A Weapon From the Past

EXECUTIVE SUMMARY

IF A TAX-SAVING IDEA SOUNDS TOO GOOD TO BE TRUE, it probably is. Corporations that invested in a Merrill Lynch tax plan discovered this the hard way when the government, applying principles from a 1935 case, Gregory v. Helverling, attacked the plan as a tax-avoidance scheme. THE MERRILL LYNCH PLAN INVOLVED A U.S. corporation’s forming a partnership with a foreign entity that paid no U.S. income tax. Over time the partnership’s transactions would generate huge losses the U.S. company could use to offset existing or expected capital gains. IN ACM PARTNERSHIP, THE GOVERNMENT DISALLOWED partnership transactions that produced large losses for Colgate-Palmolive and prohibited the use of the installment sale rules. It said they were sham transactions that created phantom losses. The Tax Court ruled in favor of the IRS, stating that the transactions lacked economic substance. The appeals court used the substance-over-form doctrine and business purpose tests from Gregory to support the lower court decision. IN ASA INVESTERINGS, ALLIED SIGNAL USED THE Merrill Lynch plan to generate losses to offset a $400 million capital gain. The IRS ruled ASA was not a valid partnership and allocated all gains and losses to Allied Signal. The Tax Court and appeals court agreed. BRUNSWICK CORP. FORMED TWO PARTNERSHIPS in an effort to offset a $125 million gain. It reported combined losses of $176 million. The IRS determined the transactions that resulted in the losses lacked economic substance. In SABA Partnership the Tax Court applied a principle from Gregory , which is that the courts are not obliged to respect the form of a transaction that was contrived to obtain an unintended tax benefit.

TINA STEWARD QUINN, CPA, PhD, is associate professor of accountancy at Arkansas State University. Her e-mail address is tquinn@mail.astate.edu . TONYA K. FLESHER, CPA, PhD, is professor of accountancy at the University of Mississippi. Her e-mail address is actonya@olemiss.edu .

here’s an adage that if a tax-saving idea sounds too good to be true, it probably is. A number of corporations that participated in a Merrill Lynch tax plan discovered this when the ruling in a 67-year-old case proved to be their downfall. Gregory v. Helvering, 293 U.S. 465 (1935) was a landmark case—the first of its kind to come before the courts on the question of whether a reorganization took place where the taxpayer had no intent to carry on the existing corporate business, only a desire to minimize taxes. Previously, the courts had always applied a literal interpretation of the law when deciding tax cases.

Other concepts that evolved from Gregory , in addition to the business purpose test, include continuity of business, the taxpayer’s right to minimize tax liability, substance over form and the step transaction doctrine. The IRS and the courts still apply many of these tests today. What follows is a discussion of three cases in which the courts applied both the doctrine of substance over form and the business purpose test to expose elaborate tax-avoidance schemes as sham transactions. The warning to CPAs and the taxpayers they represent is that the government can still use the simple concepts in Gregory as major weapons against tax shelters.

THE MERRILL LYNCH PLAN
Merrill Lynch was involved in all three cases. In 1989 and 1990 the brokerage firm marketed an investment plan to several U.S. companies that had as its primary purpose generating huge capital losses the companies could use to offset existing (or expected) capital gains. The plan required the U.S. company to form a partnership (based outside the United States) with a foreign entity that paid no U.S. income tax. In the beginning, the foreign entity would have an overwhelming majority partnership interest. The scheme relied on applying the installment sale (see “Installment Sales” ) and contingency sale rules to sales of foreign currency for cash or London interbank offered rate (Libor) notes (see “Security Definitions” ).

The plan was for the partnership to trade securities and ultimately purchase private placement notes (PPNs). Merrill Lynch marketed this plan to several large corporations. Although there were slight variations, in general it involved these steps:

The U.S. company would enter into a foreign-based partnership with a foreign entity that was not subject to U.S. income tax.

The foreign entity would have the overwhelming majority partnership interest while the U.S. company would own a distinct minority interest.

The partnership would purchase short-term PPNs eligible for the installment method of accounting. It then would sell them for a large cash down payment with the balance made up of a comparatively small amount of debt instruments (five-year Libor notes) whose yield over a fixed period of time was not ascertainable. One-sixth of the basis would be applied to the down payment. The gain from the down payment would be allocated according to the partnership interests. Therefore, the foreign partner would receive the majority of the gain.

The partnership would claim a large basis (five-sixths of the basis of the PPNs) in the Libor notes.

After the close of the first tax year, the partnership interests would become substantially reversed. (The U.S. company would acquire a majority interest by purchasing part of the foreign entity’s interest.)

The partnership would distribute cash to the foreign entity and the Libor notes to the U.S. partner in partial redemption of their partnership interests.

The U.S. company then would sell the Libor notes to a third party, accelerating the loss. Since the basis of the instruments would greatly exceed their value, the sale would result in a large “paper” loss the U.S. company would use to offset existing capital gains.

The plan was complicated and each aspect carefully arranged; however, an important step was missing. Merrill Lynch had failed to register the plan with the IRS as a tax shelter. While registration provides no protection against an IRS challenge, it is required by IRC section 6700. In August 2001 Merrill Lynch agreed to pay a substantial tax shelter penalty for its failure to register.

ALL WASHED UP
One complicated use of the Merrill Lynch plan was intended to help mitigate the tax effects of a company’s 1988 multimillion dollar capital gain. In 1989 ABN (a major Dutch bank and the foreign entity in all three cases), Colgate-Palmolive and Merrill Lynch each created a new company (referred to here as A, C and M, respectively). These three companies then formed the ACM partnership to generate capital losses Colgate could use to offset some of its 1988 capital gains. The partnership was capitalized with $205 million; A held 82.6%, C 17.1% and M 0.3%.

ACM used an elaborate series of securities transactions that ultimately resulted in its selling $175 million in PPNs for $140 million in cash and eight Libor notes with a present value of approximately $35 million. Since the total amount was based on a contingency (fluctuations in the Libor), ACM treated the transaction as an installment sale, allowing it to “recover” one-sixth of the basis each year over the term of the contract.

The $140 million ACM collected in the year of sale resulted in a $110.7 million gain, which was allocated primarily to partner A. After the close of the first tax year, Colgate purchased part of A’s partnership interest. ACM redeemed a portion, leaving Colgate as the majority partner. Subsequent installment payments resulted in capital losses allocated primarily (99.7%) to Colgate. In December 1991 the partnership sold the Libor notes, accelerating the remaining loss. Colgate reported total capital losses of more than $98 million over the course of its participation in ACM. It carried them back to offset its 1988 capital gain.

In 1993 the IRS challenged ACM’s treatment of the transaction and disallowed the use of the installment sale rules, calling it a sham transaction creating “phantom” losses. The Tax Court, in a 1997 memo decision, held for the IRS because the transaction lacked economic substance (TC Memo 1997-115).

ACM appealed to the Third Circuit Court of Appeals. In ACM Partnership (157 F3d 231 (CA-3, 1998)) the court relied on Gregory in applying the substance-over-form doctrine and the business purpose test. The Third Circuit said that where the “form of the taxpayer’s activities indisputably satisfies the literal requirements of the relevant statutory language, the courts must examine whether the substance of those transactions was consistent with their form, because a transaction that is ‘devoid of economic substance’ simply is not recognized for federal taxation purposes.”

The court referred to Gregory as the “foundational exposition” of economic principles under the Internal Revenue Code. The judges viewed the transactions as a whole, as well as each step from beginning to end, to determine if they had sufficient economic substance to be respected for tax purposes. They found ACM’s transactions had only nominal, incidental effects on the partnership’s net economic position. The court said “Gregory requires us to determine the tax consequences of a series of transactions based on what actually occurred” and affirmed the Tax Court decision that ACM’s transactions lacked economic substance.

MIXED SIGNALS
In 1990 Allied Signal used the Merrill Lynch plan to generate losses to offset a $400 million capital gain. Allied Signal and its new wholly owned subsidiary ASIC entered into a partnership with two Netherlands Antilles special purpose corporations (which ABN controlled). In April 1990 the partnership bought $850 million in PPNs. In May 1990, it sold the PPNs for $681.3 million and 11 Libor notes; the total value was approximately $850 million. The partnership reported a $539,443,361 gain on its partnership return, allocated according to the partnership interests—$485 million to the foreign entities and $53 million to Allied Signal and ASIC. The foreign entities paid no U.S. income tax. After the close of the first tax year, Allied acquired a majority partnership interest by purchasing part of the foreign entities’ interest.

In August 1990 ASA distributed the Libor notes to Allied Signal and ASIC in partial redemption of their partnership interest and cash and commercial paper to the foreign entities. The Libor notes carried an adjusted basis of $709 million (five-sixths of the basis of the PPNs). Allied sold some of the Libor notes in 1990 and the remainder in 1992, claiming a total loss of $538 million. It used the losses to offset the gain from the sale of its interest in another company. Although Allied reported a tax loss of $538 million, its actual economic profit was about $3.6 million.

The IRS audited the partnership returns for 1990 through 1992, determining that ASA was not a valid partnership and adjusting the returns to allocate all gains and losses to Allied Signal. The Tax Court focused on the purported business purpose of Allied Signal and ABN. Allied entered into the venture for the sole purpose of generating capital losses, and ABN entered into it solely to receive its expected return. Allied bore all the expenses, and ABN did not intend to, nor did it actually, share in ASA’s losses. The Tax Court concluded the relationship between the two was merely a contractual, debtor-creditor relationship—not a partnership.

In ASA Investerings (201 F3d 505 (CA-D.C., 2000), affirming TC Memo 1998-305, cert. denied 121 S.Ct. 171 (2000)), the Circuit Court of Appeals for the District of Columbia affirmed and said, “It is precisely such a sham or unreal device to perpetrate an evasion of tax liability that the Merrill Lynch scheme seeks to accomplish here for Allied Signal to obtain a $538 million tax loss from participation in a relationship that was developed and structured solely to achieve ‘phantom’ losses for a taxpayer that would incur no genuine economic loss.”

GUTTER BALL
Another case involved Brunswick Corp.’s decision to sell some of its businesses. The company expected a $125 million gain and met with Merrill Lynch to get help minimizing the tax impact. In 1990 Brunswick (with a wholly owned subsidiary) and ABN (through controlled entities) formed two general partnerships: SABA and Otrabanda. SABA was capitalized with $200 million which it used to purchase PPNs. ABN owned a 90% interest and Brunswick the remaining 10%. The partnership soon sold the PPNs for $160 million in cash and four Libor notes with present value of approximately $39 million. It reported a $126,666,667 gain of which 90% was allocated to ABN, which paid no U.S. tax. After the first tax year closed, Brunswick effectively acquired a majority partnership interest, and the partnership distributed cash to the foreign partner and the Libor notes to Brunswick. The company sold three of the notes and reported a net short-term capital loss for 1990 of approximately $85 million. Brunswick sold the fourth note in 1991 and reported a $32.6 million long-term capital loss.

In 1990 Brunswick entered into a second partnership (Otrabanda) with ABN. It had a slightly different structure but still functioned in much the same way as SABA. Brunswick contributed $15 million for a 10% interest and ABN contributed $135 million for a 90% interest. Shortly after its formation, the partnership purchased $100 million in floating-rate certificates of deposit. These CDs were not registered with the SEC and did not trade on an established securities market. One month later, the partnership sold the CDs for $80 million in cash and four Libor notes with a fair value of $19 million. Otrabanda reported a $63.3 million gain from the down payment with $57 million (90%) allocated to ABN. After the close of the first tax year, Brunswick acquired the majority partnership interest, purchasing 50% of the foreign partner’s interest. In November 1990 the partnership distributed cash to the foreign partner and the Libor notes to Brunswick. Brunswick sold them shortly afterward and reported a $60 million net short-term capital loss.

On its consolidated federal tax return for 1990, Brunswick reported a net short-term capital loss of approximately $143 million; on its 1991 return it reported a long-term net capital loss of about $33 million from partnership transactions. The IRS determined the transactions lacked economic substance and did not recognize the partnerships. It reallocated the partnership items to Brunswick and adjusted the basis of the securities.

In SABA Partnership (TC Memo 1999-359), the Tax Court applied the Gregory principle that although a business transaction may be structured in strict compliance with the law, a court is not obliged to respect its form when the record shows the transaction was contrived to obtain a tax benefit Congress did not intend. The partnerships had been organized to generate losses for Brunswick. The transactions did not change the company’s economic position, and they lacked economic substance. Therefore, the company should not recognize any gains or losses on the sales of the PPNs or CDs.

THE BEST LAID PLANS…
The Merrill Lynch scheme was clever and followed the letter of the law. Nevertheless, Gregory defeated it. While the plan was elaborate and sophisticated, what defeated it was simple. CPAs should warn clients that regardless of how elaborate a plan or how closely it follows the letter of the law, the court will look beyond the form of a transaction to its substance. If a transaction lacks a genuine business purpose, it will not pass the Gregory test. Even after 67 years this case remains a foundation of tax law practitioners must always keep in mind when helping corporate clients structure tax plans. CPAs must always heed the doctrines in the case in every tax avoidance plan. Gregory remains a powerful IRS weapon in the battle against abusive tax schemes.