How should government activities be financed? Whether or not entirely by taxation? If not, when should they be financed by borrowing, sale of services, or other means? What is the optimal revenue structure for the government? Why? How the government is financed determines both the distribution and the magnitude of the costs of government. And the answers to these questions therefore depend partly on how one thinks the costs of government should be distributed (principles of equity) and partly on how alternative financing measures affect the magnitude and distribution of government costs.
Although what is fair or equitable is a widely debated issue, there can be no doubt that political decisions about taxation should strongly be influenced by considerations of equity and fairness.
Government transfer payments to persons in the form of Social Security, welfare spending, farm subsidies, etc and redistribution of income from taxpayers to those who receive the transfer payments takes different shapes.1 How such transfer payments are to be financed depends on society's redistribution objectives. If the objective is greater equality in the distribution of income, then the transfer payments to low-income persons are to be financed by progressive taxes.
Alternatively, if the objective is merely to provide a "floor" under the incomes of persons - to bring the income of all persons up to some minimum - then taxes need not be progressive in the upper-income ranges. Although opinions about society's redistribution objectives differ, many, if not most, people would support the ideas of limiting income inequality and providing a floor to incomes. Transfers financed by progressive income taxes are the basic mechanisms for achieving such objectives.2
To the question of who should bear the cost of public goods and services, economists usually give two answers: Costs be distributed according to (1) ability to pay or (2) benefits received. Government's use of resources to provide goods and services reduces the availability of other goods and services. Government thereby imposes a cost, or burden, on each person that is measured by the value of the other goods and services. Where such services are given up, this loss of goods and services is a loss of real income, a person's ability to pay, that is, to bear the cost of the government is relatable to the person's real income.3
Another answer to the question of who should bear the costs of the government is that those who benefit should pay. To make this principle operational, benefits are defined and the amount of benefits accruing to each person is measurable. The benefit of a government-supplied service is thus the amount that a recipient would be willing to pay rather than to go without the service.
A second problem with benefit taxation is, it cannot be used to finance transfers, where transfers are based on non-Pareto criteria.4 For example, the primary beneficiary of a welfare payment to a needy family is the family itself. Financing the payment by a tax on the beneficiary (the needy family) would defeat the whole purpose of the government activity.
These problems notwithstanding, benefit taxation may be both feasible and desirable in some instances. The gasoline tax that is used to finance construction and maintenance of highways approximates a benefit tax.5 Fees and licenses (for parks, autos, trucks, hunting, etc) are also in the nature of benefit taxes when the proceeds are used to support the activity. Cities charge for sewerage and garbage collection and, less frequently, for fire protection services also approximate benefit taxes. Aircraft fuel taxes and landing fees are benefit taxes that support air transportation. Property taxes when used for improvements (sidewalks, streets) are benefit taxes too.
In each of these cases, an approximate form of benefit taxation is feasible because the persons benefiting from government services can be identified and a measure, albeit rough, of their use of the services is available. That is, the services provided are in the nature of private goods.
If equity requires the distribution of government costs according to ability to pay, then those having equal ability to pay should bear equal shares of the costs of the government. If costs are to be distributed according to benefits received, then those receiving equal benefits should bear equal shares. The principle of horizontal equity is frequently violated by existing revenue systems.
How unequal should be treated relative to one another is the question of vertical equity. How should tax burdens be related to discretionary, or taxable, income? Should a person having, say, twice the taxable income of another person bear twice, more than twice, or less than twice as large a share of government costs? That is, should the tax (revenue) system be proportional, progressive, or regressive, respectively, with respect to income? If the ratio of tax burden to income is constant, then the tax system is proportional to income. If the ratio increases as income increases, the system is said to be progressive. It is regressive if the ratio falls as income increases.6
Many, if not most, people would oppose a regressive system of distributing government costs, especially if the system imposed significant burdens on those at the lowest income levels. Beyond this, there is no clear agreement on what would be a vertically equitable distribution of costs. For some, vertical equity would entail distributing the costs so as to achieve the greatest equalization of the distribution of income. For others, the appropriate distribution would be one that placed relatively heavy loads on the very rich and relatively light ones on the very poor, without necessarily maximising the equalising effects.7
Some government policies reduce consumption in order to increase consumption possibilities. How should the cost of such investment activities be distributed? If the average income and standard of living continue to rise, as they have in the past, the ability-to-pay criterion would require that future generations bear the costs of investments and that may enhance the consumption possibilities of those generations. Such would also be the case under the benefits-received criterion.8 However, if it is thought that the future living standards will be lower than the present standards, and then the ability-to-pay rule would require that the cost of investments be borne by the present generation.
A reasonable rule in revenue system design would appear to be that of minimising the secondary (administrative, compliance, excess burden) costs of financing any particular level of government activity. However, this rule cannot be applied without exception because the revenue system that minimises the secondary costs may be undesirable in other respects. Most obviously, minimising secondary costs may conflict with equity. Stated differently, efficiency in revenue collection may conflict with equity objectives.
More generally, tax incidence differs from impact because of adjustments of market prices and incomes in response to the tax. That is, tax shifting occurs when product or factor prices are adjusted in response to taxes. If we are to conclude that tax shifting occurs, we must be able to explain why sellers (of products or factors) will respond to the tax by raising prices. This explanation must deal with the question: If it is in the sellers' interest to raise prices in response to a tax increase, why was it not also in their interest to raise prices prior to the tax increase? It is not enough, for example, simply to assert that producers will raise prices and "pass on" the tax to consumers. We are required to explain that why the producers were not charging the highest possible price before the tax increase.
Other complications arise in the assessing of incidence because the use of tax revenues, is distinguished from their collection, as it necessarily causes other changes. Specifically, when a tax is increased, the revenues gained must be used to (1) increase government spending, (2) offset revenue losses from decreases in other taxes, (3) increase government cash balances, or (4) retire government debt. The reverse changes must accompany tax decreases.
These uses of tax revenues affect the distribution of income in three distinct ways. First, the revenue used to provide goods and services or to make transfer payments directly affects the distribution of income. Second, market adjustments in prices and money incomes occur in response to one's use as well as to the collection of tax revenue. Finally, the use of revenue affects aggregate demand, employment, output, and prices. Therefore, several concepts of incidence are frequently distinguished, each of which assumes and includes the income distribution effects of a particular use of the tax revenues.
A system that distributes costs equitably may entail greater total costs than a less equitable system. Therefore, revenue system design normally encounters equity-efficiency conflicts. Resolving such conflicts may be difficult because a value must be placed on equity - that is, a trade-off rate between equity and efficiency must be determined, either explicitly or implicitly, as decisions are made. A revenue system design is complicated and consensus about what constitutes a good tax system, and the desired result is difficult to obtain because the incidence of a tax may differ from its impact and the same is difficult to determine.
The criteria for designing and evaluating revenue systems can be applied to individual elements of the system or to the system as a whole. The latter appears to be the more meaningful basis for evaluation.
To know how to modify the system, we must know how the elements of various tax forms such as sales, property, and income, of the system differ in their effect on the economy. The question, what will be the effects of a partial or complete substitution of one tax for another is also important in revenue system design.
1. This type of spending has been increasing rapidly, with transfer spending by all levels of government.
2. For more discussion, see Arthur M. Okun, Equality and Efficiency: The Big Trade-off (Washington: Brookings, 1975), p. 102.
3. To say that a person bears a cost because of government is not to say that government imposes a welfare loss on the person, since the government's activities may also be of some benefit. Instead, the sum of the two effects measures the net effect of government on the person's welfare.
4. Consider the principles of Distribution of Wealth and Income.
5. It is widely regarded as a fair tax for construction and maintenance of highways.
6. A tax may be related in some specific way to its base and related another way to income. For example, a general sales tax that is proportional to expenditures for the taxed items may be regressive with respect to income.
7. The issue of vertical equity arises in decisions about the progressiveness, or lack of it, of in come tax rates. We are concerned only with equity in the distribution of the cost of government. We are not concerned with the distribution of benefits from government expenditures, except insofar as taxes are based on benefits received. The net burden of government (cost minus benefits) may be regressive, progressive, or proportional.
8. Borrowing is a means of transferring investment costs to future generations. The burden of an investment is always borne by the current generation in the form of reduced consumption in the period of investment. However, when the investment is financed by borrowing, the present generation may be compensated for lower present consumption with greater future consumption.
(The writer is an advocate and is currently working as an associate with Azim-ud-Din Law Associates)