Economic theory suggests that consumption taxation is economically superior to income taxation, with simulations suggesting that the complete replacement of the U.S. income tax system by a consumption tax would increase long-run output by several percent. Every other industrialized country raises a significant part of its revenue from consumption taxation, as do most of the U.S. states. A shift from income to consumption taxation in the federal tax system therefore warrants careful consideration.
In this chapter, we explain the economic advantages of consumption taxation, emphasizing how it promotes economic efficiency by removing the income tax's penalty on saving and investment.
The primary economic advantage of consumption taxation is that, unlike income taxation, it does not penalize saving. The savings penalty, which is a penalty on late consumption and early work, causes economic inefficiency.
Penalty on Late Consumption. We illustrate the income tax's penalty on saving with an example drawn from Carroll, Viard, and Ganz (2008). Consider two individuals, Patient and Impatient, each of whom earns $100 of wages today. Impatient wishes to consume only today; Patient wishes to consume only "tomorrow," which is decades later than today. Savings are invested by firms in machines that produce output tomorrow. The marginal rate of return on machines—the additional return available if one more machine is constructed—is 100 percent. If financial markets are competitive, the rate of return that firms pay to savers must be equal to the marginal rate of return on machines.
In a world with no taxes, Impatient consumes $100 today. Patient lends her $100 of wages to a firm, which buys a machine that yields the 100 percent marginal rate of return and therefore provides a $200 payoff tomorrow. The firm pays Patient back her $100 loan with $100 interest, allowing her to consume $200 tomorrow.
What happens in a world with a 20 percent income tax? Impatient pays $20 tax on his wages and consumes the remaining $80, which is 20 percent less than he consumed in the no-tax world. Patient also pays $20 tax on her wages and lends the remaining $80 to the firm. On her $80 loan, she earns $80 interest and is therefore repaid $160 by the firm. However, a $16 tax is imposed on the $80 interest. Patient is left with $144 to consume tomorrow, which is 28 percent less than the $200 she consumed in the no-tax world.
The income tax has reduced Patient's consumption by 28 percent, compared to a mere 20 percent reduction in Impatient's consumption. Under the income tax, Patient faces a higher percentage tax burden than Impatient solely because she consumes later. In other words, she is penalized because she saves for future consumption rather than engaging in immediate consumption. Another way to understand the penalty is to note that the income tax reduces the after-tax rate of return on saving. Because Patient sacrifices $80 of consumption today to obtain $144 tomorrow, she receives an 80 percent after-tax return, which falls short of the 100 percent before-tax return.
In contrast, consumption taxation yields a neutral outcome if the tax rate remains constant over time. For simplicity, consider a 20 percent consumption tax that is imposed directly on individuals, with the tax being applied to income minus saving (or plus dissaving). This tax can be viewed as a personal expenditures tax, a tax that we will discuss in chapter 2. Although the X tax has a different design, we will verify in chapter 2 that it produces the same results when applied to this example.
After earning $100 of wages, Impatient consumes $80 and pays $20 tax, the same outcome as under the income tax. Patient lends her entire $100 to the firm and owes no tax because she has not yet consumed; she reports $100 of income, with an offsetting deduction for $100 of saving. On her $100 loan, she earns $100 interest, accumulating $200. She consumes $160 tomorrow and pays $40 tax; her tax is 20 percent of $200, equal to her $100 interest income plus $100 of dissaving. Each worker's consumption is reduced by 20 percent relative to a world with no taxes. Because both workers face the same percentage tax burden, the consumption tax does not distort the choice between current and future consumption.
The neutrality of this constant-rate consumption tax is confirmed by the fact that Patient earns an after-tax return of 100 percent on her savings, identical to the before-tax rate of return. When Patient makes the $100 investment, she gives up only $80 of consumption today; if she had not invested, she would have paid $20 tax and consumed only $80. Her sacrifice of $80 today provides her with $160 of consumption tomorrow, a 100 percent rate of return.
Because the after-tax rate of return is equal to the before-tax rate of return under the consumption tax, the effective marginal tax rate on saving is zero. In contrast, the income tax imposed a 20 percent effective marginal tax rate on savings, because the 80 percent after-tax rate of return was 20 percent lower than the 100 percent before-tax return.
The example assumes that the consumption tax rate remains constant over time. Consumption taxation ceases to be fully neutral if the tax rate varies over time; it penalizes saving if the tax rate rises over time and rewards saving if the tax rate falls over time. It is important to realize, though, that the income tax inescapably penalizes saving, even if the tax rate remains constant over time.
Penalty on Early Work. So far, we have described the income tax's penalty on saving as a penalty on late consumption. But it is also a penalty on early work, as can be seen from a variant of the above example. Consider two other individuals, Young Worker and Old Worker. Young Worker earns $100 of wages today, and Old Worker earns $200 of wages tomorrow. As before, the interest rate between today and tomorrow is 100 percent. Both Old Worker and Young Worker wish to consume only tomorrow. In a world with no taxes, Young Worker saves her $100 of wages, earns a $100 return, and consumes $200 tomorrow. Old Worker earns $200 of wages tomorrow, which he immediately consumes.
What happens with a 20 percent income tax? Young Worker pays $20 tax on her wages today and saves the remaining $80. She earns an $80 before-tax return, on which she pays $16 tax, and consumes $144 tomorrow. In contrast, Old Worker pays $40 tax tomorrow on his $200 of wages and consumes $160. The results fit the previous pattern, as the individual who saves (in this case, Young Worker) is hit with a 28 percent tax burden, while the individual who does not save (Old Worker) bears only a 20 percent tax burden.
As before, consumption taxation with a constant 20 percent rate results in neutral treatment. Young Worker saves her $100 of wages and earns $100 interest, which allows her to consume $160 tomorrow after paying $40 tax. With his $200 of wages tomorrow, Old Worker also consumes $160, after paying $40 tax.
Saving occurs when individuals consume later than they work. The income tax's penalty on saving therefore creates artificial incentives both to consume earlier and to work later.
Understanding the Penalties. Because the heavier tax on saving under the income tax arises from the imposition of two taxes—one on wages and one on the return to savings—it is sometimes referred to as the "double taxation of saving." Others object to the double-taxation terminology, arguing that no single event is taxed twice, as the earning of wages and the receipt of interest income are separate events.
Fortunately, we need not resolve this semantic dispute. The relevant economic reality is that income taxation places a higher effective tax rate on future consumption than on current consumption and on current work than on future work. It is irrelevant whether the higher effective tax rate arises from one event being taxed twice or from two events being taxed; it is even irrelevant that it arises from the collection of two taxes rather than from a single larger tax. All that matters is that the tax burden on future consumption and current work is larger, in percentage terms, than the tax burden on current consumption and future work.
It is sometimes thought that neutrality is attained through equal taxation of all income, whether from capital or labor, as occurs under a well-designed income tax. But that is not the case. Economic neutrality requires uniform taxation of all uses of resources, not of all income. Although someone who consumes later than she works earns additional income, that fact provides no justification for imposing additional tax.
Ending the income tax penalty on saving would improve economic efficiency and promote simplicity.
Efficiency Gains. By eliminating the penalty on saving, consumption taxation offers efficiency advantages. The consumption tax's neutral treatment of work and consumption at different dates allows individuals to choose the most efficient allocation of work and consumption across their lifetimes. Under broad assumptions about preferences, consumption taxation involves lower deadweight loss or excess burden than income taxation, as the incentive effects of consumption taxation prompt households to work later and consume earlier, thereby increasing saving.
A common argument holds that, for any given amount of revenue, consumption taxation imposes a heavier tax on work than does income taxation. Because consumption is smaller than income (in an economy with positive saving), a consumption tax generally requires a higher statutory tax rate than an income tax to raise the same revenue. Proponents of this argument assert that this higher tax rate increases work disincentives. Moving from income to consumption taxation is said to amplify work disincentives even as it eliminates saving disincentives, with the net impact on economic efficiency reflecting a trade-off between these two effects.1
Although this argument has superficial appeal, a deeper examination reveals it to be invalid, as explained by Auerbach (1997), Bankman and Weisbach (2006, 1417–30), Viard (2006), Toder and Reuben (2007, 103), Shaviro (2007b, 759–60), Weisbach (2007), and others. A revenue-neutral move to consumption taxation does increase the tax rate on working to consume today. But it reduces the effective tax rate on working to consume tomorrow, which, under income taxation, is hit with both a wage tax and a tax on income from saving. The revenue-neutral shift therefore leaves overall work disincentives roughly the same under consumption taxation as under income taxation. We explain this point further in the "Trade-off Fallacy" box (pages 18–19).
A long-standing body of literature considers the role of consumption taxation in an optimal tax system. The starting point is the result of Atkinson and Stiglitz (1976), which establishes that different consumer goods should be taxed at a uniform rate when consumers' choices among the various goods are separable from their decisions about how much to work. If this separability condition holds, uniform taxation of consumer goods is desirable even if policy makers place a strong emphasis on redistribution, because such redistribution can be advanced more efficiently by increasing tax rates on all goods rather than by singling out some goods for higher taxes than others. As Kaplow (2008b, 221–24) and others have explained, this result can be applied to the choice between income and consumption taxation by treating consumption at different dates as different goods. If consumers' choices about when to consume are separable from their decisions about how much to work, then the Atkinson-Stiglitz result states that consumption at different dates should be taxed at a uniform rate. As we demonstrated in the Patient-Impatient example, consumption taxation with a constant tax rate achieves such uniformity, whereas income taxation does not.
Of course, the separability condition is unlikely to hold precisely, and the stylized Atkinson-Stiglitz model omits some relevant features of real-world tax policy. In a more general framework, it may well be optimal to deviate from a policy of taxing consumption at all dates at exactly the same rates. But such deviations are likely to be minor and difficult to identify precisely. Moreover, the deviations may go in either direction; depending on various factors, it may be optimal to tax late consumption at either slightly lower, or slightly higher, rates than early consumption. The general analysis therefore offers little support for abandoning the uniformity achieved by consumption taxation and adopting the income-tax policy of imposing markedly higher tax rates on late consumption. Kaplow (2008b, 225–48), Bankman and Weisbach (2006), and Auerbach (2008) discuss these issues.
A number of economic simulations report substantial long-run economic gains from replacing income taxation with consumption taxation, although the size of the gains is sensitive to economic assumptions and to the design of the reform. We defer a discussion of the magnitude of the gains from reform to the concluding chapter.
Simplicity Advantages. As Slemrod (1995), Edwards (2003), and others emphasize, consumption taxation also offers powerful simplicity gains. Whereas consumption taxation requires the measurement of consumption, which is relatively observable, income taxation faces the inherently more complex task of measuring the return on saving and investment to determine the change in the household's wealth.
Income taxation must either measure accrued gains and losses or (like the current tax system) defer taxation until gains and losses are realized, a policy that penalizes asset sales and requires the tracking of cost basis. A host of tax shelters seek to realize losses without realizing associated gains or to manipulate the allocation of cost basis across assets. Income taxation must also distinguish between principal and interest on loans, requiring complicated original-issue-discount, market-discount, and imputed-interest rules. Firm-level income taxation requires rules on capitalization and amortization, depreciation, and inventory accounting. Because these complexities are inescapably required by income tax principles, they cannot be avoided even in well-designed income tax systems. All these complexities are eliminated under consumption taxation.
The U.S. Supreme Court has recognized the inherent complexity that the income tax system faces in distinguishing between capital expenditures that should be amortized and current business expenditures that should be immediately deducted. In 1933, in an opinion by Justice Benjamin Cardozo, the Court said, "One struggles in vain for any verbal formula that will supply a ready touchstone. The standard set up by the statute is not a rule of law; it is rather a way of life. Life in all its fullness must supply the answer to the riddle" (Welch v. Helvering, 290 U.S. 111, 115 [1933]). Consumption taxes allow all business expenditures to be immediately deducted, avoiding any need to seek guidance from "life in all its fullness."
To be sure, no tax can be completely simple in a complex economy. Consumption taxation retains many complexities that are present under income taxation, including the need to distinguish between consumption and costs of earning income and the need to measure the consumption services provided by financial intermediaries. But consumption taxation removes the complexities discussed above without introducing any significant complexities that are absent under income taxation, except a few complications arising from the fact that tax rate changes are more disruptive under consumption taxation than under income taxation.
Of course, any tax can be complicated by poor design. It is surely true that any actual consumption tax will be more complex than a textbook consumption tax. By the same token, though, the actual income tax is vastly more complex than a textbook income tax, adding many extraneous complications (such as the distinction between debt and equity and the distinction between corporate and noncorporate firms) to the unavoidable income-tax complexities listed above. There is no reason to expect more such extraneous complications under a consumption tax than under the current income tax.
A balanced comparison reveals that consumption taxation is significantly simpler than income taxation.
Fallacious Arguments. The real advantages of consumption taxation, as outlined above, are sometimes overshadowed by invalid arguments offered by some consumption tax supporters.
One argument assumes that consumption taxes can and will be imposed on imports and rebated on exports and concludes that a switch to consumption taxation permanently reduces the trade deficit by making domestic producers more "competitive." Despite its superficial appeal, economists have long recognized that this argument is invalid, as we will discuss in chapter 7. As explained in that chapter, we actually recommend that the X tax not be imposed on imports or rebated on exports.
Another argument holds that consumption taxes do not penalize work because tax is triggered by consuming rather than working. This claim is invalid, because individuals work in order to consume, either in the present or the future. Accordingly, as Metcalf (1996, 99), Cnossen (2009, 690), and others note, income and consumption taxes both penalize the decision to work and consume rather than to enjoy leisure. As discussed above and in the "Trade-off Fallacy" box (pages 18–19), the two tax systems impose roughly the same penalty on work at any given revenue level. Consumption taxation removes income taxation's penalty on saving, but not its penalty on work.
Still another argument holds that consumption taxation is desirable because it taxes people on what they take out of the economy rather than what they put into the economy. It is not entirely clear what this means, because budget constraints require that what each individual takes out of the economy be equal, in present discounted value, to what the individual puts into it. The actual flaw of income taxation is that it induces economic inefficiency by penalizing people for taking late consumption, rather than early consumption, out of the economy and for putting early work, rather than late work, into the economy.
A final argument holds that consumption taxation is less prone to evasion than income taxation. In reality, both systems offer the same fundamental opportunities and incentives for evasion, including the use of cash transactions and personal expenditures disguised as business expenditures. The level of evasion is likely to vary to some extent across different forms of consumption and income taxes, but there is no inherent reason why it should be smaller under consumption taxation.
These fallacies should not obscure the real efficiency and simplicity advantages of consumption taxation.
Some observers have noted that the current U.S. income tax system has certain features that alleviate the tax burden that a pure income tax would impose on saving. As detailed below, many of these features are similar to those found in consumption tax systems. The presence of these features in today's tax system may suggest that today's system already provides many of the advantages of consumption taxation, thereby diminishing the urgency of a full-fledged switch to consumption taxation.
The actual policy implications are somewhat different. Although the current tax system includes some features that resemble consumption taxation, these features provide only limited relief from the problems of income taxation, and they add problems of their own. Moreover, other features of the tax system actually increase the tax burden on saving beyond that which would be imposed by a pure income tax system.
One consumption-tax feature of today's tax system is the provision of tax-preferred savings accounts and retirement plans. Reflecting a fundamental ambivalence about taxing saving, Congress has created more than twenty types of tax-preferred savings plans and accounts, each of which is subject to different contribution limits, eligibility rules, and restrictions on withdrawals.2 The National Taxpayer Advocate (2004, 423–32) and the Joint Committee on Taxation (2001, 149–54, 163) document the complexity that these accounts have added to the tax system. The Advocate noted a study in which 30 percent of workers choosing not to participate in 401(k) plans cited complexity as the principal reason.
While adding complexity, these accounts play only a limited role in reducing the tax burden on saving. Although the accounts shield about 36 percent of household financial assets from taxes, according to the President's Advisory Panel on Federal Tax Reform (2005, 22), they probably do much less to promote saving than would a 36 percent across-the-board reduction in tax rates on saving. Tax-preferred accounts are inferior ways to encourage saving because taxpayers can reduce their tax liabilities by shifting money from taxable to tax-preferred accounts without doing additional saving. Also, the accounts offer no marginal incentive to save for those households that bump up against the maximum contribution limits, who are likely to be the households in the best position to increase their saving.
Sheppard (2011) and others have pointed to the availability of tax-preferred accounts to argue that the current tax system functions as a consumption tax for most Americans. In reality, though, the current system does not offer the simplicity advantages of consumption taxation to most Americans, as they can avoid taxes on their saving only if they comply with the intricate restrictions governing tax-preferred accounts. Moreover, the current system does not function like a consumption tax for those Americans who do most of the saving, so the economic advantages of consumption taxation remain largely unattained.
The income tax system also allows some investments to be expensed, which, as we will explain in chapter 2, effectively removes the marginal tax burden on those investments. Other investments receive accelerated depreciation, which reduces, but does not eliminate, the marginal tax burden. Also, as we will discuss in chapter 9, imputed rent on owner-occupied homes and consumer durables is not taxed, which is consistent with consumption-tax principles. But these provisions are selective and limited in scope, inefficiently favoring some investments over others. Similarly, the deferral of tax on capital gains until they are realized lowers the effective tax rate on gains but introduces a new distortion by encouraging asset holders to postpone asset sales.
Meanwhile, other features of the tax system amplify the basic saving penalty imposed by the income tax. Notably, the corporate income tax imposes an additional tax burden on investment done through C corporations, although the additional burden is partly offset by preferential individual income tax rates for dividends and capital gains. Estate and gift taxes also impose an additional tax burden on saving to pass wealth on to the next generation. At the state and local levels, individual and corporate income taxes, property taxes, sales taxes on capital goods, and inheritance taxes further penalize saving; because our proposal does not directly alter the state tax system, it may not eliminate these disincentives. Also, the current tax system does not correct for inflation in its measurement of the income from saving. For example, a taxpayer who receives 5 percent interest, when inflation is 2 percent, has real interest income of only 3 percent, but the full 5 percent of nominal interest income is taxed. The tax system also fails to correct for inflation in the measurement of taxable capital gains and depreciation allowances, which also amplifies the tax penalty on saving and investment. As Hubbard, Skinner, and Zeldes (1995) observe, the means tests used in many transfer payment programs also create saving disincentives by reducing or eliminating benefits based on the asset holdings of potential recipients; as we will discuss in chapter 4, some of these disincentives will remain in place under our proposal.
Recent laws and proposals point to an increased tax burden on saving in upcoming years. Under the March 2010 health care reform law, a new 3.8 percent Unearned Income Medicare Contribution tax on interest, dividends, and capital gains received by high-income households is slated to take effect in 2013. In late 2011, five bills that would have imposed surtaxes, ranging from 0.5 to 5.6 percent, on the adjusted gross income, including interest, dividends, and capital gains, of millionaires received majority support in the Senate, but did not win the sixty votes required for passage. President Barack Obama has also proposed that the 2001 and 2003 tax cuts be allowed to largely expire for high-income households at the end of 2012, which would increase marginal tax rates on those households' capital incomes by several percentage points (more than twenty percentage points for dividends). Moreover, the projected growth in federal spending over the upcoming decades, detailed by the Congressional Budget Office (2011b), is likely to create pressure for additional taxes on saving, pressure that may be difficult to forestall unless the United States makes a full-scale move to consumption taxation.
In summary, the current tax system includes some selective, complex, and ineffective features that ease the tax penalty on saving and other features that actually amplify that penalty. A full consideration of the current system makes clear that a move to consumption taxation is the only effective way to address the penalty on saving.
Economic theory suggests significant economic gains from moving toward consumption taxation. In this book, we explain how a particular type of consumption tax, the Bradford X tax, offers a progressive and relatively simple form of consumption taxation, and we recommend that the United States adopt this tax.
Except in chapter 10, we focus on proposals to adopt a consumption tax as a complete replacement for the federal taxes that penalize saving and investment, namely, the individual and corporate income taxes (including the individual and corporate alternative minimum taxes), the estate and gift tax (including the generation-skipping tax), and the Unearned Income Medicare Contribution tax. Complete replacement offers larger efficiency gains than partial replacement. And only complete replacement offers real simplicity gains; if the income tax remains in place as even part of the overall tax system, its complexity is still present.
We do not propose, however, to replace other federal taxes that pose little or no penalty on saving. As we will explain in chapter 4, we propose that the Social Security and Medicare payroll taxes be maintained; as we will explain in chapter 5, we recommend that the self-employment tax system be folded into the payroll tax system. We do not propose that excise taxes or customs duties be modified as part of the move to consumption taxation.
We reject the notion of allowing taxpayers to choose between the current income tax system and the new X tax system. Allowing taxpayers a choice between tax systems preserves some of the complexity of the current tax system and also makes it more difficult to meet revenue targets. Moreover, a consumption tax system generates the proper incentive effects only if all taxpayers are subject to it in all years.
In chapter 2, we survey the different types of consumption taxes and explain our preference for the X tax.
BOX
THE TRADE-OFF FALLACY
The trade-off fallacy can best be understood through an analogy. Consider an economy in which people choose between leisure, apples, and oranges. The "general" tax rate is 20 percent, but oranges are subject to an 8 percent surtax, so apples are taxed at 20 percent and oranges at 28 percent. Economists note that the surtax inefficiently favors apples over oranges and propose moving to a uniform 24 percent tax, which is revenue neutral if half of wages are spent on each fruit. A critic acknowledges that the proposed reform eliminates the bias in favor of apples over oranges, but contends that it increases the bias in favor of leisure over work by raising the general tax rate from 20 percent to 24 percent. The critic perceives a trade-off between increased efficiency in the fruit market and decreased efficiency in the labor market.
The critic's analysis is flawed. Because apples and oranges are both alternatives to leisure, work disincentives depend on the tax rates on both fruits; the oranges surtax, no less than the general tax, penalizes work. It is true that the proposed reform increases the bias in favor of leisure relative to apples by raising the tax rate on apples from 20 percent to 24 percent. At the same time, though, the proposal reduces the bias in favor of leisure relative to oranges by lowering the tax rate on oranges from 28 percent to 24 percent. The net effect on work incentives is ambiguous.
Similarly, in the Patient-Impatient example, a 20 percent income tax imposes a 20 percent tax rate on consumption today and a 28 percent tax rate on consumption tomorrow, effectively imposing an 8 percent surtax on the latter. Just as the oranges surtax increased the bias in favor of leisure over oranges and thereby added to work disincentives, so this surtax increases the bias in favor of leisure over consumption tomorrow and thereby adds to work disincentives. A person who works to consume tomorrow must pay both the wage tax and the tax on capital income, and both taxes therefore discourage work.
If half of wages are consumed today and half are saved, replacing the 20 percent income tax with a 24 percent consumption tax is revenue neutral in present-value terms. The tax rates on consumption today and consumption tomorrow are then both 24 percent. How are work disincentives affected? Although the bias in favor of leisure relative to consumption today rises from 20 percent to 24 percent, the bias in favor of leisure over consumption tomorrow falls from 28 to 24 percent. The net effect on work incentives depends on consumer preferences.
If the two types of consumption are equally complementary to leisure, the revenue-neutral switch to consumption taxation has no net effect on work incentives. If consumption today is relatively complementary to leisure, a revenue-neutral move to consumption taxation actually reduces work disincentives and increases labor supply. On the other hand, if consumption tomorrow is relatively complementary to leisure, a revenue-neutral move to consumption taxation increases work disincentives and reduces labor supply. In any case, the net change in work disincentives is likely to be small. The largest efficiency effect of the switch to consumption taxation is the gain from the removal of the saving disincentive.
In summary, the trade-off argument is invalid. Although taxing income rather than consumption permits a lower tax rate on wages, doing so does not substitute a saving disincentive for a work disincentive. Instead, it adds a saving disincentive while maintaining a roughly unchanged work disincentive. Under almost all types of consumer preferences, income taxation is less economically efficient than consumption taxation.