefinition of a "partnership." Section 761 of the InternalRevenue Code states that "the term `partnership' includes asyndicate, group, pool, joint venture, or other unincorporatedorganization through or by means of which any business, financialoperation, or venture is carried on." 26 U.S.C. § 761. Thesection further provides, "[f]or purposes of this subtitle, theterm `partner' means a member of a partnership." Id.

There can be little dispute that Castle Harbour meets thesection 761 definition of a partnership. It was an unincorporatedorganization (a limited liability company) that carried on anaircraft leasing business. There is even less doubt that theDutch Banks were partners, namely, members of the partnership.

The fact that section 761 provides the formal definition of apartnership might explain the government's inability to cite asingle case in which the debt/equity line of cases is used toreclassify a partner's interest. Of course, a partnershipinterest is generally thought to be an equity interest, andshould the question how to classify such an interest in the handsof the holder arise in a context in which the traditionaldebt/equity question arises, the distinctions drawn indebt/equity cases might be useful. Here, however, the question isthe taxation of a partnership, and that question is governed bychapter K of the Internal Revenue Code, which provides the definition ofpartnership cited above.

Section 761's formal definition is very broad and will ofteneasily be met. But that only means that the "sham transaction"doctrine — and not the debt/equity distinction — is the test bywhich a court is to scrutinize the partnership structure.

I do acknowledge that, if despite meeting the formal definitionof partnership, a partner's interest had all the characteristicsof debt, that would strongly indicate that the partner'sparticipation in the partnership was a sham. That has nothing todo with the "classification" of the interest, but is merely theresult of the principle that, when performing a sham transactionanalysis, a court looks to substance, not form. In other words,the only possible relevance of the debt/equity analysis is as anaid to performing "sham transaction" analysis.

I believe my analysis in the previous sections already explainswhy the Dutch Banks' participation as a partner in Castle Harbourwas not a sham. I do not read any of the cases on "shamtransactions," including those that specifically deal with thequestion of sham partnerships, to require me to undertake afurther analysis — using steps borrowed from a separate area oftax law — to assure myself that, not only were the partieseconomically real partners, they were also not debtors.Nevertheless, even an application of this borrowed test, leads,unsurprisingly, to the same result.40 Courts have applied varying factors to determine whether aninstrument is debt or equity, always holding that no one factoris determinative. The I.R.S., though declining to issue explicitregulations on the subject, has highlighted eight factors worthyof consideration: (a) whether there is an unconditional promise on the part of the issuer to pay a sum certain on demand or at a fixed maturity date that is in the reasonably foreseeable future; (b) whether holders of the instruments possess the right to enforce the payment of principal and interest; (c) whether the rights of the holders of the instruments are subordinate to rights of general creditors; (d) whether the instruments give the holders the right to participate in the management of the issuer; (e) whether the issuer is thinly capitalized; (f) whether there is identity between holders of the instruments and stockholders of the issuer; (g) the label placed upon the instruments by the parties; and (h) whether the instruments are intended to be treated as debt or equity for non-tax purposes, including regulatory, rating agency, or financial accounting purposes.I.R.S. Notice 94-47.

These factors are intended to aid the determination whether aninstrument characterized as debt should actually be characterizedas equity. In trying to make the reverse determination, as I mustdo here, it is apparent that several of those factors deservelittle weight. Possession of management rights by an allegedcreditor — factor (d) — indicates the creditor may really be anowner, but the reverse is not true. The average stockholder of apublically traded corporation has no management rights, but thereis little doubt he holds equity. Similarly, a loan to a thinlycapitalized entity — factor (e) — might raise the suspicion thatthe loan is actually equity, but the purchase of equity in a wellcapitalized entity is entirely ordinary and does not indicate theexistence of a debt. Finally, though a creditor with no right toenforce the payment of principal or interest — factor (b) — lookssuspiciously like an equity holder, an equity holder with a rightto force a buyout of his share is perfectly normal. In thepartnership context, the default rule is that any partner canforce a liquidation of the partnership, i.e., force herinvestment to be returned to her (plus her gains or minus her losses). SeeRevised Uniform Partnership Act § 801(1) (dissolution requiredupon notice of a "partner's express will to withdraw as apartner").

The other factors all indicate that the Dutch Banks heldequity.

Sum Certain: The Dutch Banks were not owed a sum certain.They were to receive 98% of the net Operating Income, whateverthat might be. It is true that their potential downside waslimited, but their upside was not. Thus, although they wereguaranteed a minimum return, they were not guaranteed a maximum —or, more to the point, a certain — return. The difference issignificant. An interest holder guaranteed a fixed returnresembles a debtor because he has no interest in anything otherthan solvency of the entity obligated to pay him. By contrast,even with security against downside risk, an investor withunlimited upside potential has a significant interest in theperformance of the entity in question, because performancedirectly affects the amount of her return. Moreover, this type ofarrangement has previously been found consistent with apartnership interest. See Hunt, 59 T.C.M. (CCH) at 635.

Creditors' Rights: The Dutch Banks' interest was subordinateto that of general creditors.

Identity of Debtors and Creditors: The Dutch Banks did nothave any other relationship with Castle Harbour, so this factoris immaterial.

Label Used: Although there was some evidence that the DutchBanks at times referred to their investments as debt, in generalit appears that all the parties primarily considered the banks'interest to be that of partners.

Treatment for Non-Tax Purposes: The Dutch Banks' interest wastreated as a partnership interest for two important non-taxpurposes. It was recorded as "minority equity" on GECC'sfinancial statements. It was not considered a violation of GECC's "negativepledge," which it would have been were it debt.

In short, there is nothing about the Dutch Banks' participationin Castle Harbour that leads me to conclude that labeling them"partners" was inaccurate, much less a sham.

D. Did Allocations of Castle Harbour's Income Violate the"Overall Tax Effect" Rule?

The government's final argument is that the allocation ofCastle Harbour's income violated the "overall-tax-effect" ruleset forth in Treasury Regulations § 1.704-1(b)(2), and the incomeshould, therefore, be reallocated according to each partner'sownership interest in Castle Harbour.

"A partner's distributive share of income, gain, loss,deduction, or credit shall, except as otherwise provided in thischapter, be determined by the partnership agreement." I.R.C. §704(a). "A partner's distributive share of income, gain, loss,deduction, or credit (or item thereof) shall be determined inaccordance with the partner's interest in the partnership(determined by taking into account all facts and circumstances),if — (1) the partnership agreement does not provide as to thepartner's distributive share . . . or (2) the allocation to apartner under the agreement . . . does not have substantialeconomic effect. I.R.C. § 704(b).

A "partner's interest in the partnership" signifies "the mannerin which the partners have agreed to share the economic benefitor burden (if any) corresponding to the income, gain, loss,deduction, or credit (or item thereof) that is allocated." Treas.Reg. § 1.704-1(b)(3). "In determining a partner's interest in thepartnership, the following factors are among those that will beconsidered: (a) the partners' relative contributions to thepartnership, (b) the interest of the partners in economic profitsand losses . . . (c) the interest of the partners in cash flowand other non-liquidating distributions, and (d) the rights of the partners to distributions of capital uponliquidation." Id.

An allocation lacks "substantial economic effect" if it either(a) does not have "economic effect" or (b) is not "substantial."Treas. Regs. § 1.704-1(b)(2).

Whether an allocation has "economic effect" depends on acomplicated analysis, principally of how the capital accounts ofthe partners were maintained. It is not disputed that CastleHarbour's allocations had "economic effect."

An allocation is not substantial if, among other things, itfails the so-called "overall tax-effect" rule, that is: if, at the time the allocation becomes part of the partnership agreement, (1) the after-tax economic consequences of at least one partner may, in present value terms, be enhanced compared to such consequences if the allocation (or allocations) were not contained in the partnership agreement, and (2) there is a strong likelihood that the after-tax economic consequences of no partner will, in present value terms, be substantially diminished compared to such consequences if the allocation (or allocations) were not contained in the partnership agreement.Treas. Reg. § 1.704-1(b)(2)(iii)(a).

The government argues as follows: (4) The Castle Harbour partners' interests in the partnership are their respective percentages of ownership, for example, at the start of the partnership, approximately 82% for the GECC entities and 18% for the Dutch Banks. (5) Were income allocated according to those percentages, the GECC entities would incur a $56 million tax liability, and the Dutch Banks would receive a small portion of income. (6) By allocating income 98% to the Dutch Banks and 2% to the GECC entities, the Dutch Banks received $30 million in income, while the GECC entities saved $56 million in taxes. (7) Thus, the Dutch Banks were better off by $30 million and GECC by $26 million ($56 million in saved taxes less $30 million to the Dutch Banks). Because, after taxes, everyone was better off, and no one worse off, the overall tax effect rule was violated.

The problem with the government's argument is that its premise,that the partners' interests in the partnership are theirrespective percentages of ownership, appears to be made out ofwhole cloth. A partner's interest in the partnership signifies"the manner in which the partners have agreed to share theeconomic benefit or burden (if any) corresponding to the income,gain, loss, deduction, or credit (or item thereof) that isallocated." Treas. Reg. § 1.704-1(b)(3). Contribution of capitalto the partnership is one factor that may be considered, but ithas little weight in this case when balanced against the otherfactors. The Operating Agreement explicitly allocates the DutchBanks 98% of all the partnership's Operating Income. Throughoutthe existence of the partnership the Dutch Banks always received98% of the partnership's Operating Income. When the partnershipwas liquidated, the Dutch Banks were paid the amounts remainingin their capital accounts, which reflected an increase based onallocations of 98% of the partnership's Operating Income. It istherefore crystal clear that the Dutch Banks agreed to receive —and actually did receive — the economic benefit of 98% of all theOperating Income of Castle Harbour, making their "partner'sinterest in the partnership," with respect to Operating Income,98%.

But, the government argues, allowing a partner's interest in aparticular item of the partnership to be determined by looking atthe way that item is allocated by the agreement makes the overalltax effect rule meaningless. The overall tax effect rule requiresa comparison of the partners' situation under the agreement'sallocation with their situation under the default allocation,which, under section 704(b), is made according to each partner's interest in thepartnership. The government argues that if a partner's interestin the partnership is determined by reference to the agreement,then these two items are always the same, and it is logicallyimpossible to violate the overall tax effect rule. Therefore, thegovernment contends, it cannot be that a partner's interest inthe partnership is determined by reference to the partners'agreement.

The government is mistaken. It is not true that if a partner'sinterest in the partnership is determined by reference to thepartnership agreement it will always be the same as the specificallocations contained in the agreement. To be sure, there aresituations — and this case presents one of them — where apartner's interest in the partnership is the same as theagreement's allocation. Those situations do not implicate theoverall tax effect rule. But there are other situations where thetwo items are different, and it is those situations that aregoverned by the overall tax effect rule.

Specifically, a partner's interest in the partnership is oftennot the same as the partnership agreement's allocations — eventhough that interest is determined by reference to thepartnership agreement — in cases where the agreement makesallocations based on the taxable characteristics of specificitems. In such cases, it is possible for the discrepancy betweenthe agreement's allocation and a partner's interest in thepartnership to violate the overall tax effect rule if theagreement's allocation makes one partner better off after-tax,and no partner substantially worse off.

An example, taken directly from the Treasury Regulations, makesthis clear.

Individuals H41 and J form a partnership. H is in the50% tax bracket, J, the 15% bracket. The partnership principally invests in taxable and tax-exempt debtinstruments. The partners each contribute equally to thepartnership. The partners agree to share equally gains and lossesfrom the disposition of the debt securities, however, they alsoagree that tax-exempt interest will go 80% to H and 20% to J, andtaxable interest 100% to J.

The partnership realizes $450 of tax-exempt interest and $550of taxable interest. Under the agreement H receives 80% of the$450 tax-exempt interest — $360. J receives 20% of the taxexemptinterest ($90) and all of the taxable interest ($550) — a totalof $640. For reasons I will explain in a moment, this allocationviolates the overall tax effect rule, and the income must bereallocated according to each partner's interest in thepartnership. See I.R.C. § 704(b). The regulations explain howthis is to be done: "Since under the partnership agreement [H]will receive 36 percent (360/1,000) and J will receive 64 percent(640/1,000) of the partnership's total investment income in suchyear, under paragraph (b)(3) of the section the partnership'stax-exempt interest and taxable interest and dividends each willbe reallocated 36 percent to [H] and 64 percent to J." Treas.Reg. § 1.704-1 Example 5(ii).

In this example, the partners' interests in the partnership arenot identical to the interests they are allocated under thepartnership agreement. The partnership agreement allocatestaxable interest 100% to J and 0% to H, and non-taxable interest20% to J, 80% to H. By contrast, when allocation is doneaccording to each partner's interest in the partnership, J isallocated 64% of each item, and H is allocated 34% of each item.This difference exists even though each partner's interest in thepartnership is determined exclusively by reference to thepartnership agreement. J is allocated 64% of interest income andH is allocated 34% because, in aggregate, that is how much thepartnership agreement allocates them — the agreement allocates $640 of the $1000 ofinterest income to J ($90 tax-exempt income, plus $550 taxableincome) and $360 of the $1000 to H ($360 tax-exempt income).Notably, the partner's interest is not determined by theirownership interest in the items in question (or in thepartnership as a whole).42

With that in mind, it is easy to see why the above examplefails the overall tax effect rule. Without the allocation,investment interest would be allocated according to eachpartner's interest in the partnership — H would receive 36% ofthe $450, tax free, i.e., $162 and 36% of $550 less 50% taxes,i.e., $99, for a total after-tax allocation of $261, and J wouldreceive $288 of the tax-free interest, and $352 of the taxableinterest, less $53 in taxes, for a total after-tax allocation of$587. With the challenged allocation, however, H receives $360tax free, and J receives $90 tax free and $550 less $83 in taxes,i.e., $467, for a total after-tax allocation of $557. Thus H isbetter off after taxes with the challenged allocation — $360versus $288 — and J is not substantially worse off — $587 vs.$557. Put another way, H receives a greater amount with the itemsallocated according to their taxable characteristics than hewould receive were the items allocated based simply on H'spercentage interest in them, as inferred from the partnershipagreement. Accordingly, the overall tax effect rule is violated,and the amounts are reallocated according to each partner'sinterest in the partnership as described above.

This example makes clear that (a) when the partnershipagreement contains explicit allocation provisions,43 partner's interest in the partnership isdetermined by reference to the agreement, not by reference toownership interest, and (b) this determination is not circular —it can lead to different allocations, particularly in cases whereitems are allocated based on their taxable characteristics.

In the case of Castle Harbour, the intent of the parties, andthe economic reality of the situation, was the unambiguousassignment of a 98% interest in Operating Income to the DutchBanks. Operating Income was not further differentiated based ontaxable characteristics, and there is simply no ground from whichto argue that the partners had any other interest than the 98%and 2% assigned by the agreement. Consequently, the overall taxeffect rule is not applicable because there is no differencebetween the allocations made and each partner's actual interestin the partnership. Morever, even if applicable, the overall taxeffect rule would have no effect because reassignment of incomebased on the partners' interests in the partnership would resultin the same allocation actually made.

The government is attempting to use the overall tax effect ruleto remedy its problem, not with the manner in which the DutchBanks were allocated their interest in the partnership (somethingthe overall tax effect rule might cover), but with the fact thatthe Dutch Banks were allocated such a huge interest in the firstplace. The tax benefits of the Castle Harbour transaction werethe result of the allocation of large amounts of book income to atax-neutral entity, offset by a large depreciation expense, witha corresponding allocation of a large amount of taxable income,but no corresponding allocation of depreciation deductions. This resulted in anenormous tax savings, but the simple allocation of a largepercentage of income violates no rule. The government does not —and cannot — dispute that partners may allocate theirpartnership's income as they choose.44 Neither does thegovernment dispute that the taxable income allocated to the DutchBanks could not be offset by the allocation of non-existentdepreciation deductions to the banks.45 And, as I havejust explained, the bare allocation of a large interest in incomedoes not violate the overall tax effect rule.

The truth is that the government's only real argument is itscontention that the Castle Harbour transaction was done solelyfor the purpose of, and with the sole effect of, achieving thetax benefits consequent to the 98% income allocation totax-neutral parties. In other words, the government is arguing —again — that the transaction was a "sham." That argument hasalready been addressed.

IV. Conclusion

The government is understandably concerned that the CastleHarbour transaction deprived the public fisc of some $62 millionin tax revenue. Moreover, it appears likely that one of GECC'sprincipal motivations in entering into this transaction — thoughcertainly not its only motivation — was to avoid that substantialtax burden. Nevertheless, the Castle Harbour transaction was aneconomically real transaction, undertaken, at least in part, fora non-tax business purpose; the transaction resulted in the creation of a true partnership with all participants holdingvalid partnership interests; and the income was allocated amongthe partners in accordance with the Internal Revenue Code andTreasury Regulations. In short, the transaction, though itsheltered a great deal of income from taxes, was legallypermissible. Under such circumstances, the I.R.S. should addressits concerns to those who write the tax laws.

I conclude that from 1993 to 1998, Castle Harbour properlyallocated income among its partners. The FPAAs issued by theI.R.S. were in error, and the I.R.S. must refund to TIFD III-Ethe total amount of TIFD III-E's jurisdictional deposit, plus anyinterest called for by 26 U.S.C. §§ 6226 and 6611.

The clerk will enter judgment for the plaintiff and close thefile.

It is so ordered.

1. GECC is the true party in interest and ultimate taxpayer inthis case, though many of the actions relevant to this case weretaken by GECC's wholly owned subsidiaries. I will frequently usethe name "GECC" to refer interchangeably to both GECC itself andto its wholly owned subsidiaries.

2. Citations to trial testimony are given as follows: (monthand day, last name of witness, page number). All of the trialtook place in 2004.

3. Before submitting the proposal ultimately adopted, Babcock& Brown submitted at least two other proposals to GECC, whichwere rejected. (July 21, Brickman, 212-15)

4. Legal title to the aircraft in question was held in trust.Consequently, it was the beneficial ownership in the aircraftthat GECC transferred to Summer Street. (July 22, Dull, 302)

5. These entities' compliance with the terms of the OperatingAgreement was guaranteed by GECC. (J. Ex. 3)

6. $296 million initial contribution, minus $50 millioninterest sold to the Dutch Banks.

7. $530 million in aircraft, minus $258 million non-recoursedebt, plus $22 million in rent receivable.

8. $294 million net aircraft value, plus $246 cash investment,minus $12,000 income allocation.

9. $117.5 million cash investment, plus $600,000 incomeallocation, minus $6 million distribution.

10. Additionally one aircraft was distributed back to GECC, anevent that was treated essentially as a sale. For simplicity, Iignore that transaction in this illustration.

11. Principally because of the airplane distribution gain thatthis example ignores, the actual number was higher by about $3million. (J. Ex. 37 at Bates 16344)

12. Technically, Exhibit E payments were discretionary on thepart of GECC. Nevertheless, because failure to make a scheduledExhibit E payment gave the Dutch Banks the right to force aliquidation of Castle Harbour, it was unlikely the payments wouldnot be made. (Op. Agmt. § 14.1(d), J. Ex. 1 at Bates 101499)

13. Internal rate of return is defined as the discount ratenecessary to make the net present value of a stream of futurepayments equal to zero.

14. On its consolidated financial statements, GECC did notactually record the transaction in this manner. Instead, itsimply kept the aircraft and $240 million on its books andrecorded the Dutch investment as $117.5 million of "minorityequity." (July 22, Dull, 357-58)

15. The Operating Agreement simply refers to this as "Profit"and "Loss," but the parties adopted the more convenientdesignation of Operating Income.

16. The Operating Agreement specified further Operating Lossallocations in the event losses exceeded $541 million. Theseallocations never came into play, were highly unlikely to evercome into play, and, accordingly, are not relevant. (Op. Agmt. §3.2(d)-(f), J. Ex. 1 at Bates 101436)

17. The fact that Disposition Gains and Losses were allocatedto first offset prior years' Disposition Gains or Losses meantthat Disposition Gains and Losses were allocated cumulatively.That is, if net Disposition Gains had been distributed only once,at the end of the partnership, the result would have been thesame as if — as it was actually done — gains and losses wereallocated annually in amounts sufficient to offset prior years'gains and losses. The situation was not the same with OperatingGains and Losses because Operating Gains were not allocated tooffset prior Operating Losses and vice versa.

18. As with Operating Loss, there were different allocationsthat would take place should Disposition Losses exceed $541million. (Op. Agmt. § 3.3(j)(iv)-(vi), J. Ex. 1 at Bates101440-41) Those allocations are not relevant to this case.

19. $21 million, plus $27 million, minus $74 million.

20. Note that prior years' Operating Gains were not offset,though if there had been a Disposition Gain that would haveoffset prior years Operating Losses.

21. This $6 million in Disposition Losses is the sum of the $3million Disposition Gain offset, plus $2.9 million allocated at90%, plus $100,000 allocated at 1%.

22. This $18 million in Disposition Losses is the sum of the$1.5 million Disposition Gain offset, plus $300,000 allocated at10%, plus $18 million allocated at 99%.

23. Put another way, had the Class B payments been treated asallocations followed by distributions the effect on the capitalaccount would have been the same. The capital account would havebeen increased by the allocation, but then decreased by thedistribution.

24. The amount of cash offset by depreciation did not actuallysit in Castle Harbour, but ended up being used to pay off the twosignificant expenses not included as Operating Income expenses —debt principal and Exhibit E payments.

25. In the event that the Applicable Rate was 9.03587% theInvestment Accounts would be close to zero at the end of theeight-year period, because, as noted before, the Exhibit Epayments provided an internal rate of return of 9.03587%.

26. Apparently, because of Federal Aviation Administration("FAA") restrictions, the Dutch Banks could not have been given aright to elect the managers. (July 22, Dull, 359)

27. This scenario might have occurred had there been somecatastrophic accident that exceeded insurance coverage, and ifthat accident was treated as an Operating Loss. For that reason,the Dutch Banks, prior to entering Castle Harbour, bargained hardabout the level of insurance coverage to be maintained by CastleHarbour and the degree to which such losses would come out ofOperating Income. (July 22, Dull, 354-55)

28. Moreover, were any U.S. taxes to be assessed against theDutch Banks, GECC agreed to indemnify them. (J. Exs. 4, 5) Inaddition, the Dutch Banks were prohibited, by the OperatingAgreement, from selling their interest, thus ensuring that theinterest would not be transferred to a taxable entity. (J. Ex. 1at Bates 1482)

29. Presumably GECC had to pay taxes on the amounts offset bydepreciation when income was realized upon ultimate dispositionof the assets. The record contains no evidence on what tax ratewas applied at that point. At the very least, however, thisarrangement allowed GECC to defer taxes on aircraft leaseincome.

30. The Dutch Banks were bought out after a change in U.S. taxlaw made it possible that Castle Harbour would no longer betreated as a partnership, potentially implicating GECC'sliability under the parties' tax indemnification agreement. (July22, Dull, 364-65)

31. Actually, because of the circumstances of the buyout, theDutch Banks received a slight premium of approximately $150,000for having their interest bought out early. (July 22, Dull,366-67). The details of that premium payment are not relevant tothis case.

32. This portion was significant even if, as I discussedabove, one considers the GECC entities' allocation of OperatingIncome to include the amount of the Class B payments.

33. Because GECC invested $240 million in cash as well, andmoney is fungible, it is not possible to say what part of the$117 million was used for which activity.

34. The government also appears to argue that the Dutch Banksdid not really participate in the upside of the business because"[t]he rental income, particularly in 1993, 1994, and 1995 wasvery predictable." Govt. Post-Trial Brief at 10. This argument ishardly worth addressing. An investor does not participate anyless in an investment, and an investment is not any less real,because there is some element of predictability to the return.

35. The Supreme Court in Culbertson held that thedetermination whether a partnership "is real for income-taxpurposes depends upon whether the partners really and trulyintended to join together for the purpose of carrying on thebusiness and sharing in the profits and losses or both."Culbertson, 337 U.S. at 741.

36. TIFD III-E presented a good deal of evidence on thispoint, evidence regarding the day-to-day operation of CastleHarbour's leasing business.

37. This point also was argued by TIFD III-E, primarilythrough the testimony of its expert Professor Myers, whotestified that Castle Harbour earned GECC a pre-tax internal rateof return of approximately 5.5%. I confess that I am not entirelysure of the significance of this number standing alone; internalrate of return would seem to me only to indicate profit whencompared with a company's cost of capital.

38. Though, in at least one case, a court found the underlyingbusiness purpose of a reorganized entity to be sufficient tojustify the reorganization transaction. See United ParcelService of America v. Commissioner, 254 F.3d 1014, 1020 (11thCir. 2001) ("The transaction under challenge here simply alteredthe form of an existing, bona fide business, and this casetherefore falls in with those that find an adequate businesspurpose to neutralize any tax-avoidance motive.").

39. Some language in the Boca Investerings decision supportsa reading that merely demonstrating a business purpose forforming a partnership is insufficient, and instead, a taxpayermust demonstrate a business necessity. Boca Investerings,314 F.3d at 632 ("Because the district court did not find that alegitimate, non-tax necessity existed for the formation of theBoca partnership. . . .") (emphasis supplied). I think that otherlanguage, see, e.g., id. ("We do not of course suggest thatin every transaction using a partnership a taxpayer must justifythat to form"), as well as a reading of the case as whole, showsthat the D.C. Circuit meant nothing more than that, when thereappears to be no non-tax reason for creating a separate entity toeffect a given transaction, the creation of the entity is likelya sham.

40. This is not some happy coincidence, but rather followsfrom my previous conclusion, which implicitly addressed the issueof the propriety of characterizing the Dutch Banks' interest asequity. I concluded above that the Castle Harbour transaction hadeconomic substance, in part, because the Dutch Banks reallyinvested $117 million in return for a significant portion of anyof the returns of the aircraft leasing business. For the purposeof "sham transaction" analysis, that conclusion is, I believe,sufficient to establish that the banks had an economically realequity interest in the partnership. Accordingly, it is notsurprising that a more mechanical application of the debt/equitytest yields the same result.

41. The regulations use the letter "I" to designate the firstpartner. To avoid confusion with the first person pronoun, I use"H."

42. Neither is the partners' interest determined by looking atthe way they agreed to allocate proceeds from disposition of thedebt instruments in question, i.e., equally.

43. There may be very good reasons for partners to allocateincome in amounts higher or lower than each partner's capitalcontribution. For example, allocations may reflect non-capitalcontributions such as expertise, time and energy devoted to thepartnership, and whether the partner's interest is short-term orlong-term. A partnership, in other words, can reasonablyestablish classes of partners on grounds other than their capitalcontributions.

44. Here the large allocation provided a rational method ofliquidating the Dutch Banks' partnership interests over arelatively short period of time.

45. The rules governing the allocation of, among other things,depreciation deductions in situations where the book value of acontributed asset is greater than its tax basis are set forth inI.R.C. § 704(c) and its attendant regulations. In part, theapplication of that section's "ceiling rule" leads to the taxsavings at issue here. The government, however, does not allegethat the mechanics of section 704(c) allocation were improperlyapplied in this case, so there is no need to examine them.