Abstract
Are the goals we have for our tax system mutually incompatible? This is the question that we posed to a distinguished group of tax economists at a conference hosted by the Murphy Institute and the Department of Economics at Tulane University in October 2012. More specifically, we asked whether the US economy faced a “fiscal trilemma”: Can the three goals of revenue adequacy, fairness, and promotion of economic growth all be achieved, or do we have to sacrifice one of these goals to meet the others?
Many of the traditional debates in tax policy are encompassed within this trilemma framework. For example, must a progressive tax system that meets our revenue needs hamper economic growth? Do progrowth tax systems that generate needed revenues come at the expense of progressivity? Would a progressive income tax system that managed to encourage growth provide insufficient revenues? Is there an unresolvable tension between the traditional Haig-Simons notion of a “good” tax—a broad-based, low-rate income tax base with few if any exclusions or deductions—that emerges from a largely static view of the tax system and the more growth-oriented taxes (e.g., a personal consumption tax, a value-added tax, and a retail sales tax) that emerge from a dynamic analysis of taxes?
The tensions between these three goals of tax policy—fairness, growth, and revenue adequacy—are not fully understood or quantified. Our conference brought together researchers working on these issues to explore recent advances in the field and to increase awareness of the subject. The articles in this special issue of Public Finance Review illuminate some of the key issues and tensions that arise in tax system design and provide a range of alternatives for solving the trilemma conundrum.
Earlier versions of these articles were presented at the October 2012 conference “A Fiscal Trilemma?” The articles were submitted to Public Finance Review (without any guarantee of publication), went through the regular reviewing process, and were revised in response to comments by the anonymous, external reviewers.
In this introduction, we set the stage for the articles in this volume by elaborating on the tensions inherent in tax system design and the potential for a trilemma. We also provide a preview of some of the answers offered to meet the trilemma challenge in this volume. Finally, we suggest our own conclusions about the fiscal trilemma puzzle.
Our discussion starts with the uncontroversial premise that the United States faces sharply increased revenue requirements in the future. As the articles in this issue clearly demonstrate (especially those by William Gale and Samuel Brown and by Leonard Burman), the current fiscal trajectory of the federal government is unsustainable. The combination of the aging of the population and the inexorable increases in health care costs take the United States away from its previous comfort zone of total federal taxes ranging between the postwar average of 18 to 19 percent of gross domestic product (GDP). At a minimum, we believe that taxes will need to increase by several percentage points relative to GDP in order to provide some semblance of intertemporal budget balance.
A second premise in our discussion is that we should not adopt tax policies that exacerbate existing inequalities. There is considerable evidence that inequality has increased dramatically since the 1970s, and this poses a challenge for tax policy.
A third and final premise is that we believe that promoting economic growth is essential. Even small changes in economic growth rates—through the power of compound interest—have over time profound effects on standards of living and budgetary outcomes. Increased economic growth would not only raise living standards but perhaps would also apply a balm to our political wounds as they are so often driven by budgetary pressures.
In short, revenue adequacy, fairness, and economic growth remain essential goals for tax policy. Can taxes be chosen to achieve all of these goals?
To illustrate the fiscal trilemma, it is useful to start with the status quo and consider some deviations from it. We characterize the status quo as a moderately progressive income tax system that neither severely impedes economic growth nor exacerbates inequality. To be sure, modest changes in the status quo could potentially make a slight improvement in progressivity or lead to somewhat greater economic efficiency. Still, the key problem with the status quo is that we are simply not raising enough revenue from our existing tax structure. We only meet two of the three desired goals for the tax system.
Suppose we kept the basic structure of the current system of a progressive income tax, but increased all tax rates, including those at the top. This change would raise additional revenue and potentially solve the revenue adequacy problem. It certainly would not worsen and perhaps would increase the fairness of the tax system. However, the tension here arises with the promotion of economic growth.
Suppose instead we change the types of taxes that form the core of the current system. One option is to move closer to a pure Haig-Simons income tax based on comprehensive income. Such a tax includes in the tax base the return to savings, thus providing a disincentive to savings and distorting consumption across periods. This problem has long been recognized by public finance theorists. For example, in very simple models with infinitely lived agents, the optimal tax on capital income has been shown to be zero: any positive tax on savings builds distortions cumulatively over time, and welfare can be increased through shifting taxes away from capital to labor. From this perspective, an income tax with sufficiently high rates to solve the revenue adequacy problem would almost certainly reduce economic growth and fail on one prong of the trilemma.
Pure income taxes would also tax income from capital gains as ordinary income. Under current law and long time practice, taxpayers can defer most taxation on appreciated assets by holding them rather than selling them; indeed, if held until death, capital gains taxes are avoided entirely. The current system therefore creates a lock-in effect that distorts the allocation of capital, that creates economic inefficiencies, and that reduces economic growth. With high income tax rates, the gap between capital gains and ordinary income will only increase and potentially create even larger distortions in the economy.
Suppose instead we move toward greater reliance on consumption taxes. Consumption taxes avoid these problems by exempting the returns to savings from the tax base. Could consumption taxes at sufficiently high tax rates be the solution to our fiscal trilemma? The most commonly used consumption tax in countries around the world is the value-added tax (VAT), which is essentially equivalent to a flat rate sales tax. Since the ratio of consumption to income decreases as income increases, a VAT (or a retail sales tax) will be regressive. So the most common approach to consumption taxation fails one of the prongs of the trilemma.
Are there any solutions to the fiscal trilemma problem? The articles in this issue suggest a number of approaches to solving the problem.
First, we may wish to consider the entire system of government spending and taxation as a whole and not simply focus on taxation. As the article by Rasmus Højbjerg Jacobsen, Søren Bo Nielsen, and Anders Sørensen explains, Denmark’s tax system relies on a VAT with a very high rate, but it uses the additional revenue raised by this (and other taxes) to provide both transfer programs that target the poor and public goods that benefit the middle class. Other European countries follow this model as well. Taking a comprehensive view of the fiscal system, we can look at the overall incidence of taxes and benefits and not focus exclusively on taxes. Denmark also has a modified version of the “dual income” tax system that taxes capital income at a lower rate than wage income, which tips the tax system toward reliance upon consumption taxation.
Second, we may wish to move toward greater use of consumption taxes. In the United States, the VAT has been a politically poisonous tax and, despite enthusiasm from tax economists, has little immediate political appeal. Consumption taxes in the form of retail sales taxes have long been the province of the states, and the ways in which a VAT would mesh with the tradition of fiscal federalism in the United States is an open question. Even so, as the articles by Laurence Seidman, by Burman, by Gale and Brown, and by John Diamond and George Zodrow demonstrate, consumption taxes have the potential to generate significant amounts of additional revenue, to do so in ways that maintain and even increase progressivity, and to increase short- and long-term economic growth.
For example, Seidman suggests a progressive consumption tax as an alternative. A cash flow progressive consumption tax operates like an income tax except that it allows a deduction for savings and a corresponding increase in borrowing into the tax base. Seidman envisions this tax as a supplement to the current income tax and restricted to high-income and high-consumption individuals. His hope is that, since the new tax is consumption based, it will be less distorting than simply increasing rates under the current income tax but that it will still be able to raise sufficient revenue to address the fiscal trilemma.
Similarly, Diamond and Zodrow use a dynamic computable general equilibrium (CGE) model to simulate the effects of the enactment in the United States of a temporary 10-year VAT whose resulting net revenues are used to reduce the level of national debt. They find that, relative to the current situation, such a reform can be moderately progressive both for cohorts alive at the time of reform and for future generations and that current middle-aged and elderly generations must bear a burden to confer a gain on young and future generations; in both cases, these distributional effects depend on the specific features of the VAT. They also find that this reform reduces the level of debt (and of deficits) relative to GDP, while increasing the growth rate of GDP. Gale and Brown make similar arguments for the introduction of a VAT, without quantifying the effects of this reform with a CGE model. Burman also argues for a VAT whose revenues would be earmarked for health care.
Third, there are also other, more specific consumption taxes that could address parts of the fiscal trilemma. Gale and Brown recommend that we examine carbon taxes or energy taxes, which have the advantage of potentially addressing global warming concerns while at the same time raising revenue. Unfortunately, these taxes are also unpopular in the current political environment, and their impact on progressivity is unclear.
Fourth, a very different strategy is suggested by James Hines. At present, the United States imposes substantial federal taxes on estates, gifts, and generation-skipping transfers that exceed high exemption levels (currently $5.25 million). These transfer taxes are very distortionary compared to other federal taxes, and they raise little revenue; however, their incidence is likely to be quite progressive, with tax burdens falling only on the very wealthy. Hines argues that these taxes could be eliminated, with their revenues replaced by higher top-bracket income taxes. Such a reform would raise additional federal tax revenue with improved efficiency and greater measured tax progressivity.
Fifth, one last approach to solving the trilemma problem that may be within our political reach is to tackle tax expenditures, or programs within the tax system that can be viewed as spending programs. These include subsidies to housing through the mortgage interest deductions, subsidies to many nonprofit organizations through the charitable deduction, subsidies to states through the state and local tax deduction, and subsidies to savings through exemptions from taxation from pensions and related retirement savings vehicles. However, as described by Burman and by Gale and Brown, the actual task of reforming tax expenditures is fraught with political difficulties, as these subsidies are very popular.
We believe that the articles in this special issue clearly indicate that these various strategies, taken singly or in combination, have the potential for solving our fiscal trilemma: achieving the three goals of revenue adequacy, fairness, and promotion of economic growth. Is such a comprehensive tax reform likely to occur in the current environment?
We are always hopeful, but we remain skeptical. Consider the conditions—political and economic—under which tax reform seems most likely. In our experience, we have found that the necessary (if not sufficient) conditions for a reform to be enacted are several. First, the tax system must be widely seen—by most all relevant players—as “broken.” Second, there needs to be consensus on how to fix it. Third, there needs to be a strong champion who can generate political support for reform.
Unfortunately, while there may be widespread agreement that the current tax system is indeed broken, there is no consensus about what should be done. Although economists may largely agree on the many attributes of carbon and “sin” taxes, such consensus does not extend to Congress. Reducing or eliminating tax expenditures also generates substantial agreement among economists, but the beneficiaries of these tax expenditures will not give them up without a struggle. Some form of consumption taxation is now often advocated, but there is no universal agreement even among economists on consumption versus income taxation, let alone on the specific form that consumption taxes may take. Virtually, any proposal to raise tax rates even modestly is almost certain to generate a visceral—and negative—reaction in many quarters. Attempts to combine tax reform with overall entitlement reform are often seen as a “third rail” of American politics, even if other countries like Denmark seem able to navigate this minefield.
There is also no obvious champion to push reform through Congress. Whether the current environment in Washington is more or less partisan than in other periods is debatable. Regardless, the chance of a nonpartisan agreement being pushed through the legislature by one of our political leaders seems remote in the current environment.
Put differently, the “perfect storm” surrounding the successful enactment of, say, the Tax Reform Act of 1986 (TRA86) is simply not there right now:
Unlike 1986, there is no consensus among the experts. Then, almost everyone agreed with cutting rates and expanding the base—and keeping an income base; now, there are many competing proposals with no agreement.
Unlike 1986, there are no obvious and easy “selling points” (e.g., reduce tax shelters, increase corporate tax revenues, and lower marginal tax rates).
Unlike 1986, there is no political leader who seems willing and able to generate political support for reform.
Unlike 1986, the chances of bipartisan compromises in the current setting seem unlikely.
Unlike 1986, the constraints imposed on any current reform process are likely to be far more severe and politically difficult, particularly because revenue neutrality is no longer a serious option—we need some additional revenue.
At bottom, we remain pessimistic that any tax comprehensive tax reform can be enacted in the current environment, even one that seems able to solve the fiscal trilemma.
Indeed, we are becoming increasingly convinced that perhaps we now know “too much.” We are of course not suggesting that we, as public finance scholars, should not try to examine the many and disparate effects of tax reform (or, more broadly, the effects of any government policies)—these efforts will continue, as they should. However, we wonder whether (even compared to the 1980s, with TRA86) our knowledge about the effects of any potential tax reforms paralyzes the political process. When we (as academics) are better able to identify the many distributional, revenue, and growth effects of tax reform, we essentially mobilize the losers, and this means that the political process becomes immobilized because the losers from any potential change (no matter how small) lobby to prevent the change from occurring. Further, as our research has advanced, we are better able to identify the losers (and the winners) from any changes. For example, we now know that investment subsidies decrease the returns to existing capital; that is, they punish “old capital.” Even reforms that limit the estate and gift taxes would adversely affect nonprofits unless there were some offsetting subsidies. With more models at our disposal, we are increasing the odds that we can find some model under which some losers can be identified. Since the potential losers are the ones who will make the loudest noise, we may just be expanding the class of protesters.
Of course, from an academic perspective, being able to identify the losers and the winners is obviously a plus, and we should never abandon our efforts to examine and quantify these effects. However, from the broader public policy perspective, we wonder whether we are making things better or worse. Put differently, the case can be made that the “big” policy changes that have occurred over the years have occurred because of a broader perspective that, although there are losers and winners, the overall thrust of policy change is beneficial. When we are better able to identify the losers and the winners, it often makes it harder to assemble the coalition necessary to pass the reforms.
However, tax reforms have been enacted at various times and in various states or countries. We conclude with what we believe are the main lessons from successful tax reforms around the world.
First, comprehensive reforms are often better than piecemeal reforms. Now this goes somewhat against much conventional wisdom, and there are certainly risks that disaster can result when a system is shocked too much from a comprehensive reform. Even so, under many conditions, comprehensive reform can work because of the following considerations:
Everyone recognizes that the system is broken.
The government and the taxpayers need time to absorb the shock, and the time involved in the discussion of comprehensive reform often allows this.
There is sufficient time for the tax administration to absorb the changes.
Comprehensive reforms ensure that the separate pieces of the reform fit together, so that the prices are “right.”
Comprehensive reforms ensure that everyone gains—and everyone loses—from some or another specific change, which increases the political likelihood of passage.
Comprehensive reforms ensure that everyone recognizes that the system of taxation is “broken” and needs to be fixed.
Comprehensive reforms ensure that the momentum of reform is maintained.
Comprehensive reforms ensure that the gains (and losses) are large enough for taxpayers to actually see.
Second, timing is important. The best time for comprehensive reform is—paradoxically—often in bad economic times, since this ensures that everyone’s attention is focused and that everyone recognizes the necessity of tax reform. The best time is also—not surprisingly—when there is no election looming and when the economy is performing.
Third, in many instances, base broadening is consistent with adequacy, equity, and growth concerns. Base broadening can obviously increase tax revenues. The elimination of tax preferences and more generally the broadening of the tax base can also improve both vertical and horizontal equity by ensuring that the wealthy pay their “fair” share and that equals are treated equally. Finally, base broadening reforms can enhance economic growth by improving efficiency and by reducing the incentives for income shifting activities.
Fourth, empirical analysis is often difficult but is crucial, for determining the details of any reform and in selling any reform. Data are often problematic, even in the United States, and, as we emphasized earlier, quantifying the winners and losers can mobilize the political opposition. Even so, it is essential to try to quantify the effects of tax reforms. Such quantification is especially crucial in determining the distributional effects of tax reforms.
Fifth, the administrative dimension is important, but it is necessary first to get the policy “right” before dealing with administrative problems. After all, if the reform stops with administration (e.g., no-return filing), and leaves poor policy still in place, the result will still be poor policy.
Sixth and relatedly, reforms must consider both implementation and transition issues. Any major tax change can involve the creation of a new administrative structure. A comprehensive reform can have massive impacts on asset values. Any reform also creates transitional winners and losers. These transition issues cannot be ignored.
Seventh, tax reform should pay attention to the intergovernmental dimension. Most often, the reform effort focuses exclusively on the central government. However, fixing things at the central government while leaving in place poor tax policies at the subnational level will often compromise the goals of reform. In the US context, the impact of federal tax reform on state and local taxes is a crucial consideration, especially if a VAT was introduced at the federal level and one that is not typically given appropriate consideration.
Eighth, tax reform must recognize the impact of globalization. Our current tax system was originally designed for a world in which production and consumption were primarily of tangible goods, in which the sale and consumption of these goods generally occurred in the same location, and in which the factors of production used to make the goods were for the most part immobile. In such a world, taxation was a fairly straightforward exercise: sales and excise taxes could be imposed by the government in the jurisdiction in which consumption (or production) occurred, income taxes could be imposed on factors where they lived and worked without fear that taxes would drive the factors elsewhere, and a government in one jurisdiction had no need to consider how its actions would affect the governments in other jurisdictions because tax bases were largely immobile. Globalization changes things, and changes them dramatically: globalization implies that tax bases are significantly more mobile; it implies that the measurement, identification, and assignment of tax bases are much more difficult; and it implies that the ability of any government to choose its tax policies independently of those in other countries is greatly curtailed. Such international issues must be incorporated in any tax reform discussion.
Ninth, there is no one-size-fits-all tax reform. Any tax reform must consider the institutions, the traditions, the economic policies, and especially the politics of the current situation.
Finally, tax reforms must recognize and balance the trade-offs. Any reform must balance adequacy considerations with equity considerations with growth/equity considerations. These are the essence of our fiscal trilemma. In recognizing and balancing these trade-offs, we need to remember that there is much that we do not now—and probably cannot—know.
The articles in this special issue provide many creative ideas from the frontier of public finance research for the raw material of tax reform. However, we now need sophisticated political entrepreneurs who can find a way to combine these ideas into a package and convince the public that the time is right for serious reform. The danger we face is that the pressing demands for revenue increases—particularly if we begin to face debt crises—could lead us into ill-advised policies that raise revenues and simultaneously sharply increase distortions. Thinking through the path to solve the fiscal trilemma now can help us prevent this unenviable future.