Taxation is a general concept for devices used by governments to extract money or other valuable things from people and organizations by the use of law. A tax formula contains at least three elements: the definition of the base, the rate structure, and the identification of the legal taxpayer. The base multiplied by the appropriate rate gives a product, called the tax liability, which is the legal obligation that the taxpayer must meet at specified dates. A tax is identified by the characteristics of its base, such as income in the case of an income tax, the quantity of distilled spirits sold in the case of a liquor tax, and so on. The rate structure may be simple, consisting of one rate applying to the base, such as a specified number of cents per gallon for a tax on gasoline, or complex, for example, varying rates depending upon the size of the base for a tax on personal income.

Taxes may be assessed in money or in kind. The government of Communist China imposes taxes on peasants assessed in units of grain produced, and it requires payment in grain itself. In the American Confederacy, because of the deterioration of the Confederate money during the latter phases of the American Civil War, some taxes were assessed and collected in terms of commodities. In American frontier settlements of the eighteenth and early nineteenth centuries, the local governments formed by the people in the region commonly imposed taxes by requesting that each adult male work a given number of days constructing community facilities such as roads and schools. The modern-day counterpart of this practice is conscription of men for service in the armed forces, although conscription is not generally considered as a tax. The dominant practice, however, in the contemporary world is the assessment of taxes in money and the settlement of the tax liability by the payment of money.

Taxation presupposes private ownership of wealth. If a government owned all wealth in a society, including the wealth embodied in people, it would obtain all income, and there would be nothing to tax. No government has gone to such extremes in concentrating the wealth of a society in its own hands. Even in highly socialized societies, such as the Soviet Union, people are permitted, subject to restrictions, to own themselves, household goods, savings accounts, and money. Taxation therefore becomes feasible. Nevertheless, the more wealth a government itself owns, the less is taxation necessary, because revenue from the management of assets is a substitute for tax receipts. National governments, with the exception of some of the highly socialized countries, typically find themselves on the other side of the ledger, having on balance negative net worths apart from their taxing power. Some local governments in western Europe and the United States have substantial revenues from government-owned facilities such as electric power plants, municipal water facilities, and transport systems. The profits from the management of these facilities are occasionally sufficient to permit a government to dispense with taxation altogether.

Taxes are to be distinguished from prices imposed by a government for goods and services that it provides. A price is a money payment made as a condition of obtaining goods or services. It serves as a rationing device provided that the price is positive and provided that the amount of the goods or services the buyer receives in return depends upon the price. If a government supplies water and charges according to the amount of water taken by the buyer, the device is a price and not a tax. Borderline cases arise in two types of circumstances: (1) when a charge is made as a condition of an all-or-none choice, such as a fee for a license for an automobile as a condition of operating the vehicle on any public highway; or (2) when a government imposes a requirement that the citizen use a service and then charges for the service taken, such as a requirement for a passport for foreign travel accompanied by a charge for the passport. In situations in which the element of government compulsion enters significantly, it is customary as well as reasonable to treat the charge as a tax rather than as a price.

Classification of taxes . Taxes may be classified in various ways. Since a tax is a formula of three ingredients—a base, a rate structure, and identification of the legal taxpayer—a common characteristic of any of these three elements may be employed for grouping. Thus taxes may be classified as personal or business. In such a classification, a tax on beer is a business tax because business organizations are in fact the legal taxpayers. More commonly, taxes are grouped on the basis of similarities of the tax base; for example, commodity taxes refer to all taxes in which the production or sale of commodities becomes the occasion for a government levy. Even though personal income taxes vary widely in their characteristics among countries, the presumed common element of the tax base—personal income—is used for grouping purposes.

Perhaps the single most widely used distinction is between what is called “direct” taxation and what is called “indirect.” This is a classification based on certain presumed effects of various taxes. A direct tax in this usage refers to one in which the legal taxpayer cannot shift any of the tax liability to other people, such as customers or suppliers. A clear illustration of a direct tax is a lump-sum charge levied on a person—sometimes called a head tax or poll tax. Income, death, net worth, expenditure, and sometimes property taxes are commonly classified as direct. Indirect taxes refer to those that are thought to be shifted from the legal taxpayer to others. Commonly, taxes on sales of commodities, import duties, and license fees are grouped together as indirect.

By postulating common effects of various taxes, the direct-indirect classification becomes subject to two serious defects. The effects of a particular tax device are not intuitively apparent; their discovery entails careful scientific investigation. It is thus awkward to employ a classification that begs these questions in advance. There is the further difficulty that the shifting of a tax by the legal taxpayer to others may occur in various degrees from 0 to 100 per cent. If a particular tax is proved to be shifted to others by the amount of 25 per cent, for example, the direct-indirect classification becomes irrelevant. One should have to say that the tax is 75 per cent direct and 25 per cent indirect. The difficulty arises because an all-or-none test is used when the relevant distinction is one of degree. For these and other reasons, the direct-indirect classification, although widely used in reporting revenue data, is usually avoided in scientific investigations.

Among other possible dichotomous classifications, taxes may be divided into those described as systematic means-test devices and those without this characteristic. Personal income, expenditure, and net worth are examples of means-test taxes: taxes whose base is systematically related to some relevant index of the taxpayer’s economic position. Personal income, the money gain a person experiences over a period, may be, and commonly is, looked on as a measure of his economic position. An expenditure tax treats the amount spent for personal living expenses or consumption as the index of relative economic status. Likewise, net worth, the value of assets possessed minus debts owed to others, may be used as such a measure. When the purposes of taxation include large yields and systematic redistribution of economic power, some form of means-test taxation must be employed. Although other taxes can also provide large yields, they are likely to be erratic in their effects on income distribution.

Functions of taxation . Any tax that has a yield extracts money from people or organizations and provides money for a government. As a result of a tax formula, taxpayers find themselves with less money to spend; governments, on the other hand, find themselves with more money. This transfer of money from people to government gives rise to two functions of taxation: a reduction in the spending potential of the private sector and an increase in the spending potential of the public sector.

Revenue. The negative function of reducing the spending potential of people often may be viewed as an unfortunate by-product of taxation. Few city officials, for example, would applaud the fact that as a result of the imposition of local taxes, the local citizenry has less money to spend. For local government units, it is the financial needs of the government that justify taxation. Until recent decades, this view was assumed correct for all levels of government; taxation was believed to arise solely out of the financial needs of governments rather than from a public objective of reducing citizens’ spending power.

A sovereign government with an advanced type of financial system controls the money system and, as one feature of this control, can if it wishes provide itself with unlimited quantities of money at negligible cost. This power arises from the use of national monies in the form of bank deposits and currency as opposed to commodities, such as gold and silver, whose quantity cannot be increased by government decree. A national government need no longer levy taxes in order to finance itself.

The discovery of the power of governments to free themselves from internal financial constraints has a long and complicated history. By the time of the Napoleonic Wars, the British government had discovered the convenience of having the Bank of England provide it with funds. In America, the colonies experimented rather freely with the money-issuing power; Massachusetts has the distinction of being the first government in the world to issue paper money. Yet the necessary institutions for the exercise of this power did not exist in the United States at the time of the War of 1812, and the federal government for the first and only time in its history found itself literally bankrupt. World War i was the first occasion when the power to finance government by creating money was freely used by all major belligerents. This financial power was clearly recognized by governments during World War n and was used on a vast scale. In the contemporary world, the possibility of national governments having insufficient money to finance their internal expenditures is no longer a real one. Thus, the amount of revenue to be raised by taxation depends on policy objectives rather than on government financial necessity.

Even though sovereign governments have freed themselves from financial constraints, the revenue function of taxation has not disappeared; this function becomes that of regulating private expenditures so as to stabilize employment and the price level. During periods of insufficient private expenditures, for example during recessions, a national government may allow its revenues to fall automatically and, in addition, may take steps to reduce effective rates of tax in order to increase private expenditures. Similarly, during periods of excessive expenditures, tax rates may be increased as a depressant measure. Even as late as the 1930s, few governments possessed leaders who understood the policy choices available; through fear and desperation they took steps to increase tax rates when effective recovery called for tax reduction. Thanks to the spread of economic intelligence, such serious errors in financial policy are unlikely to be duplicated should a serious world depression ever again develop.

Resource reallocation. In addition to the revenue function of taxation, taxes may alter the product-mix generated within the private sector. In Great Britain, for example, such commodities as automobiles, household appliances, liquor, and tobacco are made more expensive by taxation, whereas such items as milk, vegetables, meat, cider, and household help are made less expensive, in part through subsidies, or negative taxes. As a consequence, the British people use rather more of the latter group of commodities and rather less of the former. The tax-induced change in the product-mix comes about through the effects of taxes on prices and quantities produced. British manufacturers of electric dishwashers, for example, being confronted by a heavy tax on these goods, charge higher prices for them, and so the number these companies can profitably sell is curtailed. The labor and capital services not used to produce dishwashers as a result of the curtailed output of them are devoted instead to the production of other commodities that are lightly taxed or not taxed at all. These other commodities are therefore made more abundant and sell for lower prices. From the point of view of buyers, this alteration of the product-mix benefits those who happen to like the lightly taxed commodities and injures those who prefer the heavily taxed goods. Whether the entire consuming public as a whole can be said to be better off or worse off as a result of the alteration of the product-mix depends on whether an optimum is defined in terms of consumer preferences as expressed in the market or as expressed through political processes. [See Consumer Sovereignty.]

Almost all actual tax devices commonly used by governments display some features that may alter the pattern of productive activities in the society. Personal income taxes as found in the Western world define the tax base incompletely, leaving some gains subject to little or no tax. In the administration of net worth taxes in the Scandinavian countries, agricultural land in comparison with other types of assets is appraised lightly for tax purposes. Value-added taxes, as employed in the state of Michigan and in France, completely omit important types of value-adding activities from the tax base. In all such cases, the tax system encourages some activities over others. Although complete neutrality of a tax system can never be achieved in fact, actual tax systems become more neutral as their coverage of economic activities becomes more general.

As tax systems have developed, they have tended to favor activities of a nonmarket character, such as leisure, production of goods for personal use instead of for sale, and “do-it-yourself” projects in general. Governments have difficulty in catching such gainful activities in their tax net and, with occasional exceptions, do not attempt to do so. Consequently, and for political reasons also, tax policies in advanced countries generally favor agricultural over industrial activities.

Income redistribution. A further main function of taxation is the redistribution of economic power as measured by income or wealth. With respect to money income, a tax system is distributionally neutral if it reduces each person’s income in the same proportion. Taxation may be systematically progressive in the sense of taking an increasing proportion of income increases. Technically, a tax system is defined as progressive if the marginal rate of tax with respect to income exceeds the average rate of tax, provided marginal rates do not exceed 100 per cent. Regressive tax structures refer to the opposite case. In this context, “proportional,” “progressive,” and “regressive” describe the effect of the entire tax system on the distribution of income.

Accompanying the development of democratic political institutions in the Western world, various ethical ideas arose concerning the appropriate criteria for evaluating taxes. The dominant ethical idea that emerged is the “ability-to-pay” doctrine. This rather vague expression is intended to provide a justification for tax systems that are systematically progressive as opposed to those that are proportional, regressive, or merely erratic. Income is usually taken as the appropriate measure of personal ability-to-pay, although net worth and expenditure have also been advocated as appropriate measures. The concept of ability-to-pay implies both equal treatment of people with equal ability, however measured, and a progressive rate structure. The ability-to-pay doctrine has strong affinities to egalitarian social philosophy; both support measures designed to reduce inequalities of wealth and income.

Strict adherence to the test of ability-to-pay, when income is used as the measure of ability, would call for a monolithic tax structure restricted to personal income taxation. Logically, the idea implies systematic negative taxation as well. If a person with a modest income pays zero tax, a person with an even smaller income should pay less than zero tax—that is, receive a subsidy—to achieve appropriate differences in the treatment of people with different tax-paying abilities.

In the development of actual tax systems, only modest success can be claimed for reducing the incomes of the very wealthy by tax measures— even in countries such as the United States, Great Britain, and Australia, all of which use progressive income taxation and have been governed over appreciable periods of time by groups unsympathetic to economic plutocracy. There is little evidence to suggest that in these countries taxes have substantially reduced the wealth of the very wealthy, despite the apparent high rates of tax on large incomes and large estates.

By far the most important government measures used to reduce income inequality have been government welfare programs. Various social services, such as medical care, education, and income maintenance in the form of social security programs, have mitigated the economic hardship of low-income groups in Western countries. These programs directly raise the money incomes of the unemployed, the aged, and the incapacitated; they also potentially raise the consumption of all qualifying groups by providing some services free or at nominal cost. In some Western countries, the programs have virtually eliminated grinding poverty; they have not achieved this goal in the United States.

Shifting and incidence . A tax is said to be shifted if the legal taxpayer can by some means force others to contribute extra amounts of money to him because of the presence of the tax. Shifting is therefore achieved in degrees ranging from zero, that is, no shifting, to 100 per cent, or complete shifting. To the extent a tax is shifted from the legal taxpayer, such as the proprietor of a retail establishment in the case of a retail sales tax, other people are thereby selected to contribute to the government. Actually, full explanations of tax shifting require the determination of the true amount each person must pay to governments, including the amounts shifted to him, so that at least in principle the investigator can state precisely the amount of money a person or family contributes to government per unit of time. Nothing approaching this precision has yet been achieved in any country. It is a safe generalization that the typical citizen goes through life never knowing, even within wide limits, how much in tax he is actually paying.

The concept of tax incidence, sometimes called tax burden, is closely related to that of tax shifting. If, of the taxes imposed by a government, none are shifted at all, the incidence of a tax is said to fall on those who are the legal taxpayers—those persons who would be sued by the government for failure to pay the amounts specified by the tax formula used. In this event, the incidence of the tax holds few mysteries. The concept does involve difficult issues, however, if a tax is shifted in whole or in part. The concept of tax incidence is concerned with the identification of the persons who “finally” or “ultimately” pay the tax liabilities as opposed to those who, although legally required to pay money to the government, are acting wholly or partly as intermediaries in the tax-collection process. Thus, legislators in voting taxes on such items as liquor or cigarettes do not ordinarily assume that the vendors of these commodities are “really” paying the tax because legislators ordinarily operate on the theory that vendors can pass along the tax to buyers. Granted the validity of the theory, the incidence of the tax falls on these buyers.

Theories of tax shifting and incidence exist in great variety. Insofar as a consensus can be found, it is that means-test taxes (for example, income taxes) either are not shifted at all or are shifted only to a trivial extent and that commodity taxes, including import and export duties, are largely shifted from the legal taxpayers to others. The incidence of the general property tax imposed by local governments in the United States, company (corporation) profits taxes, taxes on the transfer of physical and financial assets, and of many minor levies is analyzed in many different ways; no definite consensus can be found among experts on the subject.

Differences in analysis of various tax devices reflect differences both in the general theoretical framework deemed appropriate to explain economic events and in the precise manner in which the investigator views the device being studied.

Tax theory has developed mainly as a by-product of classical and neoclassical economic theory, as exemplified by the works of such thinkers as Adam Smith, John Stuart Mill, W. Stanley Jevons, and Alfred Marshall. Continental general equilibrium approaches, mainly through their influence on such American thinkers as Irving Fisher, have, after a considerable time lag, become important in the explanation of the effects of various taxes. The Keynesian system of thought has a large and devoted contemporary following; it is widely used to explain the effects of entire fiscal systems. More recent theoretical work has been dominated by model building, often of a highly esoteric kind, constructed with such highly simplifying assumptions that government only rarely gets into the picture at all . Given the variety of approaches to the explanation of economic events, a generally endorsed approach to the explanation of how taxes are shifted and in what amounts cannot be expected.

Real income approach. The question of just what the investigator is attempting to explain in connection with taxation is also approached in different ways. A major disputed issue is the alternative, implicitly or explicitly presupposed, to the tax under investigation. Some students define the problem as the effects of the tax together with some government expenditure assumed to be financed by that tax. Accordingly, the problem of explaining a personal income tax is looked upon as including the effects of certain or all of some government expenditures. This approach is adopted more or less automatically by those who view economics as fundamentally a “real” system, meaning a system in which money is assumed to be absent or in which money is viewed as a purely passive device to effect exchanges, having no distorting effect on price relations among goods and services. In this approach, a tax is viewed as levied in goods and services that the government either uses directly in its affairs or trades with private individuals to obtain the goods and services used in its expenditure programs. Explanation consists of showing how the combination of income taxation and the assumed expenditures changes relative prices, quantities produced, and the amount of leisure taken. The incidence of the tax-expenditure combination is believed to be established by showing what groups experience a decline in real income.

This approach has little appeal to most students of public finance because of its remoteness from reality. (It remains, however, the dominant approach to the analysis of the incidence of import duties in pure international trade theory.) It has the further defect, apart from its restrictive assumptions, of identifying particular taxes with particular government programs, when in fact neither a government nor an individual can generally determine which expenditure is financed from a particular tax or income source. Logically, the approach is inherently incapable of isolating the effects of a tax as such, because a tax apart from expenditures is undefined. Since, in fact, government programs can take on any of a great variety of forms, including negative taxes (subsidies), the approach in principle can only give answers for each of an indefinitely large number of possible combinations.

Money income approach. A relatively recent approach, what may be called the “income theory” of tax incidence, views the basic problem of tax analysis to be the determination of the portion of each person’s income diverted to the government by a tax. This approach finds that any tax that provides a government with revenue must simultaneously make the after-tax money income of some people who work or own property smaller by the amount of the revenue. The investigator, accordingly, attempts to identify, for each tax device, those persons whose after-tax incomes are curtailed. In this approach, government expenditure for goods, services, or assets enters as a factor determining demands for current output and as analytically distinct from revenues.

This approach to taxation can be explained by illustration. A tax on cigarettes, for example, is commonly believed to be paid by cigarette smokers. According to the income theory, such a tax reduces the money incomes of certain groups. People, as buyers of services and goods, including cigarettes, are prepared to spend some dollar amount per period. The demand schedule for cigarettes being highly inelastic, a change in price results in little change in the quantity purchased—the tax raises the price of cigarettes by almost the full amount of the tax per unit, and the dollar amount spent on cigarettes increases. Given constraints on total private expenditures, the amount spent on other commodities diminishes. If these “other commodities” consist of all commodities other than cigarettes, all industries find the demands for their products lowered and all will earn lower incomes. If these “other commodities” consist of a narrow class of commodities, such as food, the industries, including agriculture, producing these goods experience reduced prices and reduced income. To only a small extent, because of the low price elasticity of the demand for cigarettes, do tobacco companies and tobacco growers also experience lower profits and wages. The product-mix changes only slightly in this case. If a tax is placed on cornflakes, the pattern of results would be somewhat different. Cornflakes, being one of a great variety of breakfast foods, has a highly elastic demand. The tax would increase the price of cornflakes, greatly reduce the amount bought, and drive resources out of cornflake production. In this case, the companies and workers in the taxed industry would experience lower incomes.

There is no shifting of a general tax on income or net worth; taxpayers experience a lowering of their after-tax income and no incentive is created to reduce further other factor incomes. Commodity taxes, import duties, retail sales taxes, and expenditure taxes are shifted more or less, depending upon the setting where they are used. Property taxes of the American type, where the tax base is mainly the assessed value of real estate, business equipment, and inventories, present complications because of the great diversity in effective rates within and among local jurisdictions and because of the benefit element of local expenditures to owners of taxed property. Property taxes can be shown to reduce property income in general and to be progressive with respect to total income.

The income theory of tax incidence applies in a symmetrical manner to negative taxes, such as subsidies to the production of some food products in Great Britain and to income-maintenance social security programs. The incidence of negative taxes refers to the identification of the ultimate recipient and the amount he receives. As is the case with ordinary taxes, the problem is to identify the private counterpart of the government’s financial transactions.

The income theory of tax shifting, as the name implies, treats government income as arising at the expense of private income. Tax revenues are treated as a form of transfer income—as are interest paid to owners of debt instruments, dividends paid by corporations, and pensions and social security payments made by governments.

Role of determinant price systems. In order to arrive at definitive conclusions, all theories ’of tax shifting need a pricing system that is determinant as opposed to one that is capricious or random. If prices of commodities depend on what executives of corporations eat for breakfast, the incidence of corporation income taxes or commodity taxes cannot be definitively ascertained. Actual price systems in Western countries exhibit capricious elements arising from market power, illustrated by the pricing of some varieties of labor services, government price regulation based on concepts of fair return and historical costs, and many others. Systematic tax theory, like the economic theory of which it is a part, assumes the orderly features of price systems and fails to the extent that the actual world lacks these characteristics. Economists differ widely in their outlook on the degree of orderliness exhibited in contemporary societies; some find that the economic world neatly illustrates the properties of a perfectly competitive pricing system, and, at the opposite pole, others find no system to explain and as a consequence deprecate economic theory.

Taxation and fiscal policy . The main financial weapons of a national government are its expenditures on goods and services, transfers (including negative taxes), taxation, public debt management, and monetary policy. Some or all of these may be manipulated to alter the level of total expenditures by all groups in the economy and at the same time may be used to alter the pattern of these expenditures.

The deliberate manipulation of taxes for the purpose of achieving full employment is subject to both political and economic constraints. The reduction of effective rates of tax, for example, may be irreversible because of political objections to tax rate increases. In addition, taxes have other functions besides revenue, such as resource reallo-cation and income redistribution, and these functions may be partly defeated by changing the tax structure for purposes of influencing private expenditures. These considerations do not imply that the manipulating of effective rates of tax poses insurmountable difficulties; only that the difficulties must be recognized and, if possible, weighed when making a final decision.

Taxes are interdependent among themselves and also interdependent with other fiscal weapons. A reduction in taxes on company profits leads to increased revenue from a personal income tax because some portion of the increase in after-tax profits will appear as an increase in dividends. Different taxes compete for the income of owners of resources; an increase in the effective rates of one tax reduces the yield of others. In selecting taxes to manipulate in influencing private expenditures, these repercussions on other tax yields must be taken into account if the desired total change in revenues is to be achieved.

Taxes are also interdependent with other fiscal devices. In Western countries, and many others as well, national monetary systems are banking systems characterized by bank creation and destruction of money, fractional reserve requirements, and central bank determination of changes in bank reserves. Treasuries must conduct their finances within this institutional framework. Effective fiscal policy presupposes cooperative central bank policies; otherwise, fiscal measures designed to stimulate the economy may be offset by monetary measures. A main problem in financial administration remains that of effective coordination of fiscal and monetary policies. They are so closely interdependent that some students prefer to speak of national financial policies rather than of two sets of policies, fiscal and monetary.

The use of taxation as a weapon to influence private expenditures becomes feasible to the extent a treasury is free from financial constraints, and freedom from constraint implies access to an unlimited amount of money. Central banks are the institutions that have the power to create money in any amount. If, then, a government decides, for example, to reduce effective rates of tax as a stimulating measure during a depression, its treasury will initially find itself depleting its cash position or, in the case of European national treasuries, will be increasingly in debt to the central bank. If the treasury department sells public debt or if the central bank does so instead, the cash released to taxpayers is reabsorbed by net sales of public debt. Depending on the circumstances, these combined actions may be perfectly offsetting, or they may on balance be stimulating or depressing with respect to private expenditures on goods and services. If a stimulating combination of measures is to be assured for a given amount of tax reduction, the maximum is achieved if no debt is sold to the public at all. In this event, with a fractional reserve system of banking, bank reserves increase at the rate of the tax cut. Such increases in bank reserves, given the practice of relatively low fractional reserve requirements or customs, would lead to a potential increase in the amount of money so exceedingly large for even modest tax reductions that central bankers would almost certainly feel obliged to offset them in part. Perhaps a more relevant definition of zero offset is a central bank response to a tax-rate change that permits the quantity of demand deposits plus currency in the hands of the public to change by the change in the yield of the tax systems. In actual practice, however, it would be rare to observe such a result. Normally, central banks and treasuries, when tax cuts are made, use debt operations to offset a sizable fraction of the tax change. For this and other reasons, faith in the efficacy of tax changes to influence the economy must be tempered; one must examine what response, in terms of changes in the size of the outstanding debt, may be expected.

World tax structures . In advanced countries, tax revenues range from a high of about 35 per cent of the gross national product in West Germany to a low of about 21 per cent in Japan; the United States government (federal, state, and local) takes an amount equal to about 25 per cent of the gross national product. Such comparisons may, however, be misleading. In advanced European countries, provision for retirement income is usually made through government programs, whereas in the United States various private pension plans supplement in substantial amounts the federal social security programs. Were retirement deductions from the remuneration of employees counted as taxes, the United States would rank closer to such high-tax countries as West Germany, Sweden, and France in effective tax rates.

International comparisons also neglect negative taxes such as family allowances, subsidies, and social security transfers, creating an impression of heavier taxation of the average household than would data showing both the amount taken in tax and the amounts received in the form of government transfers. Net tax data have unfortunately not been systematically compiled for purposes of international comparisons.

The structure of tax systems reflects the political and social characteristics of national groups. France, a country of high taxation, relies heavily on value-added taxation, whereas the United Kingdom, also a high-tax country, relies heavily on income taxes. The. United States, being a federal political system with long traditions of local financing of local functions, employs many taxes that can be administered at the state and local levels, resulting in a highly complex combination of taxes, such as federal, state, and even local income taxes, state and local retail sales taxes, and the continuation of the important, though generally criticized, local property tax. A centralized system of taxation as found in France would be alien to the mores of Americans. Tax systems, to be workable, must be in keeping with popular feelings and beliefs. This consideration explains why politicians may succeed when tax experts, especially foreign experts, fail in attempts to redesign a country’s tax system.

Of the developing nations, apart from some oil-rich countries, few are in a position to impose taxation at the effective rates found in advanced European and English-speaking countries. Mass poverty, weak public administration, and the concentration of political power in the hands of wealthy groups rule out heavy taxation. Tax systems in these countries ordinarily consist of import duties and, in a few, export duties, transaction and commodity taxes, low-rate income taxes, land taxes, and some form of death tax, usually of the inheritance type. India uses systematic income taxation, although less than 10 per cent of the population is subject to it. As these poor countries develop, their tax systems may be expected to develop as well, and in the direction of higher effective rates of tax.

The outlook generally is for continued high-level taxation where already found and increasingly high effective rates of tax elsewhere, with the possible exception of the communist countries. From a long-run point of view, revenue requirements of government are closely geared to government expenditure and transfer programs. The goods and services that governments provide are looked upon as superior to alternative private commodities, with the consequence that, even apart from military programs, government expenditures exhibit a long-run tendency to rise relative to national income. This tendency is not inevitable and may be reversed. Yet continued urbanization alone, with all that this development implies for government action, may be sufficient to assure relatively expanding government programs. In addition, the welfare state has already demonstrated its political popularity in the Western world, and, despite the lamentations of political conservatives and some economic liberals, government activities appear destined to grow both absolutely and relatively. If so, high taxation can also be expected to be an enduring characteristic of advanced societies.

Earl R. Rolph

[See also Public Expenditures

BIBLIOGRAPHY

American Economic Association 1959 Readings in the Economics of Taxation. Homewood, 111.: Irwin.

Bator, Francis M. (1960) 1962 The Question of Government Spending: Public Needs and Private Wants. New York: Collier.

Blough, Roy 1952 The Federal Taxing Process. Engle-wood Cliffs, N.J.: Prentice-Hall.

Butters, J. Keith; and Lintner, John 1945 Effect of Federal Taxes on Growing Enterprises. Cambridge, Mass.: Harvard Univ., Graduate School of Business Administration, Division of Research.

Due, John F. (1954) 1963 Government finance. 3d ed. Homewood, 111.: Irwin.

Due, John F. 1957 Sales Taxation. London: Routledge; Urbana: Univ. of Illinois Press.

Fabricant, Solomon 1952 The Trend of Government Activity in the United States Since 1900. New York: National Bureau of Economic Research.

Hall, Challis A. 1960 Fiscal Policy for Stable Growth: A Study in Dynamic Macroeconomics. New York: Holt.

Hansen, Bent (1955) 1958 The Economic Theory of Fiscal Policy. Cambridge, Mass.: Harvard Univ. Press. ? First published as Finanspolitikens ekonomisha teori.

Harvard University, International Program In Taxation 1963 Taxation in the United States. Chicago: Commerce Clearing House.

Hicks, J. R.; HICKS, U. K.; and Rostas, L. (1941) 1942 The Taxation of War Wealth. 2d ed. Oxford: Clarendon.

Holland, Daniel M. 1958 The Income Tax Burden on Stockholders. Princeton Univ. Press.

Kaldor, Nicholas 1955 An Expenditure Tax. London: Allen & Unwin.

McKean, Roland N. 1958 Efficiency in Government Through Systems Analysis: With Emphasis on Water Resources Development. New York: Wiley.

Musgrave, Richard A. 1959 The Theory of Public Finance: A Study in Public Economy. New York: McGraw-Hill.

usgrave, Richard A.; and Peacock, Alan T. (editors) 1958 Ciassics in the Theory of Public Finance. London and New York: Macmillan.

Pigou, A. C. (1928) 1956 A Study in Public Finance. 3d ed., rev. New York: St. Martins.

Prest, Alan R. 1960 Public Finance in Theory and Practice. Chicago: Quadrangle Books.

Rolph, Earl R. (1954) 1956 The Theory of Fiscal Economics. Berkeley: Univ. of California Press.

Rolph, Earl R.; and Break, George F. 1961 Pubiic Finance. New York: Ronald Press.

Schultz, William J.; and Harriss, C. Lowell (1931) 1959 American Public Finance. 7th ed. Englewood Cliffs, N.J.: Prentice-Hall.

Simons, Henry C. 1938 Personal Income Taxation: The Definition of Income as a Problem of Fiscal Policy. Univ. of Chicago Press. #^x

Vickrey, William S. 1947 Agenda for Progressive Taxation. New York: Ronald Press.

II. PERSONAL INCOME TAXES

The personal income tax is widely regarded as the fairest method of taxation yet devised. It is the major element of progression in modern tax systems and permits differentiation of tax burdens on the basis of family responsibilities and other personal circumstances of taxpayers. The yield of the tax expands or contracts more rapidly than personal income, thus imparting built-in flexibility to government revenue systems. The income tax is less burdensome on consumption and more burdensome on personal saving than an equal-yield expenditure tax, but the difference in aggregate terms is probably small for taxes of broad coverage. The effect of the income tax on work and investment incentives is unclear. Although personal income taxation has a long history, some of its major features still present numerous unsettled problems.

The first general personal income tax was introduced in 1799 in Great Britain, where it has been in effect continuously since 1842. Despite this early example, other countries were slow in adopting this tax. It was used for a brief period in the United States during and after the Civil War, and it was permanently enacted following the ratification in 1913 of the sixteenth amendment to the constitution. Austria adopted the income tax in 1849 and Italy in 1864; Australia, New Zealand, and Japan followed in the 1880s and Germany and the Netherlands in the 1890s. Elsewhere the income tax is a twentieth-century phenomenon. It spread quickly during and after World War i and became a mass tax in many countries during World War ii. Today the personal income tax raises substantial amounts of revenue in all industrialized countries of the free world and is employed, although to a lesser extent, in most underdeveloped countries.

Equity considerations

Analysis of tax equity has been concerned largely with the distribution of tax burdens among persons in different economic circumstances, i.e., with vertical equity. Questions regarding the treatment of persons in essentially the same economic circumstances—the problems of horizontal equity —have only received close attention since the 1930s.

Vertical equity . Progressive taxation appeals intuitively to most people as an equitable method of distributing the tax burden by income classes, and economists and political theorists have devoted a great deal of intellectual effort to justify it on logical grounds.

An early theory of taxation that was widely held prior to the mid-nineteenth century was that taxes should be distributed in accordance with benefits received. The benefit theorists supported a minimum of government activity, possibly including defense and police and fire protection, but not much more. The benefits of such government services were assumed to be proportionate to income, and this was regarded as a major rationale for proportional income taxation. This theory of tax distribution proved to be untenable both because of its narrow view of the role of government and the arbitrary assumption it made regarding the distribution of benefits of government services.

In the latter half of the nineteenth century, progressive income taxation was justified by the sacrifice theories that emerged from discussions of “ability to pay.” Under this doctrine, ability to pay is assumed to increase as incomes increase, and the objective is to impose taxes on a basis that would involve “equal sacrifice” in some sense. If the marginal utility of income declines more rapidly than income increases, equal absolute sacrifice leads to progression, equal proportionate sacrifice to still more progression, and equal marginal sacrifice to leveling of incomes from the top down until the required revenues are obtained. The assumptions of sacrifice theories—that the relative utility of different incomes is measurable and that the relation between income and utility is approximately the same for all taxpayers—cannot be verified by actual data or experience. Nevertheless, the ability-to-pay idea has been a powerful force in history and has undoubtedly contributed to the widespread acceptance of progressive taxation.

The basic justification for the progressive personal income tax is now probably the socioeconomic objective of reducing great disparities of welfare, opportunity, and economic power arising from the unequal distribution of income. More specifically, the justification is based on two propositions: (a) it is appropriate public policy to moderate economic inequality, and (b) taxation of personal incomes at progressive rates is an efficient method of promoting this objective, since it does not involve direct intervention in market activities. The acceptable degree of progression varies from time to time and place to place; it depends on the distribution of pre-tax incomes and the post-tax distribution desired by the voters. In practice the re-distributive effects of the income tax have been moderate in all countries.

Horizontal equity . A personal income tax conforming strictly with the “equal treatment” principle would apply to all income from whatever source derived, making allowances only for the taxpayer and his dependents. In accordance with the “accretion” concept, income would be defined as consumption plus (or minus) the net increase (or decrease) in the value of an individual’s assets during the taxable period, perhaps modified to exclude gifts and inheritances that are ordinarily subject to separate taxes and, for practical reasons, to include capital gains when realized or when transferred to others through gifts or bequests. In practice most of the income taxes now in existence depart from this standard by a wide margin.

Differentiation of tax liability on the basis of family responsibilities is ordinarily made through a system of personal exemptions for the taxpayer and other members of his family. The personal exemption was originally regarded as a device to avoid taxing individuals and families with incomes that were not adequate to provide minimum levels of subsistence. Today personal exemptions are not high enough to cover a socially acceptable minimum level of subsistence in most countries; they serve primarily to remove low-income recipients from the tax rolls and also contribute to progression in the lower part of the income scale. At higher income levels the purpose of the personal exemption seems to be to moderate the tax burden as family size increases, although the degree of moderation varies greatly among countries. Special exemptions are allowed in some countries for particular groups of taxpayers (e.g., the aged); these exemptions are subsidies that could probably be handled more equitably through direct government outlays.

A second type of differentiation employed in most countries is based on the source of income. The provisions include credits for earned income and for dividends, preferential treatment for capital gains, exemptions for transfer payments and amounts set aside for retirement, omission of the rental value of owner-occupied homes, and numerous other special benefits. The earned-income credit is regarded as a convenient method of making rough allowances for depreciation of labor skills and for expenses of earning income from personal effort which are not recognized for tax purposes. The United States abandoned the earned-income credit in 1944 for simplification reasons, but it is still in existence in the United Kingdom, Australia, and other countries. Dividend credits are designed to moderate the so-called “double taxation” of corporate profits. Preferential treatment of capital gains grew out of the English concept of income, which excluded irregular receipts from income. This treatment is now rationalized on incentive grounds and also as a procedure to avoid applying the graduated rates, in the year of realization, to incomes accrued over a period of years. Transfer payments are excluded because they accrue largely to low-income people. Payments by employers into pension plans are not included in employees’ taxable income to promote the development of private pension plans. The rental value of owner-occupied homes is untaxed in most places because it is difficult to apply the income tax to nonmoney incomes.

The third type of differentiation is based on the use of income. Deductions are required under a “net” income tax for expenditures that are essential to earning income. However, deductions for a wide variety of personal expenditures and for some forms of saving are also permitted. In one country or another, allowances are made for such items as medical expenses, charitable contributions, interest on personal loans and mortgages, state and local taxes, casualty losses, child care in families of working parents, deposits in saving associations, premiums for life, sickness, and accident insurance, and payments into annuity, pension, or other retirement plans.

Personal exemptions are an important element of a progressive income tax, but there is little justification for most of the special exclusions, deductions, and credits based on the source or use of income. Such provisions narrow the tax base and require the use of higher tax rates to raise a given amount of revenue. This puts a premium on earning or spending incomes in forms receiving preferential treatment, interferes with business and investment decisions, and distorts the allocation of resources. Since the deviations from equal treatment tend to be arbitrary, they create dissatisfaction among taxpayers who are subject to discrimination and result in pressures for the enactment of additional special benefits, pressures which legislatures find it difficult to resist. This process has been called “erosion of the tax base” in the United States, where taxable income is at least one-third lower than it would be under a comprehensive income tax. Measures to broaden the base and to use the revenues for rate reduction have been proposed by tax experts, but it is evident from the public and Congressional response that progress along these lines will be slow.

Economic effects

Three major aspects of the personal income tax may be distinguished in appraising its economic effects: first, its automatic response to changes in total personal income; second, its effects on the allocation of personal income between consumption and saving; and, third, its impact on work and investment incentives.

Automatic flexibility . The role of the personal income tax as a built-in stabilizer is one of its most significant features. In the United States, at the rates prevailing in the 1950s and in the 1960s, the personal income tax automatically offset more than ten per cent of the reductions in personal incomes during contractions. The corresponding figure for the United Kingdom was perhaps twice as large, the difference being attributable primarily to the higher starting rate in the United Kingdom. Such changes in tax liability reduce fluctuations in disposable personal income and thus help to stabilize consumption.

Built-in flexibility operates in both the expansion and contraction phases of the business cycle, so that the personal income tax moderates the growth of incomes during a business recovery just as it cushions the fall in income during contractions. This symmetrical response of the income tax (and of other stabilizers) during a business cycle is unavoidable. It should not lead to the abandonment of the stabilizers but rather to the establishment of basic tax-expenditure relationships that would be consistent with a prompt return to high employment following periods of recession. Discretionary changes in tax rates and in expenditures may be needed to implement this objective.

The responsiveness of the income tax to changes in personal incomes is a useful characteristic for underdeveloped as well as for developed countries. An increasing proportion of the nation’s resources must be devoted to public and private investment to increase the rate of economic development. Since voluntary saving is usually inadequate, the bulk of the investment funds must be provided by government. A progressive income tax automatically provides some of the financing as incomes increase. Where development is associated with rising prices, the income tax serves the dual role of moderating inflationary pressures and of increasing the rate of national saving.

Effect on consumption and saving . A personal income tax applies to the income of an individual regardless of the allocation of this income between consumption and saving. By contrast, a general consumption or expenditure tax can be postponed or avoided by delaying or eliminating consumption. It follows that an income tax is less burdensome on consumption than an equal-yielding consumption or expenditure tax which is distributed in the same proportions by income classes. In practice, where the income tax is paid by the large mass of people, much of the tax yield comes from income classes where there is little room in family budgets for reducing consumption in response to tax incentives. Under these circumstances the differential effect of the two types of taxes on total consumption and saving is likely to be relatively small.

Graduated expenditure taxes have been proposed in recent years as a method of avoiding or correcting the effects of income tax erosion, particularly in the top income brackets where exemption or preferential treatment of capital gains permits accumulation of large fortunes without tax payment. Expenditure taxation, it is felt, would discourage lavish living by people with large amounts of property and thus increase saving and risk taking without resorting to regressive taxes. Despite its apparent advantages, the expenditure tax has not been widely used. Rates in excess of 100 per cent would be required to raise significant amounts of revenue from high-income taxpayers. Moreover, the expenditure tax is more difficult to administer than the income tax and also raises much more serious problems of compliance.

Work and investment incentives . It is difficult to evaluate the effect of personal income taxes on work and investment incentives. On the one hand, high tax rates reduce the net rewards of greater effort and risk taking and thus tend to discourage these activities; on the other hand, they may provide a positive stimulus to obtain more income because they cut down on the income left over for spending. These two effects tend to offset one another, and there is no basis for deciding which is more important.

Empirical studies have shed little light on this question. The evidence suggests that income taxation does not have a significant effect on the amount of labor supplied by workers and managers. Work habits are apparently not easily changed, and there is little scope in a modern industrial society for most people to vary the hours of work or the intensity of their efforts in response to changes in tax rates.

A highly graduated income tax applying to all property incomes might reduce incentives to take risk somewhat, since it is impossible to reimburse taxpayers for losses at precisely the same rate at which their incomes are taxed. However, the income tax actually applies to a small fraction of property income in all countries. The opportunity to earn income in the form of capital gains— which are either not taxed at all or are taxed at relatively low rates—is a great stimulant to risk taking in the face of high rates on other incomes. Moreover, risk investment is to a large extent undertaken by firms operating in the corporate form; such firms are generally permitted to retain earnings after payment of more moderate tax rates than those applying to investors in the top personal income tax brackets.

Structural problems

The base of the personal income tax is determined by the definition of income, the allowable deductions, and the personal exemptions. Within wide limits, these elements can be combined with various tax rates to produce a given amount of revenue. Many of the difficult issues in most countries are an outgrowth of local problems and developments. Nevertheless, several structural problems in income taxation appear to be common to practically all countries, and these will be discussed briefly in this section.

Tax treatment of the family . Throughout most of the history of the income tax, differentiation was made among taxpayers with different family responsibilities through the use of personal exemptions. Recently, there has been a trend toward the use of different tax rates to provide additional differentiation, particularly in the middle and higher tax brackets. In the United States, France, and West Germany, this has been accomplished by the adoption of the principle of “income splitting” between husband and wife or among all family members. Other countries achieve the same objective by applying different rate schedules to taxpayers in different family situations.

In France the income of the family is divided by the number of family units, with the taxpayer and his spouse counting as one unit each and each dependent child as an additional one-half unit. The tax is then calculated as if the income of the family were divided proportionately among the family units. In West Germany and the United States, splitting is extended only to husband and wife. By contrast, the United Kingdom has made the use of joint returns by husband and wife mandatory since the early days of its income tax. Under this system the graduated rates are applied to the couple’s combined income after allowance for personal exemptions and other deductions.

Income splitting between two persons doubles the width of the taxable income brackets and thus reduces the progression in tax burdens applying to married couples. The absolute size of the benefit depends entirely on the rate of graduation; it bears no relationship to the level of tax rates. For example, if rates increased one percentage point for every $1,000 of taxable income, income splitting would reduce the tax of a married couple with taxable income of $20,000 by $1,000. This would be true whether the starting rate was 1, 10, 20, or 50 per cent.

Income splitting is generally justified on the ground that husbands and wives usually share their combined income equally. For most families the largest portion of the budget goes for consumption, and savings are ordinarily set aside for the children or for the enjoyment of all members of the family. Two conclusions seem to follow if this view is accepted. First, married couples with the same combined income should pay the same tax irrespective of the legal division of income among them; second, the tax liabilities of married couples should be computed as if they were two single persons and their total income were divided equally between them. The first conclusion is now firmly rooted in the tax laws of most countries and seems to be almost universally accepted. It is the second conclusion on which opinions—and practices— still differ.

The case for the sharing argument is applicable to the economic circumstances of taxpayers in the lower income classes, where incomes are used almost entirely for the consumption of the family unit. At the top of the income scale the major rationale of income taxation is the reduction of the economic power of the taxpayer unit, and the use made of income in these levels for family pur poses is irrelevant. Obviously, these objectives cannot be reconciled if income splitting is extended to all income brackets.

Aside from reducing progression, the practical effect of income splitting is to produce large differentials in the taxes of single persons and married couples. Differentials by marital status that depend on the rate of graduation are difficult to rationalize. However, it is difficult to justify treating single persons with families more harshly than married persons in similar circumstances. As a remedy for this problem, France grants to a widow or widower the same total number of family units for splitting purposes as if the spouse had survived. The United States permits widows and widowers to split their incomes for two years after the death of the spouse and provides half the advantage of income splitting for single persons who maintain a household for children or other dependents or support their parents in a separate household.

One of the major reasons for the acceptance of income splitting may well be inadequate differentiation provided by the traditional types of personal exemptions among taxpayers in the middle and top brackets. Single people, it is felt, should be taxed more heavily than married couples because they do not bear the costs and responsibilities of raising children. But the allowance of income splitting for husband and wife clearly does not differentiate between taxpayers in this respect since the tax benefit is the same whether or not there are children. Nor does the extension of splitting to children give the correct answer, since the benefits depend on the rate of graduation as well as on family size.

The source of the difficulty in the income-splitting approach is that differentiation of family size is made through the rate structure rather than through the personal exemptions. It would be possible to differentiate among taxpayer units by varying the personal exemptions to take account not only of the number of persons in the unit but also of the size of income, with both a minimum and maximum. If this is unacceptable, the only alternative—other than income splitting which produces anomalous results—is to vary tax rates by marital status and family size, as a number of countries have already done.

Personal deductions. In principle, the use made of a given income should have no bearing on the amount of tax to be paid out of that income. In practice, some allowances are made almost everywhere for selected items of consumption or saving. These deductions may be divided into three major types: (1) those that provide supplement to the personal exemption; (2) those that subsidize particular activities or expenditures; and (3) those that improve coordination of Federal income taxes with state or provincial and local taxes, where they exist.

A strong case can be made for allowing some deductions for large, unusual, and necessary expenditures when the personal exemptions are low. Deductions for medical expenses are the best example of this type of expenditure. They are often involuntary, unpredictable, and may exhaust a large proportion of the taxpayer’s income. Expenditures for noninsured losses due to theft, fire, accident, or other casualties are of a similar nature. In keeping with the purpose of this type of deduction, it should be limited to an amount in excess of some percentage of income, which would be high enough to exclude all but extraordinary expenditures for these purposes.

Subsidy-type deductions are most common for contributions to charitable, religious, educational, and other nonprofit organizations. In many countries heavy reliance is placed on philanthropic institutions to supplement governmental activities and in some cases to provide services which governmental units do not perform. It may be argued that private philanthropy should not be encouraged at the expense of government funds. However, few people subscribe to this view because the activities of these organizations, with rare exceptions, are considered desirable and useful.

Subsidy-type deductions are also allowed in some countries for selected items of personal saving. Great Britain has permitted the deduction of a portion of life insurance premiums since the beginning of the income tax. West Germany allows deductions for personal insurance and for deposits in building and savings associations. A number of countries have recently enacted limited deductions for amounts set aside in annuities or retirement plans by self-employed persons and employees not eligible for company pension plans. The major motivation for these deductions appears to be to promote saving, but more particularly to encourage adequate provision for retirement and for catastrophic events that entail large outlays or loss of income. The deductions for personal contributions to retirement plans are also intended to remove the discrimination resulting from the exclusion usually granted to employer contributions to employee pension plans. The growth of allowances for particular types of saving has made substantial inroads into the philosophy of income taxation; in fact, these policies constitute a substantial movement toward the expenditure tax approach.

Suggestions have been made in recent years that the tax laws should permit a deduction for the cost of higher education. These suggestions reflect the importance of higher education for economic growth and the increased costs of a college education. On the one hand, a deduction allowed to parents would give the largest benefits to the highest income classes and would therefore be inequitable. On the other hand, some portion of expenditures for higher education is an investment which is not recognized for tax purposes as an expense of earning income. The appropriate treatment would be to regard the outlay by a parent as a gift to his child and to permit the child to write off a portion of this outlay over his earning career for, say, twenty years. However, there is no basis for estimating the proportion of educational outlays allocatable to investment, and the problems of administration and compliance would be substantial. [See Capital, Human.]

Deductions for income taxes paid to overlapping governmental units are required to prevent confiscation if one or more levels of government employ high rates in the upper end of the income scale. Where the rates are moderate, it is quite appropriate to levy two taxes on the same base without coordination. However, it may be desirable to permit deductions even if the rates are not confiscatory as a device to moderate interstate differentials. For example, with a Federal rate of 70 per cent and without deductibility of state taxes, the combined tax on residents of two states with rates of 5 and 10 per cent would be 75 and 80 per cent, respectively. By permitting taxpayers to deduct the state tax on their Federal returns, the combined rates are reduced to 71.5 and 73 per cent. (If the states also permit a deduction for Federal taxes, the combined rates are further reduced to 70.5 and 71 per cent. This type of mutual deductibility is unnecessary for coordination purposes, since the coordination achieved through single deductibility is quite adequate.)

A deduction for income taxes paid to state and local governments may be a practical necessity in a Federal system, but the same justification does not hold for state and local sales, excise, and property taxes. The latter deductions defeat the purposes of taxes levied to obtain payments from taxpayers for benefits received from state and local governments and reduce the progressivity of the combined tax system.

In the United States, where personal deductions have proliferated more than in any other country except perhaps West Germany, taxpayers are granted a “standard” deduction, in lieu of the itemized deductions, of up to 10 per cent of income (with a maximum of $500 on separate returns of married persons and $1,000 on all other returns). This device was adopted in 1944 for simplification reasons, in recognition of the fact that most personal deductions are small and few taxpayers keep adequate records to support them. To an important degree, the standard deduction violates the rationale of the itemized deduction; it reduces differentiation in tax liabilities while the itemized deductions are intended to introduce such differences for the purposes selected. The existence of both a standard deduction and itemized deductions suggests that there is some ambivalence toward many of the personal deductions in the United States income tax structure.

On balance, equity would be better served by avoiding erosion of the tax base through the use of numerous costly personal deductions. This should not preclude the adoption of a restricted list of deductions for unusually large and extraordinary expenditures to prevent hardships. Subsidy-type deductions are appropriate only if they promote a significant national objective and if the deduction route is the most efficient and equitable method of achieving that objective.

Capital gains and losses . As already indicated, an economic definition of income would include capital gains in full on an accrual basis. This method is impractical for three reasons: (1) valuations of many types of property cannot be estimated with sufficient accuracy to provide a basis for taxation; (2) most people would regard it as inequitable to pay tax unless income has actually been realized; and (3) taxation of accruals might force liquidation of assets to discharge tax liabilities. Thus, where capital gains are taxable, they are included in income only when realized.

Few countries tax the capital gains of individuals, but the United States has done so since the beginning of its income tax. Realized capital gains were originally taxed as ordinary incomes, but they have been subject to preferentially low rates since 1921. The provisions applying to such gains changed frequently during the 1920s and 1930s but were stabilized beginning in 1942. In general, capital gains on assets held for periods longer than six months are subject to half the rates on ordinary income, up to a maximum of 25 per cent.

The treatment of capital gains is likely to be a compromise among conflicting objectives. From the standpoint of equity, it is well established that capital gains should be taken into account in determining personal tax liability. Moreover, low rates or exemption of capital gains encourage the conversion of ordinary income into capital gains by devices that distort patterns of investment and discredit income taxation. On the other hand, the bunching of capital gains in the year of realization requires some provision to moderate the impact of graduation. On economic grounds full taxation of capital gains is resisted because it is believed that it would have a substantial “locking-in” effect on investors and reduce the mobility of capital. It is also argued that preferential treatment of capital gains helps to stimulate a higher rate of economic growth by increasing the attractiveness of investment generally and of risky investments in particular.

The “bunching” problem can be handled by prorating capital gains over the length of time the asset was held or by adopting a general averaging system applying to other types of income as well as to capital gains. However, unless the marginal rates were fairly low, the tax might still discourage the transfer of assets. Part of the difficulty is that adherence to the realization principle permits capital gains to be transferred tax-free either as a gift or at death. The solution to this problem is to treat capital gains as if they were constructively realized as a gift or at death, with an averaging provision to spread the gains over a period of years. Great Britain adopted the constructive realization principle when it added a capital gains tax to its tax structure. Under such a system the only advantage taxpayers have from postponing the realization of capital gains is the accumulation of interest on tax postponed. Unless the assets are held for many years, this advantage is small as compared to the advantage of the tax exemption accorded to the gains transferred at death; in any event, the interest on the tax postponed is subject to income tax when the assets are transferred. Under the circumstances, the incentive to hold gains indefinitely for tax considerations alone is very greatly reduced.

Capital losses are no easier to handle than capital gains. In principle, capital losses should be deductible in full either against capital gains or ordinary income. However, when gains and losses are recognized only upon realization, taxpayers can easily time their sales so as to take losses promptly when they occur and to postpone the realization of gains. There is no effective method of avoiding this asymmetry under any system of taxation applying to realized gains and losses. In the United States, capital losses of individuals may be offset against capital gains plus $1,000 of ordinary income in the year of realization and in subsequent years for an indefinite time period. This restrictive policy is perhaps most harmful to small investors, who are less likely than those in the higher brackets to have gains against which to offset their losses. The only solution to this problem is a pragmatic one which would be as liberal as possible for the small investor without opening the door to widespread abuse and large revenue losses.

Relation to the corporate income tax. Unless corporate incomes were subject to tax, individuals could avoid the personal tax by accumulating income in corporations. Short of an annual allocation of corporate incomes on a prorata basis—a method which is excellent in theory but not in practice—the equity and revenue potential of the personal income tax can be protected only by a separate tax on corporate incomes. However, the existence of two separate taxes on the same income creates a difficult equity problem. Concern over the “double taxation” of dividends is evident in the various devices used in different countries to alleviate its alleged discriminatory effects.

On the assumption that all or a significant portion of the corporate income tax rests on the stockholder, the effect of double taxation is to impose the heaviest burden on dividends received by stockholders with the lowest incomes. Assume a corporate income tax of 50 per cent and suppose a corporation pays out $50 in dividends. The corporate income before tax from which these dividends were paid amounted to $100. If this $100 had been subject to personal income tax rates only, the nontaxable individual would have paid no tax on it; the additional burden of the corporate income tax in this case is the full $50 corporate tax. By contrast, a stockholder subject to an eighty per cent rate pays a personal income tax of $40 on the dividend, and the total tax burden on the original $100 of corporate earnings is $90. But since he would pay $80 under the personal income tax in any case, the additional burden on him is only $10.

The simplest and most effective method of dealing with this problem would be to permit corporations to deduct all or a portion of their dividends in computing taxable income. This method would apply the regular corporate tax rate to undistributed profits and would reduce or eliminate the corporate tax on distributed earnings. It would also have two additional advantages: first, dividend and interest payments would be treated more nearly alike, thus reducing the discrimination against equity financing by corporations; second, the same proportion of the corporate income tax on distributed income would be eliminated for all taxpayers regardless of their personal income tax status.

Despite these advantages, undistributed profits taxation is not used widely. The United States experimented with it in the 1930s, but the experiment created a great deal of resentment (possibly because the differentiation between distributed and undistributed profits was made by the imposition of a penalty tax on the latter rather than by allowing a deduction for dividends). The major drawback of undistributed profits taxation is that it discourages internal financing by corporations and thus may reduce saving and investment. On the other hand, some believe it is unwise as a matter of policy to permit corporations to avoid the capital markets for financing their investment programs.

If dividend relief is given at the individual level, there are three possibilities. The first is the “withholding” method, under which all or a portion of the corporate tax is regarded as having been paid at the source by the stockholder. The taxpayer includes the tax paid at the source in his income and then receives a tax credit for that amount. This method was used in Great Britain from the enactment of the 1803 income tax until 1965. Tax burdens of shareholders on distributed corporate income are the same as the burdens under the undistributed profits tax approach.

The second alternative is to permit the taxpayer to exclude some or all of his dividends from his tax return, and the third is to permit him to take a credit against his final tax liability computed at a flat percentage of the amount of dividends he receives. The United States exempts the first $100 of dividends; and Canada uses the dividend-credit approach exclusively at a rate of 20 per cent. Great Britain now makes no special allowance for dividends.

Neither the exclusion nor the credit can be regarded as a satisfactory method of removing double taxation, since neither can remove the same proportion of the excess taxation of dividends throughout the income scale. In contrast, the undistributed profits approach and the withholding method remove the same proportion at all income levels.

The desirability of doing something about the double taxation of dividends is still in dispute. First, corporations are viable economic units with characteristics and behavior patterns that have very little relationship to the income and other characteristics of their stockholders. Moreover, stockholders in large, publicly held corporations have only indirect and remote influence on management policies. On these grounds, many experts believe that a modern tax system would be incomplete without a separate tax on corporate enterprises. Second, the argument for moderating or removing the double taxation of dividends assumes that the corporate tax rests on the corporation and, ultimately, the stockholder. If the corporate income tax is shifted forward in the form of higher prices (or backward in the form of lower wages), the case for integration collapses. In the present state of knowledge, the incidence of the corporation income tax is not clear.

If integration of the corporate and personal income taxes were considered appropriate, some solution of the capital gains problem would be an essential first step. Under a system of full taxation of capital gains, including constructive realization at death, generous provision might well be made for alleviating the double tax on distributed profits. Where capital gains are either not taxed at all or are taxed at very low rates, the case for integration is weak. No country has yet resolved all of these problems satisfactorily.

Fluctuating incomes . The use of an annual accounting period combined with progressive rates results in a heavier tax burden on fluctuating incomes than on an equal amount of income distributed evenly over the years. This type of discrimination is hard to defend on equity or economic grounds. Taxpayers do not and cannot arrange their business and personal affairs to conform with the calendar. Annual income fluctuations are frequently beyond the control of the taxpayer, yet he is taxed as if 12 months were a suitable horizon for decision making. In addition, in the absence of averaging, there are great pressures for moderating the impact of the graduated rates on fluctuating incomes by lowering the rates applicable to them. Reduced rates on capital gains are often justified on this basis, although the reductions more than compensate for the absence of averaging.

There may also be a connection between the treatment of fluctuating incomes and incentives to take risk. Even with generous provisions for offsetting losses against gains, business incomes are taxed more heavily than other incomes under a progressive, annual income tax because (a) they fluctuate more than other incomes and (£>) the losses do not come off the top of the taxpayer’s income during the loss-offset period and are therefore not credited at the maximum rate. On the assumption that there is a correlation between income variability and risk, a tax system using a one-year accounting period is more burdensome on venturesome than on safe investments and thus is more discouraging to risk taking than a tax system having a longer accounting period.

Experience with general-averaging systems has been disappointing, largely because the methods used have been based on a variant of the moving average. This requires large tax payments when incomes fall below the average and small payments when they rise above it. Taxpayers properly regard such an arrangement as highly inequitable. It is now known that the payment problem may be solved by making the averaging adjustment in the form of a refund. For example, taxpayers might be permitted to average their incomes once every five years and to receive a refund (or credit) for any amount of tax actually paid in excess of 105 or 110 per cent of the tax on the average income during the averaging period. The United States adopted a variant of this method in 1964, allowing individuals to average their incomes over a five-year period where the income in the current year exceeds the average of the four prior years by more than a third and this excess is more than $3,000.

Many averaging systems, varying from cumulative lifetime averaging for every taxpayer to averaging over fairly short periods for specific types of volatile incomes, have been explored in the literature. All averaging proposals would create problems of compliance and administration and might involve substantial revenue losses, particularly if applied to the mass of taxpayers. With the advent of electronic machines, it will be possible to solve most of the administrative problems, but the revenue implications may remain serious.

The personal income tax is still in the process of development. Methods of differentiating tax liabilities of single persons and families of different size are unsatisfactory. There is increasing recognition that capital gains and losses should enter the tax base, but the equity, economic, and administrative objectives of capital gains taxation are difficult to reconcile. The appropriate relationship between the personal and the corporate income tax continues to be disputed. Little progress has been made to alleviate the excessive burden of the income tax on fluctuating income. Finally, the concept of income subject to tax departs considerably in most countries from an economic definition of income, and too many special allowances are made for specific sources and uses of income.

Despite all of these problems, the personal income tax is the best tax yet devised, and it will continue to be an indispensable and significant element of all modern tax systems for the indefinite future.

Joseph A. Pechman

BIBLIOGRAPHY

Barlow, Robin; Brazer, Harvey E.; and Morgan, James N. 1966 Economic Behavior of the Affluent. Washington: Brookings Institution.

Butters, J. Keith;

Thompson, L. E.; and Bollinger, L. L. 1953 Effects of Taxation: Investments by Individuals. Boston: Harvard Univ., Graduate School of Business Admininstration, Division of Research.

Canada, Royal Commission ON Taxation 1966 Report. Ottawa: Queen’s Printer.

Goode, Richard B. 1951 The Corporation Income Tax. New York: Wiley.

Goode, Richard B. 1964 The Individual Income Tax. Washington: Brookings Institution.

Great Britain, Royal Commission ON THE Income Tax 1920 Report. Papers by Command, Cmd. 615. London: H.M. Stationery Office.

Kahn, C. Harry 1960 Personal Deductions in the Federal Income Tax. National Bureau of Economic Research, Fiscal Studies, No. 6. Princeton Univ. Press.

Kaldor, Nicholas 1955 An Expenditure Tax. London: Allen & Unwin.

Kalven, Harry; and Blum, Walter J. (1952) 1953 The Uneasy Case for Progressive Taxation. Univ. of Chicago Press. ? First published in Volume 19 of the University of Chicago Law Review.

Lewis, Wilfred 1962 Federal Fiscal Policy in the Postwar Recessions. National Committee on Government Finance, Studies in Government Finance. Washington: Brookings Institution.

Musgrave, Richard A. 1959 The Theory of Public Finance: A Study in Public Economy. New York: McGraw-Hill.

Pechman, Joseph A. 1957 Erosion of the Individual Income Tax. National Tax Journal 10, March: 1-25.

Pigou, A. C. (1928) 1956 A Study in Public Finance. 3d ed., rev. London: Macmillan; New York: St. Martins.

Seligman, Edwin R. A. (1911) 1921 The Income Tax: A Study of the History, Theory and Practice of Income Taxation at Home and Abroad. 2d ed. New York: Macmillan.

Seltzer, Lawrence H. 1951 The Nature and Tax Treatment of Capital Gains and Losses. National Bureau of Economic Research, Fiscal Studies, No. 3. New York: The Bureau.

Shehab, F. 1953 Progressive Taxation: A Study in the Development of the Progressive Principle in the British Income Tax. Oxford: Clarendon.

Simons, Henry C. 1938 Personal Income Taxation: The Definition of Income as a Problem of Fiscal Policy. Univ. of Chicago Press.

U.S. Congress, House, Committee ON Ways AND Means 1959 Tax Revision Compendium. 3 vols. Washington: Government Printing Office.

Vickrey, William S. 1947 Agenda for Progressive Taxation. New York: Ronald Press.

III. CORPORATION INCOME TAXES

The taxation of the income of corporations has come to be one of the major sources of fiscal revenue in most countries. According to the 1965 Yearbook of National Accounts Statistics of the United Nations, corporation tax receipts in 1962 equaled or exceeded 2 per cent of the national income in 32 countries, and represented 10 per cent or more of current government receipts in 19 countries. Of the major countries, Japan places the heaviest reliance upon the corporation income tax, receipts from this tax accounting for 22 per cent of current revenues and amounting to 6 per cent of the national income. Australia, Canada, New Zealand, the Republic of South Africa, and the United States all collect more than 15 per cent of their current revenues from this source, the amounts in each case representing more than 5 per cent of national income. In western Europe corporation income taxes typically represent 3-4 per cent of national income and 6-10 per cent of current government revenues. The corporation income tax tends to be less important, relative to national income and government revenues, in the developing countries than in the more advanced economies; but this is due mainly to the fact that the corporate sector itself is less important, rather than to a failure of the developing countries to levy the tax at all or to a tendency on their part to impose the tax at significantly lower rates than those applied by the more advanced countries.

This widespread and heavy reliance on the corporation income tax testifies to its administrative feasibility and political popularity. It is highly feasible administratively because the laws under which corporations are established generally require the maintenance of accounts on a standardized basis; thus the enforcement of the tax reduces to the problem of requiring honest and accurate accounts and of resolving a series of technical issues, such as the determination of which expenditures may be expensed and which must be capitalized, and the setting of allowable rates of depreciation for specific classes of assets. These problems have been handled in most countries by administrative decrees or regulations issued by the tax-collecting authority itself, operating under broad guidelines set out in the tax legislation.

The political appeal of the corporation income tax has two roots. First, the tax obviously conforms to popular conceptions of ability to pay, since the man in the street tends to view corporations as wealthy entities themselves and as being owned predominantly by wealthy stockholders. But second, and in many ways equally important as a source of political appeal, is the fact that the corporation income tax, by definition, cannot be a source of loss to a corporation. Those corporations which have no net income pay no corporation income tax; only “profitable” companies are required to bear this levy. By contrast, other forms of business taxation can themselves be responsible for converting what would otherwise be a net profit situation into one of net loss. Hence, even within the world of business, companies in a marginal or precarious financial situation are likely to prefer the taxation of corporate net income to other forms of business taxation, and the strong opponents of corporate income taxation are likely to be the more profitable companies with the most “ability to pay.”

The administrative and political advantages of the corporation income tax do not, however, imply that it is a good tax from the economic point of view. Quite to the contrary, it is readily demonstrable that, of the major revenue sources, this tax is one of the least justifiable on economic grounds. It entails an essentially arbitrary discrimination among industries or activities, it tends to inhibit the growth of the more dynamic sectors of the economy, and it probably causes a reduction in the over-all rate of capital formation.

Efficiency effects

All the discriminatory features of the corporation income tax stem from the fact that corporate net income is the tax base. By the definition of the tax, all unincorporated activities are exempt; and even within the corporate sector of the economy, the tax falls more heavily on activities with low ratios of debt to equity (because interest on debt is a deductible expense). The consequence of these discriminations is a distortion of the economic structure, favoring noncorporate over corporate activities and, within the corporate sector, a distortion favoring those activities which can readily be financed in large measure by debt capital over those which cannot. The tax may also discriminate within the corporate sector against capital-intensive activities and favor labor-intensive activities, but the existence of this effect depends on the incidence of the tax; it may be present but need not be.

The basis for these assertions is the fact that in all economies in the modern world there is a tendency toward the equalization of the rates of return that investors receive on capital in different industries or activities. This tendency can be frustrated by restrictions on the entry of capital into given areas, can be blunted by imperfect information, can be modified by considerations of difierential risk or convenience among different investment outlets, and can be obscured by random year-to-year variations in earnings—but it is always present. Stigler (1963, p. 23) found, for example, that whereas the mean rate of return (after taxes) on invested capital in U.S. manufacturing industries averaged 7.6 per cent in 1947-1954, the standard deviation of the rates of return by two-digit industries (about this mean) was only 1.6 per cent. Moreover, he found no significant evidence of a. risk premium (either positive or negative) when he related observed average rates of return in individual industries to the standard deviation of each industry’s rate of return. Stigler’s results accord well with what one would expect a priori from a reasonably well-functioning capital market. If higher-than-average rates of return to capital exist and persist in a given activity, then one would expect investment in that activity to increase and so drive down the rate of return; if lower-than-average rates of return prevail, one would expect investment to fall off, inducing an increase in the rate of return.

The following analysis will, accordingly, be based on a tendency toward equalization of aftertax rates of return to capital in different investment uses. Given this tendency, it is clear that the corporation income tax will produce an equilibrium pattern of net rates of return among industries only through its differential impact’s being reflected in differential gross rates of return. Thus, assuming that the net-of-tax rate of return on equity would, in a given capital-market situation, tend to stabilize at 6 per cent, and assuming that the rate of return to capital in the noncorporate sector and the rate of interest on debt would also tend to stabilize at 6 per cent, we have the following possible pattern of rates of return on capital, gross of a corporation income tax at a rate of 50 per cent:

Noncorporate Industry	6%
Corporate Industry A: 2/3 debt	8%
Corporate Industry B: 1/3 debt	10%
Corporate Industry C: 100% equity	12%

The differentials in gross rates of return on capital induced by the corporation income tax have two kinds of effects: first, they are reflected in product prices and, consequently, in the levels of output of particular activities; second, they confront the different activities with different relative costs of labor and capital and, hence, induce decisions concerning the relative intensity of use of these resources which are uneconomic from the standpoint of the economy as a whole. For example, the net annual cost of $100,000 of capital, for a year, to Noncorporate Industry (see above), would be $6,000, while that to Corporate Industry C would be $12,000. If labor of a given class is paid $6,000 per year, Noncorporate Industry is induced to operate at a point where the marginal $100,000 of capital produces a yield equivalent to the marginal product of one man-year of labor, while Corporate Industry C will tend to operate at a point where $100,000 of capital will have a yield equivalent to the marginal product of two man-years of labor. Clearly, economic efficiency could be improved by a tax system which took an equal fraction of the income generated by capital in all lines of activity, regardless of whether they were corporate in structure or not, and regardless of their degree of access to debt financing.

Effect of other taxation . The foregoing sketch of the efficiency-effects of the corporation income tax implicitly viewed the tax as the only levy in the tax system that affected gross-of-tax rates of return differently in different activities. Actually, there are a variety of taxes and tax provisions in most countries which have such effects, and it is important in any analysis of real-world tax systems to consider the combined effect of all such provisions rather than attempt artificially to isolate one tax, such as the corporation income tax, from the overall structure of which it is a part.

Property taxes, for example, are often levied at different effective rates on real property of different types. More important, property taxes often are levied only on land and buildings. Thus machines, inventories, and such may escape the property tax; and corporate capital, in which machines and inventories play a larger role than in noncorporate capital, will then pay relatively less through property taxation than noncorporate capital. In this way the property tax may tend to offset somewhat the discrimination against the corporate sector that is implicit in the corporation income tax.

Similarly, in countries like the United States, where capital gains are taxed at rates lower than normal personal income tax rates, or in countries with no capital gains taxation at all, the effects of corporate income taxation as such are likely to be offset to some extent by the favored treatment of capital gains. This is so because the earnings of capital in unincorporated enterprises are taxed under the personal income tax as they are earned, at full personal income tax rates, while the personal income tax strikes only that portion of corporate earnings paid out in dividends at the full rate. Let D be the proportion of earnings paid out in dividends, t_c be the corporate tax rate, t,, the personal tax rate, and t, be the effective rate of tax on capital gains. Then $1 of corporate earnings will pay a total personal-plus-corporate tax bill equal to

t_c+ (l-t_c)Dt_p + (1 -tc)(l -D)t_g.

This can turn out to be lower than t_f, the total income tax paid on $1 of income of an unincorporated enterprise, provided that the rate of tax applicable to a marginal dollar of personal income is sufficiently higher than the corporate tax rate.

For example, assume that an individual is in the 70 per cent bracket of the personal income tax and is contemplating investing some savings in either a specific corporation, C, or a specific unincorporated enterprise, 17. Suppose that both investments are expected to have a gross-of-tax yield of 20 per cent. The net-of-tax return from the investment in U will be 6 per cent, while that from the investment in C will depend on t_c, D, and t,,. Suppose t_c is 40 per cent, D is 331/2 per cent, and t, is 15 per cent. Then, of $20 of earnings in C, $8 will be paid in corporation tax, and $2.80 in personal tax on dividends of $4. If the corporation’s savings of $8 out of earnings of $20 ultimately are fully reflected in capital gains, and if these are taxed at an effective rate of 15 per cent, then $1.20 will be paid in capital gains taxes. The total tax on $20 of income will be $12, and the net-of-tax rate of return from the investment in C will be 8 per cent—higher by 2 points than that on the investment in U.

Obviously, the effective rate of corporate-cum-personal tax on an investment will vary from individual to individual (depending on their marginal tax rates) and from corporation to corporation (depending on their dividend policies and on the degree to which their corporate savings are reflected in capital gains). Moreover, the effective rate of tax on capital gains will itself vary from situation to situation, since individuals can postpone realization of capital gains, thus postponing payment of capital gains tax and shrinking the present value of the tax paid on capital gains account. For example, if a share bought for $100 today rises in value at 8 per cent per year, capital gains tax payable upon sale r years in the future will be t_g^*[(1.08)ⁿ - 1], where t_g^* is the nominal rate of tax on capital gains, but the present value of this tax (evaluated at 8 per cent) will be t_g^*[l — (1/1.08)”]. This is what was meant above by the effective rate of tax on capital gains. It is clearly, from this example, a decreasing function of the length of time that the stock is held. In the United States, the effective rate of capital gains tax can in fact be zero, since assets held until the death of the owner pass to his heirs, who in turn are taxed only on increases in value that take place after they have inherited the property.

While the property tax and capital gains provisions tend somewhat to offset the distorting effects of the corporation income tax, the traditional treatment of income from owner-occupied housing works to reinforce the distortions implicit in the corporation income tax. Obviously, owner-occupied housing generates income in real terms, but traditionally this income has not been a part of the personal income tax base. As a consequence, this important part of the income generated by capital in the unincorporated sector of the economy pays neither corporate nor personal income tax, while the income generated in the corporate sector is subject to both.

Empirical estimation . Harberger (see Krzy-zaniak 1966) has attempted to derive rough estimates of the cost to the U.S. economy of the pattern of distortions created by the differential taxation of capital in different uses. He incorporates into a single model, which distinguishes between the corporate sector and the noncorporate sector, the effects of corporate income taxation, property taxation, capital gains taxation, and the exemption from personal income taxation of the imputed income from owner-occupied housing. Making conservative assumptions about the elasticities of response of the economy to the various distortions involved, Harberger estimates the “efficiency cost” of the U.S. pattern of taxation of income from capital at approximately $2 billion per year. This estimate concerns only the costs associated with the misallocation of a given capital stock, costs which would be zero if all income from capital were to be taxed at a given constant rate. It does not take into account the possible effects of the taxation of income from capital upon the size of the capital stock itself (through the influence of taxation on the rate of saving), nor does it fully incorporate the effects of various special provisions (e.g., percentage depletion) affecting specific industries. Hence, it is a conservative estimate in this respect as well.

Incidence

The incidence of the corporation income tax has long been the subject of debate among economists, a state of affairs which is likely to continue for some time. Underlying this debate are some genuine differences, both analytical (reflecting different assumptions about the behavior of firms) and empirical (reflecting differing views about, for example, the quantitative response of saving to the disturbances engendered by the imposition of the tax). However, expositions of the effects of the corporation income tax at times contain serious conceptual and analytical errors which should long since have been laid to rest.

Perhaps the main source of confusion has been the conception of the incidence of the tax as falling either (a) on stockholders, or (b) on consumers, or (c) on workers, or on some combination of these three. There are three errors involved in this traditional trichotomy. The first has to do with the use of the term “stockholders” rather than “owners of capital”; the second relates to the distinction between consumers and workers; and the third concerns the assumption, which is usually implicit when the trichotomy is stated, that none of the three groups will gain as a consequence of the tax.

The distinction between stockholders and owners of capital. The idea that the burden of the corporation income tax will fall on the stockholders of the affected corporations is a valid one within the confines of standard short-run equilibrium analysis. This is because in the short run, with the capital of each corporation considered as a fixed factor of production, the earnings of equity capital represent the residual share. This residual share is assumed, in traditional short-run models of competitive and of monopolistic behavior, to be maximized by the firm. So long as the demand and cost conditions facing the firm are unchanged— the conventional assumption—the output which generated maximum profit before the tax was imposed will also yield maximum profit in the presence of the tax.

Although the above analysis is correct for the short run, a major change occurs when longer-run adjustments are allowed for. Here the appropriate assumption is that the after-tax rate of return is equalized between the corporate and the noncorporate sectors. Any fall in the rate of return perceived by the owners of shares will therefore also be perceived by the holders of other kinds of titles to capital, and the isolation of stockholders as the relevant group when assessing the incidence of the tax is no longer correct. The relevant group becomes owners of capital, once attention is focused on the longer-run incidence of the tax.

The distinction between consumers and workers. Once the above is recognized, the error implicit in the distinction between consumers and workers becomes apparent. Since all income-earners in the community are owners of either labor or capital resources or both, the reduction in real income implicit in the tax must reflect the sum of the reductions in the real incomes of these two groups. That is to say, a distribution of the burden of the tax between people in their role as owners of capital, on the one hand, and people in their role as sellers of labor services, on the other, is exhaustive, leaving no room for an additional burden to be borne by consumers.

This is not to say that, within each group, different individuals will not bear different burdens because of differences in their consumption patterns. In general, those, whether capitalists or workers, who consume a greater-than-average proportion of “corporate” products as against “noncorporate” products will be relatively harder hit as a consequence of the tax than those who have the opposite bias in their consumption pattern. But the extra benefits accruing to those consumers with relatively “noncorporate” consumption patterns must, because of the deviations of these patterns from the average, exactly offset the extra burden borne by those with relatively “corporate” consumption patterns. (This statement is precisely correct if only the first-order effects of the change in tax regime are taken into account. When second-order effects are considered, there emerges an “excess burden” of the tax, deriving from the distortion of consumption patterns and resource allocation which results from the tax. Excess burden, however, is conventionally left out of account in discussions of incidence, for otherwise the sum of all burdens allocated would exceed the yield of the tax; that is, incidence is conventionally defined as dealing only with first-order effects.)

There is, nevertheless, a way in which sense can be made out of a statement like “The tax is wholly passed on to consumers.” For if analysis reveals that the real incomes accruing to labor and capital fall by equal percentages as a result of the tax, then it is equally convenient to describe the tax as being borne fully by people in their role as consumers. And if labor’s real income falls by 10 per cent as á consequence of the tax, while capital’s falls by 20 per cent, it is just as convenient to regard the tax burden as being a 10 per cent reduction of the real income of consumers as such (the percentage point fall common to the two groups), plus an additional 10 per cent reduction falling upon the owners of capital. But if this approach is taken, there is no burden to be allocated to labor in the example just cited, just as there would be none to allocate to capital if its real income fell by 10 per cent and labor’s by 20 per cent. Thus the idea of a three-way division of the burden remains illogical even when a plausible device is found for ascribing some of it to consumers.

The “no-gain” fallacy. The third error involved in typical presentations of the trichotomy—the implicit assumption that no group will gain as a consequence of the imposition of a corporation income tax—is perhaps the most serious of all, since it leads to a gross misapprehension of the nature of its incidence. It is not at all true that the share of the total burden of the tax which falls on capital must lie between zero and 100 per cent; a much more plausible range for capital’s share runs from a.* to l/b_c (where a_k is the proportion of the national income accruing to capital and b_c is the fraction of the capital stock which is occupied in the corporate sector), though even this range can easily be exceeded.

To demonstrate the plausibility of the suggested range, assume that, with fixed and fully employed stocks of labor and of capital and holding the wage rate constant as the numeraire, the net-of-tax return to capital remains unchanged as a consequence of the tax. The nominal income of both labor and capital is therefore unchanged, but the real income of both groups falls because the prices of products of the corporate sector must rise to accommodate the tax. Labor and capital must therefore suffer equiproportionally as a consequence of the tax, capital’s fraction of the total burden being a,_e, its share in the national income.

The other end of the range is generated when the gross-of-tax rate of return to capital remains unchanged as a consequence of the tax. The net-of-tax rate of return must therefore fall by the percentage rate of the tax imposed. But the equilibrium condition for the capital market assures that if the net-of-tax rate of return falls by this percentage in the corporate sector, it must fall by the same percentage in the noncorporate sector. Since the fall in the return to capital in the corporate sector just reflects the tax paid, the parallel fall in the noncorporate sector reflects that capital is bearing more than the full burden of the tax, the ratio of capital’s loss to the full burden of the tax being the ratio of total capital to corporate capital, or l/b_c In this case, therefore, labor gains an amount equal to the reduction in real income per unit of capital times the amount of capital in the noncorporate sector.

The “plausible limits” just outlined can be derived from a two-sector model with homogeneous (of first degree) production functions, on the assumption that the elasticity of substitution between labor and capital is infinite in one sector or the other. If this elasticity is infinite in the untaxed (noncorporate) sector, then so long as some production takes place in that sector in the post-tax equilibrium, the relationship between the return to a unit of capital and the wage received by a unit of labor must be the same as in the pretax equilibrium. Capital and labor therefore must bear the same percentage losses of real income as a result of the tax. When, on the other hand, the elasticity of substitution between labor and capital is infinite in the corporate sector, the post-tax gross-of-tax return per unit of capital must bear the same relationship to the wage of labor as prevailed before the tax was imposed. Hence the net-of-tax return per unit of capital must fall, in both sectors, relative to the wage of labor, by the percentage of the tax, and capital must accordingly bear (l/b_c) times the full burden of the tax.

Strikingly, these same “plausible limits” come into play when the elasticity of substitution is zero in one of the two sectors and non-zero in the other. When the corporate sector has a zero elasticity of substitution between labor and capital, the reduction in its output resulting from the tax leads to the ejection of labor and capital from that sector in the fixed proportions given by its technical coefficients of production. Suppose that the corporate sector uses labor and capital in the ratio of 1:2; as it contracts, it must therefore eject the factors in these proportions. If, now, the noncorporate sector was, in the pretax equilibrium, using the two factors in just these proportions, it will be able to absorb the “rejects” from the corporate sector without any change in relative factor prices. And since factor prices in the noncorporate sector are already net-of-tax, this means that both factors must suffer in the same proportion as a consequence of the tax, just as in the case of an infinite elasticity of substitution in the noncorporate sector.

The above result occurs when labor and capital were initially used in the same proportions in the two sectors, and it must be modified when the initial proportions differ. If the corporate sector ejects labor and capital in the ratio of 1:2, while the noncorporate sector was initially using them in the ratio 1:1, the noncorporate sector (which is assumed to have a non-zero elasticity of substitution) must alter its factor proportions so as to absorb relatively more capital. Capital’s return must therefore fall relative to labor’s, in order for equilibrium to be restored; and capital will bear more than the fraction a_k of the total burden of the tax. Conversely, if the noncorporate sector were initially more capital-intensive than the corporate sector, using the factors, say, in the proportions 1:3, the relative price of labor would have to fall so as to enable this sector to absorb the “rejects” from the corporate sector; and capital would end up bearing less than a* of the total burden of the tax.

Thus, when the elasticity of substitution between labor and capital is zero in the corporate sector, capital will bear the fraction a/, of the total burden if the two sectors have equal factor intensities; will bear more than a_k when the corporate sector is the more capital-intensive of the two; and will bear less than a/, when the corporate sector is the more labor-intensive of the two. Exactly how much more or less than a_k capital will bear depends upon the extent of the difference in factor proportions between the two industries, on the elasticity of substitution between labor and capital in the noncorporate sector (which determines the ease with which it can absorb new factors in proportions different from those initially used), and on the elasticity of substitution on the demand side between corporate products and noncorporate products (the greater this elasticity, the sharper the decline in demand for corporate products as a consequence of the tax, the larger the ejection of resources by this industry, and therefore the greater the shift in relative factor prices required to restore equilibrium).

When, on the other hand, the elasticity of substitution between labor and capital is zero in the untaxed industry and non-zero in the taxed industry, capital tends to bear more than the full burden of the tax. In this case, when the initial factor proportions are the same in both industries, the fixity of proportions in the untaxed industry assures that they will remain the same even after the tax has worked out its full effects. The relative returns to labor and capital, being governed in this case by the proportions in which the factors are used in the taxed industry, will remain the same, gross-of-tax, as they were in the pretax equilibrium. Capital’s return net-of-tax will fall by the amount of the tax, but, as in the case of infinite elasticity of substitution in the taxed industry, the reduction will occur for capital used in either industry. The total reduction in capital’s earnings will be (l/b_c) times the yield of the tax, reflecting a very substantial “overbearing” of the tax by owners of capital and a corresponding net gain to those whose income accrues principally from the sale of labor services.

The above result (for a zero elasticity of substitution in the untaxed industry) is modified when the initial factor proportions are different in the two sectors. If the corporate sector is initially more labor-intensive than the noncorporate sector, the ejection of capital and labor resources in the proportions in which the latter sector will absorb them will make the corporate sector still more labor-intensive. A readjustment of factor prices against labor and in favor of capital will have to occur, and capital will end up bearing less than (l/b_c) times the observed yield of the tax. Conversely, if the corporate sector is initially more capital-intensive than the noncorporate sector, and has a zero elasticity of substitution, factor proportions will have to alter to make the corporate sector still more capital-intensive, requiring a shift of the gross-of-tax ratio of factor prices against capital. Capital will then bear more than (l/fc_c) times the observed yield of the tax.

When capital bears 100 per cent of the burden. Falling well within the “plausible limits” of incidence defined by a_k and l/b_c is the case in which capital bears 100 per cent of the burden of the tax. This result therefore cannot be regarded as being an extreme outcome, as the conventional use of the capital-labor-consumer trichotomy implies. Added insight into the plausibility of capital’s bearing the full burden of the tax can be gained from an analysis of the case in which each industry is characterized by a Cobb—Douglas production function and in which the elasticity of substitution in demand between the products of the two sectors is unity. Letting X represent the quantity of the product of the corporate sector, Y the quantity of the product of the noncorporate sector, P_x and P_y their respective prices, and Z the national income, the unit elasticity of substitution between X and Y implies

(1) XP_x = ?Z; YP_y= (l-a)Z,

where a is the fraction of Z which is spent on X. Competitive behavior of producers of X and of Y, together with the Cobb-Douglas functions X = K^?_alphaL_alpha^(1-?), Y = K^?_yL_y^1-?), where ? and y are constants, lead to the relations

Here K_x and K_y represent the amounts of capital used in the X and Y industries, respectively, and L_x and L_y refer to the corresponding amounts of labor. The price of labor is denoted by P_L, this being the same in the two industries. The cost of the services of a unit of capital is denoted by P»» for the corporate sector and by P_k for the noncorporate sector, the former including the corporation income tax and the latter, of course, not including it. If T is the rate of corporation income tax applied to the earnings of capital in sector X, then Pt = P_tr(l — T), since the after-tax earnings (as distinct from the before-tax cost) of a unit of capital are assumed to be brought to equality in both industries through the workings of the capital market.

It can be seen from relations (1) and (2) that labor will always earn a constant fraction of the national income, regardless of whether a corporation income tax exists or not. 1’his already guarantees that exactly the full burden of the corporation tax must in this case be borne by capital. The precise way in which the burden reaches all units of capital can be seen by analyzing the relations derived from (1) and (2):

(3) K_xP_kz = ?_?Z; K_yP_k = ?(1 - ?)Z.

From these it results that [K_xP_kz/K_yP_k] is a constant equal to ??/[?(l — ?)]. But since P_k = P_kx(l - ?), this means that K,./[K_;/(1 - r)] will also be a constant—that is, the ratio (K_r/K_y) will vary directly with (1 — r). If, with no tax at all, there were 150 units of capital in each sector, a tax of 50 per cent will eventually result in there being 100 units of capital in X and 200 in Y. The 200 units of capital in Y will earn the same fraction of national income as was previously earned by the 150 units of capital in Y; hence the net-of-tax return to capital will have been reduced by a quarter, say, from $1.00 to $.75 per unit. The 100 units of capital in X will cost entrepreneurs $1.50 per unit and will therefore have the same total cost as the 150 units employed in X at a unit cost of $1.00 before the tax was imposed. But the aftertax earnings of capital in X will, like those of capital in Y, have fallen from $1.00 to $.75 per unit. Overall, capital will have lost $75, represented by the reduction of $.25 per unit spread over all 300 units, and this amount will be precisely equal to the yield of the tax to the government.

The result obtained in the above example applies not only to all cases fulfilling relations (1) and (2), which are derived on the basis of unit elasticities of substitution in demand between the two products, and in production between the two factors in each industry. It has been shown elsewhere (see Harberger 1962) that the same result obtains so long as the three critical elasticities of substitution are equal, regardless of their magnitude.

The general-equilibrium, two-sector model. All the cases presented above are special cases of a general-equilibrium, two-sector model of the incidence of taxation, in which the incidence of the corporation income tax is shown to depend in a specific way on the three critical elasticities of substitution and on the relative factor intensities of the two sectors. This model, based on the assumptions that the supplies of capital and labor are not influenced by the presence or absence of the tax, that competition prevails in both the corporate and noncorporate sectors, and that per-unit net-of-tax earnings of each productive factor are equalized between sectors, was first presented by Harberger (1962) and further elaborated by Mieszkowski (1967). They have adapted the model to explore the implications of various possible types of monopolistic and oligopolistic behavior in the corporate sector; the results of the original model have proved quite insensitive to plausible allowances for noncompetitive behavior.

The chief weakness of the model appears, at this writing, to be the assumption that the path of the capital stock through time is independent of the rate of corporate taxation. If, through a tax-induced reduction in the net rate of return on capital and/or through a tax-induced shift in the distribution of disposable income, the rate of saving is affected, the relative supplies of capital and labor will gradually diverge from the path they would have followed in the absence of a corporation income tax, with consequent effects on the distribution of income. The difficulties confronting attempts to resolve this issue are twofold. First, a dynamic rather than a comparative-static approach is required, which, while not a serious obstacle as such, involves additional parameters whose magnitudes are difficult to estimate and requires the specification of the precise nature of the dynamic structure of the economy. A great deal of further work is needed before our understanding of the economy’s workings can advance to the point where these dynamic aspects can be treated with a degree of precision comparable to that with which problems of comparative statics are handled today.

The second difficulty is conceptual rather than practical. In a comparative-static approach to incidence, excess-burden being neglected, the sum of the changes in real income of the separate groups of the economy is a global reduction in real income equal to the proceeds of the tax; this is no longer true when a dynamic framework is employed. If the rate of saving is reduced by the corporation income tax, the future incomes accruing to individuals are reduced not only because the tax has to be paid each year, but also because less has been saved in the years since the tax was introduced. But it would be wrong, in estimating the incidence of the tax, to count both (a) the full reduction of real income in the year the tax is paid and (fc) the future reduction in real income stemming from the reduction in savings induced by the tax. If one counts (a), one has already accounted for the present value of the future reduction in real income. To take explicit account of the future effects of changes in the savings pattern, one would properly have to convert the entire calculation of incidence to a consumption rather than an income basis and count (c) the current reduction in consumption resulting from the tax paid today plus (d) the future reduction in consumption occasioned by the reduction in future incomes stemming from the current tax-induced reduction in the rate of saving.

When the above difficulties are considered, it appears that the current-income approach (i.e., counting only (a) as the measure of incidence) is preferable, on grounds of both clarity and convenience, to approaches attempting to introduce dynamic responses into the measurement of incidence. Nevertheless, the dynamic responses in question here are of substantial interest in their own right, even if they are not linked to the analysis of incidence. The study of this aspect of the effects of corporation income taxation has only recently begun, the most important early efforts being those of Krzyzaniak (1966) and Sato (1967).

Arnold C. Harberger

BIBLIOGRAPHY

Goode, Richard B. 1951 The Corporation Income Tax. New York: Wiley.

Harberger, Arnold C. 1959 The Corporation Income Tax: An Empirical Appraisal. Volume 1, pages 231-250 in U.S. Congress, House, Committee on Ways and Means, Tax Revision Compendium. Washington: Government Printing Office.

Harberger, Arnold C. 1962 The Incidence of the Corporation Income Tax. Journal of Political Economy 70:215-240.

Harberger, Arnold C.; and Bailey, Martin J. (editors) 1968 The Taxation of Income From Capital. Washington: Brookings Institution.

Krzyzaniak, Marian (editor) 1966 Effects of the Corporation Income Tax: Papers Presented at the Symposium on Business Taxation. Detroit, Mich.: Wayne State Univ. Press. ? See pages 107-117, “Efficiency Effects of Taxes on Income From Capital,” by Arnold C. Harberger.

Mieszkowski, Peter 1967 On the Theory of Tax Incidence. Journal of Political Economy 75:250-262.

Musgrave, Richard A. 1959 The Theory of Public Finance: A Study in Public Economy. New York: McGraw-Hill.

Musgrave, Richard A.; and Krzyzaniak, Marian 1963 The Shifting of the Corporation Income Tax: An Empirical Study of Its Short-run Effect Upon the Rate of Return. Baltimore: Johns Hopkins Press.

Sato, Kazuo 1967 Long-run Shifting of the Corporation Income Tax. Unpublished manuscript.

Stigler, George J. 1963 Capital and Rates of Return in Manufacturing Industries. National Bureau of Economic Research, General Series, No. 78. Princeton Univ. Press.

yearbook of national accounts statistics. ? Published by the United Nations since 1958. Contains detailed estimates of national income and related economic measures for some 76 countries.

IV. PROPERTY TAXES

Property taxes are general and recurring taxes on owners or users of property, based on the capital value or the annual rental value of the assets. They are considered distinct forms of taxation, although many other taxes reach some facet of property ownership or use, including taxes on the income from property, taxes on realized appreciation in property values (capital gains), taxes in a number of European countries on net wealth, wealth transfer taxes or succession duties, and taxes on selected types of personal property, such as motor vehicles.

Property taxation is widespread and is typically used by local rather than national governments. It provides the overwhelming bulk of local government tax revenues in the United States, in the other developed English-speaking countries, in the Netherlands, and in a number of the developing countries, especially those exposed to the British tradition. The tax is also important to local governments in Belgium, Denmark, France, Germany, and Japan, countries in which it provides roughly 20 to 30 per cent of local government tax revenues.

Property taxes are most important relative to the over-all fiscal system in the English-speaking countries, where the role of local governments tends to be a large one. In Canada and the United States in recent years, the property tax has accounted for more than 45 per cent of the tax revenues of all subnational governments (including states or provinces), about 16 per cent of the total tax revenues of all governments, and more than 4 per cent of national income. It accounted for more than 13 per cent of Ireland’s total taxes; more than 11 per cent of Britain’s; and 6 to 8 per cent in Australia, New Zealand, South Africa, Denmark, and Japan.

In most countries, the tax applies to land and/or buildings only, but in the United States, some types of personal property are subject to the property tax in all but four states. Business and farm equipment and inventories are commonly taxed, and account for perhaps 75 per cent of the personal property tax base. Motor vehicles are subject to this tax in more than half the states, and household effects are taxed somewhat more rarely. Intangible personal property—securities, bank deposits, etc.—is infrequently taxed; intangibles have provided only about 2 per cent of property tax revenues recently, and all personal property about 19 per cent. In some Canadian provinces, personal property is taxed, but it provides only 1 per cent of revenues nationally; in Japan, personal property provides nearly 40 per cent of revenues.

The most common basis for taxing real property is its annual rental value, in practice usually gross rents assessed as of some earlier date, with statutory rather than actual allowances for expenses. Real property is taxed on the basis of its capital value mainly in the United States, Canada, South Africa, Germany, Austria, and Denmark. The major difference between the customary annual rental value system, such as the British, and the American-Canadian capital value variant is the treatment of land and vacant improvements. In the British system, the tax is based on the rental value of property in its present, actual use, and vacant properties therefore are not taxed. In the American system, in theory, property is valued at market value or some fraction thereof. Market value is, in an equilibrium situation, the capitalized value of expected net returns from property in its most profitable lawful use, not its present use. The capital value basis therefore tends to favor optimal use of land somewhat more than the customary annual value basis does.

The evolution of European property taxes (and their American descendant) from feudal dues into a general tax on property, and their subsequent narrowing to taxes on land and buildings, has been traced by Seligman (1895). Jensen (1931, chapter 2) gives a similar history of the American property tax. The importance of the property tax in revenue systems has declined over the years, with the growing role of national as compared with local finance. In the United States (and in Canada, as well) the property tax has also been displaced by the adoption of consumption and income taxes by state governments, beginning after 1910 but especially in response to the collapse of property values in the early 1930s. Subventions from state tax revenues increasingly replaced local property taxes, and the property tax as a proportion of total state-local taxes declined from 80 per cent in the 1920s to less than 50 per cent by 1946. However, since 1950, despite very rapid increases in the total scale of state-local finance, the property tax has maintained approximately the same relative importance (U.S. Bureau of the Census 1964).

Experience in the United States . In part, the recent buoyancy of the American property tax is related to the large role of receipts from housing in the tax. Estimates are that in 1957 housing provided about 41 per cent of tax revenues and about 44 per cent in 1962. Real estate taxes on housing amount to an average of one-sixth or more of annual rental receipts, or of cash expenditures for housing in the case of owner-occupants. This is in effect an excise tax at a rate far higher than that on any other broad category of consumer expenditure in the United States. Housing property taxes equal, on the average, about 1.5 per cent of property values (U.S. Bureau of the Census 1963a).

Slightly less than 10 per cent of American property taxes is derived from real and personal farm property. Relative to property values, farm property taxes are lower than those on nonfarm housing or business property—less than 1 per cent in recent years. However, farm property taxes equal nearly 10 per cent of net farm income, and only a slightly smaller fraction of national income originating in agriculture.

Roughly 45 per cent of the tax comes from non-farm business property. Business property taxes are especially high, however measured, for railroad, pipeline, and other public utility companies. These firms are markedly real-property intensive enterprises and are, furthermore, politically vulnerable to discriminatory taxation. In 1957, estimated property tax payments were equal to 6.3 per cent of net output for railroads and public utilities, 1.4 per cent for manufacturing, and 1.8 per cent for other nonfarm business property.

The American property tax is not a single uniform tax institution but, in reality, thousands of different taxes reflecting differences in the legal coverage of the tax, the economic tax base available, the expenditure requirements to be financed, and the resulting tax rates among the 82,000 governmental units which rely upon the property tax. The tax tends to be of least importance in state-local fiscal systems in the southeastern United States and most important in New England, the Great Lakes states and the northern Plains states. The varying role of state taxes and state aid to local government explains much of this variation, with the property tax more important in those parts of the country where the state government’s financial role is smallest.

Because urban government is costly, property tax rates are higher in urban than in rural areas and higher in the more urbanized states, notably in New England and the Middle Atlantic states, where tax rates frequently exceed 2.5 per cent of the market value of taxable property. In contrast, tax rates in most southern and Mountain states average less than 1 per cent (U.S. Advisory Commission . . . 1962, tables 37 and 41). Urbanization does not explain all the differences, however, since property tax rates in large northeastern cities are distinctly higher than in large cities elsewhere, and those in southern cities distinctly lower (U.S. Bureau of the Census 1963a). These large regional differences do not seem to have had major effects on location of industry, however, in view of the relative rates of growth of states with high and low property and other business taxes (Due 1961, p. 171).

Within large urban areas in the United States, tax rate differences are considerable among the great numbers of separate taxing jurisdictions operating in most individual metropolitan areas, and no doubt they do affect locational patterns. In the older parts of the country, per capita taxable property values tend to be lower in central cities than in their suburbs, expenditure requirements higher, and effective tax rates higher. This will tend to spur migration of business and high income residents from central city locations to suburban ones if tax differentials against the central cities are widening, as appears to be the case.

Among suburban taxing jurisdictions, property tax rates are usually lower and the level of public services higher in communities with higher property values per capita—either because they are dormitory suburbs with high-value houses or because they contain heavy concentrations of nonresiden-tial property, a situation which encourages land use planning designed to maximize the fiscal position of individual suburbs (Netzer 1962, p. 193). The results may be both inefficiency in location patterns and, to the extent that racial and other barriers limit intrametropolitan mobility, adverse effects on interpersonal equity. However, these intrametropolitan property tax differentials may be narrowing over time; the evidence is mixed in this regard.

Shifting and incidence . In theory, taxes on the value of sites—bare land—rest on the owners of the sites at the time the tax is initially levied or increased. The tax cannot be shifted forward to other users of the land, since shifting can occur only if supply can be reduced, which is not possible for land. Prospective purchasers of the sites, faced with a new or higher annual tax burden, will reduce their bids, and the higher tax will be capitalized in the form of lower land prices. There are some complications in this analysis, as Simon (1943) points out, but it is generally accepted.

In general, property taxes on improvements and on tangible personal property used in business can be expected to be shifted forward to final consumers of business services and occupants of housing. This is because the taxes will discourage new real investment in these forms, and over time the reduced supply of capital assets will raise their prices. Owner-occupants of housing will themselves bear higher property taxes because there is no way they can be shifted.

This, at any rate, is the theoretical conclusion in partial equilibrium analysis. A general tax on capital could conceivably be shifted backward to owners of capital, in the form of lowered rates of return on the whole stock of capital, provided that the supply of savings is not responsive to interest rates. Another complication is the time lag required to shift taxes on physical capital forward, since the annual increments are usually small fractions of the total stock. In addition, the partial and unequal nature of the property tax limits shifting. Firms competing in national markets are able to shift local property taxes only to the extent that these taxes are common to their competitors or reflect the value of public services financed by these taxes. But, on the whole, most business property taxes are probably shifted forward and much of the remaining portion possibly shifted backward to land owners, by reducing local land values.

Empirical studies of the incidence of the American property tax by income class based on these “shifting” assumptions have generally agreed that the property tax is, on balance, regressive when compared with current money income. Because of the forward shifting of a substantial part of business property taxes, property taxes on nonresiden-tial property are, in part, equivalent to a general consumption tax, regressive through much of the income range. Property taxes on owner-occupied housing and on rented housing appear to be even more regressive than taxes on nonresidential property. This is mainly because housing consumption outlays constitute a larger proportion of lower than of higher current money incomes.

In combination, residential and nonresidential property taxes are markedly regressive for the lowest income groups but only mildly regressive in the middle ranges of the income distribution. If no allowance is made for income tax savings due to the deductibility of property taxes, the latter are progressive for the highest income groups. When measured on the basis of a broader income definition, or one which averages income over a longer time span, the property tax is very nearly proportional in its incidence. The benefits from expenditures financed from the property tax are distinctly progressive in their incidence, notably in connection with education, as Morgan and his colleagues (1962) show.

On balance, therefore, the American property tax is no mean contributor to income redistribution from the richer income groups to the poorer ones, considered in the aggregate. However, in view of the wide dispersion about the means within income classes and of the many geographic differences, the redistributive effects with regard to individual households are highly uneven; the tax contains a substantial element of interpersonal inequity, however progressive or proportional it may be in the aggregate.

Allocative effects . In general, the American property tax (and property taxes in Canada, Britain, Ireland, and other high property tax countries) tends, over time, to shift resources in the aggregate from private construction to education and other public services. This general effect of the tax, like other economic effects, may not be visible in the empirical evidence, since it can readily be overwhelmed by other factors, such as housing subsidies and the like.

Property taxes also discriminate among inputs, encouraging the substitution of other inputs for real property; to the extent that firms and industries are limited in their opportunities to substitute, the property tax is then discriminatory among industries. Railroads are perhaps the best example of this. The competitive decline of the rails in the face of new transport technology was no doubt inevitable, but it was hastened by the property tax. Railroads are inherently real-property intensive and thus are subject to heavier taxes of this type than are their air, water, and road competitors. Rising property tax rates in the postwar years contributed to a rate of increase in rail charges which hardly assisted the carriers in their efforts to compete.

The American property tax tends to discourage housing in general, since it imposes taxes on this use of the consumer’s dollar which are markedly higher than those on most other uses. Although property taxes are frequently very high in dormitory suburbs, whether measured by house value or by personal income, the deterrent effect there may be small, since the tax is directly tied to school and other expenditure benefits realized by householders. However, in large cities, the tie to expenditure benefits is tenuous for many housing consumers, and property taxes amounting to large fractions of gross rental receipts—25 per cent or more in large northeastern cities—probably inhibit the construction of new rental housing and the rebuilding of the older cities. In any event, property tax rebates or reductions for selected classes of new housing have proved to be among the most effective stimulants yet devised.

Administration—Assessment problem. The fundamental administrative problem in property taxation is that of valuing or assessing property. In a number of countries, including Britain, valuation is done by a central government agency. In the United States, however, assessment of most classes of property is made (except in Hawaii) by local assessors; according to the U.S. Advisory Commission on Intergovernmental Relations (1963, p. 101), there are probably eighteen thousand assessment districts in the country. The quality of local assessors and assessment varies widely; it has been vigorously criticized by students of the problem since the last quarter of the nineteenth century. Some assessors are elected, part-time amateurs using primitive methods and tools; other assessment organizations are large, professional agencies applying all the technological aids available. In the best-administered jurisdictions in the United States, owners of single-family houses with similar market prices are likely to have assessments which vary by less than 10 to 15 per cent; in the worst the average variation may be far in excess of 50 per cent.

Some observers, such as the Advisory Commission, noting the important revenue role of the property tax, have urged its administrative rehabilitation. This involves limiting the coverage of the tax to classes of property which can be discovered and valued practicably, devising large enough assessment districts so that all can be served by full-time professional staffs, and greatly enlarging the role of the states in the provision of technical assistance to and supervision of local assessment.

Other observers are much more pessimistic. They note the inherent difficulties of valuing widely differing assets, only a few of which are actually sold within a short span of time and some of which—like large industrial plants—are never sold. They regard the standard of “good” assessment—assessments for similarly market-valued properties differing in the aggregate by no more than 20 per cent from the average—as an un-acceptably low level of performance as compared with sales and income tax administration. They

query whether some of the most glaring disparities in assessment practices—such as discrimination among classes of property within a city—are not in reality accommodations to a level of taxation which, if applied uniformly, would be economically and/or politically intolerable.

Prospect . Despite its inequities, its questionable impact on economic efficiency, and the poor quality of its administration (at least in the United States), the property tax persists and in revenue terms has been holding its own in the past few years in the United States and in a number of other countries. Part of the reason for this is that property tax revenues have risen rapidly in recent years, along with the level of economic activity. The market value of taxable property—the economic base of the tax—has risen almost as rapidly as gross national product in the postwar period, an apparent interruption to a long-term downward trend in capital-output ratios. Burkhead (1963, p. 70) concludes that the property tax is a far more responsive source of local government revenue than its traditional critics have allowed.

Site value tax (”single tax”). Perhaps the most vigorously advocated alternative to the prevalent systems of real property taxation is the site value tax, first propounded as “the single tax” by Henry George in 1879. The equity argument for site value taxation is that bare site values, or location rents, are created by population growth and general community improvements rather than by the actions of individual landowners, and that therefore taxation of this “unearned increment” is highly equitable. The resource allocation argument is that the site value tax applies to a surplus—the differential returns available from conducting an activity at particular sites—and therefore is economically neutral. Taxation does not reduce the supply of sites, but lowers their after-tax capitalized net returns, or price. But this neutrality is in contrast with the existing property tax, which, by applying to improvements as well as to site values, discourages new construction in general. The existing tax, moreover, tends to encourage low intensity uses, or holding of land idle for speculation, since taxes are lower if improvements are minimal. As noted earlier, the British type of property tax has this effect to a marked degree. Shifting to a site value tax would tend to foster improvements in general, and would discourage withholding of land from use, relative to present property tax practices in most places.

Site value tax advocates have tended to claim much more than this for their proposal. Some, for example, argue that site value taxation by itself can cure most of the ills of the large older cities. Opponents have presented three principal arguments against it. The first is the difficulty of separating site values and improvement values in the case of improved property; this appears to be a real difficulty administratively but not conceptually. The second is an equity argument: large windfall losses and gains would stem from a shift from the present system to the site value tax and would be intensified by the fact that many present landowners have not been the recipients of the “unearned increments” but have paid prices reflecting these to previous owners. The third is the problem of revenue adequacy. It has been estimated that to replace the present yield of American taxes on real property with a tax solely on site values would absorb more than the entire (before-tax) rent of land. These arguments suggest that a partial replacement is perhaps the maximum possibility.

Differentially higher taxation of land, or complete exemption of improvements from general ad valorem taxes on real estate, is practiced in the United States only in Pittsburgh and Hawaii but is widespread in Australia, New Zealand, South Africa, and Canada. In Australia and New Zealand, most local taxing units have exempted improvements from taxation, this trend beginning in the 1890s. In South Africa, most local authorities have either differential taxation or complete exemption of improvements. In Canada, differential taxation is widespread in the four western provinces. Because of so many environmental differences other than the property tax, it is difficult to discern whether the advantages claimed for the site value tax have been realized in these places. Most economists, however, agree that the site value tax should have better resource allocation effects than the prevalent property tax institutions.

Dick Netzer

BIBLIOGRAPHY

Burkhead, Jesse 1963 State and Local Taxes for Public Education. Syracuse Univ. Press.

Due, John F. 1961 Studies of State-Local Tax Influences on Location of Industry. National Tax Journal 14, June: 163-173.

Heilbrun, James 1966 Real Estate Taxes and Urban Housing. New York: Columbia Univ. Press.

Jensen, Jens P. 1931 Property Taxation in the United States. Univ. of Chicago Press.

Morgan, James N. et al. 1962 Property Taxes and the Benefits of Public Education. Pages 288-308 in Michigan, University of, Survey Research Center, Income and Welfare in the United States: A Study. New York: McGraw-Hill.

Netzer, Dick 1962 The Property Tax and Alternatives in Urban Development. Regional Science Association, Papers and Proceedings 9:191-200.

Netzer, Dick 1966 Economics of the Property Tax. Washington, D.C.: Brookings Institution.

Robert Schalkenbach Foundation, New York 1955 Land-value Taxation Around the World. Edited by Harry Gunnison Brown et al. New York: The Foundation.

Seligman, Edwin R. A. (1895) 1928 Essays in Taxation. 10th ed., rev. New York: Macmillan. ? See especially pages 19-65, ’The General Property Tax.”

Simon, Herbert A. (1943) 1959 The Incidence of a Tax on Urban Real Property. Pages 416-435 in American Economic Association, Readings in the Economics of Taxation. Homewood, 111.: Irwin.

U.S. Advisory Commission ON Intergovernmental Relations 1962 Measures of State and Local Fiscal Capacity and Tax Effort. Report M-16. Washington: Government Printing Office.

U.S. Advisory Commission ON Intergovernmental Relations 1963 The Role of the States in Strengthening the Property Tax. 2 vols. Report A-17. Washington: Government Printing Office.

U.S. Bureau OF THE Census 1963a Census of Housing: 1960. Volume 5: Residential Finance. Washington: Government Printing Office. ? Contains data on real estate taxes in relation to property value, income, and rental receipts. This census is taken decennially.

U.S. Bureau OF THE Census 19636 Census of Governments: 1962. Volume 2: Taxable Property Values. Washington: Government Printing Office.

U.S. Bureau OF THE Census 1964 Census of Governments: 1962. Volume 4, no. 4: Compendium of Government Finances. Washington: Government Printing Office. ? Contains comprehensive data on property tax revenues and all other federal, state, and local government financial data for 1962, by states and counties. This census is taken quinquennially, in years ending in 2 and 7.

V. SALES AND EXCISE TAXES

Taxes on the production or sale of commodities are among the oldest taxes known; they play a significant role in the tax structures of most countries of the world. Despite rapid expansion of income taxation in the last century and widespread acceptance of the argument that by usual standards such taxation is superior, the sales and excise taxes have not only maintained their position but in many countries have increased in importance. Despite the long experience with these taxes, major disputes about them continue—on such questions as shifting and incidence, relative effects on economic welfare through resource reallocation, and effects on economic development and the maintenance of full employment.

Sales and excise taxes have traditionally been classified as forms of indirect taxation, although this term has fallen into disuse because there is no generally accepted delineation between such taxes and those labeled direct. On a somewhat different basis of classification, they are designated as consumption taxes (as distinguished from income, wealth, or other taxes), under the assumption, questioned below, that their burden is distributed in relation to consumer expenditures.

The distinction between excises and sales taxes is based on the scope of coverage. Excises apply to particular commodities or related groups of commodities (such as tobacco products), while sales taxes apply to broad categories of commodities, typically to all commodities other than those specifically exempted. Obviously, a broad system of excises, such as that of Spain, does not differ basically from a sales tax and can have broader coverage than a sales tax limited to certain categories, such as the British purchase tax, or one with widespread exemptions, such as those of the Canadian provinces. However, commodity taxes usually fall clearly into one category or the other, and the distinction is useful for purposes of analysis. The terminology as outlined is not universally employed; for example, excises are sometimes referred to as selective sales taxes, and some proposals for a federal sales tax in the United States have referred to the proposed levy as a general excise tax. But the concepts given are now those most commonly employed.

Historical development . Excises are among the oldest forms of taxation, dating back, in their rudimentary form, to ancient Rome. The first use in England came in 1643. France was a major user, especially under Colbert, in the seventeenth century. Except for a few early attempts, the United States did not employ excises until the Civil War, when an extensive system was introduced for war-financing purposes. Only the liquor and tobacco taxes survived, however. Ultimately, other excises were introduced: during World War I, the depression era, and World War II. Since World War II, the taxes have slowly been reduced, and most of the remaining ones, except those on liquor, tobacco, motor fuel and motor vehicles, and telephone service, were repealed in 1965. The states have confined excises largely to liquor, tobacco, and motor fuel, and these three categories are also the major revenue producers among the federal excises and the excises of other countries. In the newly developing economies excises are typically introduced as supplements to the customs duty system when domestic production of liquor and tobacco products is first undertaken.

Sales taxation dates back to the Spanish alcabala, introduced in the fourteenth century. Because this tax was blamed for the commercial decline of Spain, it was not adopted by other countries, and the sales tax did not come into widespread use until the twentieth century. The financial problems during and immediately after World War i led Germany, France, Italy, other Continental countries, and Canada to impose the tax. More countries followed in the depression years and during World War ii; among the most recent national sales taxes are those of Sweden and Denmark. The movement in the United States began in the depression years of the 1930s, when the states were squeezed between declining revenues from other taxes and increasing expenditure needs. Following the success of Mississippi with the tax in 1932, some 29 states levied a sales tax prior to World War n, although six subsequently allowed it to expire. In the postwar era the pressures of rising expenditures led additional states to impose the tax. There has also been a trend toward higher rates and broader coverage. The sales tax movement in the provinces of Canada has been similar to that in the United States.

Forms of excise and sales taxes. Excise taxes may be collected at the manufacturing, wholesaling, or retail level; the manufacturing level is by far the most common because the relatively small number of firms facilitates control. Excises may have specific rates, applied per unit of the physical product, as, for example, motor fuel taxes; or ad valorem rates, applying to the sale price. The former are easier to administer, if the product is highly standardized, but may be regarded as less equitable, since the tax rate does not rise in relation to value, and the yield of the tax is not automatically responsive to price changes. Excise taxes are also often classified in terms of general purpose or philosophy. Those on products such as liquor and tobacco, the use of which the government seeks to penalize as a matter of policy, are known as sumptuary taxes. Typically, these are highly productive of revenue. Luxury excises are ones designed to distribute tax burden in relation to ability to pay, as measured by purchase of luxury articles. Another group of widely used excises is directly related to motor vehicle use and is designed to distribute the costs of highways on the benefit principle. The United States provides a more direct link between the yield of these taxes and the costs of financing highways than do most countries.

The most significant classification of sales taxes is on the basis of stage of collection. Multiple-stage sales taxes are those which apply at two or more stages in the production and distribution channels. The complete turnover tax version applies at all stages in production and distribution: to the sales of materials and parts, as well as to all sales of the finished products—by manufacturers, distributors, and retailers. In practice the turnover taxes in use are not entirely complete or uniform. Lower rates are sometimes applied to sales by wholesalers (e.g., Germany), and retail sales may be excluded (e.g., Belgium). The turnover tax suffers from several major defects: integrated firms are favored over nonintegrated ones, thus encouraging integration; and the over-all tax on a particular product depends upon the number of stages in the production and distribution channels through which it passes.

The single-stage taxes are confined to one stage in production and distribution and avoid the disadvantages arising from the multiple application of the turnover tax. There are three major versions. The manufacturers sales tax, as used in Canada, applies to the sale by the manufacturer of finished products. The wholesale sales tax applies to the last wholesale transaction, that is, the purchase by the retailer. The retail sales tax applies to the final sale at retail. Each of these forms of tax will operate satisfactorily. On the whole, the retail tax, while collected from a much larger number of vendors than the others, gives rise to the fewest problems, because it can be applied to the actual selling price in virtually all instances. Avoidance of discrimination among various types of distribution channels is very difficult with the other single-stage taxes, since the taxable price is influenced by the structure of distribution. With a manufacturers tax, a manufacturer selling at retail is subject to a higher tax on a given product than one selling to a wholesale distributor. Attempts to meet this problem lead to serious complications in the tax. Nonretail taxes also tend to pyramid on the way to the final consumer, because of application of percentage markups. The retail tax, however, is not suitable in a country in which most retailing is conducted on a very small-scale, noncommercial basis, through family shops and market stalls.

The most recent version of the sales tax, the value added tax, as employed in France and accepted as the ultimate standard form of sales tax for the European Common Market countries, involves the application of tax to each firm in the production and distribution channels but only taxes the value added by the firm (in practice, the tax rate is applied to the firm’s gross sales, and from this figure is subtracted the tax paid during the period on goods purchased by the firm). Thus, the evils of the turnover form of tax are avoided, since the type of distribution channel will not affect the amount of tax liability, while the direct impact of the tax is spread out over a much wider range of taxpayers than is the case with the single-stage taxes, and much of the tax is collected from large firms at stages prior to retailing. This form may facilitate exclusion of capital goods from the tax. However, it offers little if any general advantage over the retail sales tax in situations where administration of the latter is feasible.

Present use . It is not feasible to present a detailed survey of existing sales tax structures. Table 1, however, gives a general outline for the major countries. Sales taxes are now employed by all countries of western Europe except Spain (which has an extensive system of excises), although the British purchase tax is of restricted scope. The tax is used by many states in India; by Pakistan, Indonesia, and the Philippines; and by Australia and New Zealand. In Latin America, the tax is used in Brazil by both the national government and the states, and it is a significant revenue source in Chile, Argentina, Uruguay, Ecuador, and Mexico. In Canada the tax is used both by the dominion government (manufacturers sales tax) and by nine provinces (retail taxes). In the United States the retail sales tax is employed in 43 states, but the federal government uses only a limited list of excises. Countries in early stages of economic development find customs duties the most satisfactory form of commodity tax.

Virtually every country of the world uses some form of excise tax, particularly on liquor, tobacco, and motor fuel; others, especially those not using sales taxes, also apply excises to various luxury goods. No simple summary of excise systems is possible.

It is very difficult to make precise comparisons between countries of relative dependence on various

*Table 1 — Sales taxation in major countries, 1967*
	Type of tax	Basic percentage rate
Europe:
France	value added	25
Italy	turnove, to retail	3
Germany	turnove, through retail	4
Belgium	turnove, to retail	6
Netherlands	turnove, to retail	5
Luxembourg	turnove, through retail	2
United Kingdom	wholesale, selected commodity groups	varying
Eire	retail	2.5
Norway	retail	10
Sweden	retail	6
Denmark	wholesale	9
Finland	wholesale and retail	10
Switzerland	wholesale	5.4
Greece	manufacturing	6
Austria	turnover, through retail	5.25
North America:
United States (43 states)	retail	2 to 5
Canada
(federal)	manufacturing	11
(9 provinces)	retail	5 to 8
Mexico	turnover	1.8 plus state supplements
South and Central America:
Argentina	manufacturing	10
Chile	turnover	6
Brazil
(federal)	manufacturing	varying, 2 to 10
(most states)	turnover	2 to 5
Uruguay	manufacturing	5
Ecuador	turnover	1.5
Honduras	retail	3
Asia:
Pakistan	wholesale	varying
India (states)	turnover or,retail	wide variation
Philippines	retail	varying
Australia	wholesale	varying, 10 to 30
New Zealand	wholesale	20

taxes. In the United States, sales taxes generally yield about 25 per cent of the revenues of the states in which they are levied, but in a few states they yield as much as 50 per cent. Excises yield about 11 per cent of federal revenue. The Canadian federal sales tax yields about 18 per cent of total federal tax revenue; the provincial sales taxes, 26 per cent of provincial tax revenue.

The turnover tax provides 42 per cent of the German federal revenues; other figures of sales tax yield include 35 per cent in France, 21 per cent in Italy, 40 per cent in Belgium, 19 per cent in the Netherlands.

Shifting and incidence . Traditionally it has been argued that both excises and sales taxes are typically shifted forward, through price increases, to the consumers of the products and thus are borne in relation to consumer spending on the taxed commodities. In purely competitive markets, with a fixed stock of goods on hand in the market period there will be no initial change in price, and temporarily the burden will be borne by the producers. But output and supply will fall, and the market price will rise. Over a long-run period the exact amount of the tax will shift forward, if the industry is one of constant cost conditions. Under increasing cost conditions, the ultimate increase in price will be less than the amount of the tax and a portion of the burden will be borne by the owners of specialized resources used in the industry, the prices of which decline as the volume of the product sold is reduced because of the higher commodity prices.

In nonpurely competitive markets the pattern of incidence is less clear. Typically—and there is considerable empirical evidence of this—prices will be raised immediately in response to the imposition of the tax, since firms take the initiative in setting their own prices and will likely adjust prices upward when they experience a general increase in costs. As long as the various competing firms follow the same policy, the increase is likely to be profitable. There are certain to be exceptions, however. If some firms fail to increase, the others will find an increase unprofitable. The over-all demand for some products may be so elastic that increases are unprofitable. Over a longer period there will be a greater tendency for price to rise by the amount of the tax, since prices must cover average cost. Here again, however, there will be exceptions. A monopolist or a group of firms following a concerted policy and having obtained, prior to the tax, maximum excess profits for the group will find it profitable in most instances to absorb a portion of the tax, since raising price by the full amount would result in a loss in revenue greater than the reduction in cost due to reduced output. It may be argued that a general sales tax can be shifted more easily than excises, since there is less danger of a loss in sales to untaxed commodities. The common practice (often required by law), under retail sales taxes, of adding the tax to the customer’s entire bill, rather than readjusting individual prices, undoubtedly facilitates shifting.

The argument that sales taxes are borne primarily by consumers has been questioned in recent years. Rolph (1952) maintained that a sales tax is borne in the same fashion as a flat-rate income tax, namely, in proportion to factor incomes. Rolph assumed perfectly competitive markets and perfectly inelastic supplies of the factors of production, and he disregarded the use of the revenue received from the tax. Thus, factor demand and factor prices fall. His conclusions, however, have been questioned, particularly in regard to the assumption about the use of the revenues. Buchanan (1960), and Rolph in more recent writings (Rolph & Break 1961, chapter 13), have argued that regardless of the assumption made about the use of the revenue, a sales tax cannot be borne by consumers because a tax rests on consumers only if the general price level increases and general price level increases cannot be attributed to taxes but only to monetary considerations. Musgrave (1959, chapters 10, 15, 16) maintains that the distribution of tax burden depends, not upon the direction of change in prices, but rather upon the relative changes in commodity and factor prices and concludes that a sales tax confined to consumer goods is borne in relation to consumption, whether commodity prices rise and factor prices remain unchanged or factor prices fall while commodity prices remain unchanged. He argues, however, that if the tax applies to both consumption and investment goods, the burden is distributed in the same fashion as that of a proportional income tax, regardless of the direction of change in price levels. Despite this extended theoretical controversy, policy discussions relating to sales taxes generally assume that the tax is, for the most part, shifted to consumers.

The excess burden argument . For a number of years the prime criticism advanced against excise taxes and, to a lesser extent, sales taxes of restricted scope has been that of “excess burden.” A tax on a particular commodity shifts purchases to untaxed commodities, thus resulting in a loss in economic welfare without an offsetting gain to the government. The pioneer statements were those of Hotelling (1938) and Joseph (1939). Critics pointed out that the thesis was valid only if the original revenue allocation was an optimum one and the pattern of income distribution the prefererred one. Other critics, such as Wald (1945), argued that income taxes, by affecting the choice between work and leisure and choices among various economic activities, likewise adversely affected economic welfare. But in a recent study of the question, by Harberger (The Role of Direct and Indirect Taxation in the’ Federal Revenue System 1964), the conclusion is reached that, on the basis of reasonable relevant assumptions, it is likely the excess burden of excises is greater than that of income taxes, primarily because of the limited response of work effort to income tax burdens.

General evaluation . The controversy over the relative desirability of commodity and income taxation has continued for many years with no lessening of intensity. Much of the debate centers on relative economic effects. Supporters of increased reliance on sales and excise taxes argue that income taxes retard economic growth and produce unemployment by discouraging savings, investment in business expansion, and work effort, especally on the part of business executives and professional men. Since sales taxes do not have progressive rates and may be avoided by saving rather than consuming, they do not directly penalize the gains from additional effort or business expansion and give some positive incentive to save more and consume less (except when savings are made for the purchase of goods in the future with the tax still in operation).

The opponents of sales taxation question the seriousness of the adverse effects of the income tax and argue that the greater relative impact of sales taxes on consumption will reduce national income and increase unemployment in situations in which there is some tendency toward unemployment because of inadequate total spending. A sales tax, by concentrating its burden more heavily on persons spending high percentages of their incomes and by providing some limited incentive to save more, may increase the potential rate of capital formation at full employment, but the tax may make it much more difficult to attain full employment and may thus lessen the actual rate of economic growth. Furthermore, to the extent to which the income tax does have adverse effect on the economy, this may be attributed in large measure to the high progressivity of rates and may be eliminated much more simply by changes in the income tax structure than by a shift to a sales tax. The differences attributed to the two forms of taxes are largely a result of the differences in rate structure, rather than in the base of the taxes.

In recent years the emphasis of the discussion has centered on the possibility of the replacement of the corporate income tax by the value added tax. The change has been advocated in large measure on the argument that the foreign exchange position of the country would be improved. Full export rebates would be granted for the value added tax, whereas no rebate is given for the corporate income tax, nor can one be given without violating present GATT (General Agreement on Tariffs and Trade) rules. The argument that the corporate tax places American exporters at a disadvantage, however, has validity only to the extent that the tax is reflected in higher prices of the products. The change proposed would temporarily aid American exports, whether the tax is now shifted or not, but such a change could easily invite retaliatory moves by other countries, especially if the tax is now not shifted.

Increased use of commodity taxation also has equity implications, and much of the opposition to the taxes has always been based on equity grounds. The income tax can be made progressive relative to income and can be adjusted in terms of various circumstances, such as size of family, which are considered to affect taxpaying ability. On the other hand, a sales tax with a broad coverage is regressive relative to income, because the higher-income groups save a greater percentage of their income, on the average, and spend more on nontaxable services. The tax likewise tends to burden large families more heavily, compared with smaller families, at given income levels (Hansen 1962). Food exemption, however, appears to eliminate regressivity (Davies 1961), but it fails to bring the precise adjustment to tax capacity that can be attained with an income tax. Some persons have suggested that the correct basis for comparing burdens is that of permanent income [see Consumption Function], rather than actual income (Davies 1961). On this basis, even a broad-based sales tax is not regressive. But it may also be argued that actual year-by-year income is the better basis for measuring tax burdens. The significance of the equity argument is, of course, one of value judgment; to many persons the use of some regressive taxes in a tax structure that is progressive over-all is not objectionable. But in terms of usually accepted standards of equity, major reliance on such taxes is undesirable.

Sales and excise taxes are also justified on administrative grounds—as being easier to enforce than income taxes. With improved income tax administration, this argument has lost most of the merit it once had. Furthermore, since the issue is one of using a sales tax along with, not in lieu of, an income tax, the over-all administrative task is obviously greater with a sales tax than without one.

In countries with a federal government, a final argument for sales taxation is the need of the states for autonomous revenue sources, in light of federal domination of the income tax field.

On the question of the choice between sales and excise taxes, the former are less discriminatory against individuals, in terms of their preferences, than excises and are less likely to distort resource allocation. On the other hand, excises, limited to particular commodities, may be easier to administer, may accomplish certain desired goals in tax policy (such as the placing of special burdens on highway users or consumers of tobacco and liquor), and may provide a more acceptable overall distribution of burden. But to raise significant revenue, either rates must be relatively high or many commodities of widespread use must be brought within the scope of the tax, and then the excise system comes to resemble a sales tax.

John F. Due

BIBLIOGRAPHY

Buchanan, James M. 1960 Fiscal Theory and Political Economy. Chapel Hill: Univ. of North Carolina Press.

Davies, David G. 1961 Commodity Taxation and Equity. Journal of Finance 16:581-590.

Due, John F. 1957 Sales Taxation. London: Routledge; Urbana: Univ. of Illinois Press.

Due, John F. 1963a State Sales Tax Administration. Chicago: Public Administration Service.

Due, John F. 1963b Sales Taxation and the Consumer. American Economic Review 53:1078-1084.

Hansen, Reed R. 1962 An Empirical Analysis of the Retail Sales Tax With Policy Recommendations. National Tax Journal 15, March: 1-13.

Hotelling, Harold 1938 The General Welfare in Relation to Problems of Taxation and of Railway and Utility Rates. Econometrica 6:242-269.

Joseph, Margaret F. W. 1939 The Excess Burden of Indirect Taxation. Review of Economic Studies 6:226-231.

Morgan, Daniel C. 1964 Retail Sales Tax: An Appraisal of New Issues. Madison: Univ. of Wisconsin Press.

Musghave, Richard A. 1959 The Theory of Public Finance: A Study in Public Economy. New York: McGraw-Hill.

Organization For European Economic Cooperation, European Productivity Agency 1958 The Influence of Sales Taxes on Productivity, by C. Campet. Paris: The Organization.

The Role of Direct and Indirect Taxation in the Federal Revenue System. 1964 Princeton Univ. Press. ? A conference report of the National Bureau of Economic Research and the Brookings Institution.

Rolph, Earl R. 1952 A Proposed Revision of Excise-tax Theory. Journal of Political Economy 60:102-117.

Rolph, Earl R.; and Break, George F. 1961 Public Finance. New York: Ronald Press.

Sullivan, Clara K. 1965 The Tax on Value Added. New York: Columbia Univ. Press.

U.S. Congress, House, Committee ON Ways AND Means 1964 Excise Tax Compendium: Compendium of Papers on Excise Tax Structure. ... 6 parts in 2 vols. Washington: Government Printing Office.

Wald, Haskell P. 1945 The Classical Indictment of Indirect Taxation. Quarterly Journal of Economics 59:577-596.

Walker, David 1955 The Direct-Indirect Tax Problem: Fifteen Years of Controversy. Public Finance 10, no. 2:153-176.

VI. DEATH AND GIFT TAXES

Taxes upon the transfer of property at death are known as estate taxes if they are imposed on the value of the decedent’s estate as a whole with little or no regard to the status and number of heirs, and as inheritance taxes if they are imposed upon the heirs individually. The estate tax consequently employs a single rate scale applied to the entire estate, while the inheritance tax is calculated separately on the amount received by each heir. The inheritance tax commonly employs a series of rate scales that vary with the degree of relationship of the heir to the decedent.

The tax on gifts made during life (gifts inter vivos) can likewise in principle be divided into a tax collected from the donor and a tax collected from the donee. In practice, only the tax on the donor is employed, and even that is used sparingly, most taxing jurisdictions not levying a tax on gifts inter vivos at all. Some gift taxes are cumulative, in the sense that a progressive rate scale is applied to the sum of gifts made by a given donor over his lifetime, as is the U.S. federal gift tax (Harvard Law School 1963a, chapter 3). Other gift taxes apply the graduated rate scale only to gifts made during a given year, as is the case in the German Federal Republic (Harvard Law School 1963b, chapter 4).

The death tax and the gift tax could be integrated either as a cumulative tax on all transfers made by a given donor during his lifetime or as a cumulative tax on all accessions to a given donee either through gift or inheritance. The cumulative integrated tax on donors has been proposed from time to time in the United States; as yet no country has employed it. The cumulative donee tax, or accessions tax, was in force for a short time in Japan—from 1950 to 1953 (Japan, Ministry of Finance 1963, pp. 9, 91), and exists in an incomplete form in Colombia and Italy (Shoup 1966, p. 13).

History. Taxes on the transfer of property at death have a long fiscal history. The Roman vicésima heredltatum, “the twentieth penny of inheritances,” is mentioned in Adam Smith’s Wealth of Nations (1776, book 5, chapter 2, appendix to arts. 1 and 2). In the United Kingdom, the tax dates back to 1694, but not until 1779-1780 did it attain something like its modern form (Palgrave [1894-1896] 1963, vol. 1, pp. 490-493). The U.S. federal government levied an inheritance tax during the Civil War and again during the Spanish-American War (Shultz 1926, pp. 151-155). The present U.S. federal estate tax dates from 1916; many of the state death taxes have longer histories (ibid., chapters 8, 9). Virtually all of the industrialized nations now employ some form of death tax, and it is also common in underdeveloped countries (see United Nations 1954).

The death duty predates the modern type of mass income tax and also the modern general sales tax. In many instances it has a longer history even than the more restricted income taxes of the period before World War n. The widespread and early use of the death tax can be explained largely by the fact that property had to be listed and valued in any event—for transfer to the state or to feudal overlords, under prevailing doctrines regarding land tenure; or to members of the family of the deceased possessing certain minimum rights in the property; or to other inheritors. The occasion thus proved a convenient one for computing a tax base and collecting a tax. Valuation remains, however, a vexing problem with respect to much of the transferred property.

Revenue . Although the history of the death tax has been impressive in terms of longevity and spread, its revenue role has been much less so. Today it rarely accounts for more than one per cent of total tax revenues in any country, despite the fairly steep graduation that characterizes most of the rate structures. While the income tax has been transformed in some countries into a tax that strikes almost every family and while social security payroll taxes and the general sales taxes, both inventions of the twentieth century, have added enormously to fiscal revenues, the estate and inheritance taxes have remained confined to only a small percentage of the populace. Most households in most countries have little or no property, at least relative to their incomes. In the more prosperous countries death taxes have high exemptions, and the starting rates are low. Thus, in the United States in 1961, for example, only some 45,000 out of 1,400,000 adult deaths resulted in estates subject to the federal estate tax (U.S. Department of Health, Education and Welfare 1963, vol. 2, part B, pp. 9-78, table 9-3). No movement has developed in any country to convert the death tax into a mass tax imposed on virtually everyone who dies possessed of property. In any event, conversion to a mass tax would not produce the striking percentage increase in yield that has been experienced under the income tax, since wealth is far more concentrated than income.

Avoidance . Sophisticated avoidance techniques, particularly in the United States and the United Kingdom, restrict the yield of the death tax. Under Anglo-Saxon property law concepts, trusts and life estates can be so set up as to skip one or more generations in the passage of property subject to death duties. In the United Kingdom, where expiration of a life estate gives rise, in principle, to full taxation of the corpus on which the life interest is based, tax has been avoided by several devices, notably the discretionary trust. This can be so formulated that owing to the discretion lodged in the hands of trustees as to who shall receive the life payments, it is not legally certain upon the death of one life tenant that any one of the others obtains any greater interest in the property than he had before (Harvard Law School 1957, chapter 3; Wheatcroft 1965, pp. 68-69, 132-137). In the United States, expiry of a life estate or similar property right does not give rise to inclusion of the corpus in the taxable estate of the decedent. Special statistical studies made by the U.S. Treasury have shown that in the wills of wealthy decedents the life estate that skips at least one generation is common (Shoup 1966, chapter 3; Jantscher 1967, chapters 4-7). These particular avoidance techniques are apparently not available in continental European countries because of the absence of the Anglo-Saxon concept of the trust.

In many countries, including the United Kingdom, gifts made during life (gifts inter vivos) are not taxable. In some of these countries, as also in the United States, gifts made within a certain number of years before death, or deemed made in contemplation of death, are included in the taxable estate. Thus, in the United Kingdom gifts made within five years of death are included, in part, in the taxable estate.

Where gifts are subject to a separate gift tax, as in the United States, the lower rate scale of the gift tax and the opportunity for splitting the property into two parts, each of which can obtain the benefit of low brackets (gift tax and estate tax), not to mention certain other technical features, leave a broad avenue for substantial tax reduction by gifts during life. In fact, however, even the most wealthy property holders seem to avail themselves of this possibility far less than a priori reasoning might suggest; the British consequently do not appear to believe that the revenue from their death tax is appreciably imperiled by the absence of a gift tax.

Contributions to charitable, educational, religious, and similar organizations are completely exempt under the U.S. federal estate tax, in contrast to the restricted exemptions, if any, granted in other countries. Once more, the opportunity for complete escape has been utilized rather less than one might expect (Harriss 1949; Shoup 1966, pp. 60-65).

The mobility of elderly wealthy persons is another restraining influence on heavy death taxation. Recently, both the United Kingdom and the United States have altered their death tax laws to include in the tax base real estate located abroad. This change has added pressure on elderly wealth-holders to change their residence and perhaps even citizenship as they reach extreme old age. Again, the number of such decisions will probably prove to be minor compared with the prospective tax saving.

In some respects the death duty offers fewer opportunities for avoidance than does the usual income tax. In the United States, state and local securities are fully subject to inclusion in the decedent’s taxable estate even though during his lifetime the interest on such obligations is exempt from the federal income tax. Property values arising from capital gains are fully included for U.S. estate tax purposes and also for the death duties in Britain, while under the income tax they are given preferential treatment. The percentage depletion provisions in the United States that have caused so much comment with respect to the income tax are, of course, not operative for the death tax.

Effect on consumption . Per dollar of revenue, the death tax probably decreases consumption spending less than most other taxes on households. This is so because the decedent-to-be seems unlikely to decrease his standard of living appreciably in order to improve the prospects of his heirs, prospects that have been impaired by the death tax. The future heirs, in turn, seem unlikely to reduce their current standard of living merely because they are aware that they will later receive less than if no death tax were in force. With respect to the period following transfer of the property at death, it has been cogently argued by Ricardo and others that the heirs tend to look upon the capital that they should preserve as being simply the amount they receive after death tax. They thus feel under no pressure to try to rebuild the estate to a level closer to what it would have been without such a tax (Shoup [1950] 1960, chapters 3, 15). Doctrinal discussion in Anglo-Saxon economic literature over the past century and a half has centered more on the reaction of the decedent-to-be than on the heirs, prospective or actual, and some difference of opinion has developed on this score (Fiekowsky 1959, chapters 1, 3). McCulloch, for example, expressed the opinion that the property owner would attempt to build up his estate somewhat in an effort to recoup for his heirs a part of the value that would be lost by the estate tax. Present-day thought, however, does not follow McCulloch, especially in view of the apparent indifference of wealthy persons as evidenced by their failure to transfer much property during life in order to save tax money for their heirs (Shoup 1966, Appendix F).

On the other hand, it is not at all certain that this failure to take advantage of what appear to be bargain tax rates during life necessarily indicates indifference. The welfare of one’s heirs is weighed against other considerations, some more admirable than others. As modern medicine has enhanced the possibility that an elderly person may live to extreme old age, sometimes under very expensive medical and hospital care, the risk that his financial resources may be exhausted before his death has become correspondingly greater. Dread of dependence on his children and loss of flexibility in arranging for his later years, even if expensive medical care is not a problem, are powerful forces in causing a wealthy decedent-to-be to cling to his wealth, particularly when he believes that he has already given enough to his children to start them in life with substantial advantages and conjectures that further wealth would do them more harm than good. To these motives must be added sometimes a desire to retain psychological control over prospective heirs, and sometimes a gradual drift into senility before the individual can be persuaded to think about death and act on his thoughts. Simple inertia explains much, especially on the part of some elderly women who have little interest in property management, and extremely busy men of affairs who do not pause long enough even to sign a will. Family jealousies also play an occasional role in restraining gifts during life.

The transferor or the heir might recoup some of the death or gift tax by increasing his money-making efforts. The high income tax rates to which this class of persons is commonly subject make this method of capital preservation, as compared with restricting one’s consumption, a difficult one.

Effect on distribution . The distribution of wealth and income, as indicated by a Lorenz curve, has probably been made more nearly equal to a modest degree during the past thirty years or so of graduated death taxation in the United States and the United Kingdom, compared with what it would have been if the same revenue had been raised, for example, by an increment to general sales taxes (Fiekowsky 1959, chapter 3). The result seems not to have been as substantial, however, as has been hoped for by proponents of the tax, to whom a chief virtue of death duties is their presumed tendency to limit accumulation of extreme fortunes and to reduce inequality generally. The explanation for this disappointment, if such it is, lies largely in the number of avoidance devices indicated above and partly in the relatively high level of exemption (for the U.S. federal estate tax it is $60,000) and low rates in the initial ranges. No practical support has developed for the Rignano plan or its variants, which would tax especially heavily and eventually confiscate inheritances that came from inheritances, after two or three or four generations. It is instructive to recall that an inventor of one of these variants, Hugh Dalton, made no move to introduce it into the British law while he was chancellor of the exchequer in the late 1940s (Dalton 1923, pp. 114-118 in 1936 edition; p. 232 in 1954 edition).

Death taxes are said to have forced small, closely held family firms to restrict their rate of growth in order to accumulate liquid assets sufficient to pay the tax upon the death of the founder or other large family owner, or alternatively to have induced them to merge with large firms whose stock is actively traded on exchanges so that liquid assets for payment of the tax could be obtained without restricting growth of the business (Somers 1958, pp. 201-210). The extent to which these effects have in fact materialized is not clear. In the United States the law has recently been amended to guarantee the estate the privilege of a ten-year installment payment provision if the company in question meets certain tests. In any case, an extended period of payment can be granted at the discretion of the tax administration.

Present trends . Among the current trends in death and gift taxation, the most noticeable one seems to be a tendency to personalize the estate tax, so that the amount of tax will vary depending particularly upon the relationship of the heir to the decedent. In this way, the estate tax may become more and more like an inheritance tax. The U.S. federal estate tax allows exemption of up to 50 per cent of an estate with respect to transfer to the surviving spouse, and pressure is growing to exempt completely interspousal transfers and to give some tax reduction for transfers to children.

Another trend, this time working toward an increase in revenue, is evidenced in current discussions of methods by which skipping one or more generations can be reduced, through taxing expiry of life estates and inhibiting the use of discretionary trusts. The task is much more difficult than this brief discussion might indicate, because of the intricacies of property law and the consequent opportunities to avoid even the most complex anti-avoidance measures. However, additional legislation on these subjects may be expected in both the United States and the United Kingdom during the next few years. In the view of some, the death and gift tax system should be so constructed that no matter by what route property is transferred to a generation distant in time, the present value of taxes on the transfers would come to the same thing, as under the proposal by Vickrey of a bequeathing power tax (Vickrey 1947, chapter 8). To achieve this end, however, is to relinquish the relationship of heirs to decedent as grounds for differentiation of the tax.

No trend is apparent with respect to the level of exemptions and the rate and type of graduation. Both the exemption and rate structure have shown great stability over time in most countries; in the United States, for example, the present rate scale dates from 1942. Graduation by brackets, as in the income tax, is characteristic of most death taxes, but the British prefer to graduate by a series of effective (average) rates. Such graduation facilitates an equitable division of the tax between the executors of the estate and owners of parcels of property that, although not appearing in the decedent’s estate, are nevertheless aggregated with his estate in determining the tax rate applicable to such parcels and to the estate (an example is property that was transferred as a gift inter vivos within five years of death).

Jurisdictional problems either among states in a federation or among countries continue to occupy much time and thought of tax lawyers and legislators, but exert little influence on total revenues. In the United States the federal-state issue has been met by allowing up to a certain amount of state death taxes paid to be credited directly against the federal tax, with the consequence that all of the states of the United States, excepting Nevada, impose either the estate or the inheritance tax or both, sometimes indeed rather beyond the limits of the federal credit. The United States has concluded tax conventions with many other countries, chiefly to avoid double taxation of properties of nonresident aliens.

Legal and administrative complications arise through linkages of death and gift taxes with the income tax. In the U.S. federal law, a transfer of property may be an inter vivos gift for gift tax purposes but not for income tax purposes; it is not evident, however, that complete uniformity is desirable. Another linkage arises with respect to capital gains. At present, a capital gain accrued at death is not made subject to the income tax, nor is a capital loss recognized. The heirs take over the property with a new basis for computing capital gain or loss on a future sale. This basis is the value of the property in the decedent’s estate. Accordingly, a capital gain on property held until death is never subject to the income tax, and a capital loss is never allowed. An attempt by the executive to persuade Congress to eliminate this combination of loophole and hardship in the Tax Reform Bill of 1963 failed. Property given during life, on the other hand, does not have its basis stepped up (or down) in this manner; this fact helps explain the reluctance to pass on appreciated property during life rather than at death.

No taxes have had a better reputation to less effect. Favorable comments on death and gift taxation can be found in the most conservative quarters, but these taxes remain minor and of little concern to politicians and voters. In certain academic circles some doubt is beginning to arise whether many of the aims of the estate and gift taxes could not better be achieved by a low-rate annual tax on individual net wealth, which would not be vulnerable to the devices now being employed to skip generations.

Carl S. Shoup

BIBLIOGRAPHY

A Critique of Federal Estate and Gift Taxation. 1950 California Law Review 38, no. 1 (Special Issue).

Dalton, Hugh (1923) 1954 Principles of Public Finance. 4th ed., rev. London: Routledge.

Fiekowsky, Seymour 1959 On the Economic Effects of Death Taxation in the United States. Ph.D. dissertation, Harvard Univ.

Harriss, C. Lowell 1940 Gift Taxation in the United States. Washington: American Council on Public Affairs.

Harriss, C. Lowell 1949 Federal Estate Taxes and Philanthropic Bequests. Journal of Political Economy 57:337-344.

Harriss, C. Lowell 1954 Sources of Injustice in Death Taxation. National Tax Journal 7, Sept.: 289-308.

Harvard Law School, International Program In Taxation 1957 Taxation in the United Kingdom. Boston: Little.

Harvard Law School, International Program In Taxation 1963a Taxation in the United States. Chicago: Commerce Clearing House.

Harvard Law School, International Program In Taxation 1963b Taxation in the Federal Republic of Germany. Chicago: Commerce Clearing House.

Jantscher, Gerald R. 1967 Trusts and Estate Taxation. Washington: Brookings Institution.

Japan, Ministry OF Finance, Tax Bureau 1963 An Outline of Japanese Tax: 1963. Tokyo: The Bureau.

Palgrave, Robert H. (1894-1896) 1963 Death Duties. Volume 1, pages 490-493 in Robert H. Palgrave, Palgrave’s Dictionary of Political Economy. Rev. ed. New York: Kelley.

Pechman, Joseph A. 1950 Analysis of Matched Estate and Gift Tax Returns. National Tax Journal 3, June: 153-164.

Shoup, Carl S. (1950) 1960 Ricardo on Taxation: An Analysis of the Chapters on Taxation in David Ricardo’s Principles. New York: Columbia Univ. Press.

Shoup, Carl S. 1966 Federal Estate and Gift Taxes. Washington: Brookings Institution.

Shoup, Carl S. et al. 1949 Taxes on Gifts and Bequests. Volume 2, pages 143-155 in Carl S. Shoup et al., Report on Japanese Taxation. Tokyo: Supreme Commander for the Allied Powers.

Shultz, William J. 1926 The Taxation of Inheritance. Boston: Houghton Mifflin.

Shultz, William J.; and Harriss, C. Lowell (1931) 1959 American Public Finance. 7th ed. Englewood Cliffs, N.J.: Prentice-Hall. ? First published as American Public Finance and Taxation, with William J. Shultz as sole author.

Smith, Adam (1776) 1952 An Inquiry Into the Nature and Causes of the Wealth of Nations. Chicago: Encyclopaedia Britannica. ? A 2-volume paperback edition was published in 1963 by Irwin.

Somers, Harold M. 1958 Estate Taxes and Business Mergers: The Effects of Estate Taxes on Business Structure and Practices in the United States. Journal of Finance 13:201-210.

United Nations, Technical Assistance Administration 1954 Taxes and Fiscal Policy in Under-developed Countries. New York: United Nations.

U.S. CONGRESS, Joint Committee ON THE Economic Report 1956 Federal Tax Policy for Economic Growth and Stability. Hearings Before the Subcommittee on Tax Policy. Washington: Government Printing Office.

U.S. Department OF Health, Education AND Welfare 1963 Vital Statistics of the United States: 1961. Washington: Government Printing Office. ? See especially Volume 2, part B, pages 9-78, Table 9-3.

Vickrey, William S. 1947 Agenda for Progressive Taxation. New York: Ronald Press.

Wheatcroft, G. S. A. (1953) 1958 The Taxation of Gifts and Settlements, by Stamp Duty, Estate Duty, Income Tax and Surtax. 3d ed. London: Pitman.

Wheatcroft, G. S. A. 1957 Anti-avoidance Provisions of the Law of Estate Duty in the United Kingdom. National Tax Journal 10, March: 46-56.

Wheatcroft, G. S. A. (editor) 1965 Estate and Gift Taxation: A Comparative Study. British Tax Review Guides, No. 3. London: Sweet & Maxwell. ? A study of estate and gift taxation in Australia, Canada, Great Britain, and the United States.

International Encyclopedia of the Social Sciences