Menu
Tax Notes logo

GRAVELLE EXAMINES VAT, FLAT TAX IN CRS REPORT.

APR. 27, 1992

92-386 S

DATED APR. 27, 1992
DOCUMENT ATTRIBUTES
  • Authors
    Gravelle, Jane G.
  • Institutional Authors
    Congressional Research Service
  • Index Terms
    VAT
    rates, flat
    income distribution
    tax policy, progressivity
  • Jurisdictions
  • Language
    English
  • Tax Analysts Document Number
    Doc 92-3697
  • Tax Analysts Electronic Citation
    92 TNT 91-8
Citations: 92-386 S

JANE G. GRAVELLE SENIOR SPECIALIST IN ECONOMIC POLICY OFFICE OF SENIOR SPECIALISTS

April 27, 1992

SUMMARY

There has been an ongoing discussion of the possibility of some major tax change or some major new source of revenue, including proposals for a VAT or a business transfer tax, or for flat rate taxes. The issues surrounding these proposals typically relate to the effects on tax simplicity, on economic efficiency and savings, and on income distribution.

These basic issues, in turn, depend on two factors. First, what is the base of the tax -- income, consumption, or wages? Secondly, what is the rate structure (is it graduated or flat?).

The complexity of the tax system is negligibly affected by the choice of a flat rate or a graduated rate, since the consumption of the tax is typically quite simple. Rather, the complexities arise from the measurement of the tax base, and the record-keeping necessary to arrive at that measure. All of the tax bases can present some elements of complexity.

Despite claims that shifting from an income to a consumption tax base will increase saving, this effect is by no means clear. Economic evidence does not strongly support the presumption that private savings can be affected by eliminating the tax on capital income.

The distributional consequences of a tax depend on both base and rate. Consumption taxes are generally viewed as burdening the poor relative to other types of tax bases, because lower income individuals have high consumption to income levels. There are some debates about how important these differences across the tax bases are. The rate structure progressivity, however, clearly influences the distributional consequences of the tax. The income tax is quite progressive, both because of the graduated rate structure and the allowance of a refundable earned income credit. Replacement of this graduated tax with a flat rate of any type would redistribute income from the poor to the rich.

                              CONTENTS

 

 

A CLASSIFICATION OF TAXES

 

     TAX BASE: INCOME, CONSUMPTION, OR WAGE

 

     THE RATE STRUCTURE

 

 

SIZE OF THE TAX BASES

 

 

WHAT MAKES A TAX COMPLICATED?

 

 

ECONOMIC ISSUES

 

     SAVINGS, ASSET VALUES, AND ECONOMIC EFFICIENCY

 

     DISTRIBUTIONAL EFFECTS ACROSS INCOMES

 

 

APPENDIX: DIFFERENT MEANS OF ACHIEVING A CONSUMPTION TAX

 

 

NEW TAX PROPOSALS: FLAT, VAT, AND VARIATIONS

There has been an ongoing discussion of the possibility of some major tax change or some major new source of revenue. The Congressional Budget Office recently released a study of the effects of adopting a value added tax (VAT). The Treasury Department released a study of corporate tax integration, which would constitute a substantial change in taxation. There have been proposals in Congress for adopting what is termed a business transfer tax, as a substitute for corporate and payroll taxes. Presidential contender Jerry Brown has proposed a flat tax which would replace all the current Federal tax sources. There have also been proposals by economists, including the flat tax discussed by Hall and Rabushka, and the x-tax proposed by David Bradford.

This paper explains how these different taxes proposals work and what the general economic consequences of them are. It stresses that the tax characteristics of rate structure and base are different aspects of the tax system. Thus to characterize a tax requires not just information on whether the rate structure is flat or progressive, but also information on the tax base. The analysis also shows that tax proposals that appear to be very different are, in fact, similar, while proposals that are referred to by the same name (e.g., flat taxes) can be very different.

A CLASSIFICATION OF TAKES

The basic economic effects (both allocative and distributional) of taxes depend on two factors -- what is the base of the tax, and what is the rate structure (is it graduated or flat?). For example, the Brown proposal has occasionally been compared to the Hall- Rabushka proposal, perhaps because they have the same name. Yet the two proposals are completely different. The Brown proposal is partially an individual income tax and partially a consumption tax, with both taxes levied at the same flat rate. The Hall-Rabushka plan is a consumption tax base, but it is not really flat because of the exclusions allowed. The base and the rate structure are entirely different issues, although some forms of consumption tax cannot avoid a flat tax.

TAX BASE: INCOME, CONSUMPTION, OR WAGE

There are three types of broad tax bases: income, consumption, and wage bases. All of these types of taxes are currently in place. Social security taxes are based on wages; both Federal and State and local governments impose income taxes, and most States and many local governments have sales taxes, which are a type of consumption tax. 1

For a very broad based tax, the largest tax base would be income. Income in the economy can be added up two different ways -- through payments to productive factors (labor and capital) and through disposition of resources. Income is the sum of wage payments and the return to capital; it is also the sum of consumption expenditures and purchases of capital goods. These sums of income are known by the concept in the national income accounts of Net National Product (NNP); it differs from Gross National Product (GNP) because it is net of depreciation.

The income base is larger than the wage tax base because it includes capital income, and it is larger than the consumption tax base because it includes net purchases of capital. 2

Both wage and consumption taxes have the effect of eliminating the tax burden on new capital; this result is one reason for interest in the tax bases. Wage and consumption taxes are related; the consumption tax is the sum of a tax on wages and a tax on the returns to existing capital. Indeed, one can derive the consumption tax base by subtracting from income net purchases of capital goods.

It is quite possible to engage in something that looks like one kind of tax and really is another. For example, one could transform the income tax into something that is essentially a consumption tax, even though it would still look very much like an income tax, through the following steps:

(1) Businesses would suspend depreciation and inventory recovery and simply deduct all purchases of intermediate goods and capital goods instead. These purchases would be larger than the prior deductions by the amount of net investment, causing the base to fall.

(2) Interest would be neither deducted nor included in income. Dividends and capital gains would no longer be taxed, and no itemized deductions would be allowed.

Such a system would still look much like an income tax, but it would be, in fact, a consumption tax -- similar to the base for a sales tax or value added taxes. 3 Indeed, this approach, combined with substituting a flat rate schedule and some alterations in personal exemptions, is essentially the approach suggested by Robert E. Hall and Alvin Rabushka in their book The Flat Tax. 4 A similar approach (the x-tax) has been suggested by David Bradford. 5

If such a tax were designed with a flat rate structure, only one other step would be required to convert the tax base into a value- added tax (VAT): eliminate taxation of individuals completely, but do not allow businesses to deduct wages. This step would produce what is called a subtraction method VAT. The tax base for each business would be the value added by the business. That is, each business would pay a tax on the difference between gross receipts and purchases of goods from others. One more step -- applying a VAT rate to receipts but allowing, not a deduction for purchases, but a credit for the tax paid by suppliers, would transform the tax into a European style credit-invoice method VAT. And skipping the tax on intermediate suppliers and applying it only to final sales at retail (of goods and services) would convert it into a national sales tax.

Why do these approaches produce the same tax base? Again, it is appropriate to think in terms of national income identities. Total income is the sum of wages and returns to capital; a part of this income is set aside to make additional investments and a part is used to finance consumption. What the Hall-Rabushka and the x-tax approaches do is take the wage component of a consumption tax and assign it to be taxed to the wage earners, while taxing the return to old capital at the firm level. These indirect methods of getting to a consumption tax were suggested partly in order to introduce some progressivity into the rate structure, since it is possible to provide exemptions and progressive taxes for the wage component. They would also be less likely to create pressures for an overall price increase. By reducing the base by wages, the tax is much smaller, and the tax on wages is collected in the same manner as the income tax. Thus, the wage does not fall as it would under a VAT or business transfer tax with the assumption of a fixed money supply. Asset prices would, however, fall.

These different types of taxes are all applied to a consumption base, but they differ in a number of respects, as outlined in the appendix and discussed below under economic effects. The most important differentials are the effects on the market wage and the pressure to increase prices, and the ability or lack of ability to introduce progression in the rate structure.

To transform the income tax into a wage tax is more difficult in some ways and easier in others. To do that the corporate tax would have to be repealed and all rents, dividends, capital gains, and deductions and payments for interest would be eliminated. These steps are simple and would make the tax system less complicated. The most difficult problem is the measurement of wages for proprietors and active partners in unincorporated businesses, or in closely controlled corporations where wages and profits may not be readily distinguished. In order to measure the wage income in these circumstances, all the steps needed to measure income are necessary (depreciation, inventory accounting, etc.). In addition, there would need to be a method of dividing total net income from business between wages and capital income. The current payroll tax actually includes all income of these individuals in tax, which includes some capital income as well as wage income.

THE RATE STRUCTURE

Taxes are also differentiated by their rate structure. The income tax is progressive, because the rates rise as income rises. Certain types of tax structures cannot be graduated in rate because they are imposed indirectly on products or on intermediate entities rather than on individuals. For example, retail sales and value added taxes are not graduated. Nor is the corporate income tax which applies at the same rate (for large corporations) regardless of the circumstances of the owner of rights to corporate capital. 6 The social security payroll tax is a flat tax rate, but it need not be, as it could be graduated easily enough. Its graduation would have to be on an individual basis rather than a family basis, as in the income tax. The earned income tax credit may be viewed as an offset for the payroll tax for some individuals, although its mechanism is through the income tax.

In order to achieve rate progression, some or all of the tax must be levied directly on individuals or families whose overall financial circumstances can be judged. It is for this reason that some economists have proposed variants of the consumption tax that are collected not through indirect tax (VAT or retail sales) methods but which allow part of the tax to be imposed directly on families. By allowing businesses to deduct wages and individuals to be taxed on them (as in the Hall-Rabushka plan), rate progression or, at a minimum, some flat exemptions, can be allowed. Recalling that the consumption tax base is the sum of a tax on wages and a tax on old capital, the first part of the tax would be collected on individuals and could be graduated, while the second part would be collected at a flat rate on firms.

In theory, the income tax has the broadest tax. The consumption base is smaller, since consumption is smaller than income as it excludes net investment. The wage tax is the smallest. These relative sizes of the tax base will, however, be influenced by a number of adjustments that are either necessary for administrative purposes or are likely in practice to occur.

SIZE OF THE THE BASES

In order to understand the tax bases, it is useful to review some general aggregates in the economy. Gross National Product (GNP) is the sum of all of the output produced domestically plus the difference between income earned abroad and income paid to foreigners. (Gross Domestic Product is simply all output produced domestically). GNP is, however, not equivalent to total income, since some of GNP must be used to replace the capital stock that is wearing out. If these amounts are subtracted, the result is Net National Product (NNP), which is equivalent to the economist's concept of income. NNP in 1990 was 89 percent of GNP, indicating that about 11 percent of gross output goes to replace depreciating capital.

National income is NNP adjusted by certain factors, primarily indirect business taxes -- sales, excise, and property taxes. There are some other minor exclusions as well as a statistical discrepancy. National income is the closest overall concept to a full income tax base, and it is about 81 percent of GNP.

In theoretical models, consumption as a base is less than income by the amount of net investment, while wages are less than income by the amount of the return to capital. If the rate of return exceeds the growth rate, as it typically does, then the consumption base will exceed the wage base. Indeed, if we include government consumption in the total, consumption would be 94 percent of National Income, while wages would be 80 percent. In practice, however, these bases are affected strongly by certain exemptions and exclusions. In particular, a practical consumption base could be much smaller under certain types of tax structures, particularly a value added tax, because of the broad array of items which might be excluded from the tax.

The INDIVIDUAL INCOME TAX BASE under current law is reduced from 100 percent of national income to only 78 percent to arrive at adjusted gross income. (Taxable income differs from AGI because of personal deductions and standard and itemized deductions; taxable income is only 52 percent of national income.) Several items are added and subtracted from national income to arrive at adjusted gross income. AGI excludes worker fringe benefits (including the employer's portion of social security) along with earnings of pensions; these amounts are only partially offset by the inclusion of pension benefits and small amounts of social security payments. The base is also decreased by corporate profits taxes and retained earning, by earnings of other tax exempt or largely tax exempt organizations, and by imputed income (such as rents on owner occupied housing). It is increased, however, by capital gains. There are also some amounts of income missing for those not required to file a tax return, as well as errors and omissions, and some general discrepancies between tax based accounting and national income accounting (e.g. depreciation). The base for the individual income tax is also reduced because other Federal taxes are effectively deducted from income, including payroll, excise and corporate profits taxes (assuming the latter show up as dividends or capital gains). The base would probably be increased to about 86 percent of national income if these taxes were not in place. This base assumes that indirect business taxes imposed by State and local governments would still be deductible.

The CONSUMPTION TAX BASE, if one adds both private and government consumption, is 94 percent of national income. The base depends on how the tax is likely to be structured. This large base assumes that the consumption tax will apply first (prior to all other taxes including indirect business taxes). The type of base likely depends on the method used; it might apply roughly in this fashion with the mechanism of a subtraction method VAT or the method used in Hall-Rabushka, although even this broad tax base would not capture certain imputed items, such as imputed rent on owner occupied housing. If it is applied through the mechanism of a VAT or a national sales tax, the consumption base would probably be smaller. First, purchases of goods and services by the government may be excluded. Excluding government purchases reduces the consumption tax base to 73 percent of national income. The government might also make cost-of-living adjustments on transfers such as social security and welfare -- if this is done then the effective base falls to about 58 percent, in the sense that there will be an offsetting expenditure to index these payments. Finally, even a broad based consumption tax might exclude a variety of other items -- some probably necessary to exclude (such as imputed rent on owner occupied homes which cannot easily be measured or calculated) and some as "merit goods" such as expenditures on education, and medicine. Such exclusions would reduce the base to about 45 percent of national income. Some types of taxes might include further exemptions, such as food, making the base even smaller.

The WAGE TAX BASE is about 80 percent of national income. Like the income tax base, however, this tax base would be lowered by any exclusions of fringe benefits. These amounts are about 6 percent of national income, lowering the base to 74 percent. The effective base could, however, be larger if capital incomes of incorporated businesses are included as they are in the payroll tax; at the same time, the measurement of income for these businesses may be imperfect due to imperfections in the tax law.

As these illustrations indicate, the relative size of the different tax bases will depend on some specific design features and how closely the tax bases in practice conform to the underlying concepts.

WHAT MAKES A TAX COMPLICATED?

Some tax proposals have been advanced in part on the grounds that they will be simple. Some of the popularity of flat tax proposals appears to be based on the notion that they will simplify the tax law.

A graduated rate structure causes virtually no additional complexity in the tax law. Once individuals have arrived at the taxable income base, the vast majority of them can simply look up their tax in a table. Many individuals can actually have the Internal Revenue Service calculate their tax. Even if the individual has to use the rate schedule, however, the calculations are quite simple.

What is complicated about taxes for ordinary individuals, aside from having to file some sort of return in the first place, is arriving at the base, and in particular in measuring income and in taking deductions, exclusions, and credits which require record keeping. Some of these are straightforward. For example, most individuals receive reports which establish their interest and dividend income, and for individuals whose income is from these sources along with wages and who do not itemize their tax returns, the record-keeping requirements and computations are minimal. About 70 percent of returns do not itemize deductions, and many of these individuals have incomes primarily from wages, interest, and dividends.

Individuals may also receive some notification of their State income tax withholding and refunds, and in some cases mortgage interest and property taxes. For other items, such as charitable contributions, medical expenses, and so forth there must be record keeping, and in general itemizing deductions adds complications. There must also be record keeping for measuring capital gains (other than distributions from mutual funds which are generally reported). There are also a number of credits, in some cases aimed at low income individuals, which can be complicated. Child care credits and earned income tax credits are examples; the earned income credit was made considerably more complicated by legislation in 1990.

Individuals with incomes from business or from direct ownership of physical assets can have much more complicated circumstances. They have to measure depreciation, account for inventories, and they have to perform record-keeping requirements for their employees (income and payroll tax deposits). Large corporations can have extremely complicated tax returns, but, of course, will also have the resources to deal with complexity.

Calculating the base is probably easiest for an indirect consumption tax; a credit-invoice method VAT with exemptions and differential rates can, however, become quite complicated. But only businesses will have to fill out returns, and it is also possible to exempt small businesses from filing requirements without creating very much distortion in prices. This is especially the case when the businesses exempted are retail firms, since most of the tax on value added will have already been collected. The basic problem with a VAT is that it is usually proposed as a supplement, rather than a substitute, for income taxes. A subtraction method VAT might be somewhat easier since it is based on the same data as reported on the income tax, but it would be more distorting to exempt firms except at the final stage. A subtraction method VAT does not easily permit differential rates.

The Hall-Rabushka proposal is a consumption tax done through the subtraction method, but with wages also subtracted and taxed to individuals to allow some progressivity. Also pensions would be taxed only when received. This tax would be fairly simple to calculate, but all individuals as well as firms would have to file returns. It would be possible to graduate the tax on the wage portion on an individual (but not a family) basis and make the tax progressive without involving individuals in filing returns. Since the Hall-Rabushka proposal does not eliminate the payroll tax, noncorporate businesses would still have to calculate an income tax base, and, of course, all firms would still have to collect the payroll tax.

Taxing wages is relatively simple for employees. Indeed, it would be possible to require businesses to collect the tax just as they collect the payroll tax. Wage taxation is extremely complicated in the case of unincorporated businesses. For unincorporated businesses, one would have to reach the income tax base and then find some method of separating returns to capital from wages. One approach would be to tax all business profits as wages as under the current payroll tax. Or, these businesses could be taxed on a consumption tax basis as described above under the Hall-Rabushka tax, which would provide a mix of wage and consumption taxation. Or some fraction of income from the business (before the deduction of wages of other employees) could be designated as capital income, based on industry estimates, and excluded. This approach would not tax wages with precision, but might be fairer to individuals who run their own businesses than a system that taxes all income.

Income taxes can involve considerable complexity depending on how they are designed, although for many individuals who have simple forms of income and do not itemize or claim credits, income taxes are no more complicated than a wage tax. There is no way to dispense with the fundamental complications of accounting for depreciation and other forms of capital cost recover with an income tax.

The Brown proposal has been argued to simplify tax compliance; whether it would do so is not clear. Many taxpayers who currently use standard deductions (about 70 percent of the total) would have to itemize (and keep records) in order to deduct charitable contributions, rent, and mortgage interest, a move toward more complication. On the other hand, many other complicating features of the tax system (such as the earned income credit and child care credits) would be eliminated. Although corporations would be relieved of income tax accounting, unincorporated businesses would apparently continue to have to use this accounting. Firms would be relieved of payroll tax collections, but all firms would have to comply with the value added tax.

ECONOMIC ISSUES

SAVINGS, ASSET VALUES, AND ECONOMIC EFFICIENCY

One of the main arguments for switching to a consumption tax is to encourage additional savings, investment and productivity, and hence an increase in the future standard of living. A consumption tax does not tax the return to capital income. Neither, for that matter, does a wage tax.

What is perhaps less well understood is that a consumption tax, unlike a wage tax, causes the prices of assets (stock market values and the value of direct holdings of physical assets) to fall. Recall that a consumption tax is a tax on wages and old capital; when a tax such as a VAT is adopted, the consequence will be a fall in wages and asset prices. Of course, if the tax is matched by an increase in the money supply to hold nominal wages constant, then asset prices do not fall but the prices of consumption goods go up. In a simple model with no debt and no transfer payments, the real effects are identical since both asset owners and wage recipients will lose real purchasing power.

A major problem with this objective of increasing savings is that there appears to be no strong evidence that cutting taxes, or increasing the rate of return, causes individuals to save more. There is no a priori reason from economic theory to expect that increasing the rate of return will increase savings. If the rate of return goes up, individuals would wish to consume more in the future (because the "price" of future consumption in terms of foregone current consumption is less). This effect is called the price effect. This price effect might lead to an increase in savings to finance future consumption. It is, however, possible to consume more in the future and save less, since current savings will grow to a larger future amount. This effect is called the income effect. Indeed, it is quite easy to find examples where savings will decrease with a deficit financed tax incentive.

The effect on savings remains ambiguous even if the government finances the tax incentive with an offsetting tax, such as an increase in tax rates. The overall effects on savings depend on who receives the tax increases and tax cuts, and the sum of these income effects.

Since economic theory does not provide a clear answer to the effect of tax incentives on savings, researchers have turned to empirical evidence. A number of studies have examined the relationship between savings rates and after tax rates of return. Of the studies which have accounted for inflation and taxes, the first done by Boskin indicated a positive, but small relationship. 7 Subsequent studies have generally failed to find a statistically significant relationship. 8

Recently, there has been attention paid to constructing more sophisticated models of individual savings behavior which model individuals over the life cycle, account for many generations at one time, and can allow offsetting tax increases to achieve revenue neutrality. 9 One of the important implications of these life cycle models is that switching from an income to a consumption tax is more likely to increase savings than switching from an income to a wage tax. Indeed, switching to a wage tax may well decrease savings.

This presumption in favor of the efficacy of the consumption tax in increasing savings depends on certain stylized -- and uncertain -- aspects of these models affecting the distribution of tax burdens across the generations. In these life cycle models, older individuals are consuming at a high rate relative to income as they consume out of savings, while younger individuals are consuming at a lower rate because they are saving for retirement. A wage tax shifts income from the young to the old -- old individuals have little or no wage earnings. Thus, the wage tax falls more heavily on the savers and more lightly on the non-savers than an income tax. Since old individuals are dissaving, the shift in income towards them with a switch to a wage tax can cause the savings rate to fall via an income effect.

A consumption tax has the opposite effect. The old, who are dissaving to finance consumption, are burdened by the consumption tax relative to the income tax. This burden may occur either because consumption prices rise (with money accommodation to a VAT) or because asset values fall. The young, who are savers, are more lightly burdened by a consumption tax than an income tax. By shifting income from the old to the young -- and from dissavers to savers -- a consumption tax's distributional effects will reinforce the price effects that produce savings.

How relevant these results are to actual taxes which might be considered is uncertain. Unfortunately, the life cycle models tend to predict much larger savings responses than tend to be observed in the economy. This result may reflect use of responses to price which are too large and the difficulties in modeling the evolution of savings over time.

Moreover, these life cycle models are highly stylized, and there are several reasons to be skeptical that a consumption tax will actually increase savings.

First, the life cycle models do not disaggregate across incomes. All individuals are assumed to have the same lifetime incomes, and differ only because of age. The evidence indicates, however, that life cycle patterns may differ across income classes. Older individuals at higher income levels do not tend to dissave as suggested by the stylized models.

Second, these income effects across the generations may be moderated by the possibility of bequests and other private transfers across generations, which can be adjusted. Older individuals who experience declines in their asset values may, for example, react by leaving smaller bequests rather than reducing consumption.

Third, these models do not include government transfer payments which accrue to older individuals, and which may form a large part of the income of lower-income older individuals. If the money supply is not increased, or if these transfers are indexed, then the burden on older individuals who are more likely to dissave will be reduced and the income effects will be less likely to increase savings.

Finally, these models do not include financial claims to assets. Recall that the consumption tax can either be passed backward to wages and asset prices, or passed forward to consumption, depending on monetary accommodation. To the extent that the burden of a consumption tax is passed backwards in asset prices and wages, individuals who are financing consumption out of financial assets will not be burdened by the tax, since their claims will not lose value. Individuals with equity claims will experience the full reduction in the value of assets even though their equity claims are only part of the total value. Thus the proportional fall in the value of assets such as stock market shares will be much greater. If older individuals, or individuals who tend to be dissaving, are more likely to hold financial assets, the tax will burden them more lightly.

This differential between equity and debt holders occurs only when the tax is not accompanied by an increase in the money supply. If the money supply is increased to accommodate the price rise, as in a VAT, then asset values do not fall; instead product prices go up. The differences between debt and equity financing of consumption disappear. A fall in asset prices would be much more likely in the case of the Hall-Rabushka approach where the pressure to increase the money supply is much less, due to taxing wages at the individual rather than the firm level (see Appendix). Since Hall-Rabushka can be implemented with its main effect solely the reduction in the value of equity claims to capital (that is, a decrease in stock market prices and the value of physical assets), it may be less likely to increase savings than might be the case of a value added tax if dissavers are more likely to hold financial assets.

Even if substituting a consumption for an income tax increased savings, it is not clear that output would increase, or that economic efficiency would be enhanced. Since the consumption tax base is lower than the income tax base, the rates for a consumption tax would be higher, and thus the disincentive to work will be increased. The effects on output and on economic efficiency depend on the behavioral responses to those changes. The Congressional Budget Office concluded that any effects of choosing a VAT over an income tax surcharge would be quite modest. 10

Overall, the evidence of the effect of these tax substitutions on savings and on economic efficiency is unclear, even if the revenue is made up in other taxes. A more certain and more effective route to savings might be to reduce the deficit and increase government savings.

DISTRIBUTIONAL EFFECTS ACROSS INCOMES

Taxes tend to affect individuals with different incomes differently, depending on the tax base chosen, as well as the rate structure. To assess the consequences of taxes on individuals of different incomes, certain assumptions are typically made about the "incidence" of the tax. A tax can alter prices of consumption goods, wages, and asset prices and, through subsequent behavioral responses, aggregate savings, output, and rates of return. In general, consumption taxes are assumed to increase the prices of consumption goods, income taxes are assumed to fall on income, and payroll taxes are assumed to fall on labor.

Using these assumptions, if tax rates were flat, an income tax would be proportional (each family would pay the same fraction of income). A consumption tax base would be regressive (since lower income individuals consume a larger fraction of their incomes, and indeed may consume in excess of incomes). A payroll tax would impose lighter burdens at both the top and the bottom of the income scale. Some individuals at the lower end of the scale who are retired or receive transfer payments earn no wages, while those at the higher end of the scale have larger shares of income from capital. In practice, the income tax is quite progressive, due to its graduated rate structure, its exclusion of most transfer payments, and its allowance of earned income credits. The payroll tax is reasonably proportional across a wide range of incomes, although it falls below the average at the bottom and the top of the income scale, the latter effect due not only to the presence of capital income but to the ceiling on wages subject to the tax.

A value added tax tends to be quite regressive assuming the burden is passed forward in prices, although some relief can be obtained for the regressivity if transfer payments such as social security and welfare payments are indexed. Even with such indexing, the Congressional Budget Office reported that a $100 billion VAT would result in a 4.8 percent effective rate for the lowest quintile, and 3.2,2.8,2.3, and 1.5 percent for the remaining quintiles for an overall average of 2.2 percent. 11 This decline in rates reflects the higher ratio of consumption to income at lower income levels. By comparison, if the same amount of revenue were raised through a proportional increase in income taxes, the rates would be a 0.2 percent increase in the lowest quintile, and 0.7, 1.2, 1.6, and 3.0 percent respectively in the next four quintiles.

These dramatic differences between the two taxes reflect both the tax base, and the graduated rates. In addition, the tax burden under the income tax is lessened by the allowance of an earned income credit.

Some insight into the effects of major tax shifts can be seen by examining the effects of the Brown proposal, which would replace current taxes with a combination of a value added tax and an income tax with the same flat rate. According to Congressional Budget Office calculations /12/, the current income, payroll, and excise tax imposes taxes at 7.7 percent of income in the lowest quintile, and 15.2, 19.1, 21.7, and 26.7 for the next four quintiles, for an overall average of 23.1. The Brown proposal would result in tax rates of 27.7 percent in the lowest quintile, and rates of 24.5, 23.4, 22.7, and 21.7 for the next four quintiles. Thus, this proposal would involve a major shift in the distribution of the tax burden, moving from a fairly progressive structure to a regressive one.

Hall and Rabushka did not calculate any distributional tables for their proposal; nor did they provide precise estimates of the revenue costs. They suggested a 19 percent tax rate as necessary to replace individual and corporate income taxes. This rate did not take into account the loss in the tax base resulting from their exemptions; With exemptions, it would have to be considerably higher. Their proposal would replace the progressive corporate and individual income taxes with both a base (consumption) and a rate structure that is less progressive. Payroll and excise taxes would be retained. As a result the proposal will reduce the progressivity of the tax system. If the tax is modeled as a consumption tax with a wage credit, the tax would be regressive at some points, despite the exemptions, simply because the consumption tax base tends to be regressive and because lower income individuals tend to have little wage income. (Assuming an average exemption of $10,000, this type of calculation produced rates of 39 percent for the first quintile, and 25, 28, 27 and 18 percent for the next four quintiles for all Federal taxes.)

There are several different types of reservations about the standard findings of distributional studies. The first arises from the fact that these distributional studies rely on a "single snapshot" in time. Some individuals with low incomes are those who have been, or will be, higher income. Individuals may have low incomes because they are young, because they are retired or because they are in a transient period (temporary unemployment or business set backs). Indeed, there is another entire strand of distributional analysis which focuses not on income levels, but on age cohorts. In the case of the differences between the tax base, some economists would argue that it is the generational distribution issue that is more important. They would also argue that these types of distributional studies make consumption taxes look more regressive than they really are on a lifetime basis.

Secondly, it is difficult to explain the very high consumption to income ratios among those at the bottom of the income scale -- are they consuming out of assets or borrowing, being helped by family members, or are there some systematic reporting errors at these low income levels? Are there a lot of individuals who normally have high incomes but are temporarily experiencing low incomes through business or investment losses? Without being certain why these individuals report consumption at such high rates relative to income, it is difficult to assess the distributional consequences of tax changes.

A final issue has to do with the incidence assumptions mentioned above, particularly of a consumption tax substitution. Distributional analyses presume that the effect of consumption taxes is to raise the price of consumption goods. As noted in the previous section, consumption taxes can alternatively result in taxes being passed backwards to wages and assets. The effects depend on the form of the tax and the degree of monetary accommodation. In a stylized model where all assets are held in equity form and where there are no transfers, either public or private, the tax can be treated as raising the prices of consumption goods regardless of the format of the tax or the money supply response, since the real effects do not depend on whether a consumption tax results in a rise in the price of consumer goods, or a fall in the market wage and asset prices. Another way of saying this is that inflation is distributionally neutral; so that the accommodations of the tax in prices is irrelevant to the distributional effects. In practice, there are financial claims to assets, public transfers, and private transfers, and an unanticipated change in the price level is not neutral. A price increase to accommodate the tax in consumer prices will cause the burden of the tax to shift from those with equity claims to financial claims. If individuals at lower income levels tend to own their assets through financial claims, then an unaccommodated tax will be less regressive than an accommodated one. This means that the Hall-Rabushka tax may be less regressive than a value added tax simply because the pressure to inflate is less. Similarly, if transfer payments are not fully indexed (most of them, however, are), a tax substitution which does not create pressures to inflate will be less burdensome on lower income individuals. Finally, it is difficult to determine how private transfers (such as gifts between relatives) would respond to price increases.

These issues relate to the tax base. One of the important points of comparison between indirect forms of consumption tax such as the VAT is that the rate structure cannot be made progressive. The regressivity of the VAT is only slightly reduced by exempting necessities; moreover, these exclusions lead to economic distortions and administrative complexity. (Such exclusions cannot be accommodated in a business transfer tax approach). The matter of rate structure is, however, important to the distributional effects of a tax. Indeed, the current income tax is not only progressive, but it has a set of negative taxes (or tax subsidies) through the earned income credit, which offset the burden of payroll taxes. While it would be possible to increase such credits to offset the burden of a VAT on lower income individuals, the result would be a magnification of administrative burdens and compliance costs.

APPENDIX

DIFFERENT MEANS OF ACHIEVING A CONSUMPTION TAX

To explain the mechanics of how the various taxes work, consider first the simple accounting identity in a system with a flat income tax rate:

(1) C + G + nK = WL + RK = Income

which simply states that the uses of income (C is consumption, G is government spending, and nK is net investment, that is the growth rate n times the capital stock K) sum to the sources of income (WL is the wage rate times the labor supply and RK is the rate of return to capital times the amount of capital). Individuals pay income taxes out of their wages and rents, so that taxes need not be shown explicitly here. Prices are set at one. For illustrating these differences, we keep the capital and labor supply fixed, and also, therefore, the rate of return to capital.

Consider, now, a case of substituting for the European style VAT, where government spending is not subject to the tax. Also, consider the case where nominal incomes remain constant (i.e. a fixed money supply). In this case, we can write the identity as:

(2) P(1 + v)C + PG + nPK = PWL + PRK + TAX

where P is producer's prices and v is the VAT rate. Note that the right hand side must include a term for the taxes received by the government since these are now collected directly from the recipients from their factor incomes.

To continue the illustration, assign some numerical values. Suppose that WL is initially .75, RK is initially .25, G is initially .3, I is .07, and C is .63. Since G takes up 30 percent of output, the initial flat tax must be 30 percent. If we wish to hold real government spending constant, then vC must equal G or:

(3) v = C/G

In this case, v is .31/.63, or a rate of 0.476190.

If we hold nominal income constant at 1, then we can solve for the new level of producers prices:

(4) P(1 + v)C + PC + PI = 1

or P = 1/(.63(1 + 0.476190) +.3 +.07) = 1/1.3 = .77

If the money supply is fixed, both wages and the price of capital will fall to 77 percent of their original values, while the price of consumption goods will rise by about 14 percent (the new price will be (1 + v) times P, or 1.476190 times .77). Note, however, that individuals in the aggregate can still purchase the same commodities. Before, under the income tax, individuals had 70 cents of after tax income they spend on investment (7 cents) and consumption. After tax income is larger (77 cents), investment costs less, and consumption costs more.

If the money supply were to be increased to keep nominal wages constant, prices would have to rise by 30 percent. In this case, consumption prices would increase by 47.6 percent, and asset prices would not fall.

Consider now a VAT with no rebate for government expenditures. Now, the government must pay the extra tax in the cost of goods purchased and collect extra tax to finance the cost.

(5) P(1 + v)(C + G) + nPK = PWL + PRK + TAX

The tax rate would be the same as before, since:

(6) Pv(C + G) = P(1 + v)G

That is, taxes will have to be large enough to pay for government spending inclusive of the VAT, and v still equals C/G. This the tax rate will still be 0.476190 (the government collects and spends the tax simultaneously on its purchases).

P would be 1/((.63 + .3)(1 + v) + .07), and since the denominator is larger, the price (including asset prices and wages) would fall to 69 percent of their original value. Prices of consumption goods would go up by 2 percent. No real ability to purchase goods would change, and the relative prices of consumption and investment goods would stay the same. If the money supply were to be increased to keep nominal wages constant, prices would have to rise by 44 percent. Prices of consumption and government spending would rise by 47.6 percent and asset prices would not fall.

The identical results would occur for a business transfer tax, except that the tax rate would be stated as a percentage of pre-tax, rather than post-tax consumption. That is, we would write:

(7) P(C + G)/(1 - t) + nPK = PWL + PRK + TAX

The base of the tax is still consumption, since purchases of capital are subtracted from the base. This formula is identical to (5) as long as v is equal to t/(1 - t); t would in this case be approximately .32 and t/(1 - t) would be 0.476190.

The results will be somewhat different for the Hall-Rabushka approach where the wage portion is carved out and taxed separately. The dollar equilibrium in (7) for expenditure will still occur. First, that means that wages will be measured at pre-tax levels; that is the wage will be grossed up by the tax (that is, divided by (1 - t). The tax rate is identical to the business transfer tax. The base of the tax is C + G minus wages, but the revenue needed to be raised by the tax is offset by the direct tax on wages:

(8) P(t/(1 - t))(C + G - WL) = P(1 + t/(1 - t))G - PtwL/(1 - t)

This equation produces the same result as before - value of t/(1 - t) of 0.47619.

The results for consumption prices and asset prices are the same as in the business transfer tax case or the case where government spending is taxes -- consumption prices rise to 1.02310 and asset prices fall to .69. The difference is in the nominal wage, which now rises in value to 1.02310 because of the tax gross up. Individuals will still be able to purchase the same amounts of goods -- while their wages do not fall, they have to pay a direct tax. There will, however, be no need for money accommodation to keep the nominal wage from falling.

These approaches produce similar aggregate effects. They can, however, produce different results for individuals depending on their holdings of real and financial claims to assets. If there is no increase in the money supply, the tax burden will be passed backwards to wages and asset prices. When asset prices fall, however, only the equity owners of assets are burdened by the tax. However, if the tax is accommodated by a money increase to hold the market wage constant, this fall in asset prices will be offset under the VAT and business transfer tax and the tax shifted forward fully in price. One of the main differences between the VAT and business transfer tax, and the Hall-Rabushka approach is that there is no pressure to increase the money supply since the market wage actually rises in this case. Thus, this tax is more likely to favor debt over equity holders because it is less likely to induce a money supply increase.

The taxes can also produce different effects depending on whether transfer payments are indexed. If transfer payments are indexed, then there will be no burden of the tax on these types of income, but the tax rate will have to be higher. If they are not indexed, then the burden will fall more heavily on poor people, especially if the tax is of a type that causes an increase in the money supply, such as a VAT or a business transfer tax.

 

FOOTNOTES

 

 

1 Local governments also tend to rely heavily on property taxes. Property or wealth taxes are not generally discussed as a base for Federal taxation, perhaps in part because of questions of constitutionality. There are also some more minor taxes -- excise taxes on certain commodities such as gasoline, alcohol and tobacco, and estate and gift and inheritance taxes. Excise taxes are selective forms of consumption tax, and estate, gift and inheritance taxes are taxes on wealth at the time of transfer.

2 In practice, gross purchases of capital would be subtracted from gross income (income before deducting depreciation). Depreciation charges are netted out of both the income and the consumption tax base; in the former by allowing depreciation deductions, and in the latter by allowing deductions for the cost of replacement capital.

3 None of the taxes discussed in this section tend to exist, or are likely to exist, in pure form. For example, a large fraction of the return to capital is excluded in the current individual income tax because current law does not tax imputed rent on owner-occupied housing. This aspect would be carried over into a consumption tax base as well.

4 Robert E. Hall and Alvin Rabushka, The Flat Tax, Stanford: Hoover Institution Press, 1986.

5 David E. Bradford, Untangling the Income Tax, Committee for Economic Development, Harvard University Press, Cambridge, Massachusetts, 1986.

6 The corporate rate structure is graduated, with lower rates applying to small corporations, but this graduation is not related to the characteristics of the owners. Small corporations may be owned by high income people, while large corporations may be owned by moderate income individuals.

7 Michael Boskin, Taxation, Savings and the Rate of Interest, Journal of Political Economy, Vol. 86, January, 1978, pp. s3-s27.

8 See Barry Bosworth, Tax Incentives and Economic Growth, Washington D.C.: Brookings Institution, 1984; Irwin Friend and Joel Hasbrouck, Saving and After Tax Rates of Return, The Review of Economics and Statistics, Vol.65, November 1983, pp. 537-543; E. Philip Howry and Saul H. Hymans, The Measurement and Determination of Loanable Funds Savings, Brookings Papers on Economic Activity, No. 3,1978, pp. 655-705; John Makin and Kenneth A. Couch, Savings, Pension Contributions, and the Real Interest Rate, The Review of Economics and Statistics, Vol. 71, August 1989, pp. 401-407. Occasionally, the study done by Lawrence Summers, Capital Income Taxation and Accumulation in a Life Cycle Model, American Economic Review, Vol. 71, September, 1981, pp. 533-44 is cited as evidence of a high savings response. This paper is not, however, an statistical analysis but a simulation model which can be made consistent with any savings response. See Owen J. Evans, Tax Policy, the Interest Elasticity of Saving, and Capital Accumulation: Numerical Analysis of Theoretical Models, American Economic Review, Vol. 73, June 1983, pp. 398-410.

9 Auerbach, Alan and Laurence J. Kotlikoff, Dynamic Fiscal Policy, Cambridge: Cambridge University Press, 1987; Summers, Lawrence, Capital Income Taxation and Accumulation in a Life Cycle Model, American Economic Review, 71, September 1981, pp. 533-44. The Auerbach and Kotlikoff study contains a survey of the literature on the magnitude of behavioral responses.

10 Congressional Budget Office, Effects of Adopting a Value Added Tax, Washington, D.C., U.S. Government Printing Office, February 1992.

11 Congressional Budget Office, Effects of Adopting a Value Added Tax.

12 Congressional Budget Office Staff Memorandum, Distributional Effects of Substituting a Flat-Rate Income Tax and a Value-Added Tax for Current Federal Income, Payroll, and Excise Taxes, April 1992.

DOCUMENT ATTRIBUTES
  • Authors
    Gravelle, Jane G.
  • Institutional Authors
    Congressional Research Service
  • Index Terms
    VAT
    rates, flat
    income distribution
    tax policy, progressivity
  • Jurisdictions
  • Language
    English
  • Tax Analysts Document Number
    Doc 92-3697
  • Tax Analysts Electronic Citation
    92 TNT 91-8
Copy RID