Tax shelters and the Code: navigating between text and intent.

I. INTRODUCTION

Tax shelters raise difficult problems of statutory interpretation. In her interesting article, Of Lenity, Chevron, and KPMG, (1) Kristin Hickman explores one of them: the recent tendency of courts to apply the rule of lenity in civil cases, potentially leading to a narrow interpretation of the Internal Revenue Code (Code) that would undermine efforts to collect the taxes that Congress intended to impose. (2) We agree both with Hickman's articulation and analysis of this problem. In this commentary, we situate more broadly the status of tax shelters in the jurisprudence of statutory interpretation. More specifically, we show how tax shelters are a challenge to ordinary principles of statutory interpretation, whether or not lenity is expanded to make it even more difficult to combat them.

Scholars and judges have made a great deal over various differences in judicial philosophy in statutory interpretation. Most notable is the rift between those who claim to be textualists and those who advocate for more contextual analysis. (3) We acknowledge these differences, but believe that the problem of tax shelters would exist, more or less in the same form, regardless of how judges approach statutory interpretation.

At bottom, there is relatively strong consensus that the goal of statutory interpretation is to find and enforce the intended meaning of the legislature and that the best evidence of this intent is the language of the statute. Tax shelters take advantage of this consensus approach in two distinct ways. First, tax shelters take advantage of the flexibility of statutory words by structuring transactions that arguably come within statutory language that would license the avoidance of taxation, albeit in a non-prototypical sense. Second, tax shelters take advantage of gaps in legislation, where at most there is unexpressed legislative intent. Such gaps are an especially difficult problem because few, if any, statutory interpreters are comfortable construing a law beyond any fair meaning of its text.

Often, tax shelters can be characterized either way. Although they appear quaint by today's standards, it may be helpful to begin by remembering an older generation of tax shelters. A typical example might involve a passive investment by a lawyer or a doctor in an orange grove. (4) That investment would be motivated purely by the availability of tax benefits that participating taxpayers could use to "shelter" their professional income from taxation. One could frame the problem posed by such a tax shelter as a failure of a hypothetical statute to limit the sought-after tax benefits to "legitimate" farmers. (5) That failure could be a result of a gap in the underlying statute (e.g. the statute was created with farmers in mind, but fails to incorporate such a limitation) or of ambiguity in that statute (e.g. the statute limits the benefits to farmers, but employs a definition of farmer that could be interpreted either to include or exclude passive investors such as doctors and lawyers). In practice, the statutory vulnerability is likely to fall somewhere between a distinct gap and a straightforward ambiguity. For instance, the statute might explicitly refer to farmers, but provide no definition of the term. A court could invalidate such a shelter by filling the gap or by resolving the ambiguity against the taxpayer, declaring that Congress would not have intended the doctor in question to be treated as a farmer.

Alternatively, if courts were unwilling or unable to do either, Congress or perhaps the Treasury Department could intervene to prevent future revenue losses. They could do so by (i) specifying the characteristic (in this instance a lack of active participation in the orange grove's business so that the taxpayer is essentially purchasing tax benefits) that makes the doctor's ability to derive the unintended tax benefits "abusive" and (ii) using that characteristic to limit the availability of those benefits accordingly. (6)

Legislative or administrative action after the fact, however, is a second best solution in many cases, since it assumes a safe harbor for some period of time before the new laws are put in place. (7) Thus, if these sorts of shelters are to be rendered ineffective, principles beyond the interpretation of the language of the particular sections of the Code that govern the type of transaction in dispute must be invoked. (8) Hickman provides illustrations of both gap- and ambiguity-based tax shelters in her discussion of the KPMG case, and we will use her illustrations, among others, to make our points.

II. BETWEEN LANGUAGE AND INTENT: WHERE TAX SHELTERS RESIDE

For purposes of this essay, we adopt the position, taken by others, that tax shelters are generally characterized as transactions that appear to comply in a literal manner with the Code, but which are designed to reach a tax result that Congress would not have intended. (9) Transactions that take advantage of rules that intend to give favorable treatment, such as deducting mortgage interest on a principal place of residence, (10) are typically not considered to be tax shelters. (11) Nor are transactions that intentionally disobey the law. These constitute tax fraud. (12) Tax shelters, in contrast, involve the structuring of transactions to appear lawful while thwarting the intent of the legislature.

Which such devices should be allowed as a substantive matter is not a question of statutory interpretation, as Professor Weisbach points out. (13) Nonetheless, if the successful tax shelter must at least arguably comply with the Code and regulations, then it must do so in light of the various tools that courts use to determine the scope of statutes and regulations. The planner must structure the transaction to withstand judicial scrutiny despite the fact that those who wrote the law would likely have intended a different result.

A crucial question, then, is to what extent the courts take heed of the intent of the legislature in the interpretation of statutes. If courts did so to the exclusion of all other considerations, then tax shelters would never succeed. If, in contrast, legislative intent were irrelevant, they would always succeed. The reality lies somewhere in the middle. Courts are generally quite concerned with discovering and enforcing the will of the legislature. They are not in agreement, however, about what evidence of legislative intent is legitimate to consult, and legislative intent is not the only value to which they adhere. Other values, such as deference to administrative agencies--including the Internal Revenue Service (Service)--and fair notice--which underlies the rule of lenity--also play a role. These are the principles upon which Hickman relies in her analysis of tax shelters. We return to them in Part IV. First, we explore briefly the role that legislative intent plays in the interpretation of statutes.

Nearly all judges, even textualist judges who profess otherwise, attempt to ascertain the intent of the legislature and to apply the statute in a way that will further that intent. In fact, a Lexis survey of judicial decisions from the 1990s found that federal courts used intent and related words (i.e., intention, intend, etc.) within six words of Congress or legislature more than 3000 times each year during the decade, and that state courts did the same. (14) That amounts to a minimum of 60,000 instances in a ten-year period.

How do courts find legislative intent? They look first to the language of the statute. For example, in an opinion construing the innocent infringer provision of the Lanham Act, the Fifth Circuit wrote, "Our starting point in divining the meaning of a statute is the intent of Congress. The best evidence of this intent is the language of the statute." (15) Pronouncements relating language and legislative intent are easy to find in judicial opinions. (16)

This does not end the inquiry, however. It only begins to define it. Once language comes into play, judges do not always agree upon where ambiguities lie and how to resolve them. Consider Smith v. United States, (17) a case widely discussed in the scholarly literature on statutory interpretation. (18) In that case, the defendant had attempted to trade an unloaded machinegun for illegal drugs and was prosecuted for attempting to "use a firearm ... during and in relation to a drug trafficking offense." (19) A majority of six justices voted to affirm the conviction, based on the meanings of the words as revealed in a host of dictionaries, the relationship between the statute at issue and related statutes, and the likely intent of the legislature based on the proliferation of guns. The majority correctly found that the act of trading a firearm for drugs can properly be seen as using the firearm.

Justice Scalia dissented, arguing that the "ordinary meaning" of "use a firearm" means to use it as a firearm, not merely as a thing of value. (20) He, too, was correct. When we use the expression "use a firearm" we typically have in mind using it for its intended purpose. Smith illustrates an important point. Both the majority and dissent claimed fidelity to the intent of the legislature. (21) They both sought to discover that intent from the language of the statute. Yet they disagreed on what that intent was because it is possible to draw various inferences from the language, depending upon whether one asks about the outer bounds of the language or the most likely intended meaning. Thus, even those who claim to rely on a statute's language and to reject legislative intent recognize that they actually rely on language in order to ascertain intent.