Complex “derivative” financial instruments are often used in aggressive tax planning. In response, the government has implemented mark-to-market type reforms, but only partially. Considered in isolation,these incremental reforms are likely to seem well advised in measuring income more accurately. However, there is an important “second best” cost, emphasized in this Article: the ability of well-advised taxpayers either to avoid the new rule or to turn it to their advantage (here called “defensive” and “offensive” planning options, respectively). This Article uses two case studies to identify how these effects arise and to suggest ways of combating them. The first case study, Section 475, requires securities dealers to use mark-to- market accounting. Although this rule curtails tax planning by securities dealers themselves, it enables dealers to serve as accommodation parties for their clients’ tax planning: Once exempted from generally applicable rules, dealers can offer clients a tax benefit (e.g.,accelerated losses) without experiencing a corresponding tax cost (e.g., accelerated income). The second case study, the contingent debt regulations, requires lenders and borrowers to report pre-realization gains and losses based on assumed annual returns. Although this reform seems to accelerate the lender’s interest income, the rule’s narrow scope allows tax-sensitive lenders to avoid this result. Accordingly, the new rule is likely to apply only when tax-exempt entities lend to tax-sensitive borrowers, who enjoy the regulations’ accelerated interest deductions. This Article offers ways to remedy these reforms, as well as general guidance about how to implement incremental mark- to-market reforms without exacerbating the planning option.
Every stick crafted to beat on the head of a taxpayer will, sooner or later, metamorphose into a large green snake and bite the Commissioner on the hind part.
Professor Martin D. Ginsburg
The most important tax problem of recent months is the impact of aggressive tax planning on corporate tax revenue. The Secretary of the Treasury blames the “tax shelter industry,” in which tax lawyer sand investment bankers develop and market tax-motivated transactions. This Article analyzes the problem of aggressive tax planning, and recommends ways to impede it, in a context rife with opportunities for planning: the tax rules for complex financial instruments known as derivatives. While planning opportunities are prevalent elsewhere in the tax law as well, this Article focuses on derivatives because the problem is particularly acute – indeed, derivatives have been called “[t]he 900-pound gorilla in all the corporate tax shelter discussion” – and because the shelters) conventionally accepted solution has been applied (or misapplied) in ways that, ironically, may have made the problem worse. The need for reform in this area is widely acknowledged, but existing scholarly guidance is not adequate: There is a consensus that many difficulties would be solved with a comprehensive shift in tax timing rules – from realization to mark-to-market produce a precise and predictable financial result, at a known level of risk. Derivatives allow planners to negate, extend, or expand the formal arrangements and results on which tax liability is based. Derivatives turbocharge tax shelters.”) accounting or a more administrable proxy – for all assets and taxpayers. Yet there is also a consensus that such comprehensive reform is not administrable or politically feasible in the near term. The twin hurdles of administrability and politics are thought to be too high.
The question remains, then, what to do in this “second best” world? The government’s answer has been incremental mark-to-market reforms  which, considered in isolation, are each likely to seem advisable. In departing from realization, each incremental reform is likely to measure income more accurately. Accuracy is valuable for its own sake (e.g., in improving confidence in the tax system) and can also be a means of promoting the more important values of efficiency and equity. In some cases, however, the effects of these accuracy-enhancing reforms are likely to be ambiguous or even counterproductive. This counterintuitive result arises when the reform’s scope is so narrow that well-advised taxpayers can effectively opt in and out, and are governed by the new rule only when it reduces their tax burden. These new planning opportunities can prevent the reform’s promised benefits from arising, while increasing compliance costs and distortions in taxpayer behavior.
Incremental reforms vary in their effect on the planning option. I use the term “planning option” to describe the ability of well-advised taxpayers to attain better tax treatment by restructuring their transaction. This Article describes preconditions for the planning option, classifies different ways a reform can create new planning opportunities, and suggests ways of revising reforms to head off (or at least mitigate) this effect. After offering general principles, this Article illustrates them with two case studies: Section 475, which generally requires securities dealers to mark all their securities to market, while allowing securities traders and commodities dealers to elect this treatment; and the contingent debt rules of Treas. Reg. 1.1275-4, which require lenders and borrowers to report pre-realization gains and losses based on assumed annual returns, rather than market conditions. The objective of this Article is not to offer a global solution for the taxation of derivatives. Skepticism is warranted about whether there is a single magic bullet, or at least a politically attainable one. Instead, this Article offers a measure — the reform’s effect on the planning option — that should be an integral part of evaluations of proposals in this area.
In particular, this Article focuses on three ways a reform can aggravate the planning option. First, if the rule applies only to some groups of taxpayers but not others – a so-called taxpayer-based classification – the rule can enable one group to serve as an accommodation party for another. Unfortunately, in applying a new timing rule (i.e., mark-to-market) only to securities dealers and traders, Congress singled out groups uniquely suited to serve as accommodation parties. Second, when incremental-reforms are applied to some types of transactions but not to others – so-called transactional classifications – a key question is whether the new rule covers all comparable transactions. If not, taxpayers will be able to avoid the new rule, an effect here called the “defensive planning option,” which has arisen with the contingent debt rules. Finally, instead of avoiding a rule, sometimes taxpayers will deliberately qualify if the reform offers a favorable result they cannot otherwise obtain. For instance,new transactional classifications create such an “offensive planning option” in offering readier access to tax losses than was previously available, as the contingent debt rules do
In focusing on the planning option, I do not mean to suggest that this criterion always should be dispositive. It is an important component of any normative analysis, but it is not the only component. For instance, to assess whether a tax rule is efficient, we must consider not only its net effect on tax planning, but also the administrability burdens the rule imposes (e.g., compliance and auditing costs) and the overall level of tax on the activity (e.g., whether the tax being avoided is a sensible tax). A reform that creates new planning opportunities may still be justified if the reform offers off setting advantages. At bottom, the rule’s desirability requires an empirical assessment of the relative magnitude of these various effects, although empirical data may be hard to acquire. The point here is not to assemble this data or to offer a definitive judgment about any particular reform but, instead, to highlight an important cost of incremental tax reform and to understand what factors are likely to influence the magnitude of this cost.
The planning option should be studied not only because it is an integral part of any normative analysis, but also because the inquiry can be extremely difficult. Institutional details must be scrutinized,such as the way the new rule interacts with business realities (so-called frictions) and with existing law, including provisions that are quite technical and obscure. Unfortunately, over worked government officials, who in some cases have only limited transactional experience, are less likely to know these details than practicing lawyers. Yet the latter have strong incentives not to disclose new planning opportunities offered by changes in the law – indeed, I know from personal experience that clients can be quite displeased with those who do. In contrast, the tax bar usually feels free to alert the government to other efficiency costs, such as steep compliance costs that a proposed rule would impose. In evaluating a reform’s effect on the planning option, then, the government is on its own to a significant degree. An important role for academics, who are free of client conflicts, is to provide guidance on this issue.
Part I offers general principles, not unique to derivatives, about the jujitsu employed by well-advised taxpayers to turn an incremental reform to their advantage, and about potential responses the government might use. Part II returns the focus to the taxation of derivatives, and reviews the conventional case for mark-to-market-type reforms in that context. Part III considers the effects on he planning option when mark-to-market-type reforms are applied to some taxpayers, but not others: the creation of new accommodation parties, as illustrated with Section 475 and securities dealers. Part IV considers mark-to-market-type reforms that apply to a designated subset of transactions, instead of taxpayers. Using the contingent debt rules as a case study, this Part shows how a reform can offer offensive and defensive planning opportunities. For each case study, suggestions are offered about how to blunt these unintended consequences, although empirical questions must be resolved in determining which would be advisable in a each case. Part V offers the Article’s conclusions.
I. THE TAXPAYER’S PLANNING OPTION
Section A describes an inherent advantage of the taxpayer over the government – the ability to choose a transaction’s structure. Section B observes that this advantage functions as a tax reduction,implemented in a manner that is politically unaccountable, inequitable, and often inefficient. Section C describes three ways taxpayers can use this planning option to turn otherwise pro-government reforms to their advantage. Section D suggests government responses, short of comprehensive reform, that can keep a proposed reform from aggravating the planning option.
A. Taxpayer’s Structuring Advantage
In the recent public debate about tax shelters, commentators have noted the taxpayers’ process and resource-related advantages over the government. As Professor Bankman has observed, well-advised taxpayers know the probability of detection is low given declines in IRS auditing. Even if the government focuses on the transaction, a favorable settlement may be offered because of them is match in expertise and numbers between lawyers of the taxpayer and government. To address these resource-related disparities, the government has proposed to hire more auditors, has opened an office dedicated to tax shelters, has required more detailed disclosure, and has proposed stiffer penalties to even the odds in this “audit lottery.”
Even if the resource disparity can be eliminated, the taxpayer still has a valuable advantage: the ability to choose a transaction’s structure and thus, to a considerable extent, to elect the tax treatment.The government suffers from a first- mover disadvantage. It lays out precisely delineated rules, and then taxpayers are allowed to choose from this menu the transactional form most likely to reduce their tax bill.
While this right to choose is valuable, two constraints limit the planning option. First, how much must the transaction change for the tax treatment to change? The tax-advantaged course may require the issuer to forgo a business advantage. A similar tradeoff arises if the tax- advantaged course is more complicated, thereby requiring higher payments to financial and tax advisors. The tax savings is more expensive to taxpayers if they must change their behavior (thereby incurring “standard deadweight loss”) or pay experts (“avoidance costs”). If this utility loss exceeds the tax savings, the Issuer will use the more tax-expensive structure.
Even if forms are perfect substitutes, a further constraint is the need to find a counterparty willing to use the taxpayer’s preferred form. Yet a structure that reduces tax of one side (e.g., in deferring income) often increases tax of the other (e.g., in deferring a deduction). If one party’s taxes rise by as much as other party’s taxes fall, as is the case if the same rate and timing rules govern them both,there is no net tax savings for the parties to divide. Such “symmetry,” as Professor Shuldiner calls it, can severely constrain the planning option. This restraint disappears, though, if the counterparty is a tax indifferent party (such as a charity or a foreign person), since it will not object to an otherwise tax-expensive form.
To sum up, the planning option is valuable to the taxpayer if three conditions are satisfied: First, economically comparable transactions must be taxed differently, so the tax system is not consistent or continuous, to use the terminology of Professors Shuldiner and Strnad. Second, in structuring the transaction one way as opposed to another, the taxpayer must generate more tax savings than additional private costs. Third, an accommodation party must be available to engage in this preferred structure.
Since the taxpayer’s ability to choose her structure might be compared to a financial option, these intuitions may confirmed with option pricing models. Avoidance costs and standard deadweight loss should be viewed as the “exercise price” of the planning option, in that these costs represent the expenditure (whether in cash or in forgone utility) required to use the planning option, and thus to get favorable tax treatment. In options pricing models, the lower the exercise price, the more valuable the option. In addition, a financial option’s value rises with the riskiness of the underlying property. The greater the risk (or so-called “volatility”), the broader the range of possible values when the option matures. Since the option holder is protected from bad outcomes (i.e., she does not have to exercise the option), the holder prefers a broad range of possible outcomes because the chances of an especially favorable one are increased. Similarly, the planning option is more valuable as the number of possible tax outcomes increases, because the well-advised taxpayer is more likely to find an especially favorable one (while avoiding unfavorable ones).
B. Normative Assessment of Taxpayer’s Planning Option
Section A showed that the taxpayer’s ability to choose transactional structure functions as a tax reduction. While reducing tax may be good policy in some cases, tax planning is generally a poor way to achieve this goal for three reasons. First, the planning option is unappealing on political process grounds. As a tax reduction, planning is hard for the average voter to monitor and assess. Outside of the obscure domain of tax experts, few recognize that the stated tax rate is not necessarily the effective tax rate, and that this disparity arises in part from the interplay of inconsistent rules and tax-indifferent parties. Indeed, I suspect that not all members of Congress understand this issue. Those who do sometimes offer (or tolerate) planning opportunities as a boon to supporters, relatively free of scrutiny from voters at large.
Relatedly, the planning option poses a vertical equity issue. Wealthy taxpayers may benefit disproportionately, since they have better access to tax advice and can amortize the costs of identifying such strategies over potentially larger tax savings. Horizontal equity also is implicated when taxpayers with comparable incomes pay different tax bills depending upon the aggressiveness of their tax planning.
Finally, the planning option can be a significant source of inefficiency in the tax system. As the phrase is used by economists, “efficient taxes” raise revenue while creating a minimum of distortions and social waste. The planning option causes waste when taxpayers invest resources in discovering tax-advantaged strategies (avoidance costs) and when they change allocative choices to attain the desired tax result (standard deadweight loss). As planning becomes widespread, it can undermine the morale of conservative taxpayers (a form of deadweight loss) and increase the government’s administrative costs (i.e., if norms change so that taxpayers are less likely to comply voluntarily). Even so, some planning may survive as a necessary evil46 – for instance, if the cure is even more wasteful than the disease (e.g., in subjecting more taxpayers to a tax that is itself inefficient, or in prompting burdensome administrative or compliance costs or other planning opportunities). When the underlying tax is sensible and planning can be thwarted cheaply, we make the system more efficient by doing so. To assess the costs of such responses, Professors Slemrod and Yitzhaki offer a standard, the marginal efficiency cost of funds (MECF), which measures the additional social waste from raising an extra dollar of revenue through a change in tax law or administration. To use this test, we must be able to foresee, among other things, the nature and magnitude of taxpayer’s behavioral responses – that is, the new rule’s effect on the planning option. The next section notes the difficulty of anticipating planning opportunities and suggests two rules of thumb for doing so.
C. Two Ways an Incremental Reform can Aggravate the Planning Option
Given these normative concerns about the planning option, curtailing it should be a priority. If targeting planning were the sole priority, we would eliminate accommodation parties from the system while making the rules consistent and accurate, so that only one treatment was available for a given transaction by a given taxpayer. Yet the system has other priorities too, such as keeping the rules administrable, and political realities limit our choices. While more modest reforms sometimes diminish the planning option, they can also have the opposite effect in at least two ways.
1. Taxpayer-Specific Reforms: Creation of New Accommodation Parties
First, incremental reforms can create new accommodation parties, a consequence most likely to arise from reforms that affect only a subset of taxpayers (so-called “taxpayer-specific reforms”). A tax reform that applies only to some taxpayers can aggravate the planning option by causing two sides of a given transaction to be taxed under different rules — whether the difference is in rates or timing and character rules (e.g., deferred gain to one party but no correspondingly deferred loss to the other). Even if it is otherwise desirable to apply special timing or character rules to a particular group of taxpayers (e.g., to measure their income more accurately), such benefits come at a cost. The severity of this cost depends on the institutional characteristics of the affected group. Are these taxpayers likely to solicit other taxpayers for tax motivated transactions? Will they aggressively market themselves as tax accommodation parties? Is this role consistent with their skill set and perceived public role?Do they have existing customer relationships, or the institutional capacity to acquire them?
2. Transactional Reforms: Opting In and Out of New Tax Treatments
A reform that applies to certain transactions but not others — as opposed to certain groups of taxpayers but not others — can also aggravate the planning option by broadening the menu of potentially-applicable tax treatments, even if this “transactional reform” is otherwise appealing (e.g., in measuring income more accurately).
In assessing this reform’s impact on the planning option, a key question is whether this new rule replaces, or merely supplements, the old rule. If the same tax treatment now applies to any structure the taxpayer chooses, a powerful blow is dealt to the planning option. In contrast, the planning option can survive — and can even grow stronger — if the reform adds a new treatment (or extends an existing one) while leaving the old treatment in tact in some instances. If taxpayers can opt in or out of the incremental reform, they will avoid it when the rule is unfavorable (the defensive planning option) while deliberately qualifying when the rule is favorable (the offensive planning option). Two questions bear on the strength of these effects:
First, how easy is it to avoid the incremental reform? The crucial issue is whether taxpayers must give up something that matters to them.
Second, does the reform offer tax treatment that was not available before or, at least, was more difficult to attain? Thus, care must be taken not only with an incremental reform’s scope, but also with the tax consequences it offers. A reform can offer more favorable treatment than prior law, and thus an offensive planning option, in at least three ways. First, it can reduce the transaction or utility costs of a deduction that was already available under prior law. In addition, the reform can offer a new tax treatment – such as a deduction in excess of economic loss – that was not already available. Finally, there form can offer new opportunities for so-called “tax arbitrage.” In such strategies, taxpayers take two or more positions that are economically offsetting (and thus present little or no economic risk), but generate a net tax benefit.
Ultimately, the effect of any reform on the planning option turns on empirical questions. Some taxpayers will be stopped from planning: for them, the costs of avoiding the reform (e.g., standard deadweight loss and avoidance costs) are higher than the potential tax savings. This is a good result because extra revenue is collected without distorting taxpayer behavior. On the other hand, some taxpayers will change their transactions to avoid the reform, and others will change their transactions to qualify. In these cases, revenue does not increase (and, because of the offensive planning option,may in fact decline), while the tax rules do distort taxpayer behavior.
The relative magnitude of these effects determines the reform’s overall impact on planning-related waste.
3. Accuracy-Enhancing Reforms that Induce Wasteful Planning
Because of the risk that incremental reforms can create new planning opportunities, formulating effective reforms is quite difficult. For instance, although it is usually desirable to enhance the system’s accuracy by moving closer to the Haig-Simons definition of income, this goal must be pursued with care. Often accuracy is valuable as a means of promoting efficiency and equity. For instance,inaccurate rules that overtax one transaction while undertaxing others tend to misallocate resources (as taxpayers gravitate to the undertaxed transaction).
Unfortunately, though, accuracy-enhancing reforms are not always beneficial. When they are too narrow in scope — for instance, in applying only to certain transactions but not to close substitutes, or in applying only to a group of taxpayers that is likely to serve as accommodation parties – the reform can create new planning opportunities that undermine, and in some cases outweigh, the reform’s promised benefits. For example, while enhanced accuracy would otherwise promote greater respect for the system, pervasive planning has the opposite effect. Likewise, although enhanced accuracy could make resource allocations more efficient, new planning opportunities can have the opposite impact by increasing avoidance costs and, potentially, inducing new tax-motivated resource allocations.The comparative magnitude of these (and other) offsetting effects vary from case to case. Sometimes the reform will still be beneficial on balance, and other times it will not be. Either way, we should prefer reforms that offer benefits of improved accuracy without planning-related costs.
D. Modification of Reforms to Prevent Aggravation of Planning Option
Sometimes the benefits of a new rule are so substantial that they justify unavoidable costs, such as the creation of new planning opportunities. Yet often a reform can be modified to impede these newly-created planning opportunities, while retaining some or all of the new rule’s advantages. This Section offers three types of responses: abandoning the reform; modifying the rule’s scope; or modifying the treatment it offers. These broad alternatives are offered as a menu to be considered in individual cases, rather than as a recommendation of what is always preferable. In curtailing the planning option, the response usually will raise (or at least preserve) revenue and may also have appealing political-process, distributional, and efficiency effects. Yet it is hard to generalize about efficiency, as noted above, and so empirical judgments are needed on at least three questions. First, is the tax being avoided a sensible tax (although self-help is usually less efficient than repeal)? Second, what are the compliance and administrative costs of each alternative? Finally, what is the net effect on planning costs (including avoidance costs and standard dead weight loss) –considering planning opportunities that are prevented, as well as those that remain, including newly-minted ones? For instance, have we really stopped planning or just made it more expensive without deterring anyone (i.e., so that the level of social waste rises)? These empirical inquiries influence the marginal efficiency cost of each alternative. While there is rarely a perfect solution in a “second best”world, planning often can be impeded at low social cost.
1. Forgo the Reform
Ironically, it is sometimes better to stick with a flawed rule that is consistently applied, instead of fixing the rule part-way: Addition of another rule – even one that would enhance equity and efficiency if applied comprehensively – can aggravate the planning option. In some cases, the reform should not be enacted (e.g., if it creates even more planning waste than it eliminates, without offering administrability advantages). Yet before abandoning the incremental reform, the government should consider whether it can be successfully reconfigured.
2. Modify the Scope
Broadening the rule’s application can weaken the planning option. For instance, taxpayer- specific rules aggravate the planning option by applying different tax rules to parties that are likely to cooperate in tax planning – that is, those who are likely to serve as accommodation parties and those who are likely to hire them (e.g., corporate taxpayers and wealthy individuals). A reform is less likely to aggravate the planning option if it does not apply to any of these players or, alternatively, if it covers all of them. Similarly, transactional reforms should extend to all comparable transactions, so that taxpayers cannot avoid the rule (or, for that matter, qualify if they otherwise would not) without incurring prohibitive costs and utility losses.
If a reform cannot be extended this far (e.g., for reasons of administrability and politics), so that some inconsistency is inevitable, Professor Weisbach properly suggests that the boundary be set where taxpayers have a relatively inelastic preference for one side over the other. In other words, the tax system should rely on business frictions to impede tax planning, with the hope that taxpayers will choose business advantages over tax reduction when the two compete.
Yet it is sometimes hard to tell how much taxpayers care about a particular issue – and, even if they do, they may not mind making a small change at the margin to alter the tax treatment. For instance, taxpayers are likely to care about their economic return (e.g., whether it is fixed or contingent) – as opposed, perhaps, to the legal form they are using – but, even then, some taxpayers will be willing to accept a subtle change at the margin.
In addition to seeking inelastic break points, as Professor Weisbach advocates, the government should consider two further steps to curtail the planning option. The first is to make strategic use of ambiguity. If it is unclear how much the transaction must change, taxpayers will have to overshoot and, at the margin, some will simply accept the tax-expensive rule. Admittedly, this in terrorem effect is not free. Others may continue to avoid the rule, while finding it more expensive to do so. Compliance and administrative costs for all players may rise as experts are consulted about the reform’s reach and administrators are invited to make more nuanced determinations. There is also the risk of too much precautionary behavior (so that taxpayers may shy away even from “good” transactions). Even so, in many cases the net effect on social waste will be favorable (e.g., if taxpayers generally stop gaming the rule and compliance costs do not rise significantly). For this reason, it is often helpful for incremental reforms to include so-called “anti-abuse” rules, which offer the government discretion, in the guise of construing the rule’s purpose, to apply (or not apply) a rule to transactions otherwise outside (or within) its scope.
Relatedly, if the government repeatedly modifies rules retroactively in response to aggressive uses of the planning option, taxpayers will be chastened as they come to expect such reactions.
Whereas ambiguity and inelastic break points relate to the reform’s scope (i.e., which transactions are covered), the second step the government should consider is to adjust the reform’s consequences.
3. Modify the Treatment
To curtail the offensive planning option, the treatment offered by incremental reforms can be adjusted in two ways. First, a reform is less appealing to taxpayers if they cannot predict, in advance, that the new rule will reduce their tax liability. One approach, here called “market balance,” is to ensure that any tax benefit as the market moves one way (e.g., acceleration of losses) is inevitably matched by a tax detriment as the market moves the other way (e.g., acceleration of gains). For instance, if taxpayers bet “heads” in a coin toss, they should be indifferent, prior to the toss, between having tax consequences this year or next, as long the same rule governs gains and losses. While acceleration is better for losses, deferral is better for gains, and taxpayers do not yet know which they have. On the other hand, the offensive planning option survives if the taxpayer can choose her treatment after the market has moved. In addition, market balance is less effective when pretax consequences are predictable: Thus, this approach would be less effective in taxing debt (even if interest payments vary with market conditions) because, on average, a borrower is expected to have loss (interest expense)and a lender is expected to have gain (interest income).
When market balance is unlikely to be effective, an alternative is to abandon symmetry: thus a reform that imposes adverse treatment on one party, such as accelerated income, would not also offer unusually favorable treatment to the other, such as accelerated losses. With such asymmetry, the offensive planning option is significantly weakened (since the pro-taxpayer result is not permitted), although the defensive planning option may survive (as taxpayers strive to avoid the reform’s pro-government result). Given these effects, the asymmetrical reform should not reduce revenue, and may well increase it. The effect on efficiency turns on empirical questions including changes in compliance and administrative costs, if any, as well as the net effect on planning (i.e.,balancing reductions in waste as some stop planning against increases as some abstain from “good” transactions or continue to plan at higher cost).
A potential drawback of pro-government asymmetry is its “heads I win, tails you lose” quality, which appears at first blush to favor the government unfairly and to discourage “good” transactions along with tax-motivated ones. Even if this appearance is deceiving, as I believe it is in many cases, the pro-government tilt can be a political liability, since lobbyists will accuse the government of overreaching.The proper response, though, is the core argument of this Part: well-advised taxpayers have an inherent advantage in choosing their structure, which is especially potent when tax rules are inconsistent and accommodation parties are available. Adding rules that seem balanced, when evaluated in isolation, may merely aggravate this planning option. Since the government already is at a disadvantage, sometimes it must enact rules that seem unbalanced but, in reality, merely address the existing imbalance — thereby restoring conditions in which taxpayers are motivated more by business than tax considerations. Although not justified in precisely these terms, unbalanced rules have had some success in curtailing high-profile abuses. The passive activity loss rules and straddle rules, for instance, each used punitive loss deferral to foreclose aggressive tax planning. A possible political benefit of such reforms is in creating support, over the long term, for comprehensive reform that would allow for their repeal.
II. THE CASE STUDIES: TIMING RULES FOR DERIVATIVES
The issues discussed in Part I – the taxpayer’s planning option and possible government responses – apply to a broad range of problems throughout the tax law. Indeed, the mission of an expert tax practitioner is to identify and evaluate different ways to structure essentially the same transaction. Incremental tax reforms are forever being enacted throughout the tax law and, for reasons described in Part I, these reforms sometimes aggravate the planning option.
This Article applies the framework developed in Part I to the timing rules for derivatives. The planning option is particularly relevant in this context for two reasons. First, the planning option is already very powerful in this area. The same business deal can be implemented with numerous transactional forms that yield very different tax treatment. For this reason, more and more wealthy individuals and corporations are using derivatives for tax planning, and increasing numbers of tax advisors are mastering these rules. Indeed, many of the corporate tax shelters that recently have attracted attention in the media and government involve derivatives. In addition, the planning option is of critical importance for derivatives because the most commonly suggested reform in this area – enhanced accuracy through mark-to-market accounting – eliminates the planning option if applied comprehensively but, as this Article shows, this reform can aggravate the planning option when applied partially.
If implemented across the board, this reform would eliminate the planning option because the same treatment would be accorded to all structures. This accuracy-enhancing reform also has another advantage, for which it is more commonly advocated: It would remedy the inaccuracy and distortive effects of the realization rule itself (i.e., even if realization is consistently applied). By deferring and thus reducing the tax burden on gains, realization encourages taxpayers to hold appreciated property even if they otherwise would prefer to sell (e.g., to invest in something else)
— a source of distortion known as “lock in.” Moreover, in allowing taxpayers to control the timing of their tax, the realization rule gives taxpayers a “timing option”:  they can deduct losses immediately (thereby preserving the real value of the loss) while deferring tax on gains (thereby reducing the real value of the tax). These generous results create a tax preference for assets subject to the realization rule (such as growth stocks instead of bank accounts), thereby distorting allocations of capital. Since wealthy people are more likely to have investments – and thus are more likely to benefit from deferral and the timing option – realization can undermine vertical equity. Mark-to-market taxation eliminates deferral by imposing tax on gains before the property is sold, and it eliminates the timing option by denying the taxpayer control over timing.
Given these favorable effects, mark-to-market-type reforms seem very appealing. While it is not politically and administratively feasible to apply these reforms comprehensively, it is tempting to conclude that mark-to-market reforms should be extended gradually – in effect, wherever we can. The government apparently is following this model. Evaluated in isolation, each reform seems at first blush to improve accuracy, reduce deferral, and curtail the timing option. Indeed, each is a version of alternatives suggested by Professor Warren in an influential article.[87 ] Unfortunately, these reform shave created significant new planning opportunities in significant ways. The next two Parts explain how the new planning opportunities arose, and how to this planning might be impeded in the future.
III. TAXPAYER-BASED CLASSIFICATIONS: SECTION 475
Haig-Simons incremental reforms apply a mark-to-market-type rule to some situations, but not others. The dividing line could be based on an essential characteristic of either the taxpayer or of the transaction. Whereas either type of classification can aggravate the planning option, the mechanic by which this occurs is different. This Part considers a classification based on the taxpayer: Section 475,which applied mark-to-market accounting to securities dealers in 1993 and, in 1997, extended this treatment on an elective basis to securities traders and commodities dealers.
A. Motivations for Reform: The Dealer’s Timing Option and Whipsaw Concerns
The tax law defines “securities dealers” as taxpayers who regularly offer to purchase or sell securities to clients and profit from the “spread,” as do the dealer subsidiaries of Goldman Sachs and Merrill Lynch. Mark-to-market accounting was applied to dealers because of concerns about two rules under prior law, one considered too generous and the other too harsh. First, dealer inventory (i.e., all positions other than short sales and derivatives) was governed by a rule more generous than realization, known as “lower of cost or market” accounting (“LCM”). If a dealer bought stock for sale to clients (e.g., for $50) and the asset appreciated (e.g., to $100), no gain was reported until the stock was sold, as under realization. Yet unlike under realization, if the stock’s value declined (e.g., to $20),the loss could be claimed even if the property was not sold. Thus, LCM offered an inaccurate measure of a dealer’s income, since gains were under taxed. Indeed, the rule offered dealers an extra-potent timing option: While gains could be deferred, losses could be claimed without costs associated with a sale.
While the government regarded LCM as too generous, dealers protested the treatment of derivatives and short positions, which were governed by realization instead of LCM. Although realization theoretically offered dealers a timing option, two business frictions – the need to hedge and to accommodate client preferences – significantly reduced the timing option’s value. First, dealers always hedge the derivatives they sell to clients. Because the various client positions at any given time are not perfectly offsetting, dealers must take another position (usually in the public markets) to hedge this net exposure. As the mix of client contracts changes, this position will be adjusted. Thus, dealers inevitably enter into transactions that are economically offsetting but have different maturity dates. When these positions were governed by realization, the tax accounting for each occurred when the position was settled – and since otherwise offsetting positions could settle in different years, the dealer’s tax bill would not accurately reflect pre-tax profit. This mismeasurement was sometimes favorable (e.g., if the dealer realized losses and had offsetting unrealized gain) and sometimes unfavorable (if unrealized loss matched realized gain). Dealers might have been expected to manage this process to their advantage – for instance, by closing out loss positions early while retaining appreciated ones. However, since clients (the dealer’s usual counterparty) were governed by the same realization rule (at least if the clients were U.S. taxpayers), the dealer could not pursue such timing or structural choices without affecting the client’s tax position. Dictating timing and structure is at odds with the dealers’ customary mission of implementing whatever transaction the client requests. Since dealers had only limited influence over form and timing, then, the planning option was considerably less valuable to them; instead, realization offered tax uncertainty and the risk of whipsaws.
In response to this complaint from dealers, as well as its own concerns about LCM’s generosity, the government required dealers to mark all positions to market. As a result, the tax reduction implicit in LCM has been undone. Nor are dealers likely to pursue tax avoidance strategies for their own accounts, since a single accurate rule applies to all their transactions, regardless of form or the timing of settlement. This beneficial step was attained without adding significantly to compliance costs: valuation at year-end — a potentially daunting prospect for other taxpayers — is less of a burden because dealers already value their holdings for business reasons.
Likewise, securities dealers have ready access to liquidity, and thus have cash to pay tax on appreciated property even without a sale.
Nevertheless, efficiency gains from revoking a dealer’s planning option may be less impressive than they first appear. At least to an extent, business frictions already were constraining a dealer’s tax planning for positions subject to realization, as discussed above. Of course, dealers were receiving a tax benefit under LCM, but one that imposed few avoidance costs (since dealers did not have to incur transactions costs to claim a loss, although they did have to hold property to defer gains). Resources might have been misallocated if dealers were under or over-taxed relative to other businesses, but we cannot make this determination without more information about other businesses. For instance, as generous as LCM was, other dealers were also using it (e.g., car dealers) and they were not stopped from doing so in 1993. While the risk of timing mismatches under realization arguably imposed unique burdens on dealers (potentially rendering them over-taxed), this burden might have had only a limited effect on resource allocation. If timing mismatches were as likely, ex ante, to be favorable as unfavorable, then they would not reduce the dealer’s expected profitability, but would only add a source of risk, which would have impact only if dealers were risk averse. It is possible that timing mismatches were unfavorable on average, since dealers would be adversely affected when clients pursued their own tax-minimization strategies. If dealers raise their fees to pass these tax costs on to clients, demand could decline, leading to an inefficiently low level of dealer activity. On the other hand, if the tax cost to dealers precisely matches the tax benefit to clients, the level of dealer activity would not necessarily change. Even so, the point should not be overstated. Section 475 brought the system real benefits, although their magnitude can be debated. The problem is that this reform also carried a significant cost.
B. Dealers as Tax Accommodation Party
While Section 475 keeps dealers from engaging in tax planning for themselves, the reform empowers them to facilitate tax planning of their clients. In applying a new timing rule to dealers but not clients,Section 475 ensures that symmetry — and the constraint this condition imposes on tax planning – never applies to dealer transactions. The result, not understood by the Treasury or Congress when Section 475 was enacted, is that securities dealers can serve more effectively as tax accommodation parties.
For instance, assume that on January 1, 2000 a client commits to buy a share of Internet.Com from Merrill Lynch in two years for $110, because she expects the price to be higher. Under realization, she will have no tax consequences until this “forward contract” matures, unless she and Merrill terminate the derivative before then.[106 ] Yet on December 31, 2000, with a year to go on the contract,Internet.Com has plummeted to $1 a share. The client’s tax-reducing strategy is to terminate this contract early (e.g., by paying her counterparty $109 in cash) in order to report this tax loss in 2000.Yet Merrill is likely to charge a fee for this early termination, and the amount will increase if this step increases Merrill’s own tax bill. Under realization, Merrill’s tax bill would increase, since the gain would otherwise be deferred until the forward matures in 2002. If Merrill is subject to mark-to-market, in contrast, early termination does not increase Merrill’s taxes because, even without an early termination, this gain must be reported. As a result, Merrill can accommodate the client’s tax preference without increasing its own tax bill. For this purpose, Merrill functions like a tax-exemptentity.
Since tax-indifferent parties are already available to serve as accommodation parties, including charitable organizations, foreigners, Indian tribes, and corporations with net operating losses, does the addition of securities dealers have any incremental effect? To an extent, any addition to the supply of accommodation parties can reduce the cost of “hiring” one, and thus can increase the number of taxpayers who will find it worthwhile to engage in tax planning. More fundamentally, because of their institutional characteristics, securities dealers can serve as accommodation parties in cases where others cannot.
Compare securities dealers to tax exempt entities such as pension funds and charities. Even though Columbia University, for example, theoretically can serve as a tax accommodation party in derivatives transactions, it lacks the requisite expertise. Columbia’s employees do not know how to price these transactions, to hedge them, or to comply with relevant legal and regulatory requirements. Nor do most tax-exempt organizations have existing client relationships or a marketing arm to create new ones. In addition, tax-exempts are likely to incur reputational costs, as well as tax costs, in straying from their main mission. How would it look if one-third of Columbia University’s workforce was working in the securities industry? If this were the case, Columbia would be taxed on this “unrelated business taxable income” – and thus would cease to be indifferent to tax consequences.
In contrast, the very mission of a securities dealer such as Goldman Sachs is to accommodate its clients’ investment preferences. Goldman already has the requisite expertise and marketing capabilities. Indeed, when Section 475 was enacted, dealers already were the counterparty in a huge volume of derivatives transactions, and this volume has grown astronomically; all transactions in the burgeoning“over the counter” derivatives market are with dealers.
Finally, dealers not only are able to implement transactions proposed by clients, but they also have the expertise to develop new tax-reducing strategies — an expertise they have been using to powerful effect.
Two other accommodation parties have similar institutional qualities – foreign financial institutions and domestic life insurance companies – but they are hindered (although not necessarily stopped) by constraints inapplicable to U.S. securities dealers. Foreign financial institutions may be unable to implement a U.S. client’s preferred structure because of adverse consequences under their home tax regime. U.S. tax rules can also have this effect. The foreign institution must avoid the 30% withholding tax imposed by the U.S. on dividends and certain other payments to foreigners. In addition,foreign institutions will not want any profit to be “effectively connected” to the United States because this profit would be subject to U.S. tax, causing the institution to lose its usual basis for tax indifference. This wire is tripped, for instance, if the accommodation party role is viewed as dealing in securities through a U.S. office. This is not to say that foreign institutions can never serve as accommodation parties. They frequently do, but securities dealers can perform this function in instances when foreigners cannot.
The same holds for U.S. life insurance companies. As with securities dealers, a type of mark-to-market accounting allows them to accommodate timing issues of clients at no tax cost.
Yet this timing rule is available only for transactions qualifying as “life insurance” or “annuities.” Although malleable, these concepts are not infinitely elastic. The transaction must include mortality risk,or at least deny the client access to the investment for some period of time. For some potential clients, these constraints outweigh the expected tax savings. Thus, while there is a thriving market in tax-advantaged insurance products, it is not broad enough to include all the transactions that securities dealers can do.
Just as it was significant when Congress inadvertently enabled securities dealers to be accommodation parties, it was also significant when Congress added securities traders to the list in 1997. Like dealers, traders have expertise and relationships with potential clients. In addition, they may face fewer regulatory constraints than dealers. For instance, dealers may be reluctant to “park” large offsetting positions on their books because regulators may require them to hold more capital in liquid (and less profitable) investments. Such regulatory capital constraints,though, do not necessarily apply to traders. Thus, although I am not personally aware of this practice, we should expect the creation of new “trading” partnerships to serve as tax indifferent parties,relatively free of tax or regulatory constraints.
In any event, dealers already are marketing their status as accommodation parties, as evidenced by a transaction involving hedge funds that attracted media attention in recent years.
The impetus for the transaction is that hedge fund investors usually treat much of their return as short-term capital gain, taxable immediately. Using their tax indifferent status, dealers can offer clients hedge fund returns with better tax treatment. Instead of the client, the dealer invests in the hedge fund and simultaneously transfers the economic return to the client through a cash-settled derivative based on the hedge fund’s value. Under law then in effect, the investors’ return on interest-type payment). If the hedge fund is worth $350 on January 1, 2003, the investor receives $250 —the same profit she would have made if she had held the fund directly. Meanwhile, because the dealer has two positions — it owns the fund and also, through the swap, has bet against the fund’s value —the dealer is perfectly hedged. The dealer makes $250 by owning the fund and loses $250 on the swap the derivative is deferred (i.e., until it matures or is settled) and is treated as long-term capital gain. This transaction would make no sense if the dealer were taxed under the same rules as the investor, because the adverse consequences of holding the hedge fund directly — accelerated income at short-term capital gains rates — would merely be passed on to the dealer. However, Section 475 largely shields the dealer from these consequences. Having accelerated income is not a problem,because the dealer will have an offsetting loss from her position in the derivative, which is marked to market. Nor is the difference between short- and long-term capital gain significant to dealers since, as corporate taxpayers, they do not receive a reduced rate for the latter. Recently enacted legislation is supposed to deter this transaction by undoing the investor’s deferral and long-term capital gain, although it is not clear how effective it will be. If not for Section 475, however, there would be no need for the new rule, or the compliance costs and complexity it brings.
C. Responses to the Accommodation Party Concern
Whereas Section 475 has the advantage of taxing securities dealers more accurately and discouraging tax planning for the dealers’ own accounts, the rule has the disadvantage of creating an aggressive new accommodation party. What should the government have done — and, indeed, what should it do now? One possibility is to accept this increased risk of tax planning by the counterparty as an inevitable byproduct of an otherwise desirable reform for dealers, and to focus on making the system more consistent so that counterparties have fewer planning opportunities.
In other words, it is not a problem to make securities dealers accommodation parties if there is nothing for them to do. However, while it certainly makes sense to improve the rules for counterparties, the difficulty is that the tax law is vast and riddled with inconsistencies. It is not realistic to eliminate them all immediately and, in the seven years since section 475 was enacted, this obviously has not occurred. In addition to remedying the rules for counterparties, then, would it have been beneficial to modify Section 475? This Section considers the three types of responses discussed above: forgoing the reform; modifying its scope; and modifying the treatment. None is perfect, but each has advantages over the course the government took.
1. Forgo the Reform
LCM could have been replaced with the same realization rule that governed the dealers’ short and derivatives positions, as well as the counterparties. The government would still have revoked the tax reduction implicit in LCM, while preserving symmetry and thus impeding dealers from serving as accommodation parties. Although there would still be other accommodation parties in the system,they lack institutional characteristics that make dealers especially effective, as discussed above. Some of the wasteful tax planning that is now occurring, then, would not occur. More revenue would be collected, and avoidance costs and deadweight loss from this planning would be averted. While data should be collected on the empirical magnitude of these benefits, I suspect it is considerable.Indeed, the joke about Section 475 in the tax bar is that it was estimated to raise revenue when enacted, but it would also be estimated to raise revenue if repealed.
Balanced against these benefits, though, are two potential costs that are somewhat inconsistent with each other (or, at least, are potentially offsetting). If governed by realization instead of mark-to-market, dealers will have greater opportunities to engage in tax planning for themselves — although, as noted above, business frictions would help contain this impulse. On the other hand, if frictions prevent dealers from controlling form or timing, these taxpayers can become subject to whipsaws that induce precautionary planning and, possibly, an inefficiently low level of dealer activity. Put another way, it would be unfortunate if inconsistencies in the tax system impeded the growth and liquidity of the capital markets, although it is not clear that taxing dealers under realization would actually have this effect.
We are forced to pick our poison, then, and the choice turns on empirical questions on which data should be collected. My instinct is that tax whipsaws would complicate the task of an investment bank’s tax advisor (but would not chill the market for dealer services. It is more likely, in my view, that dealers would find ways around the frictions described above, and thus would engage in tax planning for their own account. Yet I suspect this planning, which is confined to a finite set of taxpayers, would generate less waste than arises when dealers market tax-motivated transactions to their clients; after all,there are more clients than dealers. In any event, the political reality is that Section 475 is unlikely to be repealed, given that dealers favor it and the case against this provision is subtle and empirically ambiguous. The question is whether we can revise this reform to preserve its advantages while mitigating the accommodation party risk.
2. Modify the Scope
Since the problem is erosion of symmetry due to the unique treatment of dealers, one solution is to restore symmetry by extending mark-to-market to the dealer’s counterparty. In other words,anyone entering into a derivative with a dealer would be required to mark the derivative to market.
Of course, to comply with this rule, the taxpayer must know the tax status of its counterparty (i.e., that it is a dealer or electing trader). A legal obligation could be imposed on the covered group to disclose their special status and its tax consequences to the counterparty (e.g., in the “confirm” or in other documentation of the transaction). Although the counterparty may also not have the requisite expertise to value the derivative each year, the dealer (or trader) already is conducting this valuation for its own return and could be required to share the information.
Whereas these compliance cost burdens are solvable, another problem is that, from the counterparty’s perspective, this approach provides a special rule for certain transactions (i.e., those done with dealers) that would not apply to other potentially similar deals (i.e., those not done with dealers). In other words, this approach presents some of the planning option pitfalls of transactional classifications. These concerns, and potential responses to them, are discussed below in Part IV.
3. Adjust Treatment
Instead of broadening Section 475’s scope, so that the rule applies to dealers’ clients as well, the government can diminish the dealer’s ability to serve as an accommodation party by modifying the treatment offered by the reform. Through an anti-abuse rule, mark-to-market treatment could be revoked, at the Commissioner’s discretion, in transactions in which dealers use their special tax status to serve as accommodation parties.
Unfortunately, it can be difficult to distinguish tax motivated transactions, in which dealers are “renting” their services as tax indifferent party, from other business transactions. Defining the “abuse” is hard because dealers will simply be following their own method of accounting, which accurately reflects their own income. The abuse largely lies in the intent of dealers and their clients, and so difficult evidentiary and conceptual problems arise. For instance, assume that a dealer stands ready to settle derivatives before they mature. Is the dealer trying to help clients trigger early tax deductions (bad intent) or to reconfigure portfolios in light of changed market circumstances (good intent)? A broad anti-abuse rule could impose unwarranted uncertainty and transactions costs on dealers, as well as intolerable administrative burdens on the government.
On the other hand, a narrow anti-abuse rule should avoid many of these costs. Although a narrow rule would fail to catch all abusive cases, clear abuses would be prevented, such as the hedge fund transaction described above, and the advantages of mark-to-market accounting would continue in nonabusive cases. In my view, three factors distinguish a clear case. First, one of the clients’ principal purposes for including the dealer in the transaction is to attain a tax benefit. Second, Section 475 allows the dealer to play this role without a corresponding tax detriment to itself. Third, the dealer has reason to know of its clients’ tax-motivated purpose. The third condition is necessary because dealers do not typically inquire about their clients’ state of mind, and imposing such a duty would add prohibitively to compliance costs. Nevertheless, the condition is easily satisfied if the dealer itself (or an affiliate) emphasized these tax benefits when marketing the structure to the counterparty.
A concern about this rule is whether dealers will in fact stop developing and marketing tax planning strategies, or will merely respond with wasteful cosmetic fixes without actually changing their behavior. Indeed, there is a tradeoff between making the anti-abuse rule broad enough to prevent such avoidance and keeping the rule focused enough to minimize compliance and administrative costs. In my judgment, the above three-prong test strikes about the right balance. The approach is likely to stop a fair amount of planning, since it is hard for dealers to market a tax avoidance strategy without providing written materials (which would trigger the rule, thereby denying the dealer any legal opinion from counsel on its own tax treatment). Yet the run-of-the- mill dealer transaction (e.g., facilitating a client’s inventory hedging) is clearly excluded.
IV. TRANSACTIONAL CLASSIFICATIONS: THE CONTINGENT DEBT RULES
The last Part showed that taxpayer classifications can aggravate the planning option by creating new accommodation parties, although adjustments sometimes can temper these effects. This Part explores a different boundary for incremental reforms — the characteristics of the transaction, rather than the taxpayer — and ways in which this type of reform can strengthen the planning option. The case study is the contingent debt regulations of Treas. Reg. 1.1275-4, which were finalized in 1996. Because this regime offers powerful defensive and offensive planning options, it is important to consider whether the advantages of this reform can be attained at lower cost.
A. Motivations for Reform: More Accurate Measurement of Income and Reductions in Tax Planning
Assume an investor pays $11,000 for a so-called contingent bond (i.e., a bond in which interest payments are based on financial facts such as stock or commodity prices). The bond does not pay a coupon. At maturity after five years, the bond pays $11,000 plus the amount, if any, by which the Dow Jones Industrial Average (the “Dow”) exceeds $11,000. Under prior law, if there had been coupon payments, these would have been included by the holder and deducted by the issuer under their usual method of accounting. On contingent payments, however, holders had no income and issuer shad no deduction until the contingency was resolved.
The government presumably had two related concerns about this outcome. First, the rule mismeasures income. Since such bonds usually offer a positive return, deferral of this income and expense was likely to be too generous to holders and too parsimonious to issuers. Put another way, as a matter of finance theory, the return on any financial asset includes interest-type compensation for use of the investor’s money, along with a risk-based return (which, ex poste,
can be positive or negative). Yet our system generally fails to tax this interest prior to maturity.
In offering the more accurate rule, the government may have hoped to curtail tax planning. This planning arose because the deferral rule for contingent bonds diverged from the “accrual” treatment of fixed rate discount bonds (although deferral was the prevailing treatment of contingent investments such as stock, forward contracts or options). Accrual is more tax- expensive for the holder (since tax is due earlier) and more tax-favorable for the issuer (since deductions come sooner). In response to this inconsistency, tax-sensitive holders would avoid the accrual rule by favoring contingent bonds over fixed-rate debt. Correspondingly, tax-sensitive issuers would choose the accrual rule by favoring fixed-rate bonds.
In response, after a series of proposed regulations, the Treasury extended the accrual
regime to contingent debt. The regulations assume that the bond appreciates by a constant yield every year, thereby affording holders ordinary income and issuers an ordinary deduction before maturity. This approach, known as the “noncontingent bond method,” resembles assumed yield proposals by Professor Shuldiner, Professors Cunningham and Schenk, and Edward Kleinbard, except that these proposals would not have applied exclusively to debt. Called “the comparable yield,” the regulations’ assumed return generally is the interest rate the issuer would pay on fixed-rate debt with the same maturity, seniority, and other relevant characteristics. As the following Table shows, the regulations dramatically accelerate the holder’s interest inclusion and the issuer’s interest deduction. The Table assumes the comparable yield is 8%, and uses the Dow-linked example described above:
TABLE 2: COMPARISON OF “WAIT AND SEE” AND REGULATIONS
Unlike in mark-to-market taxation, market conditions are irrelevant under the noncontingent bond method until realization. Even if the bond appreciates more than the assumed yield, holders do not include extra income until they sell the bond or it matures. Upon such a realization event, an “adjustment” is made: the holder has more ordinary income if the bond has outperformed the assumed yield, or an ordinary loss if the bond has yielded less than the assumed yield. Likewise, the issuer makes no “adjustment” until the bond matures or is settled early. Although less accurate than mark-to-market accounting, the regulation’s assumption of some appreciation, based on the issuer’s borrowing cost, should be approximately correct on average – and certainly should be more accurate than the deferral rule – since a lower expected yield would fail to attract holders and a higher one would be too generous from the issuer’s perspective.
In addition to improving accuracy, the Treasury presumably hoped to curtail wasteful planning. Yet this hope was probably too optimistic, as explained in the following Sections. First, the regulations significantly increase compliance and administrative costs, and probably are reducing revenue. Second, the regulations do not meaningfully impede tax-sensitive holders from seeking deferral through contingent-bond-type structures. Third, although tax-sensitive issuers who prefer to issue contingent bonds are no longer deterred, some are induced to employ contingent debt in tax arbitrage and other tax-motivated transactions. Thus, by making the system more accurate – but only in a limited way, which taxpayers can choose or avoid with relative ease – this reform created powerful offensive and defensive planning options.
B. Compliance and Administrative Costs of Reform
Unlike the deferral rule, which was relatively simple for taxpayers and government auditors to understand, the new rules are quite technical. They require computations that are at the limit of what many taxpayers can do — or, indeed, beyond it in many cases. The comparable yield must be ascertained, and then each year an amount must be reported and basis and issue price each must be adjusted. For those who buy the bond in the secondary market, complex rules operate in lieu of the usual market-discount and bond-premium rules. Additional rules specify the treatment when the contingent amount is finally paid, as well as when contingencies have become fixed but not yet paid, when tax-exempts issue contingent bonds, and when these rules interact with other regimes. Although I do not teach many of the regulations’ nuances, students generally view this regime as the most difficult in a course in on the taxation of financial instruments. Since the contingent debt regulations are hard for Columbia law students, I suspect they are not straightforward for government auditors either. It would not surprise me if these rules are misapplied and under enforced on numerous occasions. Given the inaccessibility of these rules, the subset of experts who understand them can bill at a premium. Likewise, investment bankers report that compliance burdens alone are enough to deter a significant volume of taxable holders from buying contingent bonds.
C. Effect of Reform on Revenue
Since the regulations are symmetrical, they would have no effect on revenue if all holders and issuers were subject to the same tax rate: the higher effective tax burden on holders would be matched by a tax reduction on issuers. Yet revenue would increase if holders were subject to a higher average tax rate, while revenue would decline if issuers had the higher average tax rate. Empirical evidence suggests that a loss of revenue is more likely, presumably because tax exempt entities (such as charities, pension funds, and foreign investors) are more likely to lend in the U.S. capital markets than to borrow. This effect is all the more powerful here because contingent bonds are typically held by tax-exempt entities and issued by taxable issuers, as discussed below.
D. Effect of Reform on Tax Planning
A reform that increases compliance costs and is likely to reduce revenue starts with two strikes against it. The reform might still be justified in offering some other significant benefit, such as a significant reduction in planning-related waste. Yet the regulations are unlikely to reduce the level of planning — indeed, they probably increase it, although further empirical research on the question would be helpful.
1. Defensive Planning Option for Holders
Under the regulations, tax-sensitive holders no longer are induced to buy contingent debt they otherwise would not want. Yet it is not clear how powerful this tax-motivated attraction to contingent debt ever was – and how much wasteful planning it caused – even before the regulations. Tax deferral was (and remains) relatively easy to secure for contingent returns for instance, through prefer red stock (which can offer identical terms) or prepaid forward contracts (which can offer an similar return but without principal protection). Why strain to buy contingent debt when the same treatment is equally available through other instruments?
In any event, the regulations reverse whatever planning incentive was in effect for tax- sensitive holders. Instead of being encouraged to buy contingent debt, holders now avoid it. As a practitioner, I spent many hours tweaking transactions that would be marketed to tax sensitive holders to ensure that the regulations did not apply. The task is common and usually feasible (although it requires expensive advice and sometimes wasteful restructuring) because of the regulations’ narrow scope: they apply only to instruments that qualify as debt for tax purposes. In response, tax-sensitive holders gravitate toward preferred stock if they want principal protection or, in many cases to instruments that lack principal-protection (and do not qualify as debt for tax purposes).  Since tax-sensitive holders have other choices, the likeliest clientele for securities taxed under the regulations is a tax indifferent party, such as a pension fund, insurance company, foreign investor, 401(k) orI.R.A. account, or charity – a fact that I have confirmed through investment bankers who market these transactions and practitioners who advise them.
In the rare case that taxable holders actually are subject to these regulations, moreover, they face increased incentives to engage in so-called strategic trading (i.e., holding assets with built-in gains and selling ones with built-in losses). Under realization, holders have tax incentive to sell only when the asset’s value actually declines; here, in contrast, they have this incentive as long as the bond under performs the assumed yield – even if the return is still positive. Although issuers could have had a similar incentive under the deferral rule (i.e., to settle the bond early to accelerate the deduction), the transaction costs of unscheduled retirement (e.g., call
premiums) followed by a new issuance (e.g., another underwriting fee) are far higher even in proportional terms than commissions that holders pay to sell and repurchase a contingent bond.
In sum, a robust defensive planning option is available here. For taxable holders, the regulation’s main impact is to increase compliance costs while preserving, and probably increasing, avoidance costs and deadweight loss. Nor are the regulations likely to collect additional revenue from taxable holders.
2. Offensive Planning Option for Issuers
Albeit unappealing to holders, the enhanced accuracy of the regulations is quite attractive to an issuer. As Table 2 demonstrates, instead of waiting until the bond is settled, the issuer has a deduction each year (e.g., $880 in the first year) without having to pay any cash. Even if this assumed deduction perfectly reflects the amount owed on the bond (e.g., because the Dow rises by $880), the issuer is still better off because the deduction can be claimed without the (considerable) costs of settling the bond early. The deduction is still available, moreover, even if the Dow does not rise; although such assumed losses are available for tangible assets (e.g., accelerated depreciation), they are rare for financial assets and liabilities. Eventually, the issuer must give back this excess deduction when the bond matures (i.e., by including in income the excess of the amount deducted over the amount actually paid), but the issuer still has received a significant timing benefit.
One possible justification for this acceleration of deductions — notwithstanding the loss of revenue, since holders tend to be tax indifferent — is to prevent wasteful planning: issuers who want to issue contingent bonds might be deterred by deferral of deductions or, alternatively, by tax whipsaws if the issuer hedges the bond. To see the hedging concern, assume the issuer in our Dow example does not want to bet on the Dow (i.e., that the Dow will stay below $11,000); rather, the issuer believes that holders want to place such bets, and will pay extra to do so. To fund the contingent interest obligation, the issuer may buy a financial instrument that pays an equivalent amount (i.e., the excess of the Dow over $11,000). Yet although the issuer is hedged on a pretax basis, the timing and character of payments of the bond and hedge may not match (e.g., if gains on the hedge come in an earlier year than the loss on the bond).
Although these effects could induce an inefficiently low level of contingent debt, the regulations cannot be justified as a solution to this problem for two reasons. First, it is not clear how severe this problem was, since issuers could address these concerns through relatively cheap self-help. The contingent bond could be restructured as two securities – a fixed-rate discount bond and an option.Although the pretax cash-flows were the same, the issuer’s tax treatment was better in two ways: first, the discount bond generated interest deductions prior to maturity; second, the issuer could hedge with an option identical to the one sold to the public, thereby managing the whipsaw. Admittedly, this self-help was not free. Holders needed the right to trade the option and bond separately,and so the issuer would incur costs in facilitating such separability (e.g., separately listing the pieces, keeping records of the separate owners, facing potentially adverse accounting consequences).
The government might plausibly want to allow this result while sparing issuers the transactions costs – as a way to avoid waste, even though revenue could decline. Yet the regulations were not needed to achieve this goal, which was attained through a separate regulatory action: the so-called OID integration rules of Treas. Reg. 1.1275-6. Under this regime, issuers who hedge contingent debt are,in effect, given the same tax treatment as if they had issued an investment unit: The hedge and the contingent payment on the bond are treated as canceling out, and so all that remains – and all the tax system taxes – is a fixed rate bond. As Edward Kleinbard has noted, the integration rule eliminates the tax disincentive for issuers who would otherwise issue contingent debt. Once the integration rule has solved this problem, the regulations are not also needed to address this issue.
Even though issuers need not depend on the regulations to issue contingent debt, this reform has not gone unused. As Table 2 shows, issuers receive more favorable treatment through the regulations than through an investment unit or integration (which offers the same timing of deductions as an investment unit):
TABLE 3: COMPARISON OF “WAIT AND SEE,” INVESTMENT UNIT AND REGULATIONS
Thus, the regulations offer an extra $281 of deductions in the first year and a greater disparity thereafter. This difference arises because, unlike the investment unit approach, the regulations assume growth not only in the bond, but also in the embedded option. In addition to this timing benefit, the regulations offer issuers more favorable tax character. As the Dow rises, payments made by the issuer are capital loss under the investment unit approach, but they are ordinary deductions under the regulations.
For tax planners, the enactment of yet another treatment for essentially the same transaction – even a more accurate treatment – in effect adds another weapon to the planning arsenal, as long as accommodation parties are available and the new form merely supplements, but does not replace, the others. In this spirit, issuers have used the uniquely accelerated deductions offered by the regulations to engage in tax arbitrage. For instance, while issuing contingent debt based on the Dow (and thus claiming accelerated deductions as if the Dow is rising at 8% a year), the issuer can simultaneously hold an option on the Dow (which, under realization, yields no taxable gain before the option matures). Although the pretax returns cancel out – so that, on a net basis, the issuer is not actually betting on the Dow – the issuer still is receiving a substantial tax timing benefit (i.e., accelerated deductions on the bond, matched by deferred income on the option).
Two doctrines under current law can block this arbitrage in some cases, but such arbitrage is being done, nevertheless. The first is the integration rule of Treas Reg. 1.1275-6. Although a taxpayer seeking arbitrage would not choose to integrate, the government has the power to integrate in some cases and the result, as in an investment unit, is a deduction based only on the embedded bond, not the embedded option. However, issuers avoid this rule by tweaking their hedges. The rule never applies if the hedge is stock, as opposed to an option or some other derivative. In addition, the rule applies only if there is a near-perfect match between the economic terms on the bond and the hedge. Taxpayers can avoid the rule through a sufficient gap in maturity dates or economic terms. In response, the government should clarify that it (but not the taxpayer) is free to integrate even if this gap is wide, and the government should authorize integration when the hedge is stock.
Another regime, the straddle rules, is especially well suited to address this sort of arbitrage but, for technical reasons that should be fixed, fails to do so for contingent debt. This regime was enacted to foreclose a form of tax arbitrage, common in the 1970s. Taxpayers were placing offsetting bets (a so-called straddle) such that, inevitably, one bet would win and the other would lose. They would settle the loser on December 31 (thereby generating a tax loss in the current year) while settling the winner on January 1 (thereby generating income for the following year).
To revoke this artificial timing benefit, the straddle rules generally defer a loss from one leg of a straddle until corresponding gain is recognized in the other leg. Conceptually, the same abuse is occurring with contingent debt and the hedge in our example – deductions on the bond, deferred gain on the hedge – but, technically, the straddle rules often do not apply: The problem is that a deduction for interest is less likely to be deferred than a capital loss: Unlike capital losses, interest deductions are not deferred, as a per se matter, merely because they arise from a leg of the straddle. Rather, the interest expense must be “incurred to purchase or carry” the straddle.
This requirement offers taxpayers a line of defense that often preserves the deduction. Aware of this glitch, the Treasury has twice proposed legislation to undo it, each time without success.
If this hole cannot be plugged through legislation, the Treasury should do so through regulations.
E. Empirical Evidence: DECS, and PHONES
To sum up, the regulations have competing effects on planning. Theoretically, they curtail two kinds of planning, but these achievements are less impressive than they may at first appear: first, tax-sensitive holders no longer seek to hold contingent debt instead of fixed-rate debt (but prior efforts to exploit the contingent debt form were not vigorous, since tax deferral for contingent returns was easy to secure through other structures); second, tax-sensitive issuers no longer avoid issuing contingent debt (except that this problem was mitigated by self help and then largely eliminated by separate regulatory action). On the other hand, the regulations prompt two new types of planning that have attracted considerable attention in the tax bar: the interest of tax sensitive holders in avoiding contingent debt and of tax-sensitive issuers in making tax-motivated use of this structure.
The tax bar’s enthusiasm for the latter two planning strategies is evidenced by the menu of different flavors — some to avoid the contingent debt regulations, and another to qualify – that are offered for a common transaction: issuance of public securities to hedge an appreciated equity investment. I know from personal experience that, in structuring such a transaction, a tax lawyer asks two questions. First, will the security be marketed to tax-sensitive or tax-insensitive holders?Second, does the issuer have current use for an interest deduction?
If holders are tax-sensitive, the regulations are avoided by structuring the transaction as a so-called “DECS.” Indeed, right after the regulations were finalized, practitioners spilled a great deal of ink debating whether the regulations applied to DECS, and the consensus is that they do not. Because DECS do not guarantee that holders will recover their original investment, these instruments are thought not to be debt, and thus are not governed by the regulations.
If holders are tax-indifferent, though, issuers want the regulations to apply. For this circumstance, Merrill Lynch has developed the “PHONES” structure. Taxpayers issue a thirty-year principal-protected debt instrument with an embedded option, usually on a volatile stock.
Because an option’s value rises with its term and with the volatility of the underlying, this option is quite valuable – far more so than the fixed rate debt with which it is combined. Most of the issuer’s“comparable yield” deduction derives from the option. The deduction remains even if the underlying stock does not rise (or, in the usual case, even if the issuer is holding the stock, and can defer such gains until realization so that the transaction is a form of tax arbitrage). Given this favorable result, issuers have taken a real interest. Indeed, $5 billion of PHONES-type securities were issued in 1999, and Investment Dealers’ Digest describes 1999 as potentially “the biggest year ever” for the convertible debt market because “bankers are unveiling more creative structures than ever,” including PHONES and variations on it.
This strategic use of the contingent debt regulations is supported by empirical evidence, although more systematic studies would be helpful. A lexis search of prospectuses was designed to isolate public debt securities that are exchangeable for the stock of a third party. The search identified forty-five transactions since the regulations were finalized. Of these, only nine (i.e., 20%) were governed by the regulations — and of the nine, eight were PHONES transactions. The other thirty-six transactions were DECS. Of the forty-five transactions, then, the contingent debt regulations applied only once in a transaction that was not clearly tax arbitrage. The implication is that taxpayers are able to opt in and out of these regulations in pursuit of tax planning. The search and results are described in the appendix.
F. Government Responses
Given the costs imposed by the regulations, this Section considers the three types of government responses recommended by this Article: forgoing the reform; modifying the scope; and modifying the treatment.
1. Forgo the Reform
The regulations are a potent illustration of the problem of second best. Several commentators, including Edward Kleinbard, Professor Shuldiner, and Professors Cunningham and Schenk, have proposed an assumed-yield approach to offer advantages of mark-to-market without costly valuations. Yet reforms that would be appealing if applied comprehensively (as proposed by these commentators) can yield unappealing costs when applied narrowly. Here, compliance costs rise and, because the rules apply narrowly, there are limited economies of scale in learning them. The regulations almost certainly reduce revenue and probably create more new planning costs than they eliminate, although further empirical study would be helpful to confirm these conclusions. In my view,making the system more accurate in this limited way was not worth the cost. The question is whether the benefits being sought are attainable at lower cost.
2. Modify Scope
Planning costs and revenue losses derive, in significant part, from the ability of taxpayers to opt in and out of the regulations. These effects would dissipate if the regulations were extended to all comparable transactions – including direct investments in common stock, as well the whole range of derivatives that simulate this investment. (This, of course, is Professor Shuldiner’s proposal.) Yet this comprehensive solution is unlikely, if only because of the political difficulty of taxing “mainstream” investments like common stock in this complex manner.
A more realistic alternative is to extend the assumed yield approach to other derivatives, without reaching common stock. For instance, the regime could apply to other derivatives commonly used to avoid the regulations, such as options, prepaid forwards, and preferred stock whose value is based on the price of third-party stock, indices, etc. If the regulations extend to all these structures, the defensive planning option would be weakened, but not eliminated. Investors who want a derivative would have no choice, but those willing to buy more traditional investments could avoid the regulations. The net effect turns on the elasticity of investor’s preferences for derivatives, compared to the underlying asset. If taxpayers are as happy to have the underlying, extending the rule may increase avoidance costs and deadweight loss (as tax- sensitive investors flee the derivatives market) without raising revenue (i.e., because tax-sensitive issuers could deduct interest in an even broader class of transactions). I suspect, however, that the preference of many investors is somewhat inelastic because derivatives offer more nuanced economic bets than the underlying asset.[226 ] If this is the case,extension of the regime may be worth doing (if sufficiently broad), although the question warrants further study. If expansion is promising, the current regulations might be justified as a necessary first step that imposes costs in the short run but yields benefits in the long run. Yet in the tax field, the short run can seem anything but short, as tax reform may proceed slowly and taxpayers are adept at exploiting transitions. Thus, incremental steps should be crafted with care. Since in the near term the reform will be less broad than is desirable, the treatment it offers should be modified.
3. Modify the Treatment
Assume the regulations are not viewed as a first step toward expansive use of the assumed yield approach. (The opposite assumption is considered shortly.) Without such an expansion, the rule is unlikely to apply meaningfully to tax-sensitive holders. The only attainable improvement over prior law, then, is in lifting tax constraints on issuers. Assuming this goal is worth pursuing (e.g., notwithstanding the loss in revenue it will cause), integration is adequate. The regulations are not a useful supplement because they add compliance costs and, in offering issuers more accelerated deductions and thus tax arbitrage possibilities, the reform induces additional planning costs and revenue losses. On first reflection, the government might hesitate to pair integration with the old deferral rule because the result is asymmetrical in the taxpayer’s favor: issuers receive current deductions by integrating but holders continue to enjoy deferral. Yet although the assumed yield approach is more symmetrical in theory –since holders are supposed to have current inclusions – the symmetry is illusory because tax-sensitive holders are rarely subject to the rule. If there are no plans to expand use of the assumed yield approach, then, the regulations should be repealed and integration should be retained.
Assuming the regulations are a first step toward more expansive use of the assumed yield approach, the challenge is to attain this benefit at the lowest possible cost. Compliance and administrative costs cannot be reduced, since taxpayers and administrators will have to learn the regime. The most promising avenue for reducing costs, then, is in curtailing the issuer’s offensive planning option – that is, in discouraging issuers from issuing contingent debt that, in a world without tax, they would not issue. As an evidentiary matter, it obviously is impossible to know when issuers are responding to tax instead of business considerations. Yet tax will feature prominently in at least two issuer uses of contingent debt: tax arbitrage, as well as offerings intended primarily for tax indifferent holders. In these circumstances, issuer deductions should be deferred until paid, unless the issuer integrates.
In other words, anti-abuse rules should be added or refined to block these scenarios. To target tax-arbitrage, the straddle rules and the government’s power to integrate should be refined, as discussed above. While these adjustments present auditing challenges for the government, since it can be a daunting task to find which positions are offsetting, the issuer and its tax counsel will know which positions match (particularly if a deliberate strategy is at work, which is the scenario being targeted here) and a favorable tax opinion for the transaction will not be available.
Information costs also may make it difficult for the government to know when holders are tax indifferent. Yet the issuer will have better information since, with the help of its underwriter, the issuer knows what market is being targeted. To minimize compliance burdens, at least two approaches are possible. First, deferral might apply to a presumptive percentage of each offering (e.g., 30%) unless the issuer can supply proof to the contrary (e.g., a letter from the underwriter, subject to penalties for misstatements, about the marketing efforts). Alternatively, deferral might be applied only if the instrument is part of an offering that is marketed in substantial part to tax indifferent holders. This second approach resembles an anti-abuse rule that the government actually included in the regulation, but the treatment offered by this anti-abuse rule is probably too generous and its scope is too narrow. Instead of deferring the entire deduction, the current anti-abuse rule merely reduces its size. Moreover, instead of applying whenever the issuer markets the offering to tax exempts, the existing rule has a second condition: the contingent debt instrument must be based on the value of an asset that is not publicly traded (so-called “nonmarket” information). Furthermore, taxpayers can still claim the full deduction by offering adequate evidence of their borrowing cost. Even as revised, the rule may still prove under inclusive – for example, if issuers are able to avoid it through halting efforts at marketing to taxable holders – but the rule should still have effect in clear cases. Making the anti-abuse rule broad is a sensible strategy (i.e., to discourage efforts at avoidance), since issuers are given an “out” – integration – and thus can avoid the uncertainty and compliance costs prompted by a broad and vague rule.
A more dramatic modification of the treatment is to use mark-to-market accounting instead of the assumed yield approach. If the scope of the reform is the same (e.g., debt instruments only), the defensive planning option will not be affected. Yet the offensive planning option will be weakened, at the cost of serious valuation problems in some cases.
Compared to the assumed yield approach, mark-to-market offers a weaker offensive planning option for two reasons. First, one source of the offensive planning option, tax losses (temporarily) in excess ofeconomic losses, cannot arise. Second, although both rules can supply losses without the transaction costs of a realization, mark-to-market offers an offsetting tax disadvantage — the prospect of pre-realization gains. For mark-to-market, then, the offensive planning option is constrained, at least to an extent, by market balance.
However, the classic problem with mark-to-market, valuation, can apply here with some force. If the bond itself is listed on an exchange, valuation is a simple matter. Otherwise, the bond can prove hard to value, even if the contingent interest is based on publicly traded information. The problem is that the bond contains an option, and it is not possible to value an option merely by learning the value of the underlying property. For instance, to value an option on the Dow – and, thus, the contingent bond in our example – taxpayers (and, on audit, the government) require not only the Dow’s current value (which, obviously, is easy to obtain), but also information about prevailing interest rates, the amount of time remaining until maturity and, most importantly, the Dow’s volatility.Computation of volatility requires expertise, as does computation of the option’s value once these elements are known. Changes in the issuer’s credit will also affect the bond’s value in subtle ways.Hence, valuation of these instruments is beyond the abilities of many taxpayers. The assumed yield approach has the advantage of avoiding this issue.
A frustrating reality of tax reform is that it does not always achieve what we hope it will achieve. On the drawing board, a comprehensive reform seems attractive. Once politics and administrability concerns have their way, however, the reform’s scope and effects may be substantially modified. In some cases, the revised reform is still worth enacting, but in others it is not. A key inquiry, highlighted in this Article, is whether well-advised taxpayers will be able to avoid the rule or even to turn it to their advantage.
This problem has undermined two otherwise promising Haig-Simons incremental reforms: Section 475 and the contingent debt rules. These provisions should be modified, along the lines suggested here.More generally, although moving our system towards mark-to-market accounting is promising if proper care is taken, every move toward mark-to-market is not inherently wise. We must be mindful of the planning option.
Associate Professor, Columbia University School of Law. This Article is forthcoming in the Southern California Law Review. Please do not quote this draft without the author’s permission. Particular thanks are due to Marvin Chirelstein for providing insightful comments on multiple drafts of this work and, more generally, for the invaluable advice and friendship he has offered over the years. Helpful comments were also received from William Gentry, Edward Kleinbard, Michael Knoll, Clarissa Potter, Meredith Wolf Schizer, Martin Shubick, Alvin Warren, David Weisbach, and Lawrence Zelenak, as well as from participants in the Ernst & Young Tax Policy Seminar at the Georgetown University Law Center and in a Columbia Law School faculty workshop. The excellent research assistance of Amanda Steck is much appreciated. Copyright © David M. Schizer. All Rights Reserved.
Martin D. Ginsburg, Making the Tax Law Through the Judicial Process, 70 ABA Journal 74 (1984).
In a time of record corporate profits, corporate tax revenues have actually declined. Martin A. Sullivan, Shelter Fallout? Corporate Taxes Down, Profits Up, 99 Tax Notes Today 148-2 (1999); David Cay Johnston, Corporations Taxes Are Falling Even as Individuals’ Burden Rises, N.Y. Times, Feb. 20, 2000, at A1 (corporate “profits are growing even faster than Americans’ incomes,yet the taxes they pay have peaked and have begun to fall. . . . [T]he most troubling factor [explaining this decline] is the rapid spread of tax shelters, especially among large corporations”). For discussions of the tax shelter problem, see Dana L. Trier, Beyond the Smell Test: The Role of Substantive Anti-Avoidance Rules in Addressing the Corporate Tax Shelter problem, Taxes (March 2000) (“Much ink was spilled in 1999 on what has come to be called the `corporate tax shelter problem.’”); see also U.S. Dep’t of the Treasury, The Problem of Corporate Tax Shelters: Discussion, Analysis and Legislative Proposals (July 1999); New York State Bar Ass’n Tax Section, Report on Corporate Tax Shelters, 83 Tax Notes 879 (1999) (“NYSBA Report”); Joseph Bankman, The New Market in Corporate Tax Shelters, 83 Tax Notes 1775 (1999); David P. Hariton, Sorting Out the Tangle of Economic Substance, 52 Tax Law 235 (1999); Edward D. Kleinbard, Corporate Tax Shelters and Corporate Tax Management, 51 Tax Executive 235 (1999); Janet Novak & Laura Saunders, The Hustling of X-Rated Shelters, Forbes 198 (Dec 14, 1998). In a number of widely-followed decisions, courts have been responding to aggressive tax planning. See, e.g., ACM Partnership, 157 F. 3d 231 (3d Cir. 1998); Compaq Computer Corp, 113TC , No. 17, Dec. 53,549; Saba Partnership, 78 TCM 685, Dec. 43,604 (1999); United Parcel Service, 78 TCM 262, Dec. 53,497 (1999).
See Johnston, supra note 3, at A1 (“`Corporate tax shelters are our No. 1 problem’ in enforcing the tax laws”) (quoting Secretary of the Treasury Lawrence H. Summers). The Clinton Administration has proposed a range of legislative responses, which Congress thus far has declined to enact. For a discussion, see Trier, supra note 3. The issue has featured in the 2000 presidential campaign. See, e.g., John Harwood, Bradley Targets Tax Shelters for Mining, Oil, Wall St. J., Jan. 5, 2000, at A4 (Bradley and McCain each proposed ways of targeting tax
Derivatives are financial bets, which might be about interest rates, a particular stock price, or some other financial fact. See generally Group of Thirty, Global Derivatives Study 28 (1993) (“In the most general terms, a derivatives transaction is a bilateral contract or payments exchange agreement whose value derives, as its name implies, from the value of an underlying asset or underlying reference rate or index.”). In brief, such instruments allow taxpayers to take greater advantage of economic imperfections in the tax law, such as opportunities offered by the realization rule to defer tax on gains while claiming immediate deductions for losses. For instance, in allowing taxpayers to place carefully defined financial bets, derivatives allow taxpayers to cancel out the return on other investments in a way that could trigger at least three unwarranted tax benefits: first, the taxpayer could synthesize a sale without triggering tax; second, she could synthesize a time value return without triggering its usual (adverse) tax consequences (ordinary income and loss of deferral); third, by placing two perfectly offsetting bets (a so-called straddle), she could generate a timing benefit (a loss this year, and a corresponding gain next year) without taking any economic risk. For discussion of various transactional forms that can achieve these results, see David M. Schizer, Executives and Hedging: The Fragile Legal Foundation of Incentive Compatibility, 100 Colum. L. Rev.440 (2000).
For example, the corporate tax is filled with planning opportunities. Minor formal changes can induce significantly different tax results, whose appeal depends upon the precise tax characteristics of the parties involved (e.g., corporation or individual, high-bracket or low- bracket). For instance, profits might be extracted as dividends (for corporate shareholders who benefit from the dividends received deduction), as redemptions (for individuals who desire basis offset and capital gain), as interest (deductible to the corporation and not taxable to foreign or tax exempt holders) or through tax-free spin-offs, in each case subject to various limitations. The international and partnership tax rules also offer a multitude of planning opportunities. Such issues are left for another day.
See Lee Sheppard, Slow and Steady Progress on Corporate Tax Shelters, 19 Tax Notes Int’l 231 (July 9, 1999) (“The 900-pound gorilla in all the corporate tax shelter discussion — that practitioners do not want to talk about and the Treasury report mentions only obliquely — is derivatives. . . . Derivatives make a lot of this nonsense possible because they can be designed to
See, e.g., David F. Bradford, Fixing Realization Accounting: Symmetry, Consistency and Correctness in the Taxation of Financial Instruments, 50 Tax L. Rev. 731 (1995) (noting extensive literature on financial instruments and increased concern about arbitrage and other
gaming opportunities); Daniel Halperin, Saving the Income Tax: An Agenda for Research, 77 Tax Notes 967, 967 (1997) (“What concerns me is that the growth of derivatives and other financial products makes it increasingly difficult to maintain that the tax base for capital income is even an approximate measure of true economic income from capital.”); David P. Hariton, The Taxation of Complex Financial Instruments, 43 Tax L. Rev. 731 (1988) (noting uncertainty and potential for tax evasion created by financial innovation and failure of government to revise rules); Edward D. Kleinbard, Equity Derivative Products: Financial Innovation’s Newest Challenge to the Tax System, 69 Tex. L. Rev. 1319 (1991) ( “Viewed from the perspective of a tax practitioner, the federal income tax system has not been particularly adept at coping with financial innovation in the capital markets”); Robert H. Scarborough, Different Rules for Different Players and Products: The Patchwork Taxation of Derivatives, Taxes 1031 (1994) (“Over the past 15 years, Congress and the Treasury Department have created a wide variety of different tax regimes for derivatives.”); Daniel Shaviro, Risk-Based Rules and the Taxation of Capital Income, 50 Tax L. Rev. 643 (1995) (noting “widely recognized crisis in the taxation of financial assets, resulting from the development of innovative financial products,” warning that it could extend the taxation of non fungible assets and suggesting that the crisis may render efforts at taxing investment returns unworkable); Reed Shuldiner, A General Approach to the Taxation of Financial Instruments, 71 Tex. L. Rev. 243, 243 (1992) (“The tax law has struggled to keep up with the development of new financial instruments . . . Unfortunately, the lack of a uniform theory . . . has led to rules that are often haphazard, incomplete and inconsistent.”); Jeff Strnad, Taxing New Financial Products: A Conceptual Framework, 46 Stan. L. Rev. 569 (1994) (“A recent wave of innovation in the financial markets has raised difficult tax policy questions.”); Alvin C. Warren, Jr., Financial Contract Innovation and Income TaxPolicy, 107 Harv. L. Rev. 460, 461 (1993) (“Continuous disaggregation, recombination, and risk reallocation have produced a changing array of new financial contracts that pose a serious challenge for the income tax.”); David A. Weisbach, Tax Responses to Financial Contract Innovation, 50 Tax L. Rev. 491 (1995) (“Modern financial innovation raises serious problems for the tax system”).
The realization rule defers tax on an appreciated asset until maturity. In contrast, a mark- to-market rule taxes an asset annually, regardless of whether it has been sold, based on the difference between its value in the beginning and end of the year. For discussion of the advantages of mark-to-market, see, e.g., Shuldiner, supra note 8, at 246 (“Most, if not all, of these problems could be solved by abandoning our current realization system and adopting mark- to-market accounting for financial instruments.”); Weisbach, supra note 8, at 492 (“Changing from a realization-based system to another system may be the only complete solution to problems raised by financial innovation. A mark-to-market system, for example, would solve many timing problems presented by financial contracts.”); Halperin, supra note 8, at 967-68 (“We need a more accurate measure of income, one that would be simpler, more efficient, and most importantly, fair. This can be achieved only if realization plays a less important role in the timing of both income and loss.”). Two commentators have discussed the possibility of a universal mark-to-market system. See Fred Brown, Complete Accrual Taxation, 33 San Diego L. Rev. 1449 (1996); David Shakow, Taxation Without Realization: A Proposal For Accrual Taxation, 134 U. Pa. L. Rev. 1111 (1986).
Proxies for mark-to-market include taxing an assumed gain each year prior to realization, see, e.g., Shuldiner, supra note 8 (proposing “expected value” taxation); Kleinbard, supra note 8 (proposing cost of capital allowance); Noel Cunningham & Deborah Schenk, Taxation Without Realization: A “Revolutionary” Approach to Ownership, 47 Tax L. Rev. 725 (1992), or, instead, awaiting realization and increasing the tax by an interest charge to compensate the government for tax deferral. See, e.g., Warren, supra note 8, at 482 (“Serious consideration should therefore be given to moving . . to at least some formulaic taxation of contingent returns. Of the two approaches that apply an inherent rate of return, retrospective allocation of gain seems more promising . . . . “); Mary Louise Fellows, A Comprehensive Attack on Tax Deferral, 88 Mich. L. Rev. 722 (1990); Cynthia Blum, New Role for the Treasury: Charging Interest on Tax Deferral Loans, 25 Harv. J. on Legis. 1, 12 (1988); Stephen Land,Defeating Deferral: A Proposal for Retrospective Taxation, 52 Tax L. Rev. 45 (1996); William Vickrey, Averaging of Income for Income-Tax Purposes, 47 J. Pol. Econ. 379 (1939). See also Mark P. Gergen, The Effects of Price Volatility and Strategic Trading Under Realization, Expected Return and Retrospective Taxation, 49 Tax L. Rev. 209 (1994) (describing these various approaches).
See, e.g., Strnad, supra note 8, at 574 (“[B]ecause implementation of global pattern taxation [that is, a single rule for all assets] would require systematic reform, this fact is of little comfort to administrators who must craft rules in a system arrayed with different treatments that must be taken as given.”); Shuldiner, supra note 8, at 246, 285 (noting that “it is unlikely that Congress (or the financial community) will accept wholesale use of mark-to-market accounting”).
Whereas Professor Shuldiner offers an alternative that he apparently hopes will be implemented comprehensively, he says his intent is not “to provide a set of rules that can be immediately and simply applied to all financial products.” Id. See also Weisbach, supra note 8, at 492 (“While work on complete reforms such as [a mark-to-market system] is important, complete reform is unlikely in the nearterm.”); Halperin, supra note 8, at 968 (“In short, while I recognize we cannot politically, or even practically, apply mark-to-market in all circumstances, I believe we should do so to the extent possible.”).
See generally R.G. Lipsey & Kevin Lancaster, The General Theory of Second Best, 24 Rev. Econ. Stud. 11 (1956); see also Harvey S. Rosen, Public Finance 307 (5th ed. 1999) (“[T]he theory of the second best [provides that] [i]n the presence of existing distortions, policies that in isolation would increase efficiency can decrease it and vice versa.”).
See, e.g., Section 1256 (mark-to-market accounting for certain exchange-traded derivatives): Section 475 (mark-to-market accounting for securities dealers, as well as for certain commodities dealers and securities traders); Section 1259 (one-time mark-to-market in certain hedging transaction); Proposed Section 1260 (interest-charge approach for so-called “constructive ownership” transactions);Section 1291 (interest charge approach for stock in certain foreign corporations, known as “PFICS”); Section 1296 (mark-to-market election for PFIC stock): Treas. Reg. 1-1275-4 (assumed yield approach for contingent debt). References to sections are to the Internal Revenue Code of 1986, as amended.
Robert Scarborough and Edward Zelinsky have each written thoughtfully about the risk that Haig-Simons reforms will prove unappealing when implemented partially, even if they would be desirable if implemented universally. See, e.g., Scarborough, supra note 8, at 1044 (“Piecemeal reforms, each designed to make the tax system more neutral and to reduce tax avoidance, have created inconsistencies that themselves may distort taxpayer behavior and create tax avoidance opportunities.”); Edward A. Zelinsky, For Realization: Income Taxation, Sectoral Accretionism, and the Virtue of Attainable Virtues,19 Cardozo L. Rev. 861, 865 (1997) (“[T]axing unrealized appreciation selectively is the best the accretionist program can realistically offer. Such selective accretionism would engender the same kinds of distortions and unfairness as the rule of realization while sacrificing the benefits of realization.”). This Article demonstrates that different distortions arise from different types of incremental reform, and then offers guidance about how the government can mitigate these distortions, at least to an extent, when proceeding incrementally.
The term “reform” is used here to describe changes in the substantive law, including legislation, regulations and regulatory pronouncements, as well as judicial decisions (although courts probably are least likely forum for implementing mark-to-market-type reforms for derivatives). Theoretically, changes in enforcement practices could also be relevant, although they are not the focus of this Article.
The phrase “restructuring a transaction” is admittedly imprecise. The focus here is on circumstances in which taxpayers already have decided on their business objective (e.g., funding a software venture) and modify the means of implementation in response to tax considerations (e.g., by issuing debt instead of equity). Of course, the choice of which business objective to pursue can also be influenced by tax considerations (e.g., to produce software instead of teaching mathematics, or to supply labor instead of pursuing leisure, etc.), but the emphasis here is on tax- motivated implementation.
The “planning option” described here is broader than the “timing option” described by Professor Constantinedes and others. See, e.g., George M. Constantinedes, Capital Market Equilibrium with Personal Tax, 51 Econometrica 611, 621-23 (1983). The latter phrase describes the taxpayer’s ability, under realization, to choose the timing of gains and losses, and thus to defer gain on appreciated property (by not selling) while triggering immediate losses on depreciated property (by selling). This timing option is but one variation of the broader planning option, which describes any situation in which the taxpayer can restructure the transaction to alter the tax treatment (e.g., not just in deciding whether to sell or hold, but in deciding whether to issue debt or equity, to structure an acquisition as taxable or tax-free, etc.).
Indeed, mark-to-market-type incremental reforms are likely to be unappealing to those who favor a consumption tax, and thus wish to reduce the tax burden on investment. In a prior article, I have adopted this perspective to suggest an advantage of the realization rule not previously considered in the academic literature, i.e., that it is a credible way to reduce the tax burden on savings. See David M.Schizer, Realization as Subsidy, 73 N.Y.U. L. Rev. 1549 (1998). In contrast, in this Article I generally assume that mark-to-market-type reforms are otherwise desirable (and, more fundamentally, that we want to tax the return to capital) as a way to show that partial pursuit of these goals can generate a particular type of cost, and that attention should be paid to reducing it.
Professor Shaviro has noted the informational challenges in conducting a second-best analysis. “Under the theory of second-best, one may not know anything unless one knows everything,” but this is not an excuse for “`standing by and sadly sucking our thumbs under the sign of second best.’” Daniel N. Shaviro, Selective Limitations on Tax Benefits, 56 U. Chi. L. Rev. 1189, 1218-19 (1989) (quoting E.J.Mishan, Second Thoughts on Second Best, 14 Oxford Econ. Papers 205, 214 (1962).
Cf. Michael J. Graetz, Paint-By-Numbers Tax Lawmaking, 95 Colum. L. Rev. 609, 670 (1995) (“The most difficult aspect of revenue estimating is anticipating changes in behavior that will be induced by changes in the tax law.”).
A friction is a business cost that renders an otherwise tax-favorable strategy less attractive. See Myron S. Scholes & Mark A. Wolfson, Taxes and Business Strategy (1992).
There are lawyers (as well as other players in the system, such as investment bankers) who sometimes provide this information, though. The reports of the New York State Bar Association’s tax section are a significant example, as are occasional anonymous tips. At least four motivations contribute to such disclosures, ranging from generous to self-interested. First, many are genuinely public spirited. In addition, there are reputational benefits from being perceived as an expert on a new proposal, an imprimatur provided by articles and bar reports. Third, a conservative tax lawyer (or investment banker)will be at a competitive disadvantage if more aggressive players exploit a planning opportunity she cannot recommend. Finally, a relationship with government officials has value, and an offer of assistance on one matter may secure assistance on another. Yet client disapproval and the loss of opportunities to advise on lucrative transactions are significant disincentives. My sense is that information about planning opportunities is significantly undersupplied by the private sector.
Thus, the goal of government officials is assumed here to be making good tax policy in a “second best” world (in which fundamental reform is unavailable, an income tax will be maintained, and the government has imperfect information about taxpayers behavioral responses to reforms). This Article does not focus on incentives to raise revenue through surgical “revenue raisers.” For a discussion of these “pay go” rules, see Elizabeth Garrett, Harnessing Politics: The Dynamics of Offset Requirements in the Tax Legislative Process, 65 U. Chi. L. Rev. 501 (1998). Nor is the focus on agency costs that distract government officials from making good policy, such as the desire of elected officials for votes and contributions or the desire of appointed officials to maximize their consulting incomes upon leaving government service. While these factors no doubt have influence, the threshold question is what the law should be.
For citations, see supra note 3.
See generally Bankman, supra note 3, at 1780..
 Salaries at major law firms are a multiple of government salaries. While there obviously are capable people in government, many leave after relatively brief tours of duty.
See David Cay Johnston, I.R.S. is Bolstering Efforts to Curb Cheating on Taxes, N.Y. Times, Feb. 13, 2000, at A1; U.S. IRS Temporary and Proposed Regulations Require Corporations to Disclose Tax Shelters, 20 Tax Notes Int’l 1098 (2000). While these steps are constructive, more should still be done to redress the imbalance in resources. For instance, twelve employees have been staffed to the office of tax shelters. That is smaller than the tax department at a single major law firm. For thoughtful analyses of tax enforcement, see Louis Kaplow, Optimal Taxation with Costly Enforcement and Evasion, 43J. Pub. Econ. 223 (1990) (concluding that optimal degree of enforcement depends on several factors: direct resource cost of enforcement, revenue raised, distortion it causes, distortion caused by alternative tax increases, and marginal benefit of the government expenditures); Joel Slemrod & Shlomo Yitzhaki, The Optimal Size of a Tax Collection Agency, NBER Working Paper No. 1759 6 (1985) (enforcement expenditures should increase until, at the margin, they equal the saving in evasion- related social waste, such as the risk aversion of evaders and costs).
 By “choice of structure,” I mean that business objectives can be accomplished through different legal forms (e.g., operating through a corporation or a partnership, financing via debt or equity, doing an acquisition with voting stock or cash, etc.), and these forms trigger different tax consequences.
To ensure that the transaction must change at least to an extent, the tax law relies on judicially-created “substance over form” doctrines. For instance, the government can invoke “factual sham” or“economic sham” to argue that a transaction nominally qualifying as one type should really, as a matter of substance, be viewed as another. These “substance over form” doctrines can be viewed as constraints on the planning option. For a thoughtful analysis of these doctrines, see Daniel N. Shaviro, Economic Substance, Corporate Tax Shelters, and the Compaq Case (2000) (unpublished manuscripton file with author).
I am borrowing the terminology of Professors Slemrod and Yitzhaki, who classify five social costs to taxation. They distinguish between standard dead weight loss (i.e., misallocation of resources when,in response to tax, taxpayers shift from high-tax to low-tax activities they would not otherwise prefer) and avoidance costs (i.e., time and resources invested in efforts to discover tax-reduction strategies).See Joel Slemrod & Shlomo Yitzhaki, The Costs of Taxation and the Marginal Efficiency Cost of Funds, 43 IMF Staff Papers 172 (March 1996). For discussion of the other three costs of taxation they identify (compliance costs, administrative costs and evasion costs), see infra note 48.
See Scholes & Wolfson, supra note 21, at 127-28 (“[W]e cannot emphasize too strongly the importance of these nontax costs in forging efficient tax plans . . . Because of the need to make these tradeoffs [between reducing tax and reducing transactions costs or so-called “frictions”], efficient tax planning is often quite distinct from tax minimization.”).
See Shuldiner, supra note 35, at 782 (defining symmetry); see id. at 786 (noting that symmetry can constrain tax planning). Cf. Scholes & Wolfson, supra note 21, at 5 (“[T]o organize production at minimum cost requires that the tax positions of all parties to the contract be considered.”).
Charities are not subject to U.S. tax, see Section 501, except on their unrelated business taxable income. See Section 511. Likewise, foreigners are generally not subject to U.S. tax, except on income“effectively connected” with the U.S., see Section 864, or on passive income subject to withholding tax (e.g., dividends). See Section 871. Other tax indifferent parties include pension funds, Native American tribes, taxpayers with net operating losses and, as discussed below, insurance companies and securities dealers and traders. See infra Part III.B.
See, e.g., Bradford, supra note 8, at 741 (“The existence of taxpayers in different marginal rate brackets virtually eliminates the potential to use market adjustment as a substitute for consistent rules to measure returns over time.”). Cf. Scholes & Wolfson, supra note 21, at 6 (certain tax “clienteles” are likely to transaction with each other, such as high-bracket and low- bracket taxpayers).
As Professor Shuldiner has used the term, a tax system is consistent if identical transactions are accorded identical treatment. Reed H. Shuldiner, Consistency and the Taxation of Financial Products,Taxes 781, 782 (1992) (defining consistency). As Professor Strnad has used the term, a tax system is continuous if similar transactions are accorded similar tax treatment. See Strnad, supra note 8, at 584. In a continuous system, small changes in the transaction thus would not yield large changes in the tax result.
In a sense, the third condition is a subset of the second. A counterparty almost always can be found at some price. The point is to find one who will not add prohibitively to avoidance costs.Incidentally, the term “accommodation party” is also used in commercial law, but that meaning is not intended here.
The point is that the ability to choose one’s form is analogous to an option to buy (or sell) an asset, in that holders of this right are free to use the right only when doing so is profitable. For instance, if a share of stock is worth $100, the right to buy for $200 (a “call option” with an “exercise price” of $200) is valuable because there is a chance that, before the option expires, the asset will appreciate above $200. This opportunity for gain, however remote, is not matched by a corresponding risk of loss because the option confers a choice: The option-holder can choose to buy the asset for $200, but does not have to do so. If the asset is worth $30 when the option matures, the option-holder will simply not use it. In contrast, the party granting the option – the one who agrees to sell the property for $200 if(and only if) her counterparty chooses to buy it – can only lose: The so-called option-grantor will be asked to sell for $200 only if the property is worth more. Thus, the option-grantor will charge the option-holder something, a so-called premium, for this right to choose. For an introduction to options pricing, see generally P.J. Hunt & J.E. Kennedy, Financial Derivatives in Theory and Practice (2000); Robert M. McLaughlin, Over-the-Counter Derivative Products 65-95 (1999).
Some of these expenditures might be classified as premium, instead of exercise price, depending upon whether the expenditure is already a sunk cost (as research costs might be) by the time the taxpayer decides whether to use the tax-favorable structure. Either way, the lower these costs are in the aggregate, the more appealing the planning option is to the taxpayer. The distinction may prove useful, though, in analyzing taxpayer decisions at various points in time.
For instance, since sunk cost will be ignored when taxpayers decide which structure to implement, they may choose a structure in which tax benefits are higher than costs on a going forward basis, but not on an aggregate basis (i.e., once sunk costs are factored in). The analogy is to exercise of an option at a gain that is less than the premium paid.
See Ronald J. Gilson & Bernard S. Black, The Law and Finance of Corporate Acquisitions 239 (2d ed. 1995).
See id. at 239-43 (option value increases with volatility of underlying property). To see the effect of volatility on an option’s value, assume Stable Investment is equally likely to pay either $40 or $60, whereas Risky Investment is equally likely to pay $10 or $90. Since their mean return is the same ($50), risk neutral investors would pay the same price for either asset (assuming each is an idiosyncratic gamble with no correlation to general market risk). Risky may be better when both investments appreciate, but Stable is better if things go badly. For an option, however, the latter scenario is not relevant. The focus is only on how well the investor does in good times, since the option investor is protected from the bad times (i.e., because she won’t exercise the option). As a result, an option on Risky is more valuable because it offers a greater potential payoff.
If the tax itself is unwise (so that better alternatives are available for raising revenue), tax planning might be defended on the theory that bad rules should not be enforced. Nevertheless, it is usually better to repeal the tax (or, as a less drastic step, to reduce the rate) than to rely on taxpayer self-help.
Capitalization of these tax benefits would mitigate this vertical inequity. See Shaviro, supra note 19, at 1223 (“A preference does not undo rate progressivity if incurred solely by taxpayers in the highest rate bracket and if its value is completely capitalized for those taxpayers.”). Yet capitalization occurs only if the supply of these planning opportunities is too limited to accommodate the potential demand of taxpayers in the highest bracket, such that prices are bid up, and pretax yields are bid down, to levy an “implicit tax” that cancels out the tax savings. This is not always the case. For instance, the tax savings from using securities dealers as accommodation parties has probably not been fully capitalized (although some of this savings is no doubt is shared with the dealer) since the supply of dealer services is vast: each dealer can accommodate essentially any transaction it can hedge (subject to regulatory capital limits and transaction costs), and new players can become dealers.
Of course, distributional effect can be offset by other features of the tax system, such as the progressive rate structure. For instance, it is possible that the maximum bracket is higher than it otherwise would be to compensate for the planning option’s tax-reducing effects – although Congress might fail to compensate adequately if, as noted above, the political process does not monitor planning adequately. In any event, a comprehensive distributional analysis, with attention to all features of the tax and transfer system, is beyond this Article’s scope.
While capitalization would undercut horizontal inequity, not all planning opportunities will be fully capitalized. See supra note 42. Even so, as long as taxpayers with comparable incomes have the same opportunity to plan, horizontal inequity arguably is not offended. Yet as Professor Shaviro points out, “what seems to be the same opportunity may in fact not be, if taxpayers differ in inclinations or aptitudes and some types of inclinations or aptitudes are rewarded by the tax system more than others.” Shaviro, supra note 19, at 1221.
See Rosen, supra note 12, at 292-303 (discussing theory and measurement of excess burden and why it is an important concept for evaluating actual tax systems); see also Daniel N. Shaviro, An Efficiency Analysis of Realization and Recognition Rules Under the Federal Income Tax, 48 Tax L. Rev. 1 (1992); (analyzing realization rule from efficiency perspective); David A. Weisbach, Line Drawing,Doctrine, and Efficiency in the Tax Law, 84 Cornell L. Rev. 1627, 1649-51 (1999) (arguing that efficiency, defined with reference to dead weight loss, is proper criterion for evaluating tax rules).
Cf. Shaviro, supra note 8, at 655 (noting that tax planning with derivatives can be viewed as “a step towards consumption taxation taken through taxpayer self-help” but at the cost of a large “socially wasteful excess burden”).
See David A. Weisbach, An Economic Analysis of Anti-Tax Avoidance Doctrines 7 (2000) (unpublished manuscript on file with author) (“The conclusion is that even in an optimally implemented tax system, the types of behavior associated with tax shelters and tax avoidance will occur.”).
A possible defense of planning is that those who avoid a tax through planning may have more elastic preferences for the underlying behavior than those who do not, a potentially appealing result because taxes are most efficient when levied on those with inelastic preferences.
Cf. Slemrod & Yitzhaki, supra note 30 (noting that avoidance and evasion are methods of filtering out those with elastic preferences). Yet it is usually less wasteful to add an exception to the law for these elastic players, instead of relying on costly self-help — assuming the costs of adding and administering this nuance compare favorably with the costs imposed by the self help, as I suspect they often do. Cf.Louis Kaplow, Rules versus Standards: An Economic Analysis, 42 Duke L.J. 557 (1992) (offering cost-benefit analysis of adding detail to rules and standards).
As Professors Slemrod and Yitzhaki have observed, there are several sources of social waste from taxation, and the objective is to minimize the sum of all of them – rather than one in particular. In addition to standard deadweight loss and avoidance costs, which are defined supra note 30, they list three more: administrative costs (i.e., the costs borne by the government in implementing the system);compliance costs (i.e., costs needed to comply with the tax law, even if the taxpayer is not trying to reduce taxes); and evasion costs (i.e., disutility from engaging in illegal activities and expenditures to conceal them). See Slemrod & Yitzhaki, supra note 30, at 179-82. Evasion costs do not figure prominently in the analysis here, which focuses on well- advised taxpayers who comply with the law.
The lower the ratio, the less wasteful the reform. The ratio’s numerator is the total new burden on the taxpayer – not just the tax increase itself, but also potential sources of new social waste:compliance costs and plus welfare losses from tax planning (e.g., fees to tax advisors, changes that make the transactions otherwise less appealing, etc.). The denominator is the new revenue actually raised – that is, the extra gross revenue minus the government’s added administrative costs in raising it.
The reader may wonder how marginal welfare loss is measured for the numerator. Slemrod and Yitzhaki assume that taxpayers at the margin will incur up to one dollar of utility losses to avoid paying a dollar of tax. Hence, they measure marginal welfare losses indirectly through revenue forecasts: the amount of new revenue that would be raised if taxpayers did not change their behavior is compared with the amount that actually is expected to be raised (i.e., once behavioral responses are considered). This difference, attributable to tax planning, is assumed to reveal the marginal utility loss from such planning. Although convenient, this assumption can be unreliable in two respects. First, revenue forecasts can prove inaccurate, particularly for a single reform (as opposed to a group of them, so that errors can cancel out).
See Graetz, supra note 20, at 670. Nor is the revenue loss necessarily a reliable measure of utility loss, because some who change their behavior are not marginal: although avoiding one dollar of tax will cost some (marginal) taxpayers a full dollar of utility loss, it may cost others (infra marginal ones) only a nickel.
This Article focuses on this first-order inquiry, rather than on other elements of the efficiency analysis, such as the propriety of the underlying tax or compliance and administrative costs.
 For instance, it is very unlikely that foreigners will become subject to U.S. tax in a comprehensive manner, and charitable organizations and pension funds are likely to retain their tax exempt status.Nor can consistency be achieved in one stroke, absent fundamental tax reform which, in this Article, is assumed to be politically unattainable.
Although generally applicable, these two effects are not the only ways in which “sticks” can turn into “snakes.” Others are left for another day.
The boundary between these categories is not airtight. The decision to join a group of taxpayers (i.e., so that one is subject to taxpayer-specific reforms) is, in a sense, a transaction (e.g., renouncing citizenship and becoming a foreign person, or becoming a securities dealer). As used here, “transactional” classifications are based on criteria that are less permanent, so taxpayers can position themselves on one side of a line or the other with relative frequency and ease (e.g., borrower vs. lender, buyer vs. seller, long vs. short).
The distinction between “defensive” and “offensive” here does not describe the type of planning per se, but the relationship of the planning to a given reform – that is, whether the taxpayer is avoiding the new rule (defensive) or deliberately qualifying (offensive). For instance, a taxpayer who is trying to accelerate a loss (a type of tax planning) may use a defensive planning option ( in avoiding a reform that would otherwise bar this strategy) or an offensive planning option (in exploiting a reform that facilitates this result).
See Weisbach, supra note 44, at 1662-63 (reforms should be extended to the point where close substitutes are grouped together, so that taxpayers have an inelastic preference for transactions on one side of the line).
For instance, assume a taxpayer borrows $100,000 and agrees to pay a $10,000 annual coupon. She uses the proceeds to buy a discount bond, which will grow by $10,000 a year. As an economic matter, the two cancel out. Yet if tax on the discount bond is deferred, but her deduction on the coupon bond is not, she has a $10,000 deduction each year, which can shelter other income (e.g., on salary).Tax on this other income is thus deferred until the discount bond matures, thereby generating a valuable timing benefit.
See Shaviro, supra note 29, at 5 (desirability of anti-avoidance approaches “depends on two main things. The first is the social desirability of deterring optimal tax planning in the cases that are being addressed. The second is the extent to which it succeeds in generating such deterrence rather than simply inducing taxpayers to jump through a few extra hoops before getting the desired tax consequences anyway.”); see also Kaplow, supra note 27, at 233 (while better enforcement “decreases distortion with respect to marginal individuals” who discontinue tax-avoidance, better enforcement also “increases distortion through its effect on the inframarginal cost of evasion” by forcing those who continue the tax-avoidance strategy to spend more on it).
As noted above, a comprehensive efficiency analysis would also consider other factors, such as administrative and compliance costs and the efficiency of the underlying tax rate.
The reference is to R.M. Haig and Henry Simons, who defined income as the sum of consumption and changes in the market value of a taxpayer’s property. This definition, which relies on mark-to-market accounting, is “the rallying call of tax theorists and reformers,” see William A. Klein & Joseph Bankman, Federal Income Taxation 76 (11th ed. 1997).
Professor Bradford has also noted the tradeoff between accuracy and consistency. See Bradford, supra note 3, at 736 (“[I]t may be essential that income measurement rules involving different sorts of instruments be related consistently to one another, even if the rules fail to measure income correctly.”). Professor Strnad has likewise emphasized the importance of consistency. See Strnad, supra note 3,at 572-73 (noting importance of consistency).
 For instance, a reform that covers only low income taxpayers (or omits only them) is relatively safe because these taxpayers are unlikely either to hire, or to serve as, accommodation parties. These taxpayers have less to gain from tax planning for themselves, and they usually lack the institutional characteristics needed to facilitate planning by others, such as capital and expertise.
See Weisbach, supra note 44, at 1662-63 (lines should be drawn based on cross- elasticity, so that “we should tax similar things similarly to minimize substitution costs, but not too much at the expense of direct costs”).
Professor Shaviro uses the metaphor of a taxpayer drifting downstream with a leg on two rafts, the “tax planning raft” and the “business planning raft.” “[T]he two rafts may drift sufficiently far apart that she must jump off one raft, letting it drift away while she stands entirely on the other.” See Shaviro, supra note 29, at 5.
Unfortunately, in the taxation of derivatives and also in corporate tax, empty formalisms matter a great deal. For instance, the same transaction can be documented as a swap, a forward contract, an option, or a contingent debt instrument, and the tax result will vary with the form used. For a discussion of these rules, see Lewis R. Steinberg, Using OTC Equity derivatives For High-Net-Worth Individuals,in The Use of Derivatives in Tax Planning 211, 242 n. 110 (Frank Fabrozzi ed. 1998). Likewise, an acquisition by a subsidiary is economically similar to an acquisition by the parent followed by a “dropdown,” but tax consequences can be different. For a discussion, see Martin D. Ginsburg & Jack S. Levin, Mergers, Acquisitions and Buyouts 802.6 (1999) . Because minor changes in cash flows or documentation have a significant influence on tax treatment, these boundaries are likely to be inefficient.
 A similar problem arises if the tax boundary is based on boundaries from other regimes, such as whether the relevant security is publicly traded. The advantage of such “piggybacking” is that, if the existing classification rewards transactions on one side of the line in a significant way, tax advantages are less likely to lure taxpayers to the other side. Yet which tax rule applies, for instance, if a security’s value is based only in part on publicly traded information?
In other words, sometimes discretionary “standards” should be used instead of (or in addition to) precisely-delineated “rules.” There is a sizable literature on the rules vs. standards debate, and no attempt is made here to capture it in all of its nuance. See generally, e.g, Colin S. Diver, The Optimal Precision of Administrative Rules, 93 Yale L. J. 65 (1983); Kaplow, supra note 47, at 557 (rules are generally preferable when the governed behavior is likely to be frequent, because greater up-front cost in crafting them is amortized over more uses of the rule).
See David A. Weisbach, Formalism and Statutory Interpretation, 66 U. Chi. L. Rev. 859, 875-77 (anti-abuse rules allow taxpayers the benefit of simple and predictable rules, while protecting the government from the risk that previously-uncommon, mistaxed cases will become more common).
Like the rules-standards issue, retroactivity is also the subject of a vast literature, whose subtleties are beyond this Article’s scope. Some of the normative issues are similar, in that concerns arise about those who continue to plan but face higher costs, as well as about compliance costs and excessive precaution. For an insightful discussion of retroactivity and transition, see Daniel N. Shaviro, WhenRules Change (2000).
This is the intuition behind the identification requirement in the hedging rules of Treas. Reg. 1.1221-2. Taxpayers are allowed to choose their character, but only before knowing whether they have gains or losses.
Such rules can affect the scale of the bet, as do rate changes that are symmetrically applied. If gains and losses are both deferred, the taxpayer keeps a larger share of both gains and losses (since the effect is like a rate reduction). As a result, the taxpayer may reduce the riskiness of the bet in order to keep her after-tax consequences the same. For a discussion, see Joseph Bankman & Thomas Griffith,Is the Debate Between an Income Tax and a Consumption Tax a Debate About Risk? Does it Matter, 47 Tax L. Rev. 377, 396-400 (1992) (noting ways in which taxpayers can cancel out tax rate on risk byadjusting scale of bet).
Such a strategy, known as the “timing option,” is available under the realization rule. It is discussed supra note 17 and infra in Part II.
Cf. Alvin C. Warren, Jr., How Much Capital Income Taxed Under an Income Tax is Exempt Under a Cash Flow Tax, 52 Tax L. Rev. 1 (1996) (unlike taxpayers placing a purely speculative gamble,taxpayers earning a time-value return are not indifferent to tax rates merely because parallel treatment is accorded to gains and losses).
This recommendation is consistent with the observation of Professor Strnad that, if the law is neither consistent nor continuous, one-way rules may be necessary to deter tax planning. See Strnad, supranote 8. Professor Bradford has raised a similar concern, although he offers an alternative rule (the so-called “gain recognition date” approach) meant to foreclose tax planning. See Bradford, supra note 8.Professor Knoll has recently pointed out, though, that Professor Bradford’s approach still allows tax planning. Michael Knoll, Tax Planning, Effective Tax Rates and the Structure of the Income Tax(unpublished manuscript on file with author) (arguing that Bradford’s effort to solve selective realizations and lock-in would offer taxpayer’s the incentive to engage in transfer-pricing type strategies, in which profit is allocated to assets acquired most recently).
While asymmetry can significantly weaken the offensive planning option, it may not eliminate it. A rule that is unfavorable to most taxpayers may still favor a (potentially modest) subset. Thus, while most borrowers will not want deductions deferred, those expecting to be subject to a higher tax rate in the future may prefer this result. Even so, the reform is not valuable to the latter group unless it offers a better way to defer losses than is otherwise available – an unlikely proposition, since loss deferral (unlike gain deferral and loss acceleration) is subject to few restrictions (i.e., because it generally increases revenue).
See Section 1092 (straddle rules defer losses, but not gains; they provide for short term straddle gains and long term straddle losses); Section 469 (deferring losses from passive activities in which taxpayer does not materially participate). See also, e.g., Section 1258 (conversion transaction rules convert capital gain to ordinary gain, but do not convert capital loss to ordinary loss); Section 1259(constructive sale rules cause taxpayers to recognize gain, but not loss); Section 1260 (constructive ownership rule recharacterization and interest charge applicable only to gains, not losses).
Admittedly, this political benefit has yet to surface for the straddle and passive loss rules.
[77 ] For instance, a business can be conducted through a “C corporation” (with entity level tax) or a partnership or S-corporation (without entity level tax). See Ginsburg & Levin, supra note 64, at 11-1.One firm can acquire another via a taxable asset or stock purchase (which reduces the purchaser’s future tax bill) or a tax-free reorganization (which reduces the seller’s current tax bill). Id. at 3-1.Likewise, individuals can save through a “deductible” IRA (best for those whose bracket is expected to decline at retirement) or “Roth” IRA (best for those whose bracket will not decline). See Daniel Posin,Festival of IRAS, 17 Tax Prac. 108 (1998).
For instance, the same transaction can be structured as a forward contract, a swap, a pair of options, or a contingent note. See Randall Kau, Carving Up Assets and Liabilities – Integration or Bifurcation of Financial Products, 68 Taxes 1003, 1004-05 (1990) (describing numerous ways to earn the same economic return).
See, e.g., Sheppard, supra note 7, at 232.
For a discussion, see Schizer, supra note 18, at 1610.
For a discussion of the timing option, see Constantinedes, supra note 17, at 621-23; Jeff Strnad, Periodicity and Accretion Taxation: Norms and Implementation, 99 Yale L.J. 1817, 1882-84 (1990); Gergen, supra note 10, at 211. The timing option can be viewed as a particular variation of the planning option. See supra note 17.
Under current law, a number of rules constraint the timing option, such as the capital loss limitations, the wash sale rules, and the straddle rules.
The market might capitalize the tax rule’s benefits into asset prices. Whereas this response could address equity concerns, it would not prevent an over-allocation of resources to assets subject to realization. For a recent empirical study on tax capitalization, see William M. Gentry et al., Are Dividend Taxes Capitalized Into Share Prices? Evidence from Real Estate Investment Trusts, Columbia Business School Working Paper (October 22, 1999).
Proxies for mark-to-market do not necessarily undo the timing option. For a discussion, see infra Part IV.D.1; see also Gergen, supra note 10, at 209.
See sources cited supra note 11.
For reforms in this vein, see supra note 13.
See Warren, supra note 8, at 485.
See Section 475(c)(1).
LCM is a method of inventory accounting, which first came into use in 1919. For a discussion, see Edward D. Kleinbard & Thomas L. Evans, The Role of Mark-to-Market Accounting in a Realization-Based Tax System, Taxes 788, 796 (December 1997). Securities dealers sought a special ruling that they could use this method, see Office Decision 8, 1 C.B. 56 (1919), perhaps, as Kleinbard and Evanssurmise, because of the “natural incredulity of the dealers that they had been handed so munificent a gift.” Kleinbard & Evans, supra note 89, at 796 n. 49. This accounting method was available only for “inventory,” which for tax purposes means merchandise available for sale. As a result, the rule was considered inapplicable to short positions, since these were viewed as liabilities, as well as to financial contracts, such as equity swaps, which involved a continuing contractual relationship between the dealer and client, as opposed to sale of a good. See id. at 796 (short sales not covered); id. at 797 n. 61 (equity swaps not covered).
For most taxpayers, the timing option is constrained by limits on a taxpayer’s ability to use losses, such as the capital loss limitations, the wash-sale rules, and the straddle rules. Yet these regimes do not apply to dealers. See Section 1091(a) (exempting dealers from wash sale rules, which otherwise defer losses when a taxpayer sells a position at a loss and repurchases substantially identical property within 30 days); Section 1092(e) (exempting dealers from straddle rules, which otherwise defer losses when taxpayers sell one position at a loss while retaining an offsetting position that has unrecognized gain); Section 1256(e) (deeming character of dealer gains and losses ordinary and thus exempt from capital loss limitations, which otherwise keep taxpayers from deducting capital losses until they have corresponding amount of capital gain).
 As Kleinbard and Evans describe this pro-taxpayer rule, “the only constraints on taxpayer electivity in a realization regime are, at most, transaction costs – and a lower-of-cost-or- market accounting system for inventory eliminates even those frictions.” Kleinbard & Evans, supra note 36, at 800.
LCM applied only to inventory and, as discussed above, derivatives and short positions technically did not qualify. See supra note ?.
For instance, if the dealer has more longs with clients than shorts (so that the dealer would lose money if the price declined), the dealer might do a short sale or enter into a short futures contract on a stock or futures exchange.
These timing disparities also arise when dealers use “dynamic” hedging strategies. For a description, see McLaughlin.
See Kleinbard & Evans, supra.
See Lee A. Sheppard, Who’s Marking What to Market, 97 TNT 103-3 (July 8, 1997) (Section 475 “was enacted in 1993 to repeal the outrageous lower-of-cost-or-market accounting method long permitted for securities dealers”).
Sheppard, supra note 96 (Section 475 “marked a legislative acknowledgment that the cherished realization requirement was not a realistic way to measure the income of securities dealers, a group who know the value of everything they hold at every hour of the day”).
The legislative history of Section 475 emphasizes that these usual administrability obstacles to market to market accounting are absent for securities dealers. See H.R. Rep. No. 103-111, 103d Cong., 1st Sess., 660 (May 25, 1993).
Of course, sometimes dealers presumably would have dictated timing and structure, and the relative frequency of these outcomes — and, thus, the likely vitality of the dealers’ timing and planning options — is an important empirical question on which more data would be enlightening. It is instructive that dealers themselves did not expect to control timing and character and thus preferred mark-to-market to realization, once it became clear that LCM was doomed. See Kleinbard & Evans, supra note 36, at 800 (“Once the inevitability of legislation had been accepted, the securities dealers advocated abroad application of mark-to-market to their core dealer activities . . . .”).
The precise effect, whether on capital, labor, or consumers, would depend upon incidence
This was the main argument the securities industry used to retain LCM. As one observer noted, “There is little controversy over whether or not LCM is a lopsided, pro-taxpayer accounting technique.Most of the industry’s arguments in favor of LCM boil down to `Why single us out?’ – rather than `This is a fair, accurate measurement of our income.’” See Pratt, Beltway Blackmail, Investment Dealers’Digest 22 (Sept 23, 1991).
 One might expect dealers to be relatively risk neutral given their expertise about market risk. Yet I know from personal experience that they sometimes are risk-averse about tax whipsaws. Although diversified shareholders would probably be unconcerned, traders and in- house tax lawyers often dislike such risk – particularly if they expect to be blamed for bad tax outcomes more than they will be credited for windfalls.
 For instance, assume a dealer has two client positions that are offsetting. Once the market price changes, one client will have gain (and the dealer will have matching loss) and the other client will have loss (matched by dealer gain). The latter client, but not the former, will have a tax incentive to terminate early, leaving the dealer with realized gain and unrealized loss.
Distortions arise if some clients are not deriving any tax benefit (e.g., they are tax exempt, foreigners, etc.) because the price increase will be, in effect, for a service clients do not want. The question is, then, whether dealers can discriminate (in accommodating tax preferences and in pricing) between tax-sensitive and tax-indifferent clients. In tailored transactions dealers surely do,but in standardized ones they probably do not (e.g., because the information costs of assessing the client’s tax appetites and the appropriate response are too high).
See supra text accompanying notes ? to 34.
At maturity, if the taxpayer accepts delivery of the stock, she has no tax consequences until she sells the stock. If the taxpayer instead “cash settles” the forward (by making or receiving a payment equal to the built-in value of her position), such settlement will generate taxable gain or loss. See Section 1234A.
See Section 1234A (termination gives rise to immediate tax loss).
Cf. Scarborough, supra note 8, at 1045 (noting that enactment of Section 475 “make[s] large numbers of taxpayers indifferent to realization events with respect to derivatives that they hold”);Shuldiner, supra note 35, at 789 (“Just as in the case when the second party to a contract is tax-exempt, termination symmetry becomes far less effective when the second party to the contract accounts for the contract on a mark-to-market basis.”).
Similarly, dealers can “rent” their relative indifference to character, as well as to timing, since most of their gains and losses are ordinary. For example, for taxpayers other than dealers, early termination of a swap generates capital gain or loss. See Section 1234A. Yet if certain swaps are held until scheduled payments are made, the payments are ordinary. See Schizer, supra note 5 (swaps that make periodic payments generate ordinary income and loss). The party who expects a loss will want to wait until the scheduled payment date (so the loss will be ordinary), while the party with a gain will want early termination so gains will be capital. Since a dealer is indifferent – either way, its gains and losses are ordinary – it can offer a client the choice.
See Section 511.
See Paula Froelich, OTC Derivatives Are Popular With Investors, Profitable For Brokers, Wall St. J., July 26, 1999, at B8H (“There is one hot product area that seems to be thriving regardless of the market’s fate: over-the-counter equity derivatives.”); Stephen Labaton & Timothy L. O’Brien, Financiers Plan to Put Controls on Derivatives, New York Times, Jan. 7, 1999, at C1, C3 (noting that $37 trillion worth of privately traded derivatives contracts were outstanding in January 1999, compared to only $865 billion in 1987).
For a discussion of the tax shelter industry, see generally Bankman, supra note 3. For an example of a transaction developed by dealers, see infra note 122 and accompanying text.
See Section 881.
See Section 864.
See Yaron Z. Reich, U.S. Federal Income Taxation of U.S. Branches of Foreign Banks: Selected Issues and Perspectives, 2 Fla. Tax. Rev. 1 (1994). As a securities dealer, the foreign institution can use mark-to-market accounting in filing its U.S. return, but we have come full circle. The foreign firm can do no better than a U.S. securities firm since, in this case, its ability to serve as an accommodation party also derives from Section 475, and not from its foreign status.
Specifically, when insurance companies sell policies based on the value of particular assets, they hold these assets in a “segregated account.” The company then marks-to-market these assets as well as the offsetting liability to its policy-holder, leaving the company no net tax consequences except on fees charged to the client. For instance, assume a policyholder pays $100 for a life insurance policy, whose payout is based on the value of Internet.Com (currently worth $100).The company uses the $100 to buy a share. If the share appreciates to $300, the insurance company has $200 of income on the share, and a perfectly-offsetting $200 deduction its increased liability on the policy. This tax treatment is dictated by Sections 817 and 817A. For a discussion,see Kleinbard & Evans, supra note 36, a 802 n. 97.
For instance, although the value of so-called “variable” life insurance depends upon the performance of designated investments, the policy-holder does not have access to the investment until she dies (or, at least, until some period of time passes, so she can convert the policy to an annuity).
See Section 475(f) (authoring securities traders to elect mark-to-market treatment for all their positions, except those identified as unrelated to the trading business).
For discussion of these so-called “net capital” rules, see Saul S. Cohen, Derivatives and Risk Management: The Challenge of Derivatives, 66 Fordham L. Rev. 747 (1997).
See, e.g., Chris Browning, Where There’s a Tax Cut, Wall Street Finds a Way, Wall St. J., Oct 21, 1997, at C1.
The reason is that many hedge funds trade frequently. Because U.S hedge funds usually are partnerships for tax purposes, they are not themselves subject to tax. Instead, their tax consequences flow through to the partners, who report the short term capital gains on their own returns. Unlike long-term capital gains, short-term capital gains are not eligible for a reduced tax rate. See Section 1(h)(applying 20% rate only to long term capital gain).
For example, assume the hedge fund is worth $100 on January 1, 2000. The dealer and client might enter into a swap, in which the dealer agrees to pay the amount by which the hedge fund’s value is above $100 on January 1, 2003, and the client agrees to pay the amount by which the fund’s value is below $100 on that date. (In addition, the client will have to make an annual
The taxpayer must hold the derivative for at least one year. In addition, precise structuring is needed. The tax objective is to ensure that the dealer, rather than the investor, is treated as the tax owner of the hedge fund interest. For a discussion, see the New York State Bar Association Tax Section, Report on Constructive Ownership, reprinted in 98 TNT 136-38
(1998). Note that, if the hedge fund is not profitable, this structure backfires by deferring losses (unless the investor settles the transaction early) and, in some cases, by converting ordinary deductions to less desirable capital losses.
Under Section 475, partnership gains and losses flow through to dealers, just as they would to any other partner. The character of these gains also “flows through,” and so it is the same as for other partners (e.g., short-term capital gain). Unlike other taxpayers, however, a dealer is deemed at year end to sell and repurchase its interest in the partnership, see Section 475(a), thereby in effect triggering any built-in gain or loss in the partnership’s portfolio that the partnership itself has not realized. Dealers treat these gains and losses as ordinary. See Section 475(d).
The dealer obviously can claim this loss even though there has been no realization, and so the counterparty is not including corresponding gain.
See Section 11 (tax rates for corporations). A dealer’s main character concern will be that gains and losses it includes as a partner in the hedge fund are likely to be capital, whereas gains and losses on the derivative are ordinary. As a result, these amounts cannot offset each other directly. See Section 1211 (capital losses may not be used to offset ordinary income). Even so, the dealer is likely to have capital and ordinary income and loss from other sources (as will members of the dealer’s consolidated group). If the hedge fund is profitable, moreover, the mismatch is favorable, i.e., capital gain (which can be used to offset capital losses from other sources) and ordinary loss.
See Section 1260. As I indicated in a report for the New York State Bar Association (“NYSBA”), the new rule has significant gaps. Taxpayers arguably can avoid it by entering into the transaction through an offshore corporation. See NYSBA, Comments on Constructive Ownership and H.R. 1703, reprinted in 1999 TNT 135-33 (July 15, 1999) (David Schizer, principal author) (noting that taxpayers arguably could avoid the statute by engaging in the constructive ownership transaction through a foreign corporation as to which a QEF election had been made). In addition, taxpayers can avoid the statute with a derivative that offers most, but not “substantially all,” of the opportunity for gain and risk of loss in the hedge fund. Finally, the rule does not apply if the tax-reducing strategy aims at something other than deferral and conversion (e.g., shielding tax-exempts from unrelated business taxable income or shielding foreigners from effectively connected income).
As noted above, the efficiency of this step depends in part on a comparison of the relative tax burdens on securities dealers and other businesses. It is assumed here that repeal of LCM is desirable.
The efficiency gains in blocking planning are more impressive when the tax being avoided is itself efficient. Here it is difficult to generalize about the tax being avoided, since dealers can serve as accommodation parties for avoiding essentially any tax (e.g., for both corporate and individual clients), although taxes on capital are more likely targets than taxes on wages, given constraints such as the capital loss and passive loss limitations.
See Sheppard, supra note 96.
The case against extending Section 475 to securities traders is similar, but arguably is stronger. While some traders take offsetting positions as part of their trading strategy, and thus could be subjected to whipwaws, many traders do not engage in such strategies, and so the remedial justification for mark-to-market has less force. Although traders may have more opportunities than dealers to engage in tax planning for their own account (since they need not accommodate client preferences), Section 475 does not prevent such planning because, in the case of traders, the provision is elective(though, once elected, it applies to all of a trader’s positions). There is now the risk, for example, that the same taxpayer will operate as an accommodation party through one partnership (which counts as a “taxpayer” and can elect to mark-to-market) while planning for her own account through another partnership (which does not make the election).
Such an approach is necessary when the transaction continues over more than one taxable year, as a derivative transaction could. The approach is not necessary, in contrast, for transactions (such as purchases or sales of the underlying) that do not involve continuing tax consequences for both parties over more than one year.
Unlike valuing the underlying, valuing an option requires assessments of volatility, interest rates, etc.
In fact, certain derivatives purchased through an organized exchange already are marked to market under Section 1256. The effect of the above proposal would be to broaden this class of transactions.
A broader response, which eliminates the client’s choice of transactional form, is to extend mark-to-market accounting to all positions for all those who are likely clients. Mark-to- market or a proxy could be applied comprehensively to all taxpayers who satisfy a particular income test, while leaving other taxpayers on realization accounting. However, I suspect that the compliance and administrative burdens of such an approach render it undesirable (as well as politically unrealistic). For instance, rules would be needed to distinguish those who mark-to- market from those who do not, including special rules for those whose income fluctuates from year to year.
The key point is that application of mark-to-market to some transactions, but not others, can create a defensive planning option, whose strength depends on how substitutable the alternatives are. To address this issue, the closest substitute (exchange-traded derivatives) must receive the same treatment. This would require a change in the law. Although Section 1256 requires mark-to-market treatment for some exchange-traded derivatives (e.g., exchange traded commodities or index futures and nonequity options), it does not currently apply to most exchange-listed equity derivatives. See Section 1256. Even so, the defensive planning option will remain strong unless another substitute – the underlying asset – is also marked to market, and the political prospects for this step are weak. Moreover, while partial extensions of mark-to-market can also offer an offensive planning option, it will be less strong than under proxies for mark-to- market (such as the assumed yield approach used for contingent debt, discussed below) because of market balance: the acceleration of losses under mark-to-market is not attractive, ex ante because the taxpayer usually cannot predict whether she will have losses or (accelerated) gains. See infra Part IV.F.3.c.
See text accompanying notes 121 to 127.
 For other tests based on marketing, see Section 1092(c)(3)(A)(iv) (positions are presumed to be offsetting, and thus part of a straddle, if “sold or marketed as offsetting positions”); Section 1258(conversion transactions “means any transaction substantially all of the taxpayer’s expected return from which is attributable to the time value of the taxpayer’s net investment in such transaction, and which is . . . marketed or sold as producing capital gains”).
T.D. 8647 (June 11, 1996).
 This example assumes that the Dow is somewhat less than $11,000 when the bond is issued.
Cash method taxpayers would account for them when paid, while accrual method taxpayers would do so when the issuer’s obligation to pay became fixed. See Section 446 (authorizing methods of tax accounting).
In the above contingent debt, holders and issuers had no tax consequences until the fifth year, when it could be determined whether the Dow was in fact worth more than $11,000. Assuming the Dow was at $16,163, for example, the holder had $5,163 of ordinary income in the fifth year and the issuer had a corresponding deduction. Even accrual method issuers could not claim a deduction prior to maturity, since the obligation was not fixed until the contingency was resolved.
Some commentators have suggested that the government was particularly concerned with the undertaxation of holders. See David P. Hariton, New Rules Bifurcating Contingent Debt – A Mistake? 51 Tax Notes 235 (1991) (Treasury is dissatisfied with wait-and-see because they have “focused entirely on the treatment of holders”). As evidence of the Treasury’s focus on holders, Hariton cites a sentence in the preamble of an early set of proposed regulations: “The intent underlying the regulations is that there should be no tax advantage afforded a contingent instrument as compared to separate instruments that, taken together, have similar economic effects.” Although holders had an advantage (deferred income), issuers were at a disadvantage (deferred deductions). Id.
Thus, the capital asset pricing model values assets based on the risk premium that must be offered above the risk-free interest rate. Likewise, the put call parity theorem observes that assets can be decomposed into a bond and a bet. For a discussion, see Warren, supra note 3.
In a discount bond, instead of paying interest each year (a so-called “coupon”), the issuer pays all the interest (a fixed amount) when the bond matures. Under the original issue discount rules of Section 1271-75, the system accounts for this interest as it accrues instead of waiting until the interest is paid. For example, assume the above bond no longer tracks the Dow but instead grows at eight percent per year to $16,163 after five years. The tax system would treat holders (and issuers) as if they had earned (or paid) eight percent – even though no interest had yet changed hands. Thus, the holder would include, and the issuer would deduct, $880 of interest income in the first year.
While market responses could cause the after-tax yields on these bonds to converge, there would still be a different mix of contingent and fixed-rate bonds than in the absence of taxes. As Professor Shaviro has observed, market adjustments remedy horizontal inequity but not inefficiency. See Shaviro, supra note 19
One set of proposed regulations, which “bifurcated” contingent debt into fixed rate debt and a derivative, is discussed below as an alternative. For a discussion of the other proposed regulations, see generally Keyes, supra note ?, at 7-15 to 7-55.
Debt in which the interest is based on the current value of the issuer’s own stock price, such as traditional convertible debt, is not governed by the new regulations. See Treas. Reg.1.1275-4(a)(4). Neither are bonds in which the contingent amount is paid annually as a coupon. See 1.1275-5 (rules for VRDIs).
See Treas. Reg. 1.1275-4(b).
See generally Shuldiner, supra note 8, at 301.
Cunningham & Schenk, supra note 10, at 744.
 Kleinbard, supra note 8, at 1352.
See Treas. Reg. 1.1275-4(b)(4).
Treas. Reg. 1.1275-4(b)(6)(ii)(consequences of so-called positive adjustment). For example, assume in the first year the bond appreciates by $1,000, instead of the assumed increase of $880. If Holder sells it in the first year for $12,000, she has $1,000 of ordinary income, instead of $880.
To be precise, the holder’s loss is ordinary to the extent of prior ordinary inclusions; any additional loss is capital. See Treas. Reg. 1.1275-4(b)(6)(iii)(consequences of so-called negative adjustment)
Any additional payment beyond the projected payment gives rise to an additional ordinary deduction. If the issuer ultimately pays less than the total dictated by the comparable yield, prior deductions are recaptured as ordinary income at maturity. See Treas. Reg. 1.1275-4(b)(6).
See Treas. Reg. 1.1275-4(b)(4).
See Treas. Reg. 1.1275-b(7).
See Treas. Reg. 1.1275-4(b)(9)(i). The issue here is that taxpayers who buy the bond for less (or more) than it is expected to be worth earn extra (or less of a) return. Instead of using the code’s general rules for this problem, the contingent debt regulations employ adjustments to the annual inclusion. According to Daniel Shefter, “practical application [of these rules] may prove very difficult in many situations and present opportunities for tax arbitrage.” Daniel Shefter, A Brief Intro to the Contingent Payment Debt Instrument Regs, 72 Tax Notes 479 (July 22, 1996).
See Treas. Reg. 1.1275-4(b)(6).
See, e.g., Treas. Reg. 1.1275-4(b)(9)(ii).
See Treas. Reg. 1.1275-4(d).
See, e.g., Treas. Reg. 1.1275-4(b)(9)(iv) (cross border transactions): Treas. Reg.1.1275-4(b)(v) (coordination with subchapter M); Treas. Reg. 1.1275-4(b)(9)(vi) (coordination with straddle rules).
Other difficult topics in the course include the straddle rules, the constructive sale and constructive ownership rules, the notional principal contract rules and the original issue discount rules, along with financial concepts such as option pricing.
In contrast to the regulations themselves, another regime enacted contemporaneously, the so-called OID integration regime of Treas. Reg. 1.1275-6, serves to reduce compliance costs. Because this regime, which is discussed infra in Section , could have been paired with any number of rules for unhedged contingent debt – including the old “wait and see” rule or bifurcation – its compliance cost benefits should not be deemed to offset the compliance costs imposed by the contingent debt rules themselves.
For instance, Professor Weisbach has observed that, in 1995, considerably more interest was deducted than was included – $692.5 billion versus $485 billion so that, for every dollar deducted, only 70cents were taxed. (The study covers all bonds, not just OID or contingent ones.) See Weisbach, supra note 184. For this reason, he concludes that revenue tends to decline as more cash flows are treated as interest.
A decline in revenue might be defended if the effective tax rate on the activity is inefficiently or inequitably high. For instance, the tax burdens on corporations – the issuers mot likely to benefit from the contingent debt rules, since their interest deductions are subject to the fewest restrictions – are thought to be inefficiently high. See, e.g., American Law Institute, Federal Income Tax Project:Integration of the Individual and Corporate Income Taxes (1993) (“[T]he current system has long been the subject of criticism, for which integration has often been offered as a solution.”). Yet other ways of reducing this tax burden are likely to be superior, such as integration or a limited deduction for dividends, on the theory that the boundary between debt, equity, and retained earnings probably induces more distortions than the boundary between fixed and contingent debt.
The temptation would be strongest for debt that closely resembled fixed rate debt but had a token contingency, such as a bond issued for $100 whose maturity payment was $132 plus a 50% chance of either 0 or $2, instead of simply paying $133. Yet such a bond arguably was already covered by the accrual regime under an anti-abuse rule. See T.D. 8518 (January 28, 1994), reprinted at 94 TNT 19-2. To the extent there was any doubt on this question, the government made the point clear through a revised anti-abuse rule promulgated with the contingent debt regulations. See Treas. Reg. 1.1275-2(g) (authorizing commissioner to “treat the contingency as if it were a separate position”). One wonders if the new anti-abuse rule would have been adequate to deter holder-based planning, without need for the contingent debt regulations themselves.
Although these structures offer less favorable treatment to the issuer than contingent debt – denial of any deduction, instead of deferral – tax-indifferent issuers would still supply them.
Sometimes the contingent debt regulations can be triggered by the unwary since they apply, in effect, to any debt instrument other than one qualifying for a listed exception. See Treas. Reg. 1.1275-4(a)(2) (listing exceptions). A fixed-rate bond may qualify, for instance, if there is a contingency as to the timing of payments that, for technical reasons, does not qualify for the regulatory exclusion for such instruments. See 1.1275-4(a)(2)(iii) (excluding “debt instrument subject to 1.1272-1(c) (a debt instrument that provides for certain contingencies) or 1.1272-1(d) (a debt instrument that provides for a fixed yield)”).
Investment bankers report that holders sometimes shy away from principal protection for a reason unrelated to tax: Especially in a volatile market, principal protection is expensive and holders must“pay” for it by accepting a very low coupon. Thus, in at least one recent transaction, issuers who wanted to issue contingent debt (for the tax benefits discussed below, see infra Part III.D.2) agreed to offer a higher coupon for the first three of the instrument’s thirty years as away to attract investors. See Paul M. Sherer, Eyes on the Prizes: Firm Hopes to Ease Capital-Gains Hit With Hybrid Security, Wall St. J., Nov. 18, 1999, at C24.
Even instruments that offered principal protection and were documented as debt – thereby giving the issuer deductions – arguably avoided the regulations as long as some (potentially small) portion of the yield was based on the value of a foreign currency. See Treas. Reg. 1.1275-4(a)(2)(iv) (excluding from contingent debt regulations an instrument subject to section 988, which deals with currency notes). The government recently plugged this hole by issuing a contingent-debt type regime for currency notes.
In recent field service advice, the government concluded that such an instrument was not debt for tax purposes, based on the lack of principal protection. See F.S.A. 199940007. As an example of such an instrument, assume that when the Dow is $11,000, the investor can buy a security that pays a coupon approximately equal to the Dow dividend yield and then will pay the Dow’s cash value at maturity –that is, if the Dow is at $14,000, the investor gets $14,000 (thereby participating in opportunity for gain) but if the Dow is at $3,000, the investor gets only $3000 (thereby taking more risk of loss than in the structure considered above). More typical are “DECS” that expose holders to full risk of loss, while offering a higher current yield but only part of the opportunity for gain (e.g., appreciation above 120% of the underlying’s current price). For a description of DECS, see David Schizer, Debt Exchangeable for Common Stock: Electivity and the Tax treatment of Issuers and Holders, Derivatives Report, March 2000, at 10.
Examples of this kind of security include the so-called “Trust Exchangeable” structures, such as the Amway Japan Trust PEPS transaction. See AJL PEPS Trust, Prospectus filed Nov. 16, 1995, available on Lexis (Edgar file). Although the issuer is a trust, as a matter of form, the trust is simply passing on to public investors a prepaid forward that it purchased from an individual investor(who is usually hedging an appreciated stock position). See Schizer, supra note 174, at 12-13 (describing trust structure). These individual issuers may be less interested in interest deductions, since they are subject to investment interest limitations. See Section 163(d) (interest may not be deducted except to the extent of interest income).
When tax-sensitive holders transact with tax-sensitive issuers (which do not want to issue preferred stock or prepaid forward contracts because these offer no interest deduction), the parties can compromise with other structures that offer more acceleration of interest than the deferral rule, but less than the contingent debt regulations — and also do not render the holder’s entire return ordinary interest instead of capital gain. These forms involve the pairing of a fixed- rate discount bond with a derivative contract, such as an option or forward, that provides the contingent payment. They are discussed in Part III.D.2, infra.
See Sherer, supra note 172, at C24 (noting difficulties of selling contingent debt to taxable investors and quoting one expert as saying “Phones-type securities [which are treated as contingent debt]don’t appeal to taxable investors”). Insurance firms and pension funds favor contingent notes because these taxpayers are often prevented by law from holding options, and sometimes required to invest a portion of their portfolios in debt. Contingent notes represent a way to circumvent these restrictions.
These individuals did not want their names or institutions disclosed.
See Jeff Strnad, The Taxation of Bonds: The Tax Trading Dimension, 81 Virg. L. Rev. 47 (1995); Mark Gergen, The Effects of Price Volatility and Strategic Trading Under Realization, Expected Return, and Retrospective Taxation, 49 Tax L. Rev.209 (1994).
See Weisbach, supra note 184 (although holders have trading opportunities under OID rules, “[i]ssuers, however, may have trading opportunities under the cash method not available under the OIDrules).
Tax-sensitive holders might also extract a tax benefit from the regulations by acquiring indirect ownership of underperforming bonds immediately before the bonds mature. Specifically, the bond might have been held by a tax exempt through a partnership. Transfer of the partnership interest to the tax-sensitive holder would not affect the partnership’s basis in the bond, see 743(a) (absent 754 election,sale of partnership affects only partners so-called “outside” basis in partnership interest, but not partnership’s so-called “inside” basis in bond), and thus would not prevent the bond from affording its new owner a deduction from the negative adjustment. Although I am not personally aware that taxpayers are using the contingent debt rules in this way, the strategy is a variation of one involving contingent installment sales that has recently received considerable attention. See, e.g., ACM Partnership, 157 F. 3d 231 (3d Cir. 1998). A recent proposal by the Clinton administration could block this strategy. See Treasury Explains Clinton Budget Revenue Proposals, 2000 TNT 27-27 (Feb. 7, 2000) (“The Administration proposes that Section 734(b) and 743(b) would be made mandatory with respect to any partner whose share of net built-in loss in partnership property is equal to the greater of $ 250,000 or ten percent of the partner’s total share of partnership assets.”).
As noted above, the issuer would usually have to pay a call premium and would incur a new underwriting fee upon reissuing the debt.
See Treas. Reg. 1.1275-4(b)(6).
Professor Weisbach has offered a similar defense of the contingent debt regulations. See David A. Weisbach, Should the Tax Law Require Current Accrual of Interest on Derivative Financial Instruments, Taxes (2000) (“[M]uch of what is known in the tax world as contingent debt is highly substitutable for fixed rate debt. Taxpayers would substitute fixed rate debt for contingent debt so that issuers would get current deductions. Therefore, raising revenue by repealing the OID rules for contingent debt would be an inefficient method of raising revenue.”).
Edward Kleinbard has emphasized this concern about prior law. See Kleinbard, supra (“We believe that many U.S. issuers, in particular, find that the unpredictability (or uneconomic results) of the tax law currently applicable to contingent debt obligations poses unacceptable tax risks.”).
Indeed, issuers do not usually place bets through contingent bonds. They usually hedge, with the hope that their cost of funding declines because they are providing bets that holders value.
 For instance, assume the issuer uses $3,514 of the $11,000 issue price to buy an option that will supply this payment. The issuer feels as if it borrowed $7,486 (since that is what is left over after paying $3,514 for the option) and will repay $11,000 at maturity (since any excess above that amount will be funded by the option).
For instance, if the issuer’s hedge is a so-called Section 1256 contract, the issuer must mark it to market and thus must include gains in the year before the bond matures, while the deduction is deferred until maturity.
For instance, assume the issuer issues a unit composed of an option that costs $3,514 and pays the excess of the Dow above $11,000, and a fixed rate bond that costs $7,486 and pays $11,000 at maturity. In the first year, the bond would generate an interest deduction of $599. As noted in table , infra, the regulations offer more generous deductions.
For instance, on an over-the-counter option, tax consequences would be deferred until the option was exercised (in the case of a cash settled option) or lapsed – and thus would match the timing and character on the holders’ options.
The legal right to separate the debt and option is the key fact in the case law, see e.g., Chock Full O’ Nuts Corp. v. United States, 453 F. 2d 300 (2d Cir. 1971); National Can Corp. v. United States, 687F.2d 1107 (7th Cir. 1982), and this feature is more likely to be respected if there is a meaningful possibility that holders will actually separate the two pieces.
Revenue would decline if some issuers were determined to issue contingent debt-type transactions to tax-exempt holders and would not use self-help because arranging separate trading was too costly for them (assuming the bonds were issued to tax-indifferent holders).
Under, Treas. Reg. 1.1275-6(a), integration applies to “a qualifying debt instrument with a hedge” if “the combined cash flows of the components are substantially equivalent to the combined cash flows on a fixed . . . rate debt instrument.” Id. If this condition is satisfied, the combined positions are taxed together, for instance, as a fixed rate debt instrument. See Treas. Reg. 1.1275-6(f) (describing tax treatment of integrated transaction).
See Edward D. Kleinbard, S. Douglas Borisky & Rekha Vemireddy, Proposed Regulations Affecting Contingent Payment Debt Obligations, 95 TNI 24-10 (1995) (“This integration rule may significantly reduce the tax diligence costs incurred by a nonfinancial institution issuer in evaluating a financing package . . . [F]or the first time, the issuer will be able to know for a certainty that its tax results will correspond precisely to the net cash flows.”).
Although the integration regime was proposed and finalized at the same time as the contingent debt regulations, the two are logically distinct. Indeed, once a taxpayer is governed by integration, the regulations do not apply. The integration rule could have been paired with other rules for contingent debt, instead of the noncontingent bond method. This issue is explored further infra Part IV.E.
In the example, under the investment unit approach, the 8% assumed yield applies only to the debt, whose value begins at $7,486, thereby generating $599. The assumed yield does not apply to the option, whose value is $3,514. If it did, the extra yield (.08* 3514, or $281) would bring the total yield to $880 (i.e., 599 + 281), which is the result given by the regulations. In other words, under the regulations the 8% return applies to the full $11,000, and not just the $7,486 attributable to the fixed rate bond.
If the security is an investment unit, these losses are deemed to arise from the option, and thus are treated as capital loss. See Section 1234 (loss from writing option is capital loss). In contrast, if the security is contingent debt, these losses are treated as interest. See 1.1275-4(b)(6)(“positive adjustment” is treated by issuer as additional ordinary interest deduction). Ordinary losses are preferable to capital losses because they are not subject to limitation, see Section 1211 (capital loss limitations), and because, in the case of individuals, they offset income that is taxed at a higher rate. See Section 1(h) (maximum tax rate on capital gain is 20%, whereas maximum ordinary income rate is 39.6%).
Theoretically, further acceleration of deductions was possible under prior law (although not as much as under the regulations), but the transactions costs would increase dramatically and the character was not as favorable. Instead of breaking the transaction in two (i.e., a note and a call option), the issuer would break the transaction in three: a note; a forward contract; and a put option (i.e., an option to sell). Assume the note began at $9,000 and accreted to $13,223 (i.e., an 8% accretion over five years). The forward would obligate the holder to pay this $13,223 for the Dow, and the Holder would have the right to put the Dow for $11,000. The pretax cash-flow would be the same as in the other structures: the issuer must return at least $11,000, plus the excess of the Dow over $11,000. The difference is that there is more growth in the debt (i.e., from $9,000 to $13,223, instead of from $7,486 to $11,000), and thus larger pre-realization deductions. The source of this difference is that more of the unit’s purchase price is allocated to the debt (since none must be allocated to a properly-priced forward, and less is allocated to a put than a call because the market is expected to rise). Yet further losses when the contingency was settled would be capital rather than ordinary. Moreover, this structure imposes much steeper transactions costs. For the components to be taxed separately, as the issuer desires, holders must be able to sell them separately. Yet the issuer will not want the holder to trade the forward without the bond – because the issuer would then have no collateral securing the holder’s performance on the forward. (E.g., if the Dow is below $13,223, how does the issuer know the holder of the forward will pay this amount if the issuer cannot simply apply proceeds from the bond?) This transaction can be done only if the holder is required to pledge valuable collateral (such as treasury bonds) upon separating the debt and forward. Providing for such substitution is quite cumbersome and expensive. To my knowledge, this structure has never been used in this context. Cf. FELINE PRIDES (providing for substitution of collateral – not to accelerate a deduction – but to get one that would otherwise never have been available, i.e., when issuers simulate issuance of tax-deductible common stock).
In tax arbitrage, the taxpayer holds two economically offsetting positions that are taxed under different rules, and thus generate a net tax benefit, such as accelerated deductions. For a discussion,see supra note ? and accompanying text.
For a description of these arbitrage transactions, known as PHONES, see infra note .
For example, assume Issuer issues a 5-year Dow linked bond for $11,000 and hedges by buying a 5 year Dow option for $3,714. If the OID integration rule applies, Issuer is taxed as if it issued a bond for $7,286 that pays $11,000 at maturity. This is the net result because any Dow- based payments owed on the bond are offset by gains on the hedge. The annual deductions, $599,$647, $699, $754, and$815, are the same as those on a bond in an investment unit, and are considerably lower than under the regulations. Compare Table 3. The disparity arises because no deduction is given for assumed growth in the embedded option.
See Treas. Reg. 1.1275-6(b)(3) (“Stock is not a financial instrument for purposes of this section,” and thus cannot be integrated with contingent debt).
See Treas. Reg. 1.1275-6(c)(2) (Commissioner may integrate if the combined cash flows are “substantially the same” as those required for taxpayer to integrate voluntarily); Treas. Reg. 1.1275-6(b)(2)(i) (taxpayers can integrate “if the combined cash flows of the [hedge and debt] permit the calculation of a yield to maturity” or would qualify as a VRDI).
For example, the issuer might buy a four-year hedge for a five year bond; the risk, of course, is that market conditions will make it expensive to fill the gap by buying a new one year option in four years. Alternatively, the issuer could issue a bond requiring it to pay the excess of the Dow over $11,000, while buying a hedge covering excess over, say, $12,500. While the legal standard is not clear,either of these structures plausibly avoids the OID integration rule.
The government took a positive step in this direction in Rev. Rul. 2000-12. It invoked the OID anti-abuse rule to integrate two bonds whose maturity dates were five years apart (but whose cash flows were otherwise mirror images of each other).
The government might be concerned that such integration could offer new planning opportunities to taxpayers or otherwise would raises technical issues, for instance, about whether the stock ceases to exist for other tax purposes. In response, the government might reserve for itself the right to integrate stock, without giving this right to taxpayers. To address collateral consequences, the government could say that such integration applies solely to defer deductions on the debt or to accelerate recognition of gain on the stock.
The main loss deferral rule, in Section 1092, applies only to “losses,” a term that the regulation defines as losses under Section 165. Interest expense, in contrast, is deducted under Section 163, and thus is excluded. The contingent debt regulations specifically designate as “loss” for straddle purposes the so called negative adjustment (i.e., an extra deduction when contingent payments are settled for more than the “comparable yield” methodology predicted). Yet the contingent debt regulations do not take the extra step of treating the comparable-yield based interest expense as straddle losses. Undercurrent law, the latter are addressed, if at all, under Section 263(g). For further discussion, see Wollman, supra note 217, at 15.
See Section 263(g).
Thus, the debt must enable or persuade the taxpayer to keep the other leg of the straddle (i.e., so the debt is “carrying” the straddle). If the taxpayer can show it would have kept the long position even without the borrowing (e.g., because legal requirements or business realities required it to do so), Section 263(g) probably is avoided. See Kleinbard & Nijenhaus, supra note 215, at 12 (under current law, Section 263(g) does not apply to DECS if the stock is not pledged as security for offering and proceeds are used in active business).
In the 1999 budget, the Clinton administration proposed to disallow interest deductions, as a per se matter, if they derive from a “structured financial transaction” that includes a leg of a straddle. See Treasury Releases Explanation of Administration’s 2000 Budget Proposals, 1999 TNT 21-36 (Feb. 1, 1999) (describing proposals to modify and clarify straddle rules). In the 2000 budget, the administration again proposed this change but formulated it slightly differently. See Treasury Explains Clinton Budget Revenue Proposals, 2000 TNT 27-26 (Feb. 9, 2000) (clarifying that debt can be a position in a straddle, and proposing to capitalize, as a per se matter, any interest on “straddle related debt”). Although the Treasury takes the position that this legislation is a mere “clarification,” the tax bar almost unanimously disagrees. See, e.g., Wollman, supra note 217, at25 (disagreeing with Treasury claim that proposals restate current law). As a result, the behavior of concern to the Treasury is continuing.
Under current law, two other rules might also be invoked to defer the issuer’s deduction, but they are less likely to succeed. First, the OID anti-abuse rule gives the government some discretion to disallow results inconsistent with the purposes of the regulations. This raises the question of what the purpose is. Arguably, it is to measure income and deductions on contingent debt more accurately (e.g., by using an assumed yield). If so, claiming deductions on unhedged debt is clearly consistent with this purpose, even if the debt is very long term, so that most of the deduction comes from the option. Deductions on a hedged debt instrument arguably pass muster as well, since the problem is not mismeasurement on the debt, but on the hedge (i.e., under the realization rule). It may be possible to construe the purpose sufficiently broadly to catch the hedging case (e.g., if the purpose was more accurate measurement on the debt, without allowing arbitrage), although the broader reading seems a stretch to me.
Likewise, the government cannot rely on the hedging rules of Treas. Reg. 1.446-4. This regime would not be implicated if the taxpayer bought the stock or option before issuing the debt because the rules govern only hedges of ordinary property or liabilities (and the debt, rather than the asset, would seem like the hedge). If instead the debt is issued first (so the asset is viewed as the hedge), the rules do not offer the government sufficient authority. Because they address timing on the hedge, rather than on the hedged item, the government could not change the treatment of the debt. It might want to require a pre-realization inclusion on the asset, based on the issuer’s comparable yield, but I doubt the hedging rules authorize such a dramatic departure from traditional accounting.
A hedge is supposed to offer benefits of selling the appreciated property (i.e., insulation from risk of loss, access to cash, diversification) without triggering the tax bill associated with a sale. For a discussion of hedging, see generally David M. Schizer, Hedging Under Section 1259, 80 Tax Notes 345 (July 20, 1998).
If the issuer has net operating losses, for instance, it will not value an interest deduction. The same holds for an individual who expects the interest deduction to be barred by the investment interest rules. See 163(d) (interest may be deducted only to extent of investment income).
DECS is an acronym for Debt Exchangeable for Common Stock, and is a service mark of Salomon Brothers Inc. Other securities firms offer similar securities under different service marks. The common feature of these securities is that they convert into the equity of another corporation that is unrelated to the issuer. Because this conversion is inevitable – it happens whether the holder chooses it or not –such instruments are known as “mandatorily exchangeable.” For a discussion of DECS, see Schizer, supra note 174, at 11-13..
See, e.g., Edward Kleinbard & Erika Nijenhaus, Everything I Know About Financial Products I learned From DECS, 412 PLI/TAX 461 (Oct. 17, 1996).
For a discussion, see Schizer, supra note 174, at 11-13.
For a description, see Lee Sheppard, Rethinking DECS and New Ways to Carve Out Debt, 1999 TNT 78-6 (Apr. 23, 1999); for the new “PRIZES” variation, see Sherer, supra note 172, at C24. For discussion of the tax treatment of PHONES, see Diana Wollman, PHONES: Wall Street’s Version of Call Waiting: What Are They and How Are They Taxed? 13 P.L.I. Tax Strategies for Corporate Acquisitions, Dispositions, Joint Ventures, Financings, Reorganizations and Restructurings (1999).
See Avital Louria Hahn, “Tech-communica-dia” Propels Convertible Issuance to New Heights, Investment Dealers Digest, Jan. 10, 2000, at 16 (“Issuance of the latest variety, called ZONES or PHONES,came to more than $5 billion.”).
See Avital Louria Hahn, Hi-Tech Issuers and New Structures Spark Convert Boom, Investment Dealers’ Digest, Dec. 13, 1999, at 10-11. See also Hahn, supra note 218, at 16 (“To be sure, 1999 will also be remembered as the year of new convertible structures: PHONES, PRIZES, ZONES, DECS and the like filled the calendars with megadeals.”).
See Kleinbard, supra note 8, at 1355.
See Shuldiner, supra note 8, at 250.
See Cunningham & Schenk, supra note 10, at 729.
[223 ] Some could remain from strategic trading, even if the rule applied universally.
See Shuldiner, supra note 8, at 250.
For a discussion of ways these instruments are used to avoid the regulations, see Part IV.F.1.
For instance, instead of buying the underlying stock, an investor may prefer a security that offers some, but not all, the opportunity for gain, along with a higher periodic payment. This is an attraction of a DECS security to holders. See Schizer, supra note 174, at 10.
In fact, the regulations are not symmetrical if the issuer integrates. In that instance, the holder’s inclusion is larger than the issuer’s deduction.
An issuer who is unable to integrate for idiosyncratic reasons will admittedly be in an unfortunate position, and may opt not to issue contingent debt. Yet this result may be the inevitable cost of keeping issuers from exploiting the accelerated deduction. Perhaps the most likely instance when an issuer will be unable to integrate is when hedging with stock. Yet deferral is not an obviously unjust or distortive result for those who hedge with stock (other than those who hedge dynamically, but they generally are dealers who mark to market). Indeed, allowing the comparable yield deduction is too generous, since the deduction is not matched by any inclusion on the stock.
Admittedly, issuers that are determined to evade the law can do so but, in my experience, many aggressive corporate taxpayers want at least a plausible claim for their position, and so anti-abuse rules that would subject them to liability do have effect in many cases, even if probabilities of detection are low. Adjustments in the penalty structure can help offset the low probability of detection,although this issue is beyond this Article’s scope.
In at least some cases, the government may be able to discern this information by comparing various information reported to it. Improved technology may make it possible for the government to match issuer and holder reports for a given security offering. While this strategy may prove promising, the costs and benefits of doing so warrant further study.
This formulation would impose deferral on the entire offering, even if only a portion (albeit a “substantial” one) was marketed to tax exempts. The rule might be refined to allow issuers to prove that alesser proportion of the offering should be affected.
The deduction is not based on the issuer’s “comparable yield,” but on the applicable federal rate. See Treas. Reg. 1.1275-4(b)(4)(B). The latter is a much lower rate, which is based on the borrowing cost of the U.S. government.
See id. (“A taxpayer may overcome this presumption only with clear and convincing evidence that the comparable yield for the debt instrument should be a specific yield . . . that is higher than the applicable federal rate.”). This way “out” of the penalty suggests that it was meant to target dishonest valuations, rather than the planning option.
Nor does the rule have to apply on an all-or-nothing basis. For instance, if an issuer files a separate prospectus to sell to foreign holders, the deduction on this offering would be affected, even if the rest of the offering is not.
For the same reason, mark-to-market eliminates the taxpayer’s timing option more effectively. See Gergen, supra note 10, at 209 (the timing option endures under an assumed yield approach).
As discussed in Part I.D.3, “market balance” is used here to mean that a tax benefit when the market moves one way (e.g., acceleration of losses) is matched by a tax cost if the market moves the other way (e.g., acceleration of gains). To the extent that taxpayers are unable to anticipate the market’s direction, they will not find the rule especially attractive, ex ante. Yet market balance is not a complete solution here because, on average, the issuers of debt (including, presumably, contingent debt) incur expense and holders have income (as compensation for the loan’s time value). Thus, issuers may still prefer a mark-to-market rule to realization, since accelerated timing is beneficial to those expecting a net deduction.
In order to report gains or losses based on changes in fair market value, the taxpayer must value the position at the beginning and end of the year. This burden is thought to render mark-to-market too administratively costly, at least for assets that are not publicly traded. See Schizer, supra note 18, at 1574.
Valuation is also a good deal easier if there is a public market for options comparable to the one embedded in the debt instrument, although there still may be differences based on the illiquidity of the embedded option (i.e., because it cannot trade separately from the debt). In many cases, though, contingent bonds have longer terms than options traded on exchanges, and so these options do not give reliable guidance.
In contrast, forwards are much easier to value, as their worth correlates more directly with the price of the underlying. See McLaughlin, supra note 37, at 82. Yet the instruments currently covered by the contingent debt rules, which are all principal-protected, contain embedded options rather than forwards (i.e., because they leave the holder the choice, but not the obligation, to earn a return based on the underlying property).
See generally, McLaughlin, supra note 37, at 82-84.
In some cases, it might be feasible to burden the issuer with offering periodic valuation of their bonds. There is a tradeoff between precision and cost, though. Hourly (or even perpetual) valuations would offer greatest accuracy, but less frequent valuations might prove more feasible.