Tax Analysts®Tax Analysts®

My Subscriptions:

Featured News

December 19, 2011
On Tax Increase Limitations: Part I -- a Costly Incoherence

Full Text Published by Tax Analysts®

by David Gamage and Darien Shanske

Executive Summary

David Gamage and Darien ShanskeIn this essay, the first of a series, we explore the theoretical implications of one particular type of fiscal limitation on state legislatures -- namely, special rules limiting tax increases. In this first essay we will explore the analytic soundness of these tax increase limitations (TILs). In future essays in this series we will analyze some of the consequences of TILs and in particular how they can be "evaded."1 We will argue over the course of this series of essays that because there is no meaningful content to the term "tax increase" as it is used in TILs, legislative majorities that wish to do so can readily circumvent TILs. We will then propose alternatives to TILs to better promote fiscal management at the state and local levels.


Special fiscal requirements are a common feature of state constitutions.2 In this essay we will make an analytic observation about one type of fiscal requirement: state-level tax increase limitations or TILs. By TILs we mean provisions that require a legislative supermajority for taxes to be "increased." For example, in California:

    Any change in state statute which results in any taxpayer paying a higher tax must be imposed by an act passed by not less than two-thirds of all members elected to each of the two houses of the Legislature.3

It is well known that these regimes have questionable effectiveness, at least insofar as their goal is to curb the growth of government or even simply to change the pattern of government expenditures in the applicable state relative to other states not similarly constrained.4 The dominant explanations for this failure of TILs involve the ambivalence of voters or conniving of politicians. Without casting doubt on these explanations, we think it important to make an analytic observation that we believe also contributes to the explanation of the observed phenomenon of the ineffectiveness of TILs.

Our key analytic observation is that TILs insert two conceptually vacuous notions -- "tax" and "increase" -- into the fiscal constitutions of the states that have them. It is at least in part because this combination is incoherent that TILs do not work.

We are not going to discuss related parts of state fiscal constitutions -- provisions that are often grouped together with TILs -- namely, state or local debt limitation regimes, special state and local balanced budget rules, or tax increase limitations at the local level. This is because all of these provisions, at least arguably, have a different conceptual justification (and content). Special debt limitation regimes, for example, can be justified as important for generational equity; local tax rules may reflect a reasonable concern with tax exporting or desire to enhance local democratic participation.5

The problem with tax increase limitation regimes at the state level is that these regimes must successfully define the notion of a tax increase. Yet, to borrow a striking image from Dan Shaviro, attempting to make sense of this concept is like playing a game of pin the tail on the donkey; we are all spun around and may end up pinning the tail anywhere at all.6

Seeing Through the 'Raising Taxes' Mirage

Step 1: Spending Through the Tax System

We will begin with a famous example from David Bradford.7 Bradford imagined a weapons supply tax credit granted to arms manufacturers. The arms manufacturers would get a tax credit in the amount of the value of arms they deliver to the U.S. government, say for a maximum of $100 billion.8 The U.S. government would then reduce taxes by that same amount ($100 billion). The government could then claim to have slashed taxes without compromising national security or reducing overall allocations to public services. As Ed Kleinbard observed, this anecdote illustrates the "empty formalism of our concepts of Government revenues and Government expenditures."9 It is easy enough to change the numbers so that taxes decrease and spending increases (say the credits are $150 billion) or just about any other combination one might imagine.10

And there is no need to imagine much as governments have frequently engaged in Bradford-type maneuvers. As Kleinbard notes, the Internal Revenue Code is full of tax credits awarded to private entities in return for satisfying the government's substantive policy goals; those credits are often even administered not solely by the IRS but by the federal agency with substantive expertise (for example, the Department of Energy for "qualified gasification projects").11 State tax systems are, of course, full of similar credits.12

Shaviro offers a different, real-life example of Bradford's insight.13 In 1993 the Clinton administration proposed taxing a greater proportion of a recipient's Social Security benefits under the federal income tax. The Clinton administration reasoned that this should count as a "spending cut" because, in effect, the federal government would be out less money. However, this characterization was challenged, including by the Congressional Budget Office, which claimed that this was really tantamount to a tax increase because additional revenue would be raised through the tax system rather than smaller checks cut by the Social Security Administration. In terms of policy, the issue of nomenclature was vacuous, but the issue was important in terms of politics precisely because it mattered in what ratio the administration combined spending cuts and tax increases. State constitutions, through having special rules for tax increases, essentially mandate that legislators contort themselves in similarly parsing taxing from spending.

In the end, our first analytic point relies on the fact that state tax systems, like the federal system, are riddled with so-called tax expenditures; that is, governments are spending money on desired programs through the tax code. Limiting "tax increases" thus does not limit spending through tax expenditures nor does it prevent politicians from raising more revenue by reducing tax expenditures.

Step 2: No Ideal Tax Baseline

It could perhaps be argued that this problem can be fixed. If only politicians were barred from using tax expenditures,14 they would have only one option if they wanted to fund a new program (without incurring debt): increase taxes. In such a world, TILs would have more bite. But it is not so simple. First, the search for a firm definition of what constitutes a tax expenditure has been elusive; there is no ideal baseline for any tax. Take the example of the deduction allowed for state and local tax (SALT). The federal government lists this provision as a tax expenditure,15 but it is arguably appropriate on traditional income tax grounds because it reflects the fact that some taxpayers are less well off to the extent that they pay higher state and local taxes that do not benefit them.16

Step 3: Wrong Question

Perhaps it is possible that some rough baseline could be established for "tax expenditures" and that this baseline could be made enforceable,17 and that thereby the notion of a tax increase could be given some practical substance. But the question would then become whether this notion of tax increases would be of any use; we think it would not. We will start with the broadest substantive issues made murky by the focus on "tax increases."

    Allocation and Distribution
The efficiency and equity of a unified system of taxing and spending are substantive questions. It may not matter whether taxing Social Security benefits is a tax increase or a spending decrease, but it matters a great deal as a consideration of equity whether Social Security is, in effect, means tested. Further, as a consideration of the efficient allocation of government resources, it matters a great deal whether the SALT deduction is encouraging efficient or inefficient uses of government resources. Whether these issues should be categorized as tax increases is beside the point.

    Tax System or Other Government
We have seen that bringing content to the term "tax increases" requires vilifying tax expenditures, but is this appropriate? In many cases, we think not. It can be highly desirable on both allocative and distributive grounds to use the tax system to achieve social ends that could be plausibly characterized as tax expenditures.18 In the alternative, it could be sensible to use a nontax agency to achieve a tax objective, assuming we could agree on what a tax objective would be.19

Because using the tax system for apparently nontax ends is more common and because, as discussed above, this is the expedient that is so threatening to TILs, this is the scenario we will focus on and justify, at least as a general possibility. Consider government support for higher education and suppose we would like a government program making higher education more broadly available to be administered in proportion to income. To reach that distributive goal most efficiently we might reasonably want to use the income tax system because the tax bureaucracy is already aware of a taxpayer's income. This is not necessarily the case, but it is surely plausible and will be true for some programs at least sometimes. Thus, even if we could ban tax expenditures in order to make TILs effective, we may not want to, because tax expenditures may sometimes be the best policy option, at least insofar as labeling a program a tax expenditure facilitates the program being administered through the tax bureaucracy.

    Taxation or Other Governmental
    Intervention (e.g., Regulation)?
It is well understood that regulations can act as substitutes for taxation.20 That is a specific illustration of the previous point about the continuity between the tax bureaucracy and other parts of the government. The aspect we emphasize in this subsection is the continuity between different kinds of government interventions. Sometimes it makes allocative or distributive sense to use a regulation, and at other times a tax. TILs put pressure on governments to use regulations and not taxes, but in many cases taxes might be the more desirable option on allocative or distributive grounds. Thus, for instance, economists tend to favor the use of carbon taxes to combat global warming,21 but such taxes, as "taxes," are off the table politically in part because a state such as California could not impose or increase carbon taxes without a two-thirds legislative majority.

The fixation on avoiding tax increases can do more than influence the choice of government action; it can also shape the choice of tax base. For instance, if tax rates cannot be increased without a supermajority, legislatures have an incentive to favor tax bases that show significant revenue growth. Of course, those tax bases also tend to be more volatile, encouraging a feast-or-famine pattern of state budgeting when state governments both expand and contract according to ever more severe cycles, about the least efficient result imaginable.22

Special Case of Fees: It is not controversial that the price mechanism is the gold standard for achieving allocative efficiency and, unsurprisingly, economists have urged government regulators to use the price mechanism to the extent possible -- for instance, using tolls to regulate use of a bridge. Those quasi-market levies are based on the benefit principle. That is, each user of a government service is charged in proportion to how much that user benefits. We should note right away that in many ways a toll is as much a top-down command as a regulation regarding the number of cars allowed on a bridge (say by permit) would be. Yet the toll, that is, a taxlike intervention, makes more sense because we do not want to create a new bridge permit bureaucracy (say because of the administrative expense and uncertainty regarding the optimal number of vehicles). What we want to achieve is to send a (relatively flexible) price signal about the cost of driving in order to try to cause drivers to internalize the externalities caused by their driving.

Fees, insofar as they are a regulation that raises revenue for government programs, are particularly fungible with taxes. There is no clear line between what is a tax and what is a fee. At the one end is a user fee, say for trash pickup, and at the other end a national tax, say the federal income tax. We will just stipulate that the federal income tax is not a fee, but there is a broad continuum among many other taxes and fees. For instance, take a user fee for trash collection. This user fee is an average price, not likely the cost of your trash pickup, and indeed, buried in the price of pickup may well be cross-subsidies for other users required by government regulation. Thus even this fee is not a perfect price and thus is "taxlike." And then consider local property taxes; they are more tied to specific benefits than federal income taxes, but they are less tied to a specific benefit than a trash collection fee. Even state-level taxes are tied to the benefit principle to some extent; there is at least some mobility among states and it would seem that some taxpayers move to the package of taxes and spending that they desire. This perplexity regarding the nature of state and local taxes is at the root of the difficulties in analyzing the SALT deduction using ordinary income tax principles. Making the fee-tax question so important puts enormous pressure on tax-fee jurisprudence.23

Special rules about taxes versus fees are a distraction from the hard question of whether fees or taxes are preferable in particular cases. For instance, does it make sense to advance the use of recycling by means of regulation or by fees? TILs should not be relevant to this discussion.

Step 4: Random Direction

It could be maintained that at least TILs exert some sort of pressure to shrink the size of government and should therefore be supported, even if this would require relying on crude distinctions and giving up on some desirable policy tools. Yet even this is not so. Suppose, as many critics contend, that TILs encourage the use of regulation when taxing would be more allocatively efficient -- then TILs have in effect increased the size of government by any meaningful measure.24 This is because the most rigorous definition of the size of government refers to how much government activity distorts the economy compared with an appropriate baseline, and adding new inefficient regulations distorts the economy more, not less.

This confusion extends to considering other government interventions. As we saw in Step 1, tax expenditures, which are not subject to TILs regimes, expand the size of a government both allocatively and distributively in much the same way direct spending does. Banning tax expenditures would not make the situation better, even if that were possible (Step 2). After all, as we discussed in Step 3, a well-designed credit can reduce the footprint of the government.

Costs of Confusion

Now that we have elaborated on the many levels of conceptual confusion underlying TILs, we can see why they were always doomed to be ineffective. Legislatures could always act through tax expenditures or regulations regardless of TILs, and they could do so with considerable justification in specific cases. Hence, unsurprisingly, states with TILs expand and contract their governments in response to the same political pressures as other states. Further, TILs are not benign contributions to political theater. They put enormous pressure on state law tax-fee doctrines, distract from substantive issues, and even nudge states toward inferior policy choices.

We will argue in our next essay in this series that TILs cannot prevent determined legislative majorities that are not supermajorities from effectively raising additional revenue to fund government activity. But even if TILs could succeed in reducing the scope for effective government action, that does not imply that TILs could reduce the size of government. Again, any meaningful measurement for the size of government from the perspective of those who wish to "starve the beast" must consider the aggregate effect of government action on citizens and the economy. To the extent that TILs divert government activity into less effective forms -- such as inducing the use of regulations when direct spending would be more effective -- this may reduce the scope for beneficial government action, but it will not necessarily shrink the size of government. Making government less effective is likely to increase the aggregate effect of government activity on citizens and the economy.

To end with an illustration, imagine that you obtain an injunction against your neighbor throwing noisy parties. If your neighbor responds to the injunction by instead playing loud music or turning up the television volume, your injunction may have made your neighbor worse off to the extent he would have preferred to throw parties, but you may well fail to reduce your neighbor's adverse impact on you, because the music or television may prove even more bothersome than the parties. In order to "starve the beast" of your neighbor's noise pollution, you must be able to prevent all of your neighbor's noisy activities. But when we move to the TILs context, we will argue that it is not possible to prevent all alternatives to government taxing and spending. As we will explain further in future essays, TILs thus cannot effectively starve the beast. Instead, TILs mostly serve to make governments less effective without reducing the aggregate impact of government activity.

* * * * *

Academic Perspectives on SALT is a column by David Gamage, an assistant professor at the University of California Berkeley School of Law (Boalt Hall), and Darien Shanske, an associate professor at the University of California Hastings College of the Law. During the 2011-2012 academic year, Gamage is on academic leave from Berkeley while serving as special counsel and senior Stanley S. Surrey Fellow at the U.S. Department of the Treasury, Office of Tax Policy. Nothing written in these columns should be attributed in any way to the Treasury Department, the Obama administration, or anyone other than the authors.


1 See, e.g., Darien Shanske, "Going Forward by Going Backward to Benefit Taxes," California Journal of Politics and Policy, vol. 3, iss. 2, art. 14, available at

2 For a recent catalog and critical perspective, see Richard Schragger, "Democracy and Debt," Yale Law Journal, forthcoming, available at Note that Schragger does not specifically address TILs. TILs are often proposed to be included in the U.S. Constitution. See, e.g., SJ Res. 23, section 4, 112th Cong. (2011).

3 Calif. Const. Art. 13A, section 3a. See also, e.g., Ariz. Const. Art. 9, section 22. Related regimes require voter preapproval before taxes can be raised. See, e.g., Colo. Const. Art. X, section 20(4)(a). Colorado also has a supermajority (two-thirds) requirement as to raising taxes in an "emergency." See id. at section 6(a).

4 See, e.g., Mathew McCubbins and Ellen Moule, "Making Mountains of Debt Out of Molehills: The Pro-Cyclical Implications of Tax and Expenditure Limitations," 63 National Tax Journal 603 (2010); Bruce E. Cain and George A. Mackenzie, "Are California's Fiscal Constraints Institutional or Political?" Public Policy Institute of California, Dec. 2008.

5 See, e.g., Richard Briffault, "Foreword: The Disfavored Constitution," 34 Rutgers Law Journal 907, 947-949 (2003) (on debt limitations); Kirk J. Stark, "The Right to Vote on Taxes," 96 Northwestern University Law Review 191 (2001) (on local tax limitation regimes). Neither Briffault nor Stark argues that current versions of these limitations are actually achieving these other goals. See also, e.g., Yilin Hou and Daniel L. Smith, "Do State Balanced Budget Requirements Matter? Testing Two Explanatory Frameworks," 145 Public Choice 57 (2010) (looking at the effects of some balanced budget rule regimes).

6 See Daniel N. Shaviro, Taxes, Spending, and the U.S. Government's March Toward Bankruptcy 16 (2007). See also David Gamage and Jeremy Bearer-Friend, "Managing Fiscal Volatility by Redefining Tax Cuts and Tax Hikes," State Tax Notes, Oct. 11, 2010, p. 113, Doc 2010-20216, or 2010 STT 196-1.

7 Daniel N. Shaviro, Do Deficits Matter? 101-102 (1997).

8 This simple example assumes the manufacturers have sufficient income; one can also imagine a refundable credit.

9 Edward D. Kleinbard, "The Congress Within Congress," 36 Ohio Northern University Law Review 1, 2 (2010).

10 For further discussion, see David Gamage and Darien Shanske, "Three Essays on Tax Salience: Market Salience and Political Salience," 65 Tax Law Review, at Part I.B.6.b. (forthcoming, 2012).

11 Kleinbard, supra note 9, at 2 (discussing 2009-16 IRB 802, a notice about implementing IRC section 48B).

12 See, e.g., Calif. Revenue and Taxation Code section 6010.8 (granting the California Alternative Energy and Advanced Transportation Financing Authority power to grant sales and use tax exclusions). For a full listing of tax expenditures in California, see, for example, Department of Finance, "Tax Expenditure Report 2011-12." At least 44 states provide some information on their tax expenditures; see

13 Daniel N. Shaviro, "Rethinking Tax Expenditures and Fiscal Language," 57 Tax Law Review 187, 192-194 (2004).

14 Of course, it is not at all clear how that could be achieved.

15 See, e.g., Joint Committee on Taxation, "Estimates of Federal Tax Expenditures for Fiscal Years 2009-2013," at 49, 50 (Jan. 11, 2010).

16 See generally Louis Kaplow, "Fiscal Federalism and the Deductibility of State and Local Taxes Under the Federal Income Tax," 82 Virginia Law Review 413 (1996).

17 We think not, of course. We will further discuss the problems with attempts to establish such a baseline in our next essay in this series. Discussion on this point can also be found in Shanske, supra note 1.

18 Cf. Shaviro, Bankruptcy, supra note 6, at 30-40 (applying Musgrave's allocative and distributive roles of government to analysis of taxing and spending); David A. Weisbach and Jacob Nussim, "The Integration of Tax and Spending Programs," 113 Yale Law Journal 955 (2004) (arguing for pragmatic analysis of whether a program should be administered through the tax code).

19 If it is a tax objective to lower the rate of tax on some energy investments, choosing to administer tax credits through federal or state energy agencies, as discussed above, would qualify as an example of using a nontax agency to administer a tax program.

20 For discussion, see Gamage and Shanske, supra note 10, at Part I.B.6.c.

21 See, e.g., Reuven S. Avi-Yonah and David M. Uhlmann, "Combating Global Climate Change: Why a Carbon Tax Is a Better Response to Global Warming Than Cap and Trade," 28 Stanford Environmental Law Journal 3 (2009) (summarizing arguments).

22 David Gamage, "Preventing State Budget Crises: Redefining 'Tax Cuts' and 'Tax Hikes'," 98 California Law Review 749, 757-760 (2010).

23 In California, for instance, tax limitations of various kinds have encouraged state and local governments to raise revenue with "nontaxes." When courts have upheld the use of these nontaxes, additional voter propositions have often followed. California governments are now absorbing the latest tax limitation initiative, Proposition 26, passed in November 2010. Proposition 26, which added sections to articles 13A and 13C of the California Constitution, explicitly aimed to narrow the definition of a fee, responding to one California Supreme Court case in particular. See Shanske, supra note 1. The litigation over the meaning of Proposition 26 has already begun. Kathleen K. Wright, "The Aftermath of California's Proposition 26," State Tax Notes, Nov. 14, 2011, p. 471, Doc 2011-22949, or 2011 STT 220-4. Yet this is far from a California problem; battles over the tax-fee distinction are endemic to other states with TILs. See, e.g., Barber v. Ritter, 196 P.3d 238, 248-250 (Colo. 2008); Keller v. Marion County Ambulance District, 820 S.W.2d 301 (Mo. banc 1991) (distinguishing taxes from fees for purposes of Missouri's TIL even though the TIL speaks of limiting "any tax, license or fees").

24 See, e.g., Shaviro, Bankruptcy, supra note 6, at 40.


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