Response to: Alex Raskolnikov, Revealing Choices: Using Taxpayer Choice to Target Tax Enforcement, 109 Colum. L. Rev. 689 (2009).
In Revealing Choices: Using Taxpayer Choice to Target Tax Enforcement, Alex Raskolnikov proposes replacing the current federal income tax compliance regime—which he describes as a "one-size-fits-all" system—with a two-track approach.1 When filing their annual returns, taxpayers would be required to choose between a "compliance regime" (CR) and a "deterrence regime" (DR). The two regimes would be designed so that "gamers"2 would choose the DR, while all other taxpayers3 would choose the CR. The DR "will look very similar to the current [tax compliance regime], except statutory fines . . . may be higher (perhaps much higher) than they are today."4 The CR would feature much lower penalties than the DR. It would also include a number of "separating features" that would be acceptable to non-gamers but not to gamers; the separating features would drive gamers into the DR despite the DR's much higher penalties.5 The two separating features of the CR that Raskolnikov discusses in the greatest detail are mandatory binding arbitration and the use of a strong pro-government presumption in the resolution of arbitrated disputes.6 Both of these features would make it much easier for the Internal Revenue Service (IRS) to win disputes in the CR than in the DR. Non-gamers would not object to this, because they do not take tax reporting positions that they expect the IRS to contest. Gamers, by contrast, routinely take aggressive reporting positions, and they would have to choose the DR to preserve their chances of prevailing if the IRS detects and challenges their positions.
As Raskolnikov explains in detail, the observation that different taxpayers have different attitudes toward their compliance obligations is not original with him.7 What is original is his proposal to respond to those differing attitudes by forcing taxpayers to reveal their compliance types, and then subjecting taxpayers to different enforcement regimes based on those revelations. Raskolnikov's proposal is an important addition to the scholarly literature on tax enforcement strategies. In my view, however, the current federal income tax enforcement regime has already slouched some considerable distance away from the one-size-fits-all model, in a manner not noted by Raskolnikov, by imposing stringent disclosure requirements on the most important category of gamers—the users of tax shelters. Part I of this Response describes the current tax shelter enforcement regime. Part II compares Raskolnikov's proposal with the current regime, and suggests that there are grounds for preferring the current regime to the proposal.
I. The Current Tax Shelter Enforcement Regime
Traditionally, a major attraction of tax shelters8 to gamers was the opportunity to play the audit lottery. In many—probably most—situations, gamer taxpayers using cost-benefit analysis would have decided against involvement in tax shelters if they knew the shelters would be detected and challenged by the IRS. Taking into account the low probability of prevailing in court, the likely imposition of an accuracy-related penalty,9 and the costs of litigation, the expected value of the tax shelter commonly would have been negative. The calculus changed dramatically, however, once the odds of escaping detection (either by not being audited, or by the auditor missing the issue) were factored in. The chances of escaping detection were high enough to induce gamers to participate in many tax shelters that they would have avoided if detection had been certain, or nearly certain.
In the wake of recent statutory and regulatory developments, however, taxpayers should assume that there is a near one hundred percent chance that the IRS will detect their shelters. Treasury regulations promulgated in 2003 require any taxpayer who has participated in a "reportable transaction" to file a "reportable transaction disclosure statement" with the IRS as a tax return attachment.10 Reportable transactions are defined broadly enough to include most tax shelters.11 Especially significant are the inclusions within the reportable transactions definition of "listed transactions"12 and "transactions of interest."13 In 2004, Congress enacted a series of tax shelter compliance provisions on the foundation of the 2003 regulatory reform. If a taxpayer fails to comply with the regulatory disclosure requirements, the Internal Revenue Code imposes substantial penalties—for individual taxpayers, $10,000 for each failure (increased to $100,000 per failure for listed transactions).14
If the taxpayer fails to disclose despite the threat of the nondisclosure penalty, the IRS is still likely to receive notice of the taxpayer's shelter from the tax shelter promoter. Section 6111 requires a "material advisor with respect to any reportable transaction" to disclose the details of the transaction to the IRS,15 and section 6112 requires the advisor to maintain and make available to the IRS for inspection a list of the taxpayers investing in the shelter.16 The Code imposes substantial penalties on promoters who fail to comply with these disclosure requirements.17
If either the taxpayer or the tax shelter promoter complies with the disclosure requirements, the taxpayer will assume—probably correctly—that the IRS will detect the shelter and challenge the claimed tax benefits. With the audit lottery thus nearly eliminated as a factor in the cost-benefit calculations, gamers should reject many shelters that would have attracted them before the imposition of the disclosure requirements. Of course, the disclosure requirements solve the government's audit lottery problem only if taxpayers, their material advisors, or both, comply with the requirements. The new disclosure regime would fail if gamers and their advisors were to perform cost-benefit analyses indicating that the optimal strategy is to ignore the disclosure requirements and hope the IRS remains oblivious. Judging from the comments of top-level Treasury and IRS officials, however, it appears that strategic noncompliance has not been a major problem. In 2006, then-IRS Commissioner Mark Everson told a Senate panel, "No longer are abusive tax shelters being marketed by top level accounting firms."18 Similarly, late last year outgoing IRS Chief Counsel Donald Korb remarked that "we—the IRS, Treasury, [the] Justice [Department], the whole team—really turned the corner" on tax shelters.19 Apparently the penalties for nondisclosure—which include, in addition to the civil penalties described above, possible prosecution under § 7203 for the misdemeanor offense of willfully failing to keep required tax records or supply required tax information—have been sufficient to dissuade gamers from continuing to play the tax shelter audit lottery by intentionally failing to comply with the new disclosure requirements.
II. Identifying and Evaluating the Differences Between the Current Tax Shelter Disclosure Regime and the Raskolnikov Proposal
The tax shelter disclosure requirements described above can be understood as a mechanism for separating gamers from non-gamers for the purpose of applying different enforcement regimes—in particular, different audit rates—to the two types of taxpayers. So understood, the disclosure requirements are very much in the spirit of Raskolnikov's proposal. Nevertheless, the two approaches differ in several important ways: (1) by attempting to identify only tax shelter investors rather than all gamers, the disclosure regime covers a smaller portion of the gamer population than would Raskolnikov's DR; (2) the current disclosure regime relies on very high odds of detection combined with modest understatement penalties, whereas Raskolnikov's DR would rely on very large understatement penalties combined with relatively low odds of detection; and (3) the current disclosure regime has no effect on the enforcement regime applicable to non-gamers, whereas Raskolnikov's proposal would radically change the rules applicable to non-gamers under the CR. The three differences are described and evaluated below.
A. Tax Shelter Investors as Only a Subset of the Gamer Population
Raskolnikov includes within the gamer population not only tax shelter investors, but also tax evaders who "underpay their taxes without any justification at all,"20 and non-shelter tax avoiders (including those who "take aggressive reporting positions all by themselves,"21 without professional assistance). Raskolnikov's hope is that his proposed DR will attract not only tax shelter users, but also evaders and non-shelter avoiders. By contrast, the current disclosure-based antishelter regime applies only to tax shelter users. An advantage, then, of Raskolnikov's proposal over current law is its applicability to a larger percentage of the total gamer population. It is debatable, however, whether this is a very significant advantage. Raskolnikov himself suggests that the primary targets of his proposed DR are those taxpayers "interested in using tax shelters to game the system."22
It appears, moreover, that there is no real difference between the DR and the current regime with respect to the treatment of evaders. Raskolnikov concedes that evaders will generally opt for the CR rather than the DR. In the CR they avoid exposure to the DR's high penalty rates, and they do not mind the pro-IRS tilt to dispute resolution in the CR because they have no hope of prevailing on the merits under either regime.23 Raskolnikov's response to this problem—which he acknowledges has "no perfect solutions"24—is to identify various situations in which it seems reasonable to force probable evaders into the DR by denying them the CR option.25 His examples include taxpayers who do not file returns, taxpayers who do not pay assessed tax liabilities, "informal suppliers" in the cash economy, and taxpayers previously found guilty of tax evasion in the CR.26
Something along these lines may indeed be the best that can be done with this difficult problem, but forcing some taxpayers into the DR is inconsistent with the choice-based rationale for Raskolnikov's two-regime proposal. A major justification for the two-regime approach is that the high DR penalties necessary to deter gamers are made politically acceptable by the fact that the DR penalties apply only to those taxpayers who elect the DR. In Raskolnikov's words, "Taxpayer choice, of course, is at the center of the dual regime. And so it should be. It is clear that a substantial increase in statutory fines is all but impossible if the fines are applied across the board."27 Raskolnikov's solution to the problem of evaders, however, is to impose the DR and its high penalties on certain categories of probable evaders. To the extent higher penalty rates on certain types of taxpayers are politically and morally acceptable in the absence of the taxpayer choice justification,28 higher penalties for these disfavored types could be an add-on feature of the current regime (i.e., a tax shelter disclosure regime combined with special high penalty rates applicable to specified likely-evader categories) just as easily as they could be an add-on feature of Raskolnikov's ordinarily choice-based system. In short, evaders are really outside both the current tax shelter disclosure regime and Raskolnikov's choice-based proposal, and so provide no basis for preferring one regime to the other.
The case of non-shelter avoiders is different from that of outright evaders. Many non-shelter avoiders would opt into the DR under Rasknolnikov's proposal, but they are unaffected by the current tax shelter disclosure requirements. In this respect, Raskolnikov's proposal may be superior to the present regime. It is far from clear, however, that this superiority is significant enough to justify preferring the proposal to current law. As discussed below,29 the proposal requires drastic changes in the compliance rules applicable to non-gamers, and does so primarily for the purpose of facilitating the separation of gamers from non-gamers—rather than because the CR is inherently more appropriate than current law for non-gamers. If one is not convinced of the inherent merits of applying the CR to non-gamers—for example, of denying non-gamers access to the courts, and of arbitrating non-gamers' disputes with the IRS under a strong pro-government presumption—the need to change the rules applicable to non-gamers in order to improve compliance among non-shelter avoiders may be a significant disadvantage of the proposal. In deciding whether that disadvantage is outweighed by the desirable effects of the proposal on non-shelter avoiders, it would be very helpful to have some sense of the magnitude of the non-shelter avoider problem, in terms of both the number of taxpayers of that type, and the dollars of tax avoided. However, the magnitude of this problem is not made clear in Raskolnikov's discussion.30
B. Deterrence by High Penalties Versus Deterrence by Near-Certainty of Detection
Under Raskolnikov's proposal, the government would commit to "not us[ing] the new information about taxpayers (their choice of regime) to adjust its examination strategy."31 Thus, any difference between audit rates in the DR and those in the CR would be attributable to factors other than taxpayers' regime choices. Tax shelter users in the DR would face very high penalty rates, but would have a significant chance of avoiding audit.32 Under current law, by contrast, tax shelter users face a near certainty of detection (assuming the taxpayers comply with the tax shelter disclosure requirements), but modest penalty rates. Section 6662A imposes a penalty at the rate of twenty percent on "reportable transaction understatements," with the rate increased to thirty percent in the case of an understatement with respect to a position not disclosed on the taxpayer's return.33 The usual twenty percent penalty rate under section 6662A is identical to the twenty percent penalty rate applicable with respect to non-tax shelter substantial understatements under section 6662.34
Of course, any desired level of deterrence—that is, any expected cost of tax avoidance—can be produced either by a combination of high penalties and low odds of detection (as in Raskolnikov's approach to tax shelters), or by a combination of low penalties and high odds of detection (as under the current tax shelter disclosure regime). Given the deterrence equivalence of the two approaches, the choice between them must be made on other grounds. There is a strong horizontal equity argument in favor of the low-penalties-plus-nearly-certain-detection approach of current law. Suppose the appropriate level of deterrence with respect to ten taxpayers, each of whom uses a tax shelter in an attempt to avoid $100,000 of tax, can be produced by either (1) subjecting each taxpayer to a one hundred percent chance of detection and the imposition of a twenty percent penalty (resulting in a certain tax-plus-penalty of $120,000 for each taxpayer), or (2) subjecting each taxpayer to a ten percent chance of detection and a tax-plus-penalty of $1,200,000 in the case of detection. Given the fact that all ten taxpayers are identically situated (except with respect to their good or bad luck, in the case of the second regime), the regime that treats all ten identically is fairer than the regime which permits nine to avoid tax while imposing a whopping penalty on the unlucky tenth. This is a powerful reason for preferring the tax shelter disclosure rules of current law to the proposed DR.
C. Are the Features of the CR Attractive for Their Own Sake, or Only as "Separating Features"?
In explaining the rationales for the various details of his proposal, Raskolnikov distinguishes between details intended as "separating" features and details intended as "targeting" features.35 Separating features "create incentives that compel gamers to choose DR while inducing non-gamers to opt for CR."36 Targeting features, by contrast, "facilitate[] compliance by each group of taxpayers."37 Although some aspects of the proposal might serve both separating and targeting goals, Raskolnikov explains that the two aspects of the CR that he discusses in the greatest detail—mandatory binding arbitration and a strong pro-government presumption—serve only the separating goal: "[They] do not strengthen a sense of duty, trust in government, or feelings of reciprocity. These features are unrelated to targeting."38
Mandatory binding arbitration and a strong pro-government presumption would constitute drastic changes from the current enforcement regime for those taxpayers—probably the substantial majority of all individual taxpayers—who opt for the CR.39 Arbitration and a pro-government presumption would be useful in driving gamers away from the CR and into the DR, despite the much higher penalty rates in the DR. It is not at all clear, however, that these key features of the CR are desirable apart from their facilitation of the imposition of higher penalties on gamers in the DR. Raskolnikov does not claim the features are desirable for their own sake. In fact, he acknowledges that both features are troubling when considered apart from their separating virtues:
The pro-IRS presumption . . . will make it difficult for taxpayers to challenge government's interpretation of the law. [It] will also limit taxpayers' opportunities to test the line between permissible and illegal positions. Binding arbitration will undermine the development of precedent, inhibit public articulation of values, and conceal overly aggressive enforcement from public scrutiny and judicial review.40
Having noted these drawbacks, Raskolnikov comments that the existence of the DR as an alternative "will alleviate most of these concerns."41 I am not convinced. If individual taxpayers overwhelmingly elect the CR—feeling almost compelled to do so out of fear of the high penalties in the DR, and out of concern that it is not a good idea to give the IRS the (mis)impression that one is a gamer—then the existence of the DR does little to make the separating features of the CR more acceptable.
It would be one thing to propose a greatly changed tax enforcement regime for non-gamers on the grounds that the proposed regime would be more appropriate than the current regime for non-gamers. It is a very different thing to propose a greatly changed tax enforcement regime for non-gamers in order to facilitate the imposition of a different regime on gamers (primarily, to make feasible the imposition on gamers of understatement penalties much higher than those of current law). Putting aside separation issues, I am inclined to think that access to the courts for non-gamers and a level playing field for non-gamers in disputes with the IRS, are normatively superior to the contrary features of the proposed CR. The case for the superiority of the current non-gamer regime (again, setting aside separation issues) becomes even stronger when one considers the disruption that the introduction of Raskolnikov's proposal would cause—both because any major change is inherently disruptive and because the new system would require taxpayers to make one more difficult decision each year.
All these undesirable effects of the proposal on non-gamers might be worth tolerating if the proposal were the only way to achieve an effective enforcement regime for gamers. As discussed above, however, a strong case can be made that the current tax shelter disclosure rules are a satisfactory response to the gamer enforcement problem.42 In deciding whether the current tax shelter disclosure regime or Raskolnikov's proposal constitutes the better approach to the gamer problem, a major point in the current regime's favor is that it achieves heightened deterrence for tax shelter gamers without the need for a radical revision (in fact, without the need for any revision at all) of the enforcement regime applicable to the non-gamer majority of the taxpaying population.
But this should not be the end of the analysis. If Raskolnikov's separating features accomplished their separation goal, it would then be possible to introduce targeting features into the CR that would not be appropriate for a population that included a large number of gamers; cooperative audits and cost reimbursements for prevailing taxpayers are two possibilities considered by Raskolnikov.43 Although I am not yet persuaded, it is conceivable that overall the CR—considering both the separating features and the targeting features made possible by the separating features—might be superior to the current system as an enforcement regime for non-gamers.
Conclusion
By explaining the advantages of applying different enforcement regimes to taxpayers with different motivations and by offering a detailed description of how such regimes might be designed, Raskolnikov has made an important contribution to the scholarly literature on tax compliance and enforcement strategies. In addition to the attractions of the specific proposal for a DR and a CR, Raskolnikov's article is also valuable for enabling the reader to understand the current system in a new light—as already applying a special enforcement regime to the most important category of gamers through its stringent tax shelter disclosure requirements. Whether the current system's way of dealing with gamers is preferable to Raskolnikov's proposal is a difficult question. All things considered—including the effect of each approach on non-gamers—I am inclined to think the current system is the better choice. But neither Raskolnikov's article nor this Response has attempted a comprehensive evaluation of the relative merits of deterring gamers via Raskolnikov's DR, versus deterring gamers via stringent tax shelter disclosure requirements. On that important question, much work remains to be done.
* Pamela B. Gann Professor of Law, Duke Law School.
1 Alex Raskolnikov, Revealing Choices: Using Taxpayer Choice to Target Tax Enforcement, 109 Colum. L. Rev. 689, 691 (2009).
2 Gamers are "rational actors whose marginal compliance decisions depend primarily on the expected tax penalty calculus." Id. at 692.
3 These are taxpayers who base their taxpaying decisions on any grounds other than the expected tax penalty calculus—for example, on a sense of duty or reciprocity. Id. at 691–92.
4 Id. at 691. Raskolnikov explains that the logic of his targeting approach works as long as penalty rates are higher in the DR than in the CR, regardless of the absolute penalty levels in either regime. Id. at 714–15. In theory, then, his proposal could be implemented by maintaining current (low) penalty rates in the DR and introducing new (even lower) rates in the CR. In practical terms, however, the attraction of his proposal is in the way it makes possible the imposition of higher-than-current penalty rates on gamers in the DR, without the need to impose those same rates on non-gamers in the CR. This ability to impose greatly increased penalty rates in the DR is a major theme of Raskolnikov's article. See, e.g., id. at 691, 741. In this Response, then, I assume that Raskolnikov's proposal would be implemented by maintaining current penalty rates in the CR and introducing higher rates in the DR, rather than by maintaining current penalty rates in the DR and introducing lower rates in the CR.
5 Id. at 718–22.
6 Id.
7 Id. at 694–701 (reviewing relevant literature).
8 The best non-technical definition of tax shelters comes from Michael J. Graetz: "[D]eal[s] done by very smart people that, absent tax considerations, would be very stupid." Lynnley Browning, How to Know When a Tax Deal Isn't a Good Deal, N.Y. Times, Sept. 10, 2008, Wealth & Personal Finance (Special Section), at 6.
9 See I.R.C. § 6662 (2006) (imposing twenty percent penalty on substantial understatement of income tax liability).
10 Treas. Reg. § 1.6011-4(a) (as amended in 2008).
11 Id. § 1.6011-4(b).
12 Id. § 1.6011-4(b)(2). The regulation defines a listed transaction as "a transaction that is the same as or substantially similar to one of the types of transactions that the Internal Revenue Service (IRS) has determined to be a tax avoidance transaction and identified by notice, regulation, or other form of published guidance as a listed transaction." Id.
13 Id. § 1.6011-4(b)(6). The regulation defines a transaction of interest as "a transaction that is the same as or substantially similar to one of the types of transactions that the IRS has identified by notice, regulation, or other form of published guidance as a transaction of interest." Id.
14 American Jobs Creation Act of 2004, Pub. L. No. 108-357, § 811, 118 Stat. 1418, 1575–77 (codified at I.R.C. § 6707A (2006)). For corporate taxpayers, the penalty is $50,000 for each failure (increased to $200,000 per failure for listed transactions). Id.
15 I.R.C. § 6111 (2006).
16 Id. § 6112.
17 For failures to comply with the disclosure requirements of § 6111, the penalty is $50,000 for each failure, or the greater of either $200,000 or fifty percent (seventy-five percent in the case of an intentional failure) of the material advisor's gross income from activities related to the shelter in the case of a listed transaction. Id. § 6707. For failing to make the list of taxpayer investors available to the IRS within twenty days of a request, the penalty is $10,000 for each day beyond the twentieth. Id. § 6708.
18 Deconstructing the Tax Code: Uncollected Taxes and Issues of Transparency: Hearing Before the Subcomm. on Fed. Financial Management, Government Information, and International Security of the S. Comm. on Homeland Security and Governmental Affairs, 109th Cong. 54 (2006) (written testimony of Mark Everson, Comm'r of Internal Revenue).
19 Jeremiah Coder, Korb Reflects on Long Tenure as Chief Counsel, 122 Tax Notes 20, 21 (2009); see also Browning, supra note 8, at 6 ("[T]he mass-marketed, abusive tax shelters, with names like Cobra and Son of Boss, that were sold to thousands of affluent investors from the mid-to-late 1990s to [the] early 2000[s] . . . are gone, having been sniffed out by the I.R.S. . . . .").
20 Raskolnikov, supra note 1, at 717.
21 Id.
22 Id. at 691.
23 Id. at 748.
24 Id.
25 Id. at 748–50.
26 Id. at 749–50.
27 Id. at 742.
28 For whatever it may be worth, my own sense is that it may be politically and morally acceptable to force into the DR non-filers, non-payers, and previous evaders in the CR, but neither politically nor morally acceptable to force informal suppliers into the DR merely because tax evasion is widespread in the cash economy.
29 See infra Part II.C.
30 Raskolnikov, supra note 1, at 726 (describing non-shelter avoiders in qualitative terms, but making no attempt at quantification).
31 Id. at 752.
32 Raskolnikov does not discuss what would happen to tax shelter disclosure rules under his proposal, but the logic of his proposal suggests they would be repealed. If they were not repealed, audit rates in the DR for tax shelter users—the most important category of taxpayers in the DR—would be much higher than audit rates in the CR, and Raskolnikov indicates he does not want that result. Id. at 752–53. The purpose of the high penalties in the DR is to deter taxpayers despite a significant chance of avoiding detection; the high penalties would not be needed if detection were nearly certain as a result of tax shelter disclosure requirements.
33 I.R.C. § 6662A(a) (2006) (twenty percent penalty); id. § 6662A(c) (thirty percent penalty).
34 See id. §§ 6662(a), 6662(b)(2).
35 Raskolnikov, supra note 1, at 739–40.
36 Id. at 739.
37 Id.
38 Id.
39 Raskolnikov does not predict what percentage of the taxpaying population would choose the CR, but he does suggest—quite plausibly—that most taxpayers are non-gamers. Id. at 701 ("In sum, gamers are a fairly narrow group . . . . Non-gamers, in contrast, are a broad category . . . ."). If his suggestion is correct, and if the separating features of his proposal are effective, then the substantial majority of individual taxpayers will choose the CR.
40 Id. at 741.
41 Id.
42 See supra Part II.A.
43 Raskolnikov, supra note 1, at 734–37.
Preferred Citation: Lawrence Zelenak, Tax Enforcement for Gamers: High
Penalties or Strict Disclosure Rules?, 109 Colum.
L. Rev. Sidebar 55 (2009), http://www.columbialawreview.org/Sidebar/volume/109/
55_Zelenak.pdf.
