Daniel N. Shaviro is the Wayne Perry Professor of Taxation at New York University School of Law. He is grateful to Marc Goldwein and James M. Wetzler for helpful comments on an earlier draft; to the participants at the New York City Tax Forum, where he presented this report on April 4; and to the D'Agostino-Greenberg Fund for financial support.
The Tax Reform Act of 1986 combined base broadening (such as the curtailment of tax expenditures) with a reduction in tax rates, in a manner designed to be revenue neutral and distribution neutral. It established an influential model for tax reform that continues to be cited frequently. This report argues, however, that while 1986-style tax reform was a good idea in its time, it is no longer appropriate for three main reasons. First, if tax expenditures are properly viewed as spending through the tax code, a revenue neutrality norm in which the budgetary gain from their repeal ostensibly needs to be offset by rate cuts is intellectually incoherent. Second, the long-term U.S. fiscal gap makes rate-cutting, in particular for individuals, potentially imprudent. Third, if one wants to address rising high-end income concentration in the United States since 1986, the option of raising, rather than reducing, the top marginal income tax rates may need to be squarely considered.
Copyright 2011 Daniel N. Shaviro.
All rights reserved.
Table of Contents
II. Why Base Broadening?
III. Revenue Neutrality and Budget Neutrality
A. Tax Expenditures as Disguised Spending
B. The Incoherence of Revenue Neutrality
C. Tax Expenditure Analysis Today
D. Significance of Our Fiscal Situation
IV. Distributional Neutrality and Tax Reform
A. Marginal Tax Rate Considerations
B. Distributional Changes Since 1986
C. Changes in Thinking Since 1986
The Tax Reform Act of 19861 is widely viewed as one of the great success stories of modern legislation. Cynics tried their best to undercut its appeal,2 and nearly everyone understood that the politics of enactment had required it to be quite imperfect.3 Yet TRA 1986 has remained a canonical symbol of high-minded legislative achievement.
The reasons for the 1986 act's ongoing good press are not limited to the drama of its unexpected enactment4 along with its having seemingly verified Washington's two supreme sentimental tropes concerning the merits of bipartisanship and of leaders who "get things done." Apart from that, many in the policy community, both in Washington and beyond, understood and admired the act's central animating idea that tax reform means broadening the income tax base while lowering rates so that the package as a whole is revenue neutral and distribution neutral. Tax reform did not previously mean that. Instead it was mainly about stand-alone tax preference repeal, often with the goal of increasing the tax system's progressivity. For good reason, the revised model was widely viewed as having both clarified the intellectual case for tax reform -- by putting to one side arguably distinct issues about the right level of progressivity -- and increased its practical political feasibility.
Against this background it comes as no surprise that as time passed and the grand bargain of base broadening for rate reduction slowly unraveled, people interested in improving the U.S. federal tax system began to muse about doing it all over again. Perhaps those who know history are doomed to try to repeat it, despite Heraclitus's dictum that you cannot step into the same river twice. More baffling, however, has been the great non sequitur of 1986-style reform's revival as a widely discussed policy option in response to the enormous budget deficits and long-term fiscal gap that have drawn public concern in recent years. Thus, for example, the president's fiscal commission, led by Erskine Bowles and Alan Simpson, recently proposed significant income tax rate reductions that would be financed by repealing many or all of the identified tax expenditures in the Internal Revenue Code.5 Although the proposed tax changes would still have raised significant net revenue,6 critics unsurprisingly responded by asking why an ambitious fiscal responsibility effort would apparently treat rate reduction as more important than deficit reduction.7
Just think about it: We are told (and I would agree) that the code is riddled with costly and wasteful special tax benefits, while we face grave, rising, and unsustainable fiscal shortfalls.8 And the proposed response is to accompany the elimination of those bad rules with significant rate cuts that eliminate most of the budgetary gain from repeal! In other words, massive political capital -- a scarce resource in gridlock-prone and interest-group-dominated Washington -- should be diverted from deficit reduction, the ostensible goal, and deployed instead to achieve a 1986-style tax reform that to a significant extent is merely revenue neutral. Even if tax rate reduction is unambiguously desirable, one would think more justification was needed for why this particular use of so much of the budgetary gain from repealing costly and wasteful tax preferences should prevail over competing possibilities.9
President Obama's suggestion in his 2011 State of the Union address that Congress eliminate "loopholes" in the corporate tax and "use the savings to lower the corporate tax rate"10 is subject to the same critique. Indeed, the only difference is that here one might argue there are strong independent reasons for wanting to lower the corporate rate, irrespective of the general desirability of base broadening, particularly the rising gulf between U.S. statutory tax rates for corporations and those imposed elsewhere. Starting with the goal of cutting rates and then deciding how to pay for the change is arguably different from starting with the base-broadening impetus and then (more or less gratuitously) giving away the budgetary improvement. However, the critique that we should not use up low-hanging fruit (in the sense of clearly bad existing rules) without improving our long-term fiscal situation can reasonably be offered here, no less than in other tax reform settings.
I can think of only one recent occasion on which Washington policy discourse has embraced as stark a non sequitur as that underlying the supposed link between deficit reduction and 1986-style tax reform. In 2005 the Bush administration, ostensibly in response to Social Security's long-term financing problems, proposed a package of wholesale design changes involving a shift from the existing benefit structure to a system using private accounts. The administration conceded that the changes would have no direct effect on the program's long-term solvency. Whether or not this was a good package,11 the obvious lack of any linkage between the diagnosis and the proposed cure helped to ensure its largely hostile reception. This time may prove no different.
While the biggest problem with attempting 1986-style tax reform today is that it would detract from effectively addressing our dire long-term fiscal situation, two further points are worth making. First, defining tax reform as revenue neutral rather than as overall budget neutral was never intellectually coherent -- although pretending otherwise proved convenient in 1986. Indeed, it contradicts the principal contemporary rationale for base broadening, which is that the eliminated provisions are actually "backdoor spending hidden in the tax code."12 If they really are spending, why must other tax cuts reduce the budgetary gain so that overall revenues will not commensurately rise?
The mistaken focus on revenue neutrality, as distinct from just budget neutrality (and perhaps some better metric of overall government activity), reflects a broader conceptual confusion that also can be seen, for example, in the fiscal commission's insistence on limiting revenues to 21 percent of GDP,13 and in the House's new "cut as you go" rules that exempt all tax cuts from needing to be offset budgetarily in the manner of spending increases. These naïve formulations guarantee that identical policies can either satisfy or violate the intended constraint, depending purely on form.
Second, the idea that base broadening must be packaged with rate cuts so that the overall tax reform can be distribution neutral has likewise lost practical appeal since 1986. Statutory income tax rates remain significantly lower than they were before 1986, while income concentration at the very top of the distribution has greatly increased. Against that background, a 1986-style swap, such as the fiscal commission's proposal that the top individual rate might drop as low as 23 percent,14 may reasonably be questioned by anyone who is highly concerned about income concentration. This is all the more so in light of the long-term fiscal gap, which makes significant rate reduction (presumably in response to having otherwise overshot the mark of restoring solvency) seem gratuitous. Moreover, while on the left distributional neutrality may not go far enough, on the right it has faded altogether from view as a conceivable tax reform objective. The House Republicans' 2012 budget proposal, for example, unsurprisingly makes no mention of it and emphasizes instead the view that tax policy should emphasize low rates and economic growth.15
Part II of this report discusses the changing rationales for supporting base broadening and explains why the 1980s shift from a progressivity rationale to an efficiency rationale made so much sense then. Part III discusses revenue neutrality, showing that the concept is incoherent and that the concept of budget-neutral base broadening, while at least coherent, makes considerably less sense today than it did in the 1980s. Part IV discusses distributional neutrality, with particular reference to the significance of greatly increased high-end income concentration since 1986 and of new rationales for increasing progressivity. Part V offers a brief conclusion.
Special tax preferences for particular types of income have been endemic since the earliest days of the U.S. federal income tax, and they have attracted focused criticism from tax reform proponents since at least the 1930s.16 The main critique long emphasized horizontal equity, or the belief that those with equal incomes should pay the same tax. To a considerable degree, however, the concern lay with tax preferences' effect on progressivity. Thus, a much-discussed example for many decades was that of Mrs. Anna Thompson Dodge, who inherited $59 million in 1920, invested it entirely in tax-exempt municipal bonds, and was said to have earned an average of $1.5 million per year on which she never had to pay federal income tax.17 A horizontal equity perspective might suggest that this was merely unfair to other comparably rich individuals who were paying significant taxes. In fact, however, the grievance evidently had more to do with comparing Mrs. Dodge to people who were poorer yet paying significant taxes.18
Base broadening's close connection in practice with a liberal political agenda was well known.19 For example, consider Stanley S. Surrey, the prominent Harvard law professor and Kennedy administration official who (in addition to pioneering tax expenditure analysis) was tax reform's leading proponent, in both the public and academic sectors, from the 1950s through his death in 1984. It was no secret that Surrey "felt very strongly that the tax system should be sharply progressive, and . . . regarded 'all the Mickey Mouse stuff in the Code as attenuating the progressivity of the rate structure.'"20
There was good reason for this linkage at the time. The top marginal rate for individuals exceeded 90 percent for 20 years (from 1944 through 1963) before declining to 70 percent, where it remained through 1981. Effective rates for wealthy individuals were far lower, however, thus focusing progressives' attention on bringing the two closer together. What is more, with marginal tax rates being so steeply graduated, deductions and exclusions were worth far more per dollar of taxable income to high-income than low-income taxpayers, prompting Surrey's critique that they offered "upside-down subsidies,"21 increasing along with the taxpayer's income for no evident reason.
Obviously, if the main rationale for base broadening is to increase progressivity -- although Surrey's complaints about tax expenditures were by no means limited to this issue -- any notion of combining it, 1986-style, with lower rates would be beside the point. By the late 1970s, however, the intellectual climate surrounding tax reform had begun to change. Arguments for significantly lower tax rates than those that had prevailed since the 1940s were gaining ground both politically and intellectually. Few were arguing in the late 1970s that the 70 percent top rate for individuals should head back toward its pre-1963 level in excess of 90 percent. Instead, it was lowered to 50 percent in 1981.
Meanwhile, in tax policy circles, horizontal equity was "being supplemented, even supplanted, by efficiency as the goal of economic policy"22 and as the main source of complaint about tax preferences. If Activity A received a tax preference that Activity B was denied, resources would tend to shift from B to A until their after-tax returns were equalized, thus eliminating the horizontal equity problem but leading to resource misallocation.23 The new focus on efficiency and lower rates also affected the critique of tax preferences as overly benefiting the rich. To view 70 percent, or even 50 percent, as an undesirably high rate undermined support for the earlier tax reformers' ideal of bringing the effective rate closer to the statutory rate. If statutory rates were too high, perhaps it was not all bad that effective rates were lower, even if one disliked the means (that is, inefficient tax preferences) by which the gap persisted.
Against this background, there appeared to be a compelling logic to a reconfigured version of tax reform in which base broadening would be accompanied by rate reduction to yield an overall product that was budget neutral and perhaps also distribution neutral but presumably more efficient than existing law. In December 1981 the Hall-Rabushka flat tax became the first widely discussed reform proposal to follow this pattern, although it did not purport to be distribution neutral. Sen. Bill Bradley then took the next step. As early as 1979, he had "met with his senatorial staff and said he wanted to draft a major tax-reform bill that eliminated loopholes and imposed a lower, flatter tax-rate system."24 With considerable help from Washington tax experts, Bradley gradually fleshed out this idea. Soon thereafter, having recruited a House-side cosponsor, Richard Gephardt, Bradley saw to the introduction of the Fair Tax Act of 1982,25 better known as Bradley-Gephardt. It proposed unprecedentedly comprehensive and ambitious base broadening, accompanied by across-the-board tax rate reduction (and flattening of the rate schedule) so that the top rate was only 30 percent.
Although Bradley and Gephardt were both Democrats, Republicans swiftly took notice of the new tax reform approach. Many concluded that it was both substantively appealing and potentially a political bonanza for the Democrats if they should succeed in co-opting it as distinctively their issue. Soon two prominent congressional Republicans, Rep. Jack Kemp and Sen. Robert Kasten, had introduced a generally similar comprehensive tax reform bill. Meanwhile President Reagan ordered a Treasury study of the subject, in part to head off the threat that Walter Mondale might make the new model of fundamental tax reform a central issue in the 1984 presidential election. This was conveniently timed for post-election release, ensuring that voters would not hear beforehand about the proposed repeal of popular tax preferences.
The story of how the December 1984 publication of the "Treasury I" tax reform plan led by stages to the enactment of its clearly recognizable (if watered-down) descendant in October 1986 has frequently been told26 and need not be recounted again here. However, two main points are worth noting. First, the political leaders who were shepherding it to enactment soon learned to their disappointment that "it's not that good a story," as Gephardt coldbloodedly put it as early as the summer of 1985.27 Tax reform survived -- and barely, at that, on several occasions -- only because of what the economist Henry Aaron has called the "dead cat" problem.28 Despite a lack of public enthusiasm about tax reform, any political leader who let it die on his doorstep risked being viewed not only as a failed leader but as having sold out to nefarious special interests.29 TRA 1986 therefore presented the striking irony of legislation that leaders felt politically constrained to support once the process had gathered momentum, even though voters generally greeted it with indifference at best.30
Second, while efficiency was the prime intellectual rationale for the 1986 act (at least among tax policy experts), selling it to the public depended heavily on the outrage generated by stories of millionaires and profitable corporations paying no tax.31 Politicians repeatedly emphasized this as a key aim, and it strongly affected the act's design, such as by motivating substantial reliance on ambitious and extensive alternative minimum taxes on both individuals and corporations.32 Without this impetus -- which created concern that non-enactment would not merely be a one-day story but would continue to haunt the killers of tax reform in subsequent years -- it is very plausible that the legislation would have been permitted to die at any of the many moments throughout the enactment process when its pulse almost stopped.
From the perspective of 2011, neither of these lessons from 1986 is especially encouraging. The dilemma Gephardt noted, that tax reform becomes unpopular the moment its details emerge, is if anything keener today than it was back then. Today the big money from base broadening lies far less in special business tax breaks, or in cracking down on tax shelters, than in politically popular items that are claimed by millions of taxpayers. Thus, as Howard Gleckman notes:
Tax reform is one of those issues that sounds fabulous at 30,000 feet. After all, what could be more popular than a clarion call for tax fairness and simplicity? . . . The trouble comes when you start to dig in the muck of the tax code. That's when you wrestle with issues such as what to do about the immensely popular mortgage interest deduction, or whether workers should have to pay tax on the cost of their employer-sponsored health insurance. And that's when the issue becomes highly polarizing.33
This dilemma helps to explain, for example, why the Bush administration in 2005, after convening a highly touted bipartisan tax reform panel, swiftly concluded that the panel's recommendations, although well within the contours that one might have expected from a proposed 1986-style reform,34 were best left "gathering dust on some bookshelf in the Treasury Department."35 And perhaps it helps to explain why the Obama administration has been slow in pursuing even corporate, much less individual, tax reform. Indeed, even the House Republicans' 2012 budget proposal, which advocates politically controversial changes to programs such as Medicare and Medicaid, settles in the tax reform area for stating that unspecified "large tax expenditures" should be eliminated at the same time rates are lowered.36
With efficiency having only limited mass appeal, there remains a need for a more salient sales pitch to the general public. Concern about wealthy individuals and highly profitable corporations paying zero tax lacks the high profile it had in 1986, perhaps in part because it remains less common than it was then.37 But this also reflects what I would call a shift in populist cultural style. Tea Party energies, for example, have a too reflexively anti-government flavor for not paying tax -- as distinct from corporations getting taxpayer-funded bailouts -- to become a main focus of voter outrage.
Support for base broadening does, however, potentially draw strength from two changes since 1986. First, the urgency of our long-term deficit problems has grown far greater in the last 25 years. Even in the mid-1980s many observers, such as Mondale during the 1984 campaign and Gephardt when he cold-shouldered tax reform in 1985, argued that deficit concerns made revenue-neutral tax reform an unwise divergence of effort and political will. Today the potentially calamitous projected path of public debt relative to GDP (discussed below in Part III.D) makes this argument considerably stronger. It makes a case, however, for pre-1986-style tax reform -- tax preference repeal to raise revenue that is not handed back to taxpayers via lower rates -- even though the prime motivation is now budgetary rather than increasing progressivity as with 1970s tax reformers.
Second, and potentially more favorably to the cause of 1986-style reform, the claim that the provisions being addressed are tax expenditures and hence disguised spending has greater bite today than it had in 1986. To be sure, the 1980s featured prominent political battles about the size of government, epitomized by Reagan's oft-stated critique of "tax-and-spend" Democrats. Nonetheless, the notion that spending is out of control and needs to be reined in -- whether for budgetary reasons or on general principles -- is far more central to political debate now. Thus, the House Republicans' 2012 budget proposal, while silent on the names of particular tax provisions that it would eliminate, endorses tax expenditure analysis and urges that its more general anti-government spending philosophy be applied to the provisions (albeit with simultaneous tax rate cuts).38 The question I address next, therefore, is how these sentiments affect the tax reform battlefield as a matter of political salience and policymaking logic.
A. Tax Expenditures as Disguised Spending
To explain why special tax preferences are widely, and to a large extent appropriately, viewed as tax expenditures and hence disguised spending, it is useful to start with what I and others regard as the canonical illustration, David Bradford's fictional example of the weapons supplier tax credit (WSTC).39 Bradford posits something like this: The government is planning to spend $10 billion for a particular set of weapons programs. There is broad sentiment, however, for cutting federal taxes and spending. Congress accordingly legislates as follows. First, the scheduled $10 billion weapons appropriation is canceled. Second, the companies that would have supplied the weapons are given $10 billion worth of WSTCs, which they can claim on the condition that they furnish the federal government the same weapons that would otherwise have been procured through direct spending.40
By conventional budget measures, this would indeed reduce both tax revenues and federal outlays by $10 billion. In fact, nothing significant would have changed by using the WSTC in lieu of direct appropriations.41 Both the companies and the federal budget would end up in exactly the same place as otherwise, and the federal government would end up with the same weapons. So anyone who thought that this apparent tax and spending cut had made the government meaningfully smaller would be sadly deluded.
One way to resolve the conundrum is to conclude that the WSTC is not really a tax provision at all, despite having been placed in the tax code and administered by the IRS. After all, the tax code aims to apportion tax burdens, presumably in relation to some notion such as ability to pay, and the WSTC has nothing to do with that. It is merely a disguised and indirect government purchase of items for military use.
I have argued elsewhere that while this logic is basically correct, it is better put in terms of the distinct (although frequently overlapping) distributional and allocative functions of government. Distribution is about who gets what, while allocation is about how societal resources are used. The income tax is conceived primarily as a distribution system -- it aims to assign greater burdens to those who are better off, as opposed to being intended to (although it does) discourage taxpayer decisions to work and save.42 Accordingly, smuggling into it a clearly allocative rule that is designed to equip the U.S. military with particular resources leads to public confusion and misinformation if it is equated with for example reducing income tax revenues by the same $10 billion via a reduction in marginal rates.
This, in my view at least, increases the coherence of the explanation of why the WSTC is really a tax expenditure and disguised spending. It does, however, suggest that prevailing tax expenditure lists are in some respects mistaken. For example, I would call the child tax credit a primarily distributional measure aimed at adjusting how tax burdens should relate to family size. This is clearly a distributional issue; for example, one cannot reasonably view the credit as primarily a pro-natalist policy tool designed to increase the birthrate. And even if one did so view it, this would leave the question of why personal exemptions, unlike child tax credits, ostensibly are not tax expenditures. Likewise, the earned income tax credit is in my view primarily a distributional instrument aimed at increasing low-wage workers' after-tax incomes by temporarily reducing the overall marginal tax rates they face.43
Another vexing characterization problem relates to items that rather than reflecting a clearly allocative policy, relate to the divide between income tax and consumption tax approaches to the distribution of tax burdens. Tax benefits for pension saving and IRAs, for example, could be viewed either as allocative pro-retirement saving policies (from an income tax perspective) or as the partial implementation of a consumption tax approach to the distribution of tax burdens. From a consumption tax standpoint, it is hard to view eliminating the extra tax burden that an income tax places on deferred, as compared to immediate, consumption use of one's earnings as an example of disguised spending through the tax code.
Still, for many items commonly listed as tax expenditures, the "disguised spending" framework is indeed robust and compelling. Thus, consider the deductibility of home mortgage interest. To be sure, a purist might say that the real tax preference is the exclusion of homeowners' imputed rental income from not having to pay the market rent they would charge someone else and that disallowing the home mortgage interest deduction would merely amount to partial indirect reversal of that exclusion.44 Yet, it arguably comes close enough as a shortcut to call the interest deduction the tax preference and to classify it as disguised spending that serves to subsidize home ownership.
Or consider the exclusion for employer-provided health insurance. Because provision of the benefit is indisputably compensatory, the only remotely plausible ground for arguing that the exclusion serves to help measure income properly rather than to subsidize employer provision (and employee take-up) of the benefit would relate to the old dispute regarding whether medical expenses generally should be deductible.45 However, the view that consumption of healthcare is not part of income generally has not prevailed, either in academic debate46 or, for the most part, in the code.
If we thus accept the tax expenditure view of the home mortgage interest deduction and the employer-provided health insurance exclusion, we are effectively saying that they are just like the WSTC (or at least enough like it to be similarly categorized). If a weapons supplier built a $10 billion weapon for the Pentagon and paid zero tax by using the WSTC to offset in full what would otherwise have been a $10 billion federal income tax liability, one would likely arouse little controversy by arguing that what really happened was that the company both paid $10 billion of tax and received a $10 billion appropriation -- as would have been self-evident absent the subterfuge.47 The same conclusion holds if one agrees that the home mortgage interest deduction and the exclusion for employer-provided health insurance are tax expenditures that constitute disguised spending through the tax code.
Thus, suppose I have taxable income of $150,000 and owe $40,000 of federal income tax but that (1) I deducted $20,000 of home mortgage interest and excluded $10,000 worth of health insurance, and (2) I faced a 35 percent marginal tax rate. What really happened under the disguised spending view is as follows: First, I had income of $180,000, rather than $150,000, on which I owed federal income tax of $50,500 (given the 35 percent marginal rate on the extra $30,000, adding $10,500 to the overall bill). Second, the federal government paid me $10,500 for doing things that it wanted to subsidize. Thus, taxes and spending were actually $50,500 and $10,500 respectively, rather than the simplistically observed $40,000 and zero.
In other words, just as in the WSTC example, we can recharacterize what seemingly happened in a way that better fits the underlying rationales for the relevant rules. We thereby avoid making the federal government look smaller, if one is applying the widespread (if naïve) view that government size is simply a function of tax levels or spending levels. Congress therefore cannot mislead us (and perhaps itself) simply by embedding allocative or spending policies in the tax code and thereby short-circuiting the need for reciprocal and offsetting cash flows.
While putting it this way may seem a bit stark, there really is no resisting the above conclusion if one accepts both (a) that this is the right way to view the WSTC example and (b) that the home mortgage interest deduction and employer-provided health insurance exclusion similarly are spending or allocative rules buried in the tax code, rather than normal adjustments (like deductions for business expenses) for better measuring income. Leonard Burman and Marvin Phaup have recently proposed treating tax expenditures in exactly this way for various budgetary and government accounting purposes.48
Conservatives sometimes resist this logic, such as on the grounds that tax expenditure analysis reflects the "sinister premise [that one should] think of all income as virtual state property, and forbearance to tax away every last penny of it as itself a tax expenditure."49 As the House Republicans' 2012 budget proposal makes clear, however, this is faulty thinking. The proposal does indeed note that "with spending, the government collects the money first in the form of taxes from those who earned it, and reallocates the money elsewhere," whereas "with tax expenditures, government agrees not to collect the money as long as it is put to a government-approved use."50 However, the proposal offers no reason to regard this distinction as significant, and it emphasizes instead that whether direct spending or tax expenditures are being used, the result is the same: "Instead of markets directing economic resources to their most efficient uses, the government directs resources to politically favored uses."51 It therefore urges that tax expenditures generally be repealed so that they will "stop diverting economic resources to less productive uses."52
Yet, even avowed proponents of base broadening, who seemingly agree that tax expenditures are disguised spending that should generally be eliminated, appear to forget this when they start arguing for 1986-style tax reform.53 They evidently fail to see (or expect others to fail to see) that by their own logic, the notion of revenue-neutral tax reform -- as distinct from budget-neutral tax reform, which is a coherent concept -- rests on the very conceptual error that motivates the drive for tax reform.
B. The Incoherence of Revenue Neutrality
Keeping firmly in mind the view that tax expenditures (or at least the clear cases) are merely disguised spending, let's reconsider the 1986 package of tax expenditure repeal plus fiscally equivalent rate reduction. Clearly, the package improved prior law if we posit that the repealed provisions generally were bad policy.54 However, one can always improve prior law by enacting a package that includes the repeal of bad rules, unless the rest of the package is bad enough to undo the net benefit -- a criticism that would be hard to sustain regarding reducing marginal tax rates from their high levels at the time while keeping distribution relatively constant. This still leaves the question, however, of why Congress chose this package, with rate cuts accompanying the base broadening, instead of some other (or stand-alone tax preference repeal). Three closely linked answers are germane here.
1. Argument for lower rates. Lowering tax rates, in the end from 50 to 28 percent for high-income individuals and from 46 to 34 percent for corporations, was widely recognized as an appealing policy move. While there may also have been many other appealing policy moves to consider, this one happened to involve the same congressional committees as repealing tax expenditures, and in the public's view this was on the same topic given that the disguised spending being targeted was statutorily and administratively part of federal income tax law.
Reducing tax rates without a pay-for would have made budget deficits far larger. Unfinanced rate cuts were not a plausible policy move in the prevailing mid-1980s political climate, which featured high levels of concern about deficits and a Republican Party that had not yet decisively embraced continual advocacy of unfunded tax cuts. Tax expenditure repeal was not just a very lucrative pay-for. For the reasons I discuss below (relating to revenue neutrality), tax expenditure repeal could not readily be preempted by any "funding predator" seeking to use it on behalf of some other appealing but costly policy change.55
2. Argument for budget neutrality. The above helps explain why Congress in the mid-1980s could not (or would not) simply pass income tax rate cuts without paying for them and why it was also unlikely to use tax expenditure repeal to pay for something else. Enacting tax rate cuts, if at all, only on a fully financed basis therefore made both political and policy sense. A further question, however, is why Congress did not strongly consider using at least some of the budgetary gain from tax expenditure repeal to reduce the deficit. A norm of budget neutrality meant that one was rejecting net fiscal gain, no less than net fiscal loss. Yet Congress had recently, both in 1982 and 1984, enacted major deficit-reducing legislation that increased revenues (as officially defined) in addition to reducing direct spending.
As noted above, there were indeed people who thought that tax expenditure repeal should raise revenue. The logic of their view was easy to follow. If deficit reduction is bound to be politically painful, why preemptively remove tax expenditures from the list of potential contributors to the process? This would inevitably increase the pain that Congress would have to impose elsewhere to provide the same quantum of overall budgetary improvement.
One response to this view was that the bitter medicine of repealing tax expenditures was already politically daunting even if firmly linked to the sweetener of cutting rates. Providing no offsetting benefit to current voters (a challenge that deficit reduction also separately faced) would make it harder still. A second response, however, concerned the political advantages of tabling, for purposes of considering tax reform, disagreements about exactly how deficit reduction (even if accepted as necessary) ought to be accomplished.
Democrats and Republicans largely agreed about the terms of their disagreement regarding what the federal budget should look like once it had been placed on a more sustainable path. Democrats, even if bristling at Reagan's "tax and spend" critique, avowedly wanted a bigger federal government in the fiscal realm than did the Republicans. And both parties accepted that this meant Democrats favored higher levels of taxes and spending than Republicans did. For this reason, all recent bipartisan deals to reduce current or future budget deficits had been carefully negotiated to involve a mutually acceptable balance between tax increases and spending cuts.
Taking this dispute off the table for purposes of concocting a tax reform deal that leaders in both parties could accept necessarily meant that the overall legislation should be not just budget neutral but specifically revenue neutral. That is, it had to hold constant not only budget deficits but also officially measured net revenues. Yet, in agreeing about this and mutually believing that it had some coherent relationship to their underlying philosophical dispute, Democrats and Republicans alike were guilty of the very fallacy that the term "tax expenditure" tries to address -- as I discuss next.56
3. Argument for revenue neutrality. Suppose we agree that the bipartisan pursuit of comprehensive tax reform requires tabling the long-running dispute between Democrats and Republicans concerning the preferred size of government, even though both sides recognize that the issue is unavoidable when they are taking on deficit reduction. Does revenue neutrality, via a budgetary match between the fiscal gain from repealing tax expenditures and the fiscal loss from lowering tax rates, provide this mutually desired size-of-government compromise?
The correct answer is clearly no if one accepts the tax expenditure framework. After all, that framework posits that tax expenditure provisions are actually "spending through the tax code," as the fiscal commission report puts it.57 Again, consider the fictional WSTC, supposing it to be an annual $10 billion weapons program. As we saw above, all real-world outcomes and the true set of policies that the government is following are effectively the same whether the government annually (1) swaps $10 billion checks with the weapons suppliers for income taxes on the one hand and the weapons fee on the other or (2) permits the two offsetting payments to be netted via a credit that is listed on the companies' tax returns. Either way, resource allocation and wealth distribution end up exactly the same.
Now suppose the federal government decides to stop ordering the weapons and that Democrats and Republicans agree not just that the change should be budget neutral rather than reducing the deficit, but also that it should keep the size of government the same. Further, they agree to implement this decision through 1986-style revenue neutrality under the view that net revenues (and perhaps also total spending) offer a useful proxy for the size of government. If the weapons program was funded via the WSTC, this means that tax rates should be cut. But if direct appropriations were used, other spending can presumably increase by $10 billion. In other words, two alternative but effectively identical means of funding weapons procurement lead to conflicting answers regarding how one would keep the size of government the same.
Anyone who accepts tax expenditure analysis and still thinks about revenue-neutral tax reform this way -- as, in my anecdotal experience, many people do -- has unwittingly performed a kind of psychic double entry bookkeeping. After criticizing tax expenditures as disguised spending, one swiftly reverts to thinking about them as if they were nothing of the sort but indeed the equivalent of having a lower rather than a higher tax rate.
A more logically consistent way of thinking would follow Burman-Phaup and regard not only the fictional WSTC but the actual home mortgage interest deduction and the exclusion for employer-provided health insurance as involving -- consistently with their distributional and incentive effects -- the implied collection of income taxes followed by an offsetting government outlay. The Joint Committee on Taxation recently scored the home mortgage interest deduction as costing $484.1 billion for 2010 through 2014.58 Under a Burman-Phaup approach, which seemingly implements the view of tax expenditures taken by 1986 tax reformers (among others) and in the fiscal commission report, immediate repeal of this preference would reduce spending rather than increase tax revenues.59 It would therefore be nonsensical to regard a size-of-government compromise as requiring that this spending cut be offset by reduced tax rates, thereby causing "true" tax revenues to decline.
Seen from this perspective, the 1986 tax reform was not actually revenue neutral in any meaningful sense. Rather, it was a budget-neutral reduction in both taxes and spending, as reformulated to take account of the insights offered by tax expenditure analysis. The 1986 act therefore effectively made the government smaller on both sides of the standard budgetary ledger. Of course, this does not mean Democrats should have opposed it. After all, the underlying goal presumably is to have the government do good things, rather than more things as an end in itself.
Likewise, consider again the proposal in the fiscal commission report that tax expenditures be slashed but that most of the resulting budgetary improvement be used to finance lower tax rates. This amounts to a proposal to cut spending a lot (even without regard to what the commission says about the rest of the budget) and to simultaneously cut "true" revenues a lot, albeit not quite as much given the net budgetary improvement.60
Tax expenditure analysis therefore contradicts any claim that a serious long-term budget plan should include significant tax rate cuts just so that overall revenues do not excessively increase. Of course, one could still argue (as does the House Republicans' 2012 budget proposal) that rate cuts are sufficiently desirable on policy grounds to be an important part of such a plan. Likewise, they might reasonably be part of the price of reaching a bipartisan agreement, if such a thing seems attainable. But the notion that revenues as such should be at a given level in the new budgetary equilibrium, as a proxy for the size of government or for the amount of net economic distortion that the income tax system is causing, makes absolutely no sense analytically. Despite its rhetoric about disguised spending through the tax code, the fiscal commission report appears to ignore this point when it urges that revenues be capped at 21 percent of GDP.61
C. Tax Expenditure Analysis Today
Just as zombies and vampires are "undead" rather than truly either dead or alive, and just as viruses "straddle the definition of life [lying] somewhere between supra molecular complexes and very simple biological entities,"62 so tax expenditure analysis occupies a strange middle ground in public consciousness between full acceptance and rejection. On one hand, in some respects it is more accepted than ever. Twenty years ago official tax expenditure lists were routinely dismissed as falsely presuming "that some of us should be deemed to know the answers better than others" and as making no more sense than would "substantial research into the question whether panda bears are truly bears or merely raccoons."63 Today the official lists are if anything too uncritically accepted. Thus, consider the fiscal commission report, which, while suggesting that the child tax credit and EITC might present stronger cases for retention than classic tax expenditure items such as the home mortgage interest deduction, declines to argue that they are different in kind -- even though, among these three rules, only the interest deduction resembles the WSTC in directing economic activity to a particular resource use (homeownership). This is a true difference in kind, making the home mortgage interest deduction, but not the two credits, plausibly "disguised spending through the tax code," wholly without regard to which of these three rules one likes least or most. Obviously, however, had the fiscal commission report tried to draw those distinctions (necessarily with only a brief analysis, given that its main analytical concerns lay elsewhere), it would have been accused of playing games with the official tax expenditure lists simply to advance a particular policy agenda.
At the same time, however, the framing effect that results from putting spending in the tax code (leading to WSTC-style netting of offsetting cash flows that would otherwise have been separately observable) evidently remains too powerful for even most proponents of tax expenditure repeal to resist entirely. As in 1986, when Democratic and Republican tax reform proponents alike seem to have deluded themselves that revenue neutrality meant they were taking off the table the issue whether the government should be bigger or smaller as measured by observed tax collections, no one could today win general acceptance of the claim that repealing special items in the code is in fact a spending cut rather than a tax increase.
Martin A. Sullivan recently argued that it is up to Tea Party-style foes of government spending to decide whether tax expenditures can politically be challenged in the current political environment.64 In my view, the verdict was in before the question was asked, because of a broadly shared outlook that extends far beyond Tea Party circles. Just imagine trying to explain to Tea Party activists -- or to any other voters -- that, if we repeal their home mortgage interest deductions and thus cause their April 15 tax bills to increase, we have not actually raised their taxes and indeed have made the government smaller by keeping taxes the same and cutting spending.
Accordingly, while the concept of revenue neutrality, unlike that of budget neutrality, fails to offer a meaningful description of 1986-style reform, and while this ought to be apparent to anyone who thinks rigorously about the logical implications of the widely half-accepted tax expenditure framework, we have not reached a point where substance truly prevails over form. Tax expenditure repeal likely will continue to be widely coded as a tax increase, thus preserving some of the political logic that could make it convenient to pair it with tax rate cuts. But given that one is not truly holding constant something that is meaningful via "revenue-neutral" reform, the decision to embrace it can only be defended as a deliberate tactical choice in which rate cuts are at least partly substituted for reducing the fiscal gap as a primary objective.
In some instances, the tactical choice to make a revenue-losing change and rely on tax expenditure repeal to make the overall enactment revenue neutral can reasonably be defended. Consider the Obama administration's consideration of applying the 1986 model to corporate tax reform. In an era characterized by intensifying global tax competition, by strong incentives and opportunities for multinational firms to shift taxable income out of the United States, and by socially costly debt bias in the code, the case for cutting the U.S. statutory corporate rate so that it is no longer an outlier among the rates applied in peer countries has much to recommend it. Thus, a U.S. corporate rate cut is worth considering even though, considered alone, its enactment would presumably worsen the long-term U.S. fiscal gap while also providing existing firms an arguably anomalous transitional windfall gain. Against this background, contemporaneously financing the change by repealing corporate tax expenditures might also make sense, even though it would prevent using that repeal to lower the fiscal gap.
Similar arguments may apply to other budget-worsening proposed changes (like corporate rate cuts) that can be rationalized as important structural improvements. Consider proposals either to reduce debt bias in the code, such as by providing a debt-like "allowance for corporate equity" deduction,65 or to reduce or eliminate the residual U.S. tax on resident multinationals' foreign-source active business income. Even in the face of an enormous long-term fiscal gap, any such change that one deems sufficiently desirable should presumably be enacted if possible, and the immediate budgetary impact could reasonably be addressed via simultaneous tax expenditure repeal.
The point is not that one should never consider 1986-style packages, but that they need strong justification given the overall fiscal situation. And while using tax expenditure repeal to keep a particular set of enactments budget neutral may be preferable to providing no financing, it can only do so much to ease concern about the fiscal impact. Each time one does this, less low-hanging fruit remains available to help address budgetary concerns in the future -- unless one thinks of the newly lowered rate (or other tax benefit) as itself merely a temporary sweetener, to be targeted in due course itself. This, however, would contradict the stated rationale for enactment.
What does this tell us about using tax expenditure repeal to finance a significant reduction in tax rates for individuals? Here, while the logical structure of the argument is clearly the same as that for, say, 1986-style corporate tax reform, I would argue that the bottom-line merits are very different -- at least, barring the view that individual rate cutting is merely a temporary expedient, intended to be reversed promptly. Substantially lowering individual rates would not only be very costly -- the fiscal commission report, for example, contemplated dedicating to this cause more than 80 percent of the estimated $1.1 trillion annual budgetary gain from comprehensively repealing tax expenditures66 -- but arguably lacks persuasive motivation.
However, assessing this choice requires more fully considering two issues. The first is the nature and severity of the long-term fiscal situation, which I discuss next. The second, which I address in Part IV, is the desirability, under current circumstances, of 1986-style reform that uses significant rate cuts to keep a base-broadening exercise distributionally neutral.
D. Significance of Our Fiscal Situation
Current policy debate about the U.S. fiscal situation typically focuses on near-term budget deficits. However, interest rates for newly issued Treasury debt remain extremely low, showing that the U.S. government can still readily borrow. Moreover, a few years of large deficits that merely reflected continuing aftershocks from the 2008 financial crisis would clearly be affordable. The continuing economic slowdown will at some point presumably end.
The real source of anxiety regarding the U.S. fiscal situation concerns the long-term picture, which shows our budget policy to be on an unsustainable course. The biggest contributor to this problem is the projected growth of the healthcare sector, which affects the budget through such channels as Medicare, Medicaid, and the cost of the employer-provided health insurance exclusion.67 But other questions about future budget policy, such as whether we will extend the Bush tax cuts and curtail the projected rapid growth of the AMT, can also have major effects on the projections.
The long-term fiscal sustainability problem manifests itself most visibly in the form of a projected exploding ratio between U.S. public debt and GDP. This ratio has only once reached 100 percent: briefly at the end of World War II when it was clear that we were in a position to change course with the end of combat operations. Today, under a Congressional Budget Office long-term projection based on a realistic interpretation of current policy, the ratio will again reach 100 percent by 2023, but this time without any comparable slowdown. The projection shows the ratio reaching 200 percent by 2038 and then continuing to grow exponentially -- for example, to 700 percent by the 2070s.68 The projection gets worse still -- with a 700 percent debt-to-GDP ratio emerging by the early 2040s -- if one assumes, quite plausibly, that the interest rate for U.S. government borrowing will start rising at some point as the ratio grows.69
A rising debt-to-GDP ratio can affect both our actual and expected ability to avoid government insolvency. Moreover, with a rising debt overhang, the obligation to pay interest not only grows ever more burdensome but itself becomes a key source of instability. Suppose the interest rate on federal debt rises because debtors are beginning to worry about our long-term creditworthiness. Enough of our public debt is short term70 that the increased borrowing cost on rolling over existing instruments may itself "lead investors to raise their estimate of the likelihood of a future default," requiring the interest rate to rise again (and then again). This can lead to "an upward spiral of interest rates and eventually a sudden stop in all international lending to the United States."71
The timing of such a credit market event is inherently unpredictable, since it may depend on chance events and future market actors' expectations. However, a crisis can occur long before one's debt-to-GDP ratio has risen off the charts. Some countries have experienced fiscal crises when the ratio still remained as low as 40 percent.72 But when and if a U.S. fiscal crisis comes to pass, the consequences are potentially gruesome. In a pessimistic but not implausible scenario, we may suddenly find that "no one is willing to lend to the U.S. government at any interest rate. Essentially, the Treasury holds a debt auction and no one shows up."73
What would be the next step? While it is hard to know, in one possible scenario the Federal Reserve Bank would be instructed to print money and use it to buy the bonds, thus literally giving the government continued money to spend but via Weimar Germany-style finance. This in turn might lead to hyperinflation, worldwide rejection of the dollar as an acceptable international currency (much less the world's reserve currency), and massive waves of financial failure by firms and individuals holding U.S. debt that had now lost substantial value. What is more, the U.S. government would be unable to provide fiscal backup as it did in 2008, because this time it in effect would be the party wearing the Lehman Brothers dunce cap (albeit without facing actual bankruptcy proceedings). The resulting shock to the entire world economy would likely feed back to increase the domestic economic carnage while also potentially undermining global (and perhaps even U.S.) political stability. In sum, a U.S. government credit event could make the 2008 financial crisis look like a Sunday school picnic.
Alarming as all this may sound, it is merely a hypothetical scenario. The real question is how concerned we should be that it will actually happen. In assessing this question, an important starting point is to note that the entire rising debt-to-GDP scenario relies on what I call a "statement about statements."74 That is, it involves computing expected future cash flows under a particular interpretation of current policy as applied to a set of economic and demographic assumptions.
Suppose that current policy is easy to change, with Congress simply not having gotten around to it yet. That we are seemingly headed off a fiscal cliff might then be of no greater concern than the likelihood that, if you use a GPS to drive from New Orleans to Miami, the system will for a while direct you due east, seemingly toward a rendezvous with the Atlantic Ocean hundreds of miles north of your destination. In that case, however, you know that the GPS has reasons for using an indirect route compared with how the crow flies. Moreover, you can be confident that in due course its soothing if robotic voice will indeed tell you to turn south. In the fiscal setting, by contrast, neither of the analogous certainties holds. There may be no good reason for diverging from a pre-announced sustainable budget path, and the wheel might be stuck when the need to turn it can no longer reasonably be delayed.
In this sense, it would be a mistake to call the long-term fiscal problem fundamentally technological and demographic, even if those factors are the main drivers of unsustainable health expenditure growth. The problem and its causes are well known, and even now there is still plenty of time to adjust. The question is simply whether the U.S. political system will be able to respond rationally, without a fiscal shock. I personally am skeptical on political economy grounds relating to the pervasiveness of chicken games between political antagonists, the apparent disappearance of 1980s-style centrist bargaining,75 and the likely political difficulty of cutting healthcare expenditure growth (regardless of whether it should be cut). Only time will tell if this pessimism is justified. In any event, the likelihood of a federal government credit event depends on how the political system functions over the next decade or more, not simply on the budgetary projections themselves.76
The degree of concern we should have about using up low-hanging fruit through 1986-style tax reform very much depends on the political economy question. Not only do we still have time to address long-term U.S. fiscal problems, but there are many different ways of doing so. For example, the House Republicans' 2012 budget proposal, by substantially reducing Medicare and Medicaid spending growth, could indeed significantly lower the long-term fiscal gap if future Congresses kept to the same course. Moreover, even insofar as one accepts that substantially increased revenues are needed, they need not come out of the income tax. For example, a great deal of revenue could be raised by enacting a VAT, a carbon tax, or a financial transactions tax.77
If the U.S. political system can and will respond appropriately in due course, there may be no particular reason to worry about forgoing the opportunity to use tax expenditure repeal to improve our budgetary situation. By contrast, under a more pessimistic view, it may be vital to avoid large-scale fiscal reform efforts that fail to address the risk of a long-term crisis and to avoid wasting opportunities (such as from tax expenditure repeal) for major budgetary improvement.
The second major component of 1986-style tax reform is distributional neutrality. To see how distributional thinking in relation to tax reform has changed since 1986, it is useful to start by comparing two moments that were seemingly parallel but turned out not to be. In the first, Sen. Packwood, the chair of the Senate Finance Committee in 1986, let it be known that he was radically changing course, in the middle of a committee process that seemed to have gone badly awry, by offering a plan that might permit lowering the top individual rate to as little as 25 percent.78 This attracted excited attention because it promised so dramatically to reduce the distortions and tax-planning incentives created by the existing income tax. Several leading Democrats, as well as Republicans, on the Finance Committee expressed immediate sympathy and offered strong if conditional support.
Fast-forward to late 2010 when the fiscal commission report outlined an approach whereby tax expenditure repeal could fund lowering the top individual rate to as little as 23 percent, or at the most 29 percent, while still being better than budget neutral (and more progressive than current policy).79 This failed to generate anything approaching similar excitement on the center through the left of the political spectrum, although tax rate cutting remains politically popular on the right.80 While there were obviously many other important differences between the two occasions, sharply reducing the top individual rate simply lacked the broad resonance in 2010 that it had in 1986. This in turn reflected that a lot had changed economically, politically, and intellectually between 1986 and 2010. Before discussing these changes, however, a brief background discussion of tax rate design issues may be useful.
A. Marginal Tax Rate Considerations
The choice of a top marginal tax rate depends on two distinct considerations. The first is how high rates are generally, while the second is how graduated they are as income rises. In neither dimension are lower rates necessarily better, however, even assuming sufficient tax revenues are generated to pay for public goods. Rather, assessing how high rates should be, either in general or at the top, requires trading efficiency considerations (which generally call for lower and flatter rates) and distributional considerations.
Suppose initially that all the government did with tax revenues was pay for public goods such as national defense. If efficiency were the only issue, the optimal marginal tax rate at all income levels would be zero despite the need for revenue. Rather than discouraging productive economic activity with even a low flat rate tax, the government's best course would be to levy a lump sum tax (that is, one that is entirely invariant to taxpayer decisions), such as a uniform head tax.81
As soon as one concedes that this is distributionally undesirable, one has abandoned any claim that tax rates ought to be as low as possible on efficiency grounds and instead has entered the realm of considering trade-offs between efficiency and distributional concerns. Even flat tax proponents who are intellectually rigorous enough to rule out a zero bracket (like that in the Hall-Rabushka flat tax, which is therefore actually a graduated rather than a flat tax) have effectively conceded this basic point and thus are merely "haggling over the price," as George Bernard Shaw might have put it.82 Nonetheless, taking this first step along the road to advancing distributional goals at the expense of efficiency does not require that one keep going. Leaving aside the widespread confusion about how to classify tax expenditures, there is ample room for reasonable disagreement concerning whether the general tax rate structure, in a sustainable long-term budget scenario, should be pitched relatively low or relatively high.
Part of the dispute is about efficiency. As between policy analysts, those who are more conservative tend not only to reach higher estimates of the efficiency costs of taxation but also to see market failure as less pervasive, or the government's capacity to address it effectively as more limited, than do those who are more liberal. Thus, consider healthcare, the main long-term driver of dire budget scenarios beyond the current recession. The extent to which the government ought to fund private healthcare expenditure, whether through a single-payer plan or the current hodgepodge of Medicare, Medicaid, and the employer-provided health insurance exclusion (among other elements), depends in part on whether one views the private healthcare market as at least potentially well-functioning. It also depends on how effectively one believes the government can address market failures.
However, the choice between higher and lower government spending levels -- determining how much revenue is needed over the long run -- also partly turns on distributional considerations. Suppose one has a flat rate income tax that funds uniform benefits, whether cash grants or government goods and services that have the same value to everyone. The overall effect is redistributive because higher-income individuals pay more tax to fund uniform benefits. Thus, say that Ann earns $20, Bob earns $30, Camilla earns $50, and the government taxes all three at a flat 30 percent rate (raising $30 of revenue) to fund a uniform cash grant (or alternatively to fund uniform benefits worth exactly the amount spent). Ann pays $6 of tax, Bob pays $9, and Camilla pays $15, but each gets $10 back, resulting in progressive redistribution -- at an efficiency cost because the tax affects work incentives.83 Conservative and liberal policy analysts may reasonably disagree about how to evaluate this trade-off between redistribution and inefficiency, with implications for how high tax rates generally should be.
While higher government spending (as long as the benefits skew lower on the income distribution than the financing) is one way to make the fiscal system more progressive, the better known and perhaps more obvious way to do so is by causing marginal tax rates to be (more) progressively graduated. This too will likely have efficiency costs. For example, it can induce wasteful tax planning to exploit rate differences between taxpayers, such as income shifting between the members of a family group or the emergence of clientele effects regarding the ownership of tax-favored assets.84 But increasing tax rate graduation, by reducing rates at low-income levels, while raising them at high-income levels tends to have net efficiency costs, even without interactions between taxpayers in different brackets. For one thing, deadweight loss from behavioral responses to taxation tends to rise faster than the tax rate.85 Consequently, efficiency losses at the high end should exceed efficiency gains at the low end when one thus tweaks the rate structure. For another, high-income individuals typically have greater liquidity and tax planning flexibility, as well as access to better advice, than those at low-income levels, potentially making the former more tax responsive. So the question of how to value both the efficiency costs on one side of the ledger and the progressive distributional effects on the other side naturally leads to disagreement between conservative and liberal policy analysts here as well.
One last factor that weighs against rate graduation, whether one is on the liberal or conservative side of the spectrum, is worth briefly mentioning. Suppose one could increase the marginal tax rate at either a lower point or a higher one in the income distribution -- for example, by creating a new higher bracket that applies as one's taxable income increases either (a) from $40,000 to $50,000 or (b) from $400,000 to $500,000. The former tax rate increase would affect the tax liability of everyone who earns at least $40,000. After all, even if your overall income is far higher, you will pay more in toto if the portion of your income that falls in that bracket is now taxed at a higher rate. For people earning considerably more than $50,000, however, this tax increase has no effect on marginal incentives to generate taxable income, which depend purely on the rates applying in the brackets in which they might realistically end up. By contrast, a tax bracket in the $400,000 to $500,000 range is likely to affect incentives at the margin for a significantly higher percentage of the people to whom it potentially applies. As a result of this difference, the lower-bracket rate increase has a huge efficiency advantage over the higher-bracket rate increase. For a large percentage of the taxpayers to whom it applies, it imposes an effectively infra-marginal or quasi-lump sum tax -- one they cannot escape through any realistic or potentially appealing change in their behavior.
The analytical importance of this consideration appears to have been largely (or even wholly) unappreciated until the economist James Mirrlees in a landmark article86 founded what is now called the theory of optimal income taxation (OIT). Mirrlees examined how a utilitarian or other welfarist who believes that only individuals' subjective well-being (always counted positively) matters to the assessment of social outcomes would determine what rate structure would maximize aggregate social welfare (in the context of a highly stylized model) if using specified assumptions about the declining marginal utility of income and labor supply responses to taxation. The basic goal was to pursue progressive redistribution until the point where the efficiency losses outweighed the redistributive gains. As he set out, Mirrlees expected to find that his assumptions generally would support significant rate graduation, but instead he kept finding that relatively flat rates were optimal under a variety of specifications.87 The key reason, offsetting the declining marginal utility assumption's tendency to support highly graduated rates, was that the marginal tax rates applying at lower income ranges raised revenue from higher-income taxpayers without affecting their marginal incentives.88
Subsequent OIT literature has even found that the very top rate, if one could hand-tailor it to the single highest-income taxpayer in the entire society, would actually be zero under theoretically reasonable (albeit impractical) assumptions. Joel Slemrod explains the underlying intuition as follows:
First consider an income tax schedule in which the marginal rate applicable to the highest observed income is positive. Now consider a second tax schedule which is identical to the first except that it allows the highest-earning household to pay no taxes on any excess of income over what it would have earned under the first tax schedule. When faced with the second tax schedule this household is certainly better off, works more hours, and pays no less tax than under the first schedule. All other households are at least as well off. . . . In other words, raising the marginal tax at the top above zero distorts the labor supply decision of the highest earner but raises no revenue.89
One should keep in mind that the importance of this hypothetical result is open to question. In plausible models, zero may be "a poor approximation [of the optimal rate] even for the top 0.1 percent."90 At the same time, however, "economics has some difficulty in making a theoretical case for increasing marginal tax rates,"91 by reason of the point that tax rates further down the distribution are effectively inframarginal for a higher percentage of the taxpayers who are subject to them.
The discussion in this section has concerned the trade-off between efficiency and distributional concerns that may underlie determining how high the top marginal tax rate -- reflecting both overall tax levels and rate graduation -- ought to be. Yet, isn't this irrelevant to the merits of 1986-style tax reform in which lowering and flattening the rate structure is merely part of a larger, distribution-neutral package? Arguably, the answer to this question depends in part on what one thinks about the current distribution of tax burdens. As discussed next, for many observers, it is likely that changed circumstances since 1986 greatly reduce the appeal of taking current distribution as a given, and they instead encourage renewed consideration of the Stanley Surrey-era tax reform idea of using base broadening to narrow the gap between preexisting statutory rates and actual effective rates.
B. Distributional Changes Since 1986
To address how circumstances have changed, it is useful to start by briefly reviewing the key findings in important recent empirical work by Thomas Piketty and Emmanuel Saez regarding income inequality in the United States during the 20th century,92 periodically updated in recent years by Saez.93 Piketty and Saez find that the top decile in the U.S. income distribution typically earned 40 to 50 percent of all households' income from the 1920s through World War II, varying from year to year and with the details of the measure used (an issue that I discuss further below). With World War II, however, the top decile's percentage of overall income experienced a huge negative shock, declining to below 35 percent and then remaining flat through the early 1970s, when it began to increase again. In other words, the United States from the 1940s through this time was substantially less unequal on the high end than it had previously been. This period of relative income equality lasted long enough that influential economic theories emerged depicting it as an inevitable byproduct of economic advancement.94
By the late 1970s this trend had reversed, and the top decile's income share had begun rising again. By the time TRA 1986 was being debated, the top decile's income share was approaching 40 percent, but the trend had not yet attracted enormous attention, nor were there yet obvious grounds for concluding that we were in a fundamentally new period. In the years since 1986, however, the trend toward greater income concentration at the top has continued and even accentuated, and the top decile's income has returned to the approximately 50 percent level that had been unknown since the 1920s (before the Great Depression). Moreover, while the 2008 financial crisis has had a minor negative impact on those households' income share, Saez concludes from the most recent available data (2008) that "the Great Recession is unlikely to have a very large impact on top income shares and will certainly not undo much of the dramatic increase in top income shares that has taken place since the 1970s."95
Table 1. Top Decile (and Above) Income Shares Without Capital Gains
Top 10% Top 5% Top 1% Top 0.1% Top 0.01%
Income Income Income Income Income
Year Share Share Share Share Share
1986 34.57 22.59 9.13 2.87 1.0
2008 45.6 33.36 17.67 7.77 3.34
Percent increase in
share since 1986 31.9 47.7 93.5 170.7 234
Source: Facundo Alvaredo, Tony Atkinson, Thomas Piketty, and Emmanuel Saez,
"The Top Incomes Database" (data for United States), available at
Table 2. Top Decile (and Above) Income Shares With Capital Gains
Top 10% Top 5% Top 1% Top 0.1% Top 0.01%
Income Income Income Income Income
Year Share Share Share Share Share
1986 40.63 29.49 15.92 7.4 3.34
2008 48.23 36.52 20.95 10.4 5.03
Percent increase in
share since 1986 18.7 23.8 31.6 40.5 50.6
Source: Facundo Alvaredo, Tony Atkinson, Thomas Piketty, and Emmanuel Saez,
"The Top Incomes Database" (data for United States), available at
Although the modern trend toward greater high-end income inequality had already begun to manifest itself during the period when TRA 1986 was being debated, it has gone considerably further since then. Moreover, the changes since 1986 appear even more striking when one looks at how income shares at the top have increased for narrower slices than the entire top decile. Table 1, derived from Saez's most recent online updating of available data, provides a snapshot comparison.
Since Table 1 excludes capital gains (although not other income from capital, such as interest and dividends), it largely reflects changes in wage and salary income. While this is informative in itself, the particular rationale for excluding capital gains is the classic bunching problem in income measurement. Suppose I hold an asset for 20 years and then sell it for a $1 million capital gain. If this is unusual for me (that is, I don't have many other such assets that I also regularly sell), including the capital gain in the year of sale would give a distorted picture of my actual long-term situation.96 The preferred answer might be to include unrealized appreciation in the annual income measure while ignoring capital gains realizations as such. The data for that approach are unavailable, however. Table 2 therefore offers the same comparison of 1986 to 2008 as does Table 1, but this time with capital gains being included.
One should keep in mind that in one key respect, including capital gains may not go far enough to capture true accrued economic income at the high end. Consider founders of hugely profitable companies who are made rich by the stock appreciation but can live well by borrowing against the stock value and thus have no need to either pay themselves high salaries or sell their stock before death. Table 2, no less than Table 1, may miss their contribution to the rising high-end concentration of economic income.97
Despite these concerns, Table 2, considered in light of Table 1, supports two main conclusions. First, the basic story of dramatically increasing income concentration at the top since 1986 remains clear, although with capital gains included it does not appear quite as extreme as in Table 1. Second, the trend's being less extreme under the expanded measure helps to illustrate (subject to the founders' point) the key role played by sharply rising wages and salaries at the top.
As Saez notes, "the income composition pattern at the very top has changed considerably over the century [with] an explosion of top wages and salaries. . . . Top income earners today are not 'rentiers' deriving their incomes from past wealth but rather are 'working rich,' highly paid employees or new entrepreneurs who have not yet accumulated fortunes comparable to those accumulated during the Gilded Age."98 As it happens, however, the "dramatic increase in top wage incomes has not been mitigated by an increase in mobility at the top of the wage distribution. . . . [Instead], the probability of staying in the top 1 percent wage income group from one year to the next has remained remarkably stable since the 1970s."99
Rising income concentration at the very top has already visibly changed the terms of public debate about distribution issues. This was tellingly illustrated by the political battles waged during most of 2010 over whether the expiring Bush tax cuts, which both parties agreed should generally be extended for low- and middle-income taxpayers, should also be extended for the top 2.5 percent of individual filers, that is, those earning $250,000 or more (in the case of a joint return). President Obama argued that this group's tax rate should be permitted to rise from the expiring 33 and 35 percent levels to the pre-001 top rates of 36 and 39.6 percent. He often posed the issue as whether, in light of the long-term budget situation, we should extend "tax cuts for millionaires and billionaires."100
Many responded, however, that if the goal was to permit tax rates to increase solely for the rich, $250,000 was too low. This in turn reflected that even within the top 2.5 percent, those at the very top have pulled far away from those who are closer to the proposed cutoff point. Accordingly, there was widespread discussion of having tax rates increase only for those earning more than $1 million.101 No similar concern had been widely raised during the 1986 process, when the use of two top income groups to measure distributional neutrality -- those earning from $100,000 to $200,000, and those earning $200,000 or more -- had been widely accepted without any evident concern about the difference between those who were barely in the top group and those who were millionaires or the super-rich.102
The newly felt distance between those merely in the top 2 to 3 percent in the income distribution and those at the very top manifested itself especially notably in the widely reported controversy over a blog post by University of Chicago law professor Todd Henderson103 -- worth briefly reviewing here because it so naïvely yet straightforwardly captures broader attitudes. Henderson began his critique of the Obama administration's support for permitting the Bush tax cuts to expire for those earning more than $250,000104 by positing that the president must view the Henderson family's "lifestyle [as] capable of financing the vast expansion of government he is planning."105
Henderson then went on to explain that even a family such as his with two high earners (his wife is a doctor) simply "can't afford" a tax increase, as it is barely getting by. However, his definition of barely getting by includes owning a private home that generates $15,000 in annual property taxes, sending three children to private school, saving and investing for retirement, and hiring both a housekeeper and a lawn service. Henderson predicted that if the top tax rate increases from 35 to 39.6 percent as urged by the Obama administration, he would find it necessary not only to fire the housekeeper and lawn service (and to "cancel our cell phones and some cable channels, as well as take our daughter from her art class at the community art center"), but also perhaps even "sell our house -- into an already spiraling market of declining asset values -- and our cars, assuming someone will buy them." Given those parlous circumstances, he urged that the president's plan to increase high-end taxes be confined to the true "super rich" -- except that they "don't pay taxes -- they hide in the Cayman Islands or use fancy investment vehicles to shelter their income."
To place this heart-rending story in perspective, in 2007 the median cash income for an adult who worked full time was $41,425106 -- perhaps around 10 percent of the Hendersons' joint total. Yet Henderson's sense of his relative position is one that many readers who are reasonably well-paid professionals may find entirely recognizable emotionally -- as do I -- even if they are less stunningly oblivious to the broader world than he is (or at least less indiscreetly oblivious). The changes in high-end wealth distribution over the last 25 years, along with the fact that those near the top often are keenly aware of circumstances at the very top, have for some time made it extremely common for high earners in the Hendersons' range "not [to] consider themselves rich because they compared themselves to the statistically small segment of the people who earned more than them, rather than the much larger segment who made less."107
Consider a recent survey of more than 1,000 millionaires (defined as people with at least $1 million in investable assets, excluding real estate and retirement accounts). Of the people surveyed -- whose mean holdings of investable assets totaled $3.5 million -- 42 percent responded that they do not feel wealthy. An official at Fidelity Institutional Wealth Services explained: "They compare themselves to their peer group . . . and they are also thinking about the long period they will have in retirement and want more assets" to fund their preferred lifestyles. Wealth, for those individuals, apparently is defined as holding at least $7.5 million in investable assets -- again, excluding real estate and retirement accounts.108
Even if the changes in high-end distribution since 1986 made no substantive policy difference in choosing an appropriate top rate, they might affect what distributional neutrality means to people who care about it. Thus, in 2011, unlike 1986, a much more fine-grained analysis may be necessary in assessing whether an agreed-to distributional neutrality constraint is satisfied by a package that combines base broadening with rate cuts. It is therefore unsurprising that the fiscal commission report examined how an illustrative tax reform proposal would affect after-tax income by members of the following high-income groups: the 80th to 90th percentile, 90th to 95th, 95th to 99th, the top 1 percent, and the top 0.1 percent.109
Making this type of change in the analysis potentially complicates achieving the stipulated distributional neutrality. For tax preferences the usage of which is likely to grow more slowly than income at the high end (for example, the home mortgage interest deduction110), a trade-off between base broadening and rate reduction that achieves distributional neutrality at, say, the 98th percentile will presumably fail to do so at the 99.9th percentile. Thus, achieving it at the very top may require stringently addressing items that do not exhibit this pattern.
As the fiscal commission report shows, this is indeed potentially feasible. Its illustrative distributional analysis of proposed 1986-style tax changes found a greater negative percentage change in after-tax income for the top 0.1 percent of the income distribution than any other group. However, this reflected the proposal's including such stringent and potentially controversial changes as converting the deductions for home mortgage interest and charitable contributions into 12 percent tax credits and eliminating the lower rates for capital gains and dividends.111 The proposed dividend change, of course, is not necessarily true base broadening, given the entity-level tax on corporate income. And insofar as the tax rate increases for capital gains and dividends are being relied on to achieve distributional neutrality (or better) for the top 0.1 percent, the impact may be uneven, with the "working rich" whom Saez identifies, such as highly compensated bankers, potentially being big winners.
Even if one can enact enough changes with comparatively great high-end impact to achieve the requisite distributional neutrality at the very top, developments since 1986 may affect whether that is enough, as compared with more definitively using base broadening to address rising high-end income concentration. I therefore next address how distributional thinking has changed since 1986, including changes by reason of the observed trends.
C. Changes in Thinking Since 1986
Even if high-end distribution had not changed since 1986, the fact that the top federal income tax rate for individuals is now 35 percent (or 39.6 percent if the Bush tax cuts are finally allowed to expire) rather than 50 percent, as it was back then, might reduce enthusiasm for rate cuts. The economic distortion caused by the lower rate is presumably significantly less.
However, the OIT literature, with its explication of the rationale for flat or even declining rates at the very top (to address inefficiency that is not offset by increased redistribution) is better known today than it was in 1986, at least in legal tax policy circles. Although Mirrlees's landmark article appeared in 1971, I cannot recall ever hearing about it during the 1986 tax reform debate (which I followed closely as a JCT staffer), and indeed I know of no reference to it in the legal literature before the publication of a now-celebrated article by Joseph Bankman and Thomas Griffith in December 1987.112
The OIT critique of high rates at the very top would be even more on point if a new top bracket were proposed for income at, say, the $1 million level. Compared with the top rate under current law, which (for 2011) starts at income of $379,150, this one would be inframarginal a smaller proportion of the time. Nonetheless, under 2011 circumstances a powerful case can be made for a relatively high top rate. And even if this ideally would involve a relatively flat rate structure, set high enough to achieve progressivity through the use of redistributive tax-plus-spending packages, high-end rate graduation is potentially a fallback if the preferred approach is politically unavailable.113
1. The direct case for a highly progressive rate structure. Saez, who in addition to studying rising income inequality has written extensively in the OIT literature,114 recently proposed a plan for comprehensive income tax reform that was explicitly grounded in optimal tax theory.115 Under his plan, the tax rate would rise to 45 percent for adults in the top 1 percent of the income distribution (earning $280,000 or more) and then to 60 percent for those in the top 0.1 percent (earning $1.325 million or more).116 The proposal thereby departs significantly from the norm in OIT analysis.
In support of its consistency with OIT theory, Saez notes the theory's normative implications for tax rates if we assume that as income keeps rising, the social marginal utility of a dollar converges to zero.117 Once one literally reaches the zero point, extra dollars earned at the very top are (by definition) worth nothing socially. This causes the optimal tax rate at the top to be whatever would raise the most revenue. In effect, the type of reasoning that underlies the hypothetical case for a zero top rate gets turned in a very different direction. Rather than saying of the people at the very top, "Why charge them anything further, if the people below won't gain anything?" we say, "Why not get the very most we can from them, since, even if they respond by earning and keeping less, this has no social welfare cost?" These two statements are in fact entirely consistent logically,118 even though they seem to push in different directions.
However, Saez's analysis raises the question of how one could ever reach the zero point for the social value of high-end income in a welfarist analysis or one in which individuals' subjective well-being (always counted positively) is all that matters to the assessment of social outcomes. One can almost get there by applying a highly egalitarian social welfare function in which further increasing well-being at the top verges on having zero weight. To get all the way to the zero point, however, one must assume complete satiation or that high-end earners derive no utility from extra income once they have ascended high enough. This is plausible in a world where not everything of value (for example, perpetual life or other people's love) can in fact be bought. It does, however, violate the standard non-satiation assumption of welfare economics, which holds that more is always better.119 And it raises the question of why people at the high end seek higher salaries -- although, as we will see shortly, there is in fact a possible explanation rooted in their competing for status via pretax salary levels.
Other design goals that underlie Saez's 60 percent plan seem less reconcilable with a welfarist framework. He argued that unspecified principles of fairness dictate that "gains from economic growth should be distributed evenly across income groups,"120 thus keeping the degree of income inequality constant (or at least non-rising) over time. He noted that while one might want to apply this principle to all economic gains since the 1970s, using progressive income taxation to achieve this would require a top tax rate exceeding 80 percent, which might be too high even under a revenue-maximizing principle. Saez thus instead advocated pursuing the more modest goal of achieving equal distribution of all economic gains since 1989, just after the end of the Reagan administration and at a point when income inequality had already risen significantly.121
This approach appears to depart from treating only individuals' subjective well-being -- as opposed to comparing the present with the past -- as relevant to the assessment of social outcomes. Under a pure welfarist approach, income distribution in 1970 or 1989 seemingly would have no bearing on what it should be in 2011 (given efficiency-equity trade-offs) unless people's memory of past circumstances affected their current well-being or behavior in ways that Saez does not specify. However, one reason for adopting a benchmark based on past income distribution might be the view that extreme income concentration at the top, of the sort that has been reemerging in the United States over the last 40 years, has important tax policy implications that standard economic analysis often ignores. I therefore next examine the other reasons the current U.S. level of high-end income concentration may be thought to support imposing higher or more graduated rates than one might favor under different circumstances.
2. 'Nonstandard' assumptions that may support higher rates, whether or not more graduated. What I will call standard economic models, including those most commonly (although not universally) used in the OIT literature, assume that people derive utility solely from their own consumption of market goods and leisure.122 Those models also take people's preferences as given, not as being affected by others' observed circumstances. Finally, standard models assume that the economy is functioning efficiently (the effects of the income tax aside) and thus that the only reason for a non-lump-sum tax is to promote distributional goals at the price of creating some deadweight loss. However, rising high-end inequality over the last 25 years has encouraged new work (and renewed appreciation of older work) questioning each of the three standard assumptions, with potential implications for the conclusion that low or flat tax rates are generally best. The following briefly reviews some of the main challenges to each assumption.
a. Utility from one's own consumption only. We are a highly social species in which status competition plays a vital role, surely reflecting evolutionary origins in which one's relative status in the group could powerfully affect one's access to scarce resources and rate of gene transmission. It therefore comes as no surprise that extensive social science research shows that "relative material payoffs affect people's well-being and behavior."123 Research subjects "exhibit a strong and robust aversion against disadvantageous inequality" in which they are worse off than someone else, and some "also exhibit an aversion to advantageous inequality although this effect seems to be significantly weaker."124 The literature sometimes calls these two preferences envy and altruism, respectively,125 although to my mind, the former term has an unduly dismissive "ripped from the opinion pages of The Wall Street Journal" quality that ignores how universal and fundamental it may be.
Concern for relative status implies that, when some people in a given reference group become materially better off, this may directly make others worse off. A key question, of course, is who is in one's reference group. One can try to choose what economist Robert Frank calls the "right pond," where one is insulated from too much awareness of those who are doing significantly better.126 However, as the examples of Henderson and the millionaires who do not feel rich may help to show, this is not always so easy. Consider the experience that many urban professionals may have when going to an expensive restaurant, queuing up for an airplane flight when one is not in the first or business class cabins, or sending one's children to a private school where there are more affluent families.
What is more, the emotion need not be mere peevishness in the sense of Gore Vidal's notorious remark, "It is not enough to succeed; others must fail." People above one in the social ranking may be all too eager to make sure one notices their higher position. Nor is envy just an evolutionary goad to keep one fighting for a place at the top. It may also reflect awareness that one's welfare can end up being more tangibly reduced if others' resources so far exceed one's own that their voice in collective societal decision-making is much stronger. Consider the possible political economy implications of a radically unequal U.S. society in which the super-rich might exercise an iron grip on feasible policy choices through their superior ability to influence politicians.127 One may reasonably fear that as economic elites pull away from the rest of society, they will find themselves increasingly able to demand policies that entrench and privilege their position and that permit them to extract resources from others. Indeed, there is evidence consistent with this fear. Important recent political science research, initially by Martin Gilens128 but expanded on in a recent book by Jacob Hacker and Paul Pierson,129 found that both low- and middle-income individuals' support for a policy change had virtually no impact on its chance of being enacted. "By contrast, when more of the well-off supported a change, it was substantially more likely to happen."130
Even if one does not endorse this entire story, the relevance of relative material payoffs contradicts the standard economic assumption that rising earnings at the top (or anywhere else) make the recipients better off and no one worse off. This can have significant effects on the optimal trade-off between efficiency and distributional concerns that underlies the choice of tax rate structure. Frank, for example, posits that most people would prefer World B, in which "you earn $100,000 per year [and] others earn $85,000," over World A, in which "you earn $110,000 per year [and] others earn $200,000."131 While this does not itself make the case for World B -- given the hypothesized others' presumably opposite preference -- it may readily translate to a setting in which higher tax rates (whether or not more graduated) that fund uniform benefits, leading to reduced earnings across the spectrum by reason of the substitution effect away from productive work, nonetheless yield aggregate welfare gains because after-tax income is more evenly distributed. In fact, OIT models that incorporate utility functions in which people care about relative position generally find that this supports imposing higher, even if not more graduated, rates.132
A further possible implication of relative status concerns goes to the labor supply effects of taxing high earners. In a standard economic model, people work to finance market consumption. If they are taxed more, they will tend at the margin to work less, substituting toward leisure, although the income effect of having less after-tax income can push in the other direction. Many high earners, however -- for example, corporate executives and star athletes -- appear to place high value on getting paid as much as possible, even if it seems unlikely that they will get to spend the extra money. One explanation is that they are competing with their peers over relative status.133 Insofar as this is the operative motivation, increasing the tax rate that high earners face as a group may not discourage high-end labor supply (a key reason for favoring flat or falling tax rates at the top), since all are affected in tandem and their relative positions, as measured by the salaries they command, remain the same.
b. Preferences as given, rather than as affected by others' observed circumstances. The relative status concerns discussed above operate at the macro level by causing one to care about how well one is doing overall compared with other people. However, positional concerns also operate at the micro level by affecting the utility that people derive from particular consumer goods. Here the literature goes back at least to Thorstein Veblen's classic Gilded Age analysis of "conspicuous consumption."134 In 1976 economist Fred Hirsch coined the term "positional goods," or goods whose observed consumption levels affect social rankings,135 thus causing one's utility to depend on their relative consumption.136 In recent years, Frank has prominently argued that positional goods impose negative externalities, no less than pollution, because one person's use of them makes others worse off.137
Indeed, as Frank noted, this can happen even when people think they just want the normal and accepted things for themselves, as opposed to being passionately committed to the competitive rat race. "When top earners build larger mansions, for example, they shift the frame of reference that defines an acceptable house for those just slightly below them on the income scale. And when those people respond by building bigger houses, they in turn shift the frame of reference for those just below them, and so on, all the way down."138
The analogy to pollution reflects that if I build a large house, it may make others need to incur the cost of acquiring larger houses simply to recover the status level and degree of contentment that they had before. To be sure, this leaves the lasting material benefit of now living in a larger house, albeit financed by a shift of budgetary resources that one previously had preferred to use differently. Even this benefit, however, may lose subjective value over time because of habituation (although this is a different problem, which could arise even in a one-person society).
If positional goods create negative externalities, there is a standard Pigovian argument for taxing them, just as with pollution.139 While the exact cost imposed -- and thus the optimal level of the tax -- is hard to determine, this does not mean that the tax imposed should be zero. The question of which consumer goods to classify as positional and thus to subject to the Pigovian tax has long been considered more daunting, however. Real-world luxury taxes, which in the spirit of Veblen can easily be thought of this way, inevitably miss substitute positional goods that consumers can use to avoid the tax without altering the costs imposed on others.
Frank argues, however, that there is no need to make such fine-tuned determinations of relative positionality, because market goods are generally much more positional than leisure. Thus, he asserts that while people would commonly prefer (i) a 3,000-square-foot house when everyone else has 2,000 square feet to (ii) a 4,000-square-foot house when everyone else has 6,000, they generally do not think this way about vacation time. Four weeks' vacation when everyone else has six weeks is still preferable to two weeks' time when everyone else has only one week, if the main focus is simply on experiencing, say, beach or spa time in lieu of office time.140 Thus, in his view, a consumption tax or an earnings tax can increase efficiency rather than simply serving to raise revenue in a distributionally preferred manner at the cost of inefficiency.141
c. No efficiency rationale for taxing earnings. Optimal income tax models confirm that if market consumption is classified as status-seeking in the above sense, the resulting affirmative efficiency rationale for taxing earnings, while not eliminating the traditional concern about distorting labor versus leisure choices, provides support for higher marginal tax rates, although not necessarily for greater rate graduation.142 Even without greater rate graduation, however, the resulting fiscal system is more progressive, because taxes that rise with earnings can fund benefits that are enjoyed more uniformly.143
A further possible efficiency rationale for taxing earnings that has generated discussion in recent years concerns the view that labor markets increasingly have a winner-take-all or tournament character in which a few winning participants receive huge payoffs while all others in effect draw losing tickets.144 Among a group of contestants, therefore, the one who either is just slightly more able or works slightly harder than all the others may end up receiving a hugely disproportionate payoff. The phenomenon clearly appears to help explain the huge returns now enjoyed in some fields, such as by entertainers, star athletes, and corporate executives, although its broader applicability is disputed.145
A view of labor markets as increasingly tournament-like can lend support to greater tax progressivity on familiar distributional grounds. It may suggest, for example, that ability dispersion in the labor market has effectively increased or that risk-averse lottery participants would benefit ex ante from a tax on the winners to offer compensation to losers.146 Robert Frank and Philip Cook, however, argue that the tournament model also supports additional efficiency arguments for progressive taxation.147
One of their contentions is that high-return labor market tournaments draw too many contestants from a social standpoint. Suppose I am choosing between (1) entering a winner-take-all tournament in which I will earn either a huge return or nothing and (2) finding a distinctive economic niche in which my income is certain and does not reflect competition with anyone else. In the former case but not (at least by hypothesis) the latter, there is an externality problem that I have reason to ignore, insofar as my entering the tournament makes others less likely to win.148 The problem of having too many contestants grows even worse if people are systematically subject to overconfidence regarding their own abilities.149 But imposing high taxes on tournament winners may reduce over-entry.
A second efficiency problem that may result from labor market tournaments is the incentive to over-invest in marginally increasing one's productivity, if doing so is costly and yet potentially generates a disproportionate reward. That is, by improving just slightly, one potentially wins the entire prize instead of nothing. The disproportion may lead to costly "arm's races" between contestants who cannot agree to cooperate by mutually forgoing investment that would otherwise have a negative payoff. The issue presented by the classic example of professional athletes taking harmful steroids in the quest for a competitive edge150 arises as well when people work overly long hours because their competitors are doing so. When more direct means of addressing these problems (such as a steroid ban) are unavailable, taxing the winners once again may help address the incentive problem.
The question of how best to evaluate these arguments -- in common with many others that have been discussed in this section -- remains very definitely unresolved. Moreover, if tournaments reflect imperfectly competitive labor markets, it may be hard to predict the distributional impact of a higher tax on the winners. A recent OIT paper on tournaments suggested the surprising possibility that, if they reflect monopsonistic labor markets (in which high-skill individuals wield significant market power), a redistributive tax might conceivably end up increasing, rather than reducing, after-tax inequality.151
Nonetheless, the traditional economic efficiency case for keeping marginal tax rates as low as possible clearly has been seriously challenged in recent years, potentially supporting greater redistribution than might have been viewed as optimal during the 1980s. Those who view any of the above-described arguments as persuasive, or who simply are concerned about the radical increase in high-end income inequality since the late 1970s, may reasonably conclude that Stanley Surrey-style, rather than 1986-style, base broadening is now in order. From that perspective, addressing the gap between current (or even increased) statutory rates and actual effective rates may have more appeal than attempting a distribution-neutral 1986-style trade all over again.
The startling success of the 1986 tax reform process continues to transfix modern-day proponents of income tax base broadening. However fortuitous and indeed unlikely the ultimate passage of TRA 1986 may have been -- and however prescient, from a coldblooded political standpoint, Gephardt was when he concluded early on that it wasn't "that good a story" -- it would have had no chance whatsoever if not for two important intellectual departures from the previously dominant tax reform model.
The first departure was to frame base broadening as part of a revenue-neutral process rather than using the revenues to reduce budget deficits. While tax expenditure analysis shows that the revenue-neutral frame is intellectually meaningless, and while the associated decision to make tax reform budget neutral could reasonably be questioned -- even in the 1980s, when the long-term U.S. fiscal situation was far less dire than it is today -- the decision to proceed along these lines clearly had political benefits in making consensus on distinct issues possible.
The second departure was to make tax reform distribution neutral (at least as officially measured) rather than treating tax preference repeal as an occasion to increase progressivity. This addressed intellectual confusion between distinct issues and permitted prospective tax reformers who disagreed about progressivity to work together on tax base issues when their views were compatible.
Today, however, the merits of 1986-style tax reform look very different. An extreme uphill slog politically even in the best of times, it also would represent a costly diversion of effort from addressing more pressing problems. Our dangerous long-term budget situation makes lowering tax rates almost a frivolity under current circumstances, even if in principle desirable. As long as the fiscal situation remains bleak, a precondition for seriously considering any rate cutting should be that there are strong independent arguments for the change (unrelated to the revenue neutrality myth).152 Also, developments over the last 25 years have strengthened the case for seeking to increase the tax system's progressivity and for questioning the magnitude of the efficiency payoff from statutory rate reduction.
One possible strategy that I have not addressed herein is that of enacting merely temporary rate cuts (in political fact, whether or not they have an express sunset date), as a device to ease the enactment of base broadening. This is a bait-and-switch scenario under which, once the hard political work of broadening the base has been accomplished, rates can rise again. In 1986 some conservatives who were uneasy about tax reform predicted that exactly this would happen, and to a degree they were right (since the top individual rate went back up to 39.6 percent, albeit not all the way to 50 percent).
Insofar as this can happen, reduced individual rates, no less than bad income tax preferences, are themselves low-hanging fruit that can be plucked in due course. However, suspicions that this was the plan might require very strong commitments against doing it that would be difficult to disavow later on. And the scenario is a long shot anyway given the political obstacles faced by even a sincere version of 1986-style reform.
Changing times call for changing approaches. Today, no less than in the early 1980s, it is time to significantly revise the prevailing notion of what a tax reform effort should look like. However long the odds may be against accomplishing anything major that is genuinely constructive in a deeply dysfunctional national political environment, my own inclination would be to harness income tax base broadening to long-term deficit reduction, and to address the progressivity of the U.S. fiscal system rather than hope lightning will strike in the same way twice.
1 P.L. 99-514.
2 See, e.g., Richard L. Doernberg and Fred S. McChesney, "Doing Good or Doing Well? Congress and the Tax Reform Act of 1986," 62 N.Y.U. L. Rev. 891 (1987), arguing that the 1986 act was merely congressional business as usual, designed to force interest groups to buy back the tax preferences that had just been repealed. I refuted this article's pat cynicism about the political motivations underlying enactment of the 1986 act, in favor of a more nuanced and contextually informed cynicism, in Daniel N. Shaviro, "Beyond Public Choice and Public Interest: A Study of the Legislative Process as Illustrated by Tax Legislation in the 1980s," 139 U. Pa. L. Rev. 1 (1990).
3 For example, the 1986 act used an expanded alternative minimum tax in lieu of more extensive regular tax base broadening -- a choice that bore ill fruit in future years as the AMT began applying to increasing numbers of individuals. The 1986 act also relied on increased corporate tax revenues to achieve overall revenue neutrality while offering most individuals an apparent tax cut. See Linda Sugin, Tax Expenditures, Reform, and Distributive Justice, 19 (draft of Dec. 31, 2010), calling this a "cynical gesture" because "the corporate tax . . . is hidden from view of the individuals who bear the burden of it."
4 See Jeffrey H. Birnbaum and Alan S. Murray, Showdown at Gucci Gulch: Lawmakers, Lobbyists, and the Unlikely Triumph of Tax Reform (1987).
6 The fiscal commission report, id. at 15, estimates an annual net revenue gain that would rise to $180 billion by 2020, a number that may seem large in absolute terms but arguably is small if tax expenditures as a whole already cost $1.1 trillion per year, as the report asserts. Id. at 28.
7 See, e.g., Paul Krugman, "The Hijacked Commission," The New York Times, Nov. 11, 2010 ("The goals of reform, as Mr. Bowles and Mr. Simpson see them, are presented in the form of seven bullet points. 'Lower Rates' is the first point; 'Reduce the Deficit' is the seventh.").
8 Under the Congressional Budget Office's "alternative fiscal scenario," for example, federal debt held by the public, which currently stands at about 50 percent of our GDP, may reach 200 percent of GDP while being on a steep vertical path by the mid-2030s. See CBO, "Federal Debt and the Risk of a Fiscal Crisis" (July 27, 2010), at 2, Doc 2010-16701, 2010 TNT 144-27. See infra discussion in Part III.D.
9 The case presented by the House Republicans' 2012 budget proposal is somewhat different. While the proposal advocates revenue-neutral (if not distribution-neutral) tax reform, it does so on the grounds that rate reduction and eliminating tax expenditures are both independently desirable, and it relies on other budgetary changes (for which it argues on similar grounds) to pivot toward long-term fiscal balance. See House Committee on the Budget, "The Path to Prosperity: Restoring America's Promise" (Apr. 5, 2011), at 17-18 and 50-52, Doc 2011-7151, 2011 TNT 66-40 (House Republicans' 2012 budget proposal).
10 See President Obama, State of the Union Address (Jan. 25, 2011), available at http://www.whitehouse.gov/the-press-office/2011/01/25/remarks-president-state-union-address.
11 See Shaviro, Making Sense of Social Security Reform (2000), for a relatively sympathetic discussion of increasing or at least clarifying the relationship between Social Security taxes paid and benefits received, and perhaps increasing investors' portfolio choice regarding their accruing benefits under a private accounts-like structure that I called "progressive privatization."
12 Fiscal commission report, supra note 5, at 28.
13 Id. at 14.
14 Id. at 29.
15 See House Republicans' 2012 budget proposal, supra note 9, at 5.
16 See, e.g., Henry C. Simons, Personal Income Taxation: The Definition of Income as a Problem of Fiscal Policy, 170-184 (1938) (devoting an entire chapter to criticizing the exclusion from income of municipal bond interest).
17 See "Mrs. Horace Dodge Dies at 103; Among World's Richest Women," The New York Times, June 4, 1970. This presumably was poor tax planning on Mrs. Dodge's part, because if the tax-exempt bonds paid lower pretax returns than those that were taxable, she should have placed enough of her fortune in the latter category to benefit until her taxable income had ascended into a high enough rate bracket for the two to offer equal after-tax returns.
18 See, e.g., Hearings on Flat Tax Proposals Before the Senate Committee on Finance, 104th Cong., 2d Sess. (May 18, 1995) (Statement of Prof. Michael J. Graetz of Yale Law School, recalling the "outrage generated in 1969 by Treasury Secretary Joe Barr's revelation that 154 people that year had more than $200,000 in adjusted gross income and paid no taxes, [and] the beating taken by Mrs. Dodge, in particular, who had $1 million of tax-exempt interest and no tax liability.").
19 See Shaviro, "Rethinking Tax Expenditures and Fiscal Language," 57 Tax L. Rev. 187, 204 (2004).
20 Jonathan Barry Forman, "Origins of the Tax Expenditure Budget," Tax Notes, Feb. 10, 1986, p. 537.
21 See, e.g., Stanley S. Surrey, Pathways to Tax Reform: The Concept of Tax Expenditures, 134-138 (1973).
22 See Boris I. Bittker, "Equity, Efficiency, and Income Tax Theory: Do Misallocations Drive Out Inequities?" 16 San Diego L. Rev. 735, 736 (1979).
24 Birnbaum and Murray, supra note 4, at 26.
25 H.R. 6944, 97th Cong., 2d Sess.; S. 2817, 97th Cong., 2d Sess. (introduced Aug. 5, 1982).
26 Birnbaum and Murray, supra note 4, remains the fullest and best-known account of the enactment of 1986 tax reform. I draw extended lessons from the story, similar to some that I will mention here, in Shaviro, "Beyond Public Choice and Public Interest," supra note 2.
27 Quoted in Birnbaum and Murray, supra note 4, at 110. Rather than remaining committed to fundamental tax reform, Gephardt favored trade protectionism and deficit reduction. Id.
28 I was present at a panel discussion some years ago during which Aaron said this. See http://danshaviro.blogspot.com/2010/12/two-perspectives-on-tax-reform.html.
29 See Birnbaum and Murray, supra note 4, at 111 (noting "paranoia on the Hill" that opposing tax reform would "appear to be caving to special interests"); Shaviro, "Beyond Public Choice and Public Interest," supra note 2, at 27 (noting that then-Senate Finance Chair Robert Packwood was dubbed "Senator Hackwood" by The New Republic when he appeared to be letting tax reform fail).
30 Eric M. Patashnik, Reform At Risk: What Happens After Major Policy Changes Are Enacted, 47 (2008). Patashnik notes that a Gallup poll taken on the eve of enactment of the 1986 tax reform found that "less than one in three respondents expected it to improve the nation's economy, make the tax system less complicated, or promote a fairer distribution of the tax burden." Id.
31 See Birnbaum and Murray, supra note 4, at 88; Graetz statement, supra note 18, recalling "the expressions of outrage generated in 1986 when the laborers on General Electric's assembly line paid more taxes than the company."
32 See Shaviro, "The New Alternative Minimum Tax: Perception, Reality, and Strategy," 66 Taxes 91 (1988).
33 Howard Gleckman, "Will Obama Call for Tax Reform in the State of the Union?" Tax Vox, the Tax Policy Center Blog (Jan. 12, 2011), available at http://taxvox.taxpolicycenter.org/2011/01/12/will-obama-call-for-tax-reform-in-the-state-of-the-union/.
36 House Republicans' 2012 budget proposal, supra note 9, at 51. The discussion here makes it clear to the discerning reader that the provisions marked for possible elimination include, for example, the home mortgage interest deduction and the exclusion for employer-provided health insurance. Otherwise, its reference to "over $1 trillion in forgone revenue" from tax expenditures would be wide of the mark. However, the discussion does not name any particular item, and it notes that "the exact definition of a 'tax expenditure' is open to debate." Id. The House Republicans' budget proposal also refers in the corporate tax area to accompanying rate reduction with "eliminating or modifying deductions, credits, and special carve-outs that leave many companies paying no tax at all," but here, too, it names no specific provisions. Id. at 53.
37 There are exceptions, however, such as the recent controversy over Google's and General Electric's aggressive tax planning. See, e.g., Kaja Whitehouse, "Tax Breaks -- Artful Dodger Google Saves $3B in 3 Years," New York Post, Oct. 22, 2010, at 43; David Kocieniewski, "G.E.'s Strategies Let It Avoid Taxes Altogether," The New York Times, Mar. 24, 2011, at A-1.
38 See infra text accompanying notes 50-52.
39 See David Bradford, "Reforming Budgetary Language," in Sijbren Cnossen and Hans-Werner Sinn (eds.), Public Finance and Public Policy in the New Century (2003).
40 To ensure that the credits can be claimed in full without needing to be refundable (i.e., allowable even to the extent in excess of a claimant's pre-credit federal income tax liability), Bradford suggests that they be made freely tradable to other taxpayers who could then use them.
41 To make the equivalence complete, one might need to include the credits in taxable income (as would have been done with the direct appropriations), or else adjust them to reflect includability under the direct spending approach.
42 As discussed in Part IV.C, infra, there are potential allocative rationales for taxing income or earnings, but these generally are not regarded as first-order considerations.
43 See Shaviro, "Rethinking Tax Expenditures," supra note 19, at 224-225.
44 See Shaviro, "Selective Limitations on Tax Benefits," 56 U. Chi. L. Rev. 1189, 1195 (1989).
45 See, e.g., William Andrews, "Personal Deductions in an Ideal Income Tax," 86 Harv. L. Rev. 309 (1972).
46 See, e.g., Louis Kaplow, "The Income Tax as Insurance: The Casualty Loss and Medical Expense Deductions and the Exclusion of Medical Insurance Premiums," 79 Calif. L. Rev. 1485 (1991).
47 To make the equivalence exact, one would have to adjust pretax amounts as needed to create equivalent after-tax income in each case.
48 See Leonard E. Burman and Marvin Phaup, Tax Expenditures, the Size and Efficiency of Government, and Implications for Budget Reform (2011).
49 Charles Fried, "Whose Money Is It?" The Washington Post, Jan. 1, 1995, at C-7.
50 House Republicans' 2012 budget proposal, supra note 9, at 51.
52 Id. at 51-52.
53 The House Republicans' 2012 budget proposal comes closer than most to avoiding this fallacy. It does indeed insist that the repeal of tax expenditures not increase total revenues but instead be offset by lower rates (id. at 51), whereas when calling for lower direct spending it does not similarly emphasize this linkage. But it does not claim that revenue neutrality, as such, is an important concern. Instead, it urges that budget analysts "start, not by asking what is the 'right mix' of tax increases and spending cuts to balance the budget, but by asking what is the purpose of government and then raising only as much revenue as the federal government needs to efficiently fund those missions that rightly belong in its domain." Id. at 50.
54 However, two general features of the 1986 act that some supporters of broad-based tax reform might criticize were its (1) moving away from the consumption tax pole and closer to the income tax pole and (2) strengthening double taxation of equity-financed corporate income.
55 On "funding predators" under the pay-as-you-go rules' requirement of budget neutrality, see Elizabeth Garrett, "Harnessing Politics: The Dynamics of Offset Requirements in the Tax Legislative Process," 65 U. Chi. L. Rev. 501, 515 (1998).
56 Of course, this is not to deny the shrewdness of the tactical choice to hold revenues constant, which may have been critical to the favorable reception afforded tax reform proposals by leaders in both parties.
57 Fiscal commission report, supra note 5, at 15.
59 The net five-year budgetary improvement would likely be considerably less than $484.1 billion, however. Even absent any macroeconomic disruption from the repeal, such as by reason of increased mortgage defaults, people's response to the changed incentives would presumably reduce home ownership or the use of mortgage debt relative to that assumed for purposes of the JCT's computation, which assumed that the deduction would be retained.
60 Cf. Martin Feldstein, "The Tax Expenditure Solution for Our National Debt," The Wall Street Journal, July 20, 2010 (arguing that the best way to cut government spending, under current political circumstances, is by repealing tax expenditures).
61 Fiscal commission report, supra note 5, at 14.
62 dFORM, "An Introduction to the Bacteriophage T4 Virus," available at http://www.dform.com/projects/t4/virus.html.
63 Douglas A. Kahn and Jeffrey S. Lehman, "Tax Expenditure Budgets: A Critical View," Tax Notes, Mar. 30, 1992, p. 1661.
65 For a brief discussion suggesting that the case for enacting an allowance for corporate equity deduction may be stronger than that for lowering the U.S. corporate rate, see Shaviro, "Should Corporate Tax Reform Emphasize an 'Allowance for Corporate Equity' (ACE)?" available at http://www.law.nyu.edu/ecm_dlv4/groups/public/@nyu_law_website__faculty__faculty_profiles__dshaviro/documents/documents/ecm_pro_068214.pdf.
66 See fiscal commission report, supra note 5, at 28. The report noted, however, that some tax expenditures could be added back relative to this baseline, leading to less tax rate reduction but also to a lower percentage of "giveback" of the revenue loss from rate reduction.
67 See Alan J. Auerbach and William G. Gale, "Déjà Vu All Over Again: On the Dismal Prospects for the Federal Budget," 63 Nat'l Tax J. 543, 554 (2010). In response to claims by some that "the long-term fiscal gap is not a general fiscal problem, but a medical spending problem," Auerbach and Gale respond that "the fact that healthcare growth is important in an accounting sense does not imply that attempts at closing the fiscal gap should focus solely or even primarily on reducing health care spending growth. The issue is which spending reductions and/or tax increases, relative to the baseline projections, would have the least negative effects on the economy and society." Id.
68 Burman et al., "Catastrophic Budget Failure," 63 Nat'l Tax J. 561, 564-565 (2010). This projection arises under the CBO's "alternative baseline," which assumes permanent extension of the Bush tax cuts, curtailed AMT growth, and a continued rapid rise in healthcare expenditures levels. Id.
70 Id. at 572.
71 Id. at 574.
72 Id. at 571.
73 Id. at 574.
74 Shaviro, Taxes, Spending, and the U.S. Government's March Toward Bankruptcy 86 (2007).
75 During the Reagan administration, the important bipartisan tax and budget deals that were enacted included a deficit-reducing tax increase in 1982, a Social Security fix with benefit cuts and tax increases in 1983, a deficit-reducing tax increase in 1984, the Gramm-Rudman Hollings Balanced Budget Act in 1985, comprehensive tax reform in 1986, and the Medicare Catastrophic Coverage Act in 1988 (although the parties were forced by widespread public opposition to repeal this act the following year).
76 Yet another source of uncertainty relates to financial markets. A sudden and disastrous credit market event regarding U.S. government bonds, in lieu of gradual adjustment such as slowing rising interest rates as the debt-to-GDP ratio grows uncomfortably high, would strongly suggest ex post that the pre-crash bond market had been experiencing a bubble. See Burman et al., supra note 68, at 567-570. As of today, however, one could plausibly view still-low interest rates on long-term U.S. bonds as reflecting either herd behavior and myopia or a prescient judgment that Congress will eventually fix the problem.
77 A typical financial transactions tax such as a "Tobin tax" would be levied on securities trades and perhaps also on the use of derivatives to replicate those trades. See, e.g., Thornton Matheson, "Taxing Financial Transactions: Issues and Evidence," IMF Working Paper 11/54 (Mar. 2011).
78 See Birnbaum and Murray, supra note 4, ch. 9 ("The Two-Pitcher Lunch"). The 25 percent plan, drawn up by JCT staffers at Packwood's direction, was supposed to be revenue neutral but not distribution neutral. Id. at 212.
79 See fiscal commission report, supra note 5, Figure 8, at 32 (suggesting that the commission's proposals as a whole, with a 28 percent top individual rate, would reduce the after-tax incomes of people in the top 0.1 percent of the income distribution most of all, and those in the top 1 percent second most).
80 The House Republicans' 2012 budget proposal, for example, would reduce both the top individual rate and the top corporate rate to 25 percent. While it would accompany the rate cuts with tax expenditure repeal, there is no suggestion that this part of the package (much less its overall proposal) be tested for distributional neutrality. See House Republicans' 2012 budget proposal, supra note 9.
81 At least one recent commentator has advocated such a tax. See Jeffrey A. Schoenblum, "Tax Fairness or Unfairness? A Consideration of the Philosophical Bases for Unequal Taxation of Individuals," 12 Am. J. Tax Pol'y 221, 258-271 (1995).
82 Indeed, this point even holds for the House Republicans' 2012 budget proposal, which emphasizes growth while generally not mentioning distributional concerns but nonetheless favors retaining at least modest rate graduation in the income tax.
83 The efficiency cost is even greater if the goods or services that the government provides are worth less than $30, as may happen if the only reason for using them in lieu of cash grants is to ease enactment of the package.
84 Suppose, for example, that there are three marginal tax rates (20, 30, and 40 percent) and that tax-exempt municipal bonds pay 70 percent of the pretax interest rate on otherwise identical bonds that are taxable. Taxpayers in the 40 percent bracket have a strong reason to hold municipal bonds, and those in the 20 percent bracket have a strong reason to avoid them, requiring each to trade off maximizing expected after-tax returns against optimal diversification.
85 See, e.g., Jonathan Gruber, Public Finance and Public Policy 600 (3d ed. 2010).
86 James A. Mirrlees, "An Exploration in the Theory of Optimum Income Taxation," 38 Rev. Econ. Stud. 175 (1971).
87 Id. at 207; Shaviro, "The Minimum Wage, the Earned Income Tax Credit, and Optimal Subsidy Policy," 64 U. Chi. L. Rev. 405, 467 (1997).
88 See Shaviro, "Welfare, Cash Grants, and Marginal Rates," 59 SMU L. Rev. 835, 850 (2006).
89 Slemrod, "Optimal Taxation and Optimal Tax Systems," 4 J. Econ. Persp. 157, 164 (1990).
90 Ravi Kanbur and Matti Tuomala, Relativity, Inequality, and Optimal Nonlinear Income Taxation 11 (Feb. 2010).
91 Norman Ireland, "Optimal Income Tax in the Presence of Status Effects," 81 J. Pub. Econ. 193, 194 (2001).
92 Thomas Piketty and Emmanuel Saez, "Income Inequality in the United States, 1913-1998," 118 Q. J. Econ. 1 (2003).
93 See, e.g., Saez, "Striking It Richer: The Evolution of Top Incomes in the United States (Updated with 2008 Estimates)."
94 See Piketty and Saez, supra note 92, at 1, noting Simon Kuznets's hypothesis that "income inequality should follow an inverse-U shape along the development process, first rising with industrialization and then declining, as more and more workers join the high-productivity sector of the economy" (citing Simon Kuznets, "Economic Growth and Economic Inequality," 45 Am. Econ. Rev. 1 (1955)).
95 Saez, supra note 93, at 2.
96 Including salary income can pose a similar problem, even absent bunching of many years' accrual into a single year of realization, if my high earnings in a given year are a significant outlier relative to my usual circumstances. This, however, may be less common than the outlier problem with capital gains, apart from the general life-cycle issue of people typically having annual earnings profiles that peak in late middle age.
97 Those individuals might also escape the income tax, both directly during their lives and via the step up in asset basis at death, unless it applies with sufficient rigor at the entity level.
98 Saez, supra note 93, at 4 (footnote omitted).
99 Id. at 4, n.4.
100 See David Kocieniewski, "Definition of Rich Now a Tax Issue," The New York Times, Sept. 30, 2010, at B-1.
101 See id.
102 See, e.g., Treasury, "Tax Reform for Fairness, Simplicity, and Economic Growth," vol. 1, Overview, at 47-49 (Dec. 1984) (distribution tables for Treasury I plan showing estimated effects of the proposal).
103 The Henderson blog post titled "We Are the Super-Rich," was deleted by its author after the controversy erupted but can still be found at http://delong.typepad.com/sdj/2010/09/todd-henderson-we-are-the-super-rich.html.
104 Henderson begins by complaining: "The rhetoric in Washington about taxes is about millionaires and the super rich, but the relevant dividing line between millionaires and the middle class is pegged at family income of $250,000 (I'm not a math professor, but last time I checked $250,000 is less than $1 million.)." Henderson evidently did not think to check an English language dictionary, which would have informed him that, in common usage, a "millionaire" is one whose wealth, as opposed to annual income, exceeds $1 million.
105 Henderson appears to be unaware of the underlying U.S. fiscal situation, as well as of the lack of any evidence (such as from contemporaneous budget proposals) that the Obama administration was contemplating significantly increasing the projected future path of U.S. government spending.
106 See Kenneth Scheve, "Envy and Altruism in Hard Times" (Aug. 2010), at 13.
107 Kocieniewski, supra note 100 (attributing this observation to University of California, Berkeley, economics Prof. J. Bradford DeLong).
108 See Reuters, "Survey: 42 Percent of U.S. Millionaires Don't Feel Wealthy" (Mar. 14, 2011), available at http://www.reuters.com/article/2011/03/14/us-fidelity-survey-idUSTRE72D3RK20110314.
109 See fiscal commission report, supra note 5, at 32. A further change since 1986 is that people might care about distributional neutrality over a longer time frame. In 1986 the distributional estimates went only five years forward and relied on adverse transitional treatment of tax shelters that wealthy individuals had purchased before enactment to achieve the requisite distributional neutrality within that time frame. Looking further down the road, high-income taxpayers would presumably gain from continuing to have the lower rates but no longer facing this transition hit on an ongoing basis.
110 Even if unlimited home mortgage interest deductions were allowed, their usage might be likely to grow more slowly than income at the high end. For example, a multimillionaire whose income in the top bracket doubled might in any event be unlikely to double her debt-financed homeownership, because at some point one may approach satiation at that margin. But doubling is more specifically impeded by the $1.1 million ceiling on mortgage principal that generates deductible interest, and the provision that only one's first two homes can be qualified residences for purposes of the deduction. See section 163(h)(3) and (4).
111 See fiscal commission report, supra note 5, at 31-32.
112 Joseph Bankman and Thomas Griffith, "Social Welfare and the Rate Structure: A New Look at Progressive Taxation," 75 Calif. L. Rev. 1905 (1987).
113 See Lawrence Zelenak and Kemper Moreland, "Can the Graduated Income Tax Survive Optimal Tax Analysis?" 53 Tax L. Rev. 51 (1999) (explaining that, if "demogrants" or uniform cash benefits are ruled out, rate progressivity may be the best available means of advancing distributional goals).
114 See, e.g., Saez, "Optimal Progressive Capital Income Taxes in the Infinite Horizon Model," National Bureau of Economic Research Working Paper 90546 (July 2002); Henrik Kleven et al., "The Optimal Income Taxation of Couples," 77 Econometrica 537 (2009).
115 Saez, "From Optimal Tax Theory to Practice: Comprehensive U.S. Income Tax Reform" (Jan. 2010), available at http://www.aeaweb.org/aea/conference/program/retrieve.php?pdfid=542.
116 See id., Summary Table: Baseline Versus Proposal.
117 See id. at 15.
118 Under the Slemrod hypothetical, the zero rate at the very top is, by definition, consistent with revenue maximization. One could thus theoretically modify the Saez plan to charge a zero marginal rate at the very top if one knew that the revenue cost of doing so was zero.
119 See, e.g., Gruber, supra note 85, at 27.
120 Saez, "Striking It Richer," supra note 93, at 8.
121 Id. at 10.
122 For exceptions to the standard approach in the OIT literature, see Michael J. Boskin and Eytan Sheshinski, "Optimal Redistributive Taxation When Individual Welfare Depends Upon Relative Income," 92 Quarterly J. Econ. 589 (1978); Matti Tuomala, Optimal Income Tax and Redistribution 1, 122-135 (1990); Andrew J. Oswald, "Altruism, Jealousy, and the Theory of Optimal Non-Linear Taxation," 20 J. Pub. Econ. 77 (1983); Ireland, supra note 91; and Kanbur and Tuomala, supra note 90.
123 Ernest Fehr and Klaus M. Schmidt, "A Theory of Fairness, Competition, and Cooperation," 114 Quarterly J. Econ. 817, 821 (1999). For a comprehensive recent review of the evidence suggesting that relative position matters, see Andrew E. Clar et al., "Relative Income, Happiness, and Utility: An Explanation for the Easterlin Paradox and Other Puzzles," 46 J. Econ. Lit. 95 (2008).
124 Id., citing George F. Loewenstein et al., "Social Utility and Decision Making in Interpersonal Contexts," 57 J. Personality & Social Psych. 426 (1989). From an evolutionary standpoint, it is an interesting question why people often dislike advantageous inequality. One explanation could be that it excites concern about the risk of becoming low-status, perhaps enhanced by an intuition that status has declining marginal utility. Another explanation could be that those below one in rank are potential allies against those higher in rank, an alliance that feelings of sympathy may help one to form.
125 See, e.g., Scheve, supra note 106.
126 See Robert H. Frank, Choosing the Right Pond: Human Behavior and the Quest for Status (1985).
127 There may also be positive spillover effects of more concentrated wealth at the top, however. Thus, suppose new medical treatments or consumer products are initially developed for the benefit of the richest individuals but then become cheaper and more widely available.
128 Martin Gilens, "Inequality and Democratic Responsiveness," 69 Pub. Opinion Quarterly 778 (2005).
129 Jacob S. Hacker and Paul Pierson, Winner-Take-All Politics: How Washington Made the Rich Richer -- and Turned Its Back on the Middle Class (2010).
130 Id. at 12.
131 Frank, "Positional Externalities Cause Large and Preventable Welfare Losses," 95 Am. Econ. Rev. 137 (2005).
132 See Boskin and Sheshinski, supra note 122, at 594; Oswald, supra note 122, at 86; Ireland, supra note 91, at 27; Kanbur and Tuomala, supra note 90, at 11.
133 See, e.g., Christa H.S. Bouwman, The Geography of Executive Compensation (2010) (finding that geography affects executive compensation by reason of spatially limited status competition).
134 See Thorstein Veblen, The Theory of the Leisure Class (1899).
135 See Fred Hirsch, Social Limits to Growth (1976).
136 Frank, "Are Positional Externalities Different From Other Externalities?" 92 J. Pub. Econ. 1777 (2008).
137 See, e.g., Robert H. Frank and Philip J. Cook, The Winner-Take-All Society (1995).
138 See Frank, supra note 136.
139 A Pigovian tax can be used, for example, to make polluters bear the cost imposed by their emissions, causing them to engage in the optimal level of polluting but otherwise valuable economic activity.
140 See Frank, supra note 131, at 138.
141 Frank asserts that this favorable efficiency argument does not apply to income taxation, apparently on the view that saving is non-positional. See Frank and Cook, supra note 137, at 213-214. This may be incorrect; however, if one views saving as merely a means of purchasing future consumption, that is no less positional than current consumption. Also, there may be current positional benefits to holding large savings, even if people do not directly observe one's bank balance.
142 See Ireland, supra note 91; Kanbur and Tuomala, supra note 90.
143 In OIT models, the benefit side commonly involves demogrants. However, positing that the government instead furnishes goods and services that are enjoyed relatively uniformly does not greatly change the analysis, except that one needs to ask the further efficiency question of how the value of these items compares with their cost. Kanbur and Tuomala, supra note 90, while "find[ing] support for greater progressivity in the tax structure as relative concern increases [find] that this incremental impact is less at higher levels of inequality." Id. at 1.
144 See Frank and Cook, supra note 137. Labor market tournaments were introduced to the economics literature by Edward P. Lazear and Sherwin Rosen, "Rank-Order Tournaments as Optimum Labor Contracts," 89 J. Pol. Econ. 841 (1981).
145 See, e.g., Zelenak and Moreland, supra note 113, at 88-90.
146 See Agnar Sandmo, Inequality and Redistribution: The Need for New Perspectives 9 (2005) (noting that a given tournament model "implies that the marginal tax rate should be higher, the larger is the random element in the determination of wage differentials").
147 See also Martin J. McMahon Jr. and Alice G. Abreu, "Winner Take All Markets: Easing the Case for Progressive Taxation," 4 Fla. Tax Rev. 1 (1998).
148 See Frank and Cook, supra note 137, at 106-109. Zelenak and Moreland, supra note 113, at 89-90, disagree that this necessarily leads to excessive participation in tournaments if, at the same time, risk aversion discourages entry. However, the externality problem may lead to inefficiency even if risk aversion keeps the overall number of tournament participants from being too high, since in particular cases it may induce someone to make the socially wrong decision.
149 Frank and Cook, supra note 137, at 103-105.
150 Id. at 170.
151 Mats Persson and Agnar Sandmo, "Taxation and Tournaments," 7 J. Pub. Econ. Theory 543 (2005).
152 The burden of making a case is more likely to be met for corporate tax rates than for individual rates, given the global pressure of tax competition. But even on the corporate side, the case for revenue-neutral, 1986-style tax reform would be stronger if the United States were closer to long-term budgetary balance.
END OF FOOTNOTES
About Tax Analysts
Tax Analysts is an influential provider of tax news and analysis for the global community. Over 150,000 tax professionals in law and accounting firms, corporations, and government agencies rely on Tax Analysts' federal, state, and international content daily. Key products include Tax Notes, Tax Notes Today, State Tax Notes, State Tax Today, Tax Notes International, and Worldwide Tax Daily. Founded in 1970 as a nonprofit organization, Tax Analysts has the industry's largest tax-dedicated correspondent staff, with more than 250 domestic and international correspondents. For more information, visit our home page.
For reprint permission or other information, contact firstname.lastname@example.org