2   Why Tax Shelters Matter

Abusive tax shelter activity at the turn of the twenty-first century cost the U.S. Treasury billions of dollars and sent a number of lawyers and accountants to prison. To appreciate the complexity of the problem of abusive shelters, it is important to understand what a tax shelter is and what in principle distinguishes a legitimate shelter from an abusive one. Understanding this fundamental distinction is critical to appreciating exactly what happened in the most recent tax shelter wave and in assessing the extent to which the tax system remains vulnerable to new types of abusive tax shelters. A basic difficulty is that such shelters are often more easily recognized after the fact than precisely defined in advance. Nonetheless, this theoretical distinction—as difficult as it is to apply in practice—separates transactions that constitute permissible tax avoidance, approved by Congress, and those that can serve as the basis for civil or even criminal liability.

Legitimate and Abusive Tax Shelters

An important goal of the income tax system is to measure a taxpayer’s economic income and tax that income at specified rates. Consistent with this goal, the Internal Revenue Code contains an expansive definition of gross income—“all income from whatever source derived”—and lists a range of items to be included in it. It then permits “ordinary and necessary” business expenses to be deducted from gross income to arrive at a lower figure. Known as “taxable income,” or net income, this is the amount of income actually subject to income taxation. Determining net income thus requires numerous judgments about which expenditures can be deducted from income because they are associated with earning it.

The concept of net income also comes into play when a taxpayer buys and later sells an asset. For tax purposes, the income that the taxpayer receives from the sale is reduced by the cost that he or she incurred in originally purchasing it. This is known as the “basis.” Although the basis of an asset may be adjusted upward or downward after purchase, the concept of basis is intended to limit taxable income to the net economic gain that the taxpayer enjoyed from buying and then selling the asset. The higher the basis, the lower the taxable income from the sale.

Tax shelters reflect the fact that many provisions in the tax code seek to influence taxpayers by treating income and expenses connected with certain types of behavior more favorably than other income or expenses. The law does this by taxing some income at a lower rate or not at all, and by allowing deductions for certain types of outlays but not others. The tax laws encourage savings for retirement, for instance, by exempting certain contributions to retirement plans from taxation until the taxpayer withdraws the money upon reaching retirement age. They seek to promote home ownership by making most home mortgage interest deductible. They encourage charitable giving by making many such gifts deductible. They provide for employment “cafeteria plans” that allow employees to pay certain medical expenses from pretax dollars. Participation in these activities enables taxpayers to “shelter” some portion of their income from taxes. The government is willing to forgo revenues from taxing the sheltered income because it wants to use the incentive of lower taxes to encourage people to engage in certain socially useful activities.

Legitimate shelters are those that involve the kinds of activities that Congress wants to encourage. Because Congress intends to benefit taxpayers who participate in these activities, the fact that an individual might be motivated by a desire to lower taxes is not a concern. On the contrary, that’s the point of these types of provisions: harnessing the desire for lower taxes to promote valuable economic and social behavior.

A useful way to distinguish legitimate from abusive tax shelters is to ask the question: Does the shelter in which the taxpayer is engaging involve the kind of socially useful behavior that Congress wants to encourage by creating a tax incentive? Would Congress have considered this type of taxpayer behavior valuable enough to incur the loss of tax revenues? If the answer is no, then a taxpayer is seeking tax-favored treatment for activities that Congress did not intend to subsidize and is engaging in an abusive tax shelter. There is no reason in this instance to allow taxpayers to pay lower taxes. Indeed, permitting this would be counterproductive. Tax policy experts generally agree that behavior like this is wasteful and does not contribute to social welfare. In other words, the resources invested in abusive tax shelters represent a deadweight loss to society. Professor Michael Graetz famously described such a tax shelter as “a deal done by very smart people that, absent tax considerations, would be very stupid.”¹

In some cases, statutory requirements and external checks curtail the extent to which taxpayers can receive unintended tax benefits. In the retirement savings context, for instance, laws limit the amount that can be sheltered for this purpose, require the monies to be contributed to special accounts—employer-sponsored retirement plans or IRAs—and impose penalties if the monies are withdrawn before a specific age. Retirement savings contributions are confirmed by third parties, such as banks and employers. These mechanisms, which are written into the law and enlist the verification of third-party institutions, prevent taxpayers from using tax-favorable provisions to avoid paying taxes in ways that are not consistent with their purpose.

But there is a limit to the constraints that Congress can build into legislation, and a clever taxpayer and her tax advisor can often seek to take advantage of interactions among various Code provisions to obtain a benefit inconsistent with Congressional intent. The facts of Knetsch v. Commissioner, a case that made it to the United States Supreme Court, provide a good example.² In 1953, Karl Knetsch was a successful sixty-year-old businessman who earned $200,000 a year, equivalent to more than $1.5 million in 2010. Knetsch purchased $4 million dollars’ worth of thirty-year term annuities from the Sam Houston Insurance Company. Under the terms of the annuities, Knetsch would earn 2.5 percent interest, or $100,000 a year, until his ninetieth birthday. When Knetsch reached ninety, the annuities would begin to pay him $43 a month until his death.

Knetsch funded the purchase of the annuities with a loan from Sam Houston Life Insurance Company at an interest rate of 3.5 percent a year. As a consequence, the interest rate on his debt was higher than the interest rate on his return. Knetsch was paying $140,000 a year, or $40,000 more than he was receiving from the annuities. Given his outlay over the thirty-year maturity period, it would take more than 2,000 years for him recoup his investment once he began receiving annuity payments.

It is hard to imagine that Knetsch, or anyone else for that matter, would find such a deal attractive. Who would be willing to pay $40,000 a year until he reached age ninety for the right to receive $43 a month thereafter? Not surprisingly, the transaction had other benefits that explained its appeal. The interest income that Knetsch earned from the annuities was not taxable. In addition, the loan was non-recourse—secured only by the cash value of the insurance policy and not his personal assets—so Knetsch was not at risk of losing his property if he defaulted on the loan.

Most important, under the language of tax laws then in effect, Knetsch could deduct from his taxable income the interest on the loan that he was paying the insurance company each year. In the 1950s, marginal tax rates—the amount of tax paid as income rises—were very high. By deducting the interest on the loan against his income, Knetsch believed that he could reduce his tax liability by more than $110,000 per year. Assuming tax rates remained the same, his yearly $40,000 pretax investment (making $140,000 in interest payments minus the $100,000 income he received from annuities) would produce a $70,000 after-tax profit for the next thirty years. Without considering the tax benefits, the transaction made no economic sense. But when the tax consequences were factored in, Knetsch had every reason to think he had a very good deal.

At the time Knetsch bought his annuities, there was no language in the Code explicitly stating that he could not deduct his loan interest payments in this type of arrangement. To the contrary, the Code stated that interest paid on indebtedness could be deducted, and did not indicate that debt incurred to buy this type of insurance should be treated differently from any other debt.³ Nonetheless, when the Supreme Court decided the case, it inquired into whether the interest payments that the taxpayer made to the insurance company “constituted ‘interest paid . . . on indebtedness’ within the meaning of” the tax laws. The Court observed that “Knetsch’s transaction with the insurance company did “not appreciably affect his beneficial interest except to reduce his tax.” Thus, the Court concluded, “there was nothing of substance to be realized by Knetsch from this transaction beyond a tax deduction.”⁴

Couldn’t Knetsch argue that the ability to deduct interest payments on his annuity reflected Congressional desire to encourage people to save for retirement? The response is that Congress intended only to benefit people who were “really” saving for retirement. Even though Knetsch purchased a retirement annuity, the transaction as a whole was not genuinely designed to provide retirement savings. As the Court put it, Knetsch’s savings plan for retirement was “a sham.”⁵ No economically rational person would pay $40,000 a year in order to receive $43 a month when he reached age ninety. Given this fact, the Court concluded that Knetsch’s only reason for engaging in these activities was to gain tax benefits that Congress had not intended to confer. His conduct therefore constituted an abuse of the tax law. In Knetsch, the Court reaffirmed the principle that even if taxpayers comply with the letter of a provision that provides tax benefits for certain activity, they will not receive those benefits if their behavior is not the type of activity the provision is designed to reward.

In Knetsch, as in earlier and later cases, the Court did not rely solely on the literal language of the statute. Rather, it determined that in permitting the deducting of interest on loans, Congress had intended only to cover actual indebtedness. Because the Court did not have a statement of purpose or specific wording on which to rely, it was articulating what it understood to be Congress’s intent. There was no reason to reward Knetsch under the tax code for spending the time and money to engage in these transactions because he wasn’t doing so to save for retirement in any meaningful sense. Without the tax benefits he sought, neither Knetsch nor society would be better off from him engaging in these transactions. Society, in fact, would be worse off—Knetsch would have wasted assets for no good reason. Why then, should the government reward him for doing so?

Anti-Abuse Doctrines

Overview

Because determining Congress’s intent can be difficult, courts over the years have developed what are known as “anti-abuse” doctrines to help distinguish legitimate from abusive shelters when Congressional intent is unclear. These doctrines reflect the idea that the costs associated with a transaction that is genuinely intended to produce income should be respected for the purpose of reducing taxable income. This is consistent with the principle of taxing net income. By contrast, unless they result from an activity that clearly falls within a category that Congress intended to favor, costs that are incurred only to obtain tax benefits cannot be used to reduce taxes. The fundamental issue underlying these doctrines is whether a claimed tax loss corresponds to an actual economic loss that makes the taxpayer worse off.

In disallowing the tax treatment Knetsch sought, the Court applied an anti-abuse doctrine that has become known as the “economic substance” doctrine. The origin of the doctrine is generally attributed to the 1935 U.S. Supreme Court case Gregory v. Helvering.⁶ In that case, a taxpayer owned an interest in a corporation that owned 1,000 shares of stock in another company. The taxpayer created a third corporation and transferred shares to it. Six days later, she dissolved this corporation and received the 1,000 shares. The taxpayer claimed for tax purposes that she had received the shares as a result of a corporate reorganization that resulted in taxation at a lower rate than had the shares been distributed to her via a dividend payment. The Supreme Court rejected the taxpayer’s claim that the shares were transferred to her as part of a reorganization, despite the fact that the steps taken conformed to the terms of the tax code. The Court empasized,

The whole undertaking, though conducted according to the terms of subdivision (B), was in fact an elaborate and devious form of conveyance masquerading as a corporate reorganization, and nothing else. . . . [T]he transaction upon its face lies outside the plain intent of the statute. To hold otherwise would be to exalt artifice above reality and to deprive the statutory provision in question of all serious purpose. ⁷

As tax scholar Leandra Lederman has noted, the economic substance doctrine requires that where a taxpayer claims tax-favored treatment for a transaction, if that transaction is not one that Congress obviously intended to encourage, it must have economic substance to obtain the taxpayer’s desired treatment.⁸ That is, the taxpayer’s economic position must change in a meaningful way as a result of the transaction. Over the years, the doctrine has received various formulations that have emphasized different factors. Among the factors the courts have considered are: whether a transaction holds a reasonable possibility for pretax profit,⁹ whether a transaction involves meaningful financial risk to the taxpayer,¹⁰ whether there is a reasonable ratio between a taxpayer’s investment and the tax benefit sought, and whether a taxpayer’s sole or primary purpose is to avoid taxes.¹¹

The economic substance doctrine and other anti-abuse doctrines flow from the focus of the income tax system on measuring net income. Consistent with this idea, costs associated with transactions that are intended, based on objective and subjective criteria, to produce income, are respected for purposes of taxation. By contrast, costs incurred as a result of transactions that are not designed to produce income—and which have not been singled out by Congress for favorable treatment—are not respected for tax purposes.

Anti-abuse doctrines created by the courts have engendered a great deal of controversy. Applying these doctrines involves difficult judgments and courts often seem to arrive at inconsistent results based on highly nuanced contextual differences. Determining Congress’s intent regarding an unanticipated situation often seems like guesswork, especially when the transaction in question has the same form as the category of transactions Congress has identified for favorable tax treatment. As a result, courts have produced a large body of cases with no easily discernible coherent thread, and no scholar has succeeded in coming up with a unifying account that has commanded general acceptance among tax specialists. These difficulties give rise to a reasonable objection: If divining some hypothetical legislative intent is such a challenge, wouldn’t it be preferable for courts to recognize the tax benefits claimed by a taxpayer in a given case and allow Congress to make its purpose explicit by amending the Code? Courts adopt this strategy all the time in other regulatory contexts.

As an alternative, why not at least require the IRS to identify those shelters that it regards as abusive? The agency has the authority to issue notices that describe those transactions for which taxpayers will not be permitted to claim tax losses. Until that occurs, why not permit taxpayers to claim benefits based on their literal compliance with the Code?

Judicial doctrines that are hard to define and apply raise deeper normative issues as well. Isn’t it unfair to taxpayers like Knetsch to disallow the tax benefits they claim when they relied on the language of a statute in planning their affairs? Don’t taxpayers have a right to try to minimize the taxes they owe? Disallowing the tax benefits Knetsch claimed on the ground that his transaction lacked economic substance—it didn’t involve actual “indebtedness” in the sense intended by Congress—was one approach to deal with his clever tax planning technique. Another approach would have been to allow Knetsch to prevail and invite Congress to make its intent clear by amending the statute. Indeed, by the time the Knetsch case had reached the Supreme Court, Congress had already changed the law—closed the loophole—to bar prospectively the technique Knetsch used. One can argue that such an approach is preferable because it defers to Congress to clarify its intent by enacting a provision that applies to the type of transaction at hand.

Notwithstanding their imprecision, however, doctrines that aim to prevent the use of abusive tax shelters play a critical role in limiting the vulnerability of the tax system to taxpayers’ efforts to avoid their obligations. While deference to Congress is often an appropriate approach in other regulatory contexts, in the tax shelter area economic substance and other anti-abuse judicial doctrines are necessary to safeguard the income tax system. This is true because the complexity of the tax law creates nearly limitless opportunities for taxpayers to structure their transactions to take advantage of formal features of the law, and because the income tax relies on a “self-assessment” system in which taxpayers are expected to calculate how much tax they owe.

Rules, Complexity, and Judicial Doctrines

The tremendous complexity of the Internal Revenue Code reflects in part the fact that it represents tradeoffs among several different considerations. In addition to the goals of taxing net income and encouraging certain behaviors, these considerations include economic efficiency, ease of administration, and horizontal and vertical equity—equity among taxpayers at the same income level and across income levels, respectively. The more complex a given law, the more finely it distinguishes among different individuals or transactions and the better it may be tailored to achieve its underlying objectives. As legal scholar David Weisbach notes, in tax “complex law reduces under- or over-taxation of particular activities relative to the desired amount as compared to simple law,” which uses broader categories.¹² In attempting a nuanced approach, the tax laws end up with a bewildering number of classifications and categories that cover the full spectrum of personal, professional, and business transactions and relationships that exist among U.S. taxpayers.

As Weisbach emphasizes, the complexity of the Code is also a function of taxpayers’ capacity to change the form of their transactions to obtain favorable tax treatment without changing the underlying economics. In this regard, the income tax differs significantly from other regulatory schemes. In most areas of law, it makes sense to draft the law to fit the most common circumstances rather than to attempt to devise a rule for every possible situation. As Weisbach observes, “If the law fits only common circumstances, rare and unusual facts might be mis-regulated, but under- or overdeterrence in these cases is not costly because the transactions are unusual.” That is, at some point, the costs of fitting the law to rare circumstances exceed the benefits of doing so. “More importantly,” Weisbach notes, “the uncommon transactions will remain uncommon . . . so one may be confident in the assessment of the costs and benefits” of the amount of specificity in the law.¹³

In tax law, in contrast, as Weisbach suggests, “uncommon transactions that are taxed inappropriately become common as taxpayers discover how to take advantage of them.”¹⁴ Congress therefore can’t afford to ignore the fact that these transactions will receive tax treatment that it regards as undesirable. The response to this situation in tax law typically is for Congress to draft additional rules to apply to the once uncommon transaction that has now become more common.

Weisbach describes one example of this dynamic. Taxpayers A and B each own properties that they would like to exchange with one another. If they do so, however, the exchange would constitute a “realization” event that would trigger tax liability for each. Put simply, were they just to exchange properties, the exchange would be considered essentially a sale and each taxpayer would have to pay tax on his or her profit. Therefore, instead of making that exchange, they each contribute their properties to a partnership, which then distributes to them each other’s property. Under now-repealed partnership tax rules, the contribution of the properties and their distribution in this “mixing bowl” transaction were both tax-free events. Thus, as Weisbach notes, A and B could avoid taxation “by using a partnership to do indirectly what would have been a realization event had it been done directly.”¹⁵

In response, Congress amended the partnership rules to specify precisely which transactions would be treated as contributions and distributions and which would not. The result, as Weisbach observes, is “a highly complex set of rules that distinguishes between partnership transactions and ‘mixing bowl’ sales, with safe harbor periods, presumptions, exceptions, required risk allocations, allowable preferred returns, accommodation partner rules, and dozens of examples.”¹⁶ These rules serve to classify transactions with various characteristics as equivalent either to a tax-free partnership transaction or a taxable sale or exchange.

Tax law makes hundreds of such distinctions, such as between debt and equity, sale and lease, and independent contractors and employees. In doing so, the tax law generally “must, on some basis, choose between the two extremes rather than characterize transactions in the middle as part of each.” This inevitably provides for significantly different treatment of activities just on each side of the dividing line. This discontinuity creates incentives for taxpayers to change the features of a transaction only enough to gain desirable tax consequences—to move the transaction just over the line to obtain tax preferred treatment. This type of arbitrage occurs to some extent under any regulatory scheme. It may be especially pronounced in tax law, however, because taxpayers can receive tax benefits by changing the form of a transaction without significantly changing its underlying economics.¹⁷

Given the vast and unforeseeable number of transactions that conceivably can occur, any system of rules that aims to classify these activities will tend to evolve toward ever-greater complexity. One feature of this complexity will be what Weisbach calls “interaction costs,” which “are the costs of ensuring that the various provisions of the law work together—that they do not conflict or have unintended gaps or loopholes.”¹⁸ As the number of rules increases, interaction costs rise dramatically. “[I]n the tax law context[,]” Weisbach notes, “the law must anticipate all of the interactions, even for rare transactions. Otherwise, rare transactions will become more frequent as taxpayers discover the tax benefits. Thus, as tax rules become complex, interaction costs increase rapidly and rules quickly become unmanageable.”¹⁹ Furthermore, as “technical complexity yields greater ambiguity,” this “creates more planning opportunities for those inclined to operate in the gray area of the law.”²⁰

Weisbach focuses on the theoretical reasons that adding rules gives rise to new opportunities for tax arbitrage. In practice, Congress’s frequent amendments to the rules have, if anything, exacerbated the problem. Typically, Congress focuses on fixing one problem at a time. As a result, tax authorities try “to shore up current law with thousands of stopgap measures,”²¹ and new provisions are often drafted with less than ideal clarity.²² In the meantime, other sections of the Code lie undisturbed for long periods, waiting for a creative shelter designer to mine their potential.²³ New abusive tax shelter opportunities pop up routinely.

Given these near limitless opportunities, relying on Congress or the Treasury to enact more rules is not a solution each time a new tax shelter comes to light. Anti-abuse doctrines are necessary to prevent taxpayers who want to take advantage of tax benefits that Congress has not expressly ruled out in the law. Otherwise, every year before Congress has a chance to act, well-off taxpayers who can afford sophisticated tax planning will be able to reduce or erase their tax liabilities. Consider again the transaction in Knetsch, likely the earliest mass-marketed abusive tax shelter on record. The total number of taxpayers who bought the transaction is not known, but we do know that Sam Houston Life Insurance regularly advertised the transaction in the Wall Street Journal and the business press and sent out unsolicited direct mailings to potential customers.²⁴ Sam Houston, moreover, wasn’t the only insurance company offering the same deal. Depending on how far and how fast these transactions spread, the Treasury was at risk of losing enormous amounts in tax revenue if it had to wait for Congress to amend the Code to disallow these shelters.²⁵

Self-Assessment and the Role of Opinion Letters

One other characteristic of the U.S. tax system contributes to tax shelter activity—the self-assessment nature of the regulatory scheme. It is a “self-assessment” system because the government relies on taxpayers to sort their transactions into the appropriate categories and determine how much tax they owe, rather than doing the work for the taxpayer. The IRS rarely audits tax returns to verify that these calculations are correct. In assessing their own taxes, taxpayers have strong incentives to understate what they owe.

In the United States and most other countries, it is the taxpayer’s responsibility to organize receipts and expenditures into the appropriate categories, determine their tax treatment, and calculate the amount of tax that is due. Taxpayers have much more information about their activities than does the government, which makes self-assessment cost-efficient as long as taxpayers provide honest information about their obligations. In addition, self-assessment can be seen as consistent with the principle of political self-determination, under which citizens impose obligations on themselves that the government generally trusts them to honor.²⁶ Although audit rates of wealthy Americans have risen since the late 1990s, at the turn of the twenty-first century the IRS reviewed less than one percent of returns to evaluate their accuracy. Then as now, in the overwhelming majority of cases, a taxpayer’s return was treated as presumptively accurate. Yet, as Professor Michael Doran observes, “[t]he federal government collects more than $1 trillion each year through the federal income tax, and those collections are premised on the notion that, for the most part, taxpayers will determine how much they owe and will remit those amounts to the government.”²⁷

Under this scheme, taxpayers are not simply passive objects of regulation whose obligations are determined by government officials. Rather, they play a crucial role in ensuring that the system functions by determining their own obligations themselves. When the taxpayers signs their return, they make the following statement: “Under penalties of perjury, I declare that I have examined this return, including accompanying schedules and statements, and to the best of my knowledge and belief it is true, correct, and complete.”²⁸ The statement expresses the idea that the government ultimately is heavily dependent upon taxpayers to act in good faith when they submit their returns.

Despite the expectation that taxpayers will honestly calculate their liability, tax law reflects a considerable amount of tolerance for self-serving behavior. Taxpayers are subject to minimal disclosure obligations and are given wide latitude to take positions that favor themselves at the expense of the government. They are not required, for instance, to disclose information that may weaken their self-assessment unless failure to do so would be fraudulent. Instead, the law simply imposes a higher penalty for substantial understatement of taxes if relevant information has not been disclosed. This gives taxpayers an opportunity to submit returns that reflect an interpretation of facts that best promote their interests without alerting the government to facts that may undermine that interpretation.

Contrast this with, say, environmental regulation, which imposes detailed reporting requirements on companies and subjects them to periodic inspections. If environmental law were structured like tax law, companies would submit conclusions about their compliance with the law to the government, but without necessarily providing information necessary to evaluate those conclusions. We would effectively have an “honor system” for pollution control—which, for the most part, is how the tax system works. Such a system makes the government vulnerable to taxpayers who are not willing to act in good faith in calculating their tax liabilities.

Although taxpayers are subject to penalties for understating the taxes they owe, these are generally believed to be too low to be effective deterrents. Taxpayers who engage in transactions whose primary purpose is to shelter income can avoid penalties if they satisfy certain objective and subjective criteria. Taxpayers must have “substantial authority” in the law for their claims that their assessment of their taxes is correct. This standard has been interpreted to mean that their legal position has at least a 35 percent chance of prevailing on the merits.²⁹ As a general rule, to avoid penalties for understating their taxes, taxpayers must act in good faith and their belief in the accuracy of their return must be reasonable.³⁰

Under regulations in effect in the late 1990s, the requirement that a taxpayer’s belief be “reasonable” was satisfied if a taxpayer relied in good faith on an opinion from a professional tax advisor who analyzed the pertinent facts and legal authorities and concluded unambiguously that “there [was] a greater than 50 percent likelihood that the tax treatment of the item will be upheld” if challenged by the IRS. The penalty provisions shifted the onus of determining whether a shelter was legitimate from the taxpayer to a legal advisor.

In a world in which tax professionals gave independent, well-researched, and thoughtful advice, this penalty scheme would function to prevent many abusive transactions. But financial incentives can easily distort the market for legal opinions. During the tax shelter boom, it was widely believed among tax practitioners and government officials that an opinion from any law firm protected a taxpayer from understatement penalties (as well as other civil and criminal liability). In tax shelter promoter circles, opinion letters became known as “get out of jail free cards.”

The tax self-assessment regime, which relies on the good faith of taxpayers and the objectivity of their advisors, is highly vulnerable to abuse. When shelters were widespread during the 1990s, the risk of penalties from participating in an improper shelter was minimal. To begin with, although it is not permissible to take into account the chance of audit in determining whether to take a position on a return, a sophisticated taxpayer was well aware that there was a very small likelihood that the return would be audited. Even if the return was audited and the shelter discovered, a taxpayer could avoid penalties by producing an opinion letter from a law firm that argued that a court would more likely than not uphold the transaction. In the highly unusual situation where an opinion letter was deemed inadequate to provide penalty protection, the taxpayer would be subject to, at most, a 20 percent penalty on the understatement of taxes. So taxpayers who used an abusive tax shelter to understate their taxes by $100,000—if caught and found not to have reasonable cause—would owe the government $120,000 plus the interest that had accrued on the amount due.³¹

Fairness and Other Normative Considerations

Arguments based on the distinctive features of the income tax laws explain how tax shelters arise and why judicial anti-abuse doctrines are necessary to combat them. They do not, however, address the fairness of applying these doctrines to taxpayers who rely on the language of a particular tax provision when they enter into transactions, only to discover after the fact that the tax benefit they counted on receiving has been disallowed, they will have to pay interest on the taxes due, and in the occasional case they may be subject to penalties.

The response is that tax is not the only regulatory scheme in which courts do not rely solely on the language of statutes but interpret them with a view to their purpose. Since Congress first began to legislate, federal courts have been required to apply its mandates to unforeseen circumstances. The issue of fair notice is no different in the tax context than in other regulatory contexts where courts have looked beyond the terms of a statute to determine how it might apply to a situation that Congress had not anticipated.³²

In addition, as in other regulatory contexts, in arranging their affairs, taxpayers are expected to determine not only what statutes say but also how courts have interpreted them. A taxpayer contemplating a transaction whose sole purpose is to avoid taxes—and which confers no economic benefits apart from its tax benefits—is presumed to be on notice that courts have invoked the economic substance doctrine to deny tax-favored treatment to those types of transactions. Knetsch and cases like it that apply the economic substance doctrine are “an inescapable part of the tax landscape.”³³

It is, of course, a fiction that taxpayers will be able on their own to determine how courts might treat a transaction for tax purposes. Penalties for understatement of tax are intended to encourage taxpayers who are unsure about their obligations to consult with tax practitioners to obtain advice about how courts are will likely treat shelter transactions if challenged by the IRS. If taxpayers rely on professional advice that their position more likely than not will be upheld, they won’t be subject to penalties. The worst that can happen to taxpayers who acted in good faith is that their lawyer was wrong and they owe interest and more taxes than they originally paid. Not a pleasant prospect for taxpayers who discover this long after the fact, but not a particularly unfair result, either.

Some commentators nonetheless might object that the tax system is different from other regulatory regimes and that taxpayers are entitled to use the formal categories provided in the tax laws to engage in transactions that minimize their taxes, regardless of their underlying economics.³⁴ This claim is captured in the oft-repeated quote from Gregory v. Helvering that “[t]he legal right of a taxpayer to decrease the amount of what would otherwise be his taxes, or altogether avoid them, by means which the law permits is not to be doubted.”³⁵

It is not clear what exactly the Court meant in stating that a taxpayer had “a legal right” to minimize her taxes. Immediately following this sentence, the Court went on to observe that: “the question for determination is whether what was done, apart from the tax motive, was the thing which the statute intended.” In addition, no right to minimize one’s taxes exists in the Constitution; nor has Congress created a right to avoid paying taxes subject to special solicitude or protection. As the Supreme Court’s subsequent decisions suggest, it was not identifying a specific right or entitlement; rather, it was simply noting that the motive to avoid taxes by itself does not render a transaction legally (or morally) suspect.³⁶ Indeed, as discussed earlier, many provisions in the Code seek to induce certain behaviors by enlisting taxpayers’ desire to avoid paying taxes. Anti-abuse doctrines are important to prevent major tax revenue losses and do not present particularly worrisome fairness concerns. They are also important from a social welfare perspective. If anti-abuse doctrines motivate tax professionals to give conservative advice that steers wide of the line between legitimate tax planning and abusive transactions, nothing valuable to society is lost.³⁷

Curbing abusive transactions therefore is an important social objective and anti-abuse doctrines can play a major role in accomplishing it. In the late twentieth century, however, tax shelters arose on an unprecedented scale. The first wave of shelters in the 1970s and 1980s had distinctive characteristics, and eventually were thwarted by new legislation. The wave that followed involved more varied and complex shelters that presented a more difficult challenge.

Tax Shelters Grow Up

While taxpayers undoubtedly have found ways to shelter income from taxation since the income tax was established in the early twentieth century, tax shelters became a widespread and more prominent activity in the 1970s and 1980s. One reason was the growing complexity of the tax law as a result of the cumulative effect of additional rules. Another was an increase in the use of the tax code to favor certain activities, as tax law became more popular as a way for Congress to promote social purposes while limiting budgetary outlays. This popularity continues to this day. As one experienced tax practitioner comments, “It is amazing how many things we are running through the Internal Revenue Code and through the Internal Revenue Service. We’re running education, we’re running energy, we’re running health care, we’re running retirement and I can go on and on.”³⁸ The result, according to one tax scholar, is “a variety of subsidies and economic stimuli that are unrelated to the administration and collection of taxes.”³⁹ This can create confusion about which tax benefits Congress has or has not intended to provide, as taxpayers structure their activities to comply with the formal requirements for receiving benefits.

In the early years of the tax shelter industry, promoters designed products mainly for high-wealth investors facing high marginal tax rates. From the end of World War II until 1981, individuals were subject to federal income tax rates that could reach 70 percent or more. Indeed, the taxpayer in the Knetsch case was subject to a marginal federal income tax rate exceeding 90 percent. As more taxpayers became interested in sheltering income, promoters expanded their market to upper middle-class taxpayers by syndicating tax-favored investments for groups of investors. That is, the shelter promoter formed a limited partnership to engage in these investments in which the taxpayers were limited, or passive, investors who did not participate in the business operations. The partnership invested in activities such as oil and gas exploration, real estate, and equipment leasing, as well as more exotic enterprises like llama breeding. Syndicators marketed these partnerships “largely to self-employed professionals such as medical doctors, dentists, architects, lawyers, and others who had substantial income from their professional services.”⁴⁰

These investments often produced deductions that taxpayers could use to defer taxation but not completely eliminate it. When the deferral ended and the time came to pay the tax, the income that had been sheltered could be treated not as so-called “ordinary income” but as capital gains, which is taxed at a lower rate. In a real estate investment, for instance, investors made modest contributions to a building project and the partnership borrowed the rest of the money. As two journalists described, “By deducting the interest and taking rapid depreciation write-offs, the partnership was able to show huge paper losses, which the investors divided among themselves and wrote off on their income tax returns.”⁴¹ Thus, for an investment of $10,000, a person in the 50 percent marginal tax bracket might acquire paper losses of $20,000 or more, thereby saving $10,000 in taxes. When the building finally was sold, the income from the sale would be taxed at the capital gains rate.

Taxpayers gained tax advantages not only from investments that deferred taxation and eventually taxed income as capital gains. They also took advantage of the growth in tax-exempt products such as state and local bonds, as well as opportunities to exclude from taxation earnings on the invested cash value of life insurance policies.⁴² As a result, “[d]emand for tax advantaged investment structures grew rapidly through the 1970s and 1980s as tax avoidance began to displace investment quality as a primary investment goal.”⁴³ In 1980, IRS Commissioner Jerome Kurtz reported that nearly 200,000 returns representing 18,000 shelters were involved in the examination and IRS appeals process. These returns, Kurtz maintained, involved almost $5 billion in “questionable deductions.”⁴⁴ Indeed, between 1979 and 1986, the annual number of tax shelter returns under examination by the IRS increased from 182,731 to 426,634.⁴⁵ Kurtz warned, “The great abuse we are finding in this area could result in a serious decline in taxpayers’ perception of the fairness and evenhandedness of our administration of the tax system and consequently in the level of voluntary compliance.”⁴⁶

The level of tax shelter activity during this period is reflected in data presented by Professor George Yin. From 1975 to 1986, the amount of net losses reported by individual taxpayers from their partnership and subchapter S corporation⁴⁷ investments grew by more than 525 percent, while the amount of reported net income increased by only a little over 136 percent.⁴⁸ Consistent with this, the number of individual income tax returns reporting net losses from such investments increased during this period by about 171 percent, while the number reporting net income grew by only 36 percent. As Yin observes, “The increase in the number of returns reporting losses is contrary to normal market expectations and suggests that some amount of the losses may have been non-economic or tax shelter losses.”⁴⁹

The government used every weapon in its arsenal to combat the shelters that arose in the 1970s and 1980s. It disallowed them on the ground that they had no economic substance or were not otherwise allowed under the tax regulations. It imposed new reporting and record-keeping requirements for tax shelters, adopted more stringent constraints on tax practitioners giving advice on tax shelters, and amended the tax law to change the treatment of certain activity associated with many of the shelters. The consensus, however, seems to be that tax shelter activity abated only when Congress adopted the “passive activity loss” and at-risk rules in the Tax Reform Act of 1986. This legislation addressed a characteristic common to most of the shelters, which was using passive investment losses to shelter ordinary income. Under those rules, losses from passive investments could only be used to offset income from those same investments, not other types of income.⁵⁰ In addition the rules required certain claimed losses to be limited to amounts a taxpayer had at risk in a transaction.⁵¹ The rules were “a major factor, if not the single most critical factor,” in curbing tax shelter activity. Data indicate that after 1986, increases in net income from investment in entities traditionally used to shelter income increased dramatically while increases in net losses were much smaller.⁵²

Congress seemed to have figured out how to stop the widespread abusive tax shelter activity of the 1970s and 1980s with the passive loss rules and other reforms in the 1986 Act. The Act also significantly lowered the highest marginal tax rate. As a consequence, tax policy analysts believed that high-bracket taxpayers would have much less of an incentive to engage in shelter transactions. For several years afterward, shelters existed on a much smaller scale than they had before. Many observers assumed that the government had shut down the tax shelter industry for all practical purposes.

But conditions were brewing for a revival. Intensified global economic competition, stunning wealth from the Internet boom, and greater business demands on law and accounting firms all helped create a new wave of shelters in the 1990s and early 2000s. These forces intersected to bring forth sophisticated abusive shelters on a scale unlike anything anyone had seen before.