Whether income is an accurate measure of taxpaying ability depends on how income is defined. The only definition that has been found to be completely consistent and free from anomalies and capricious results is “accrued income,” which is the money value of the goods and services consumed by the taxpayer plus or minus any change in net worth during a given period of time. (Tax experts commonly call this the Haig-Simons definition of income, based on work by American economists Robert M. Haig and Henry Simons.) This definition cannot be applied without important modifications. First, many tax codes do not consider as taxable income those changes in net worth resulting from gifts, bequests, and other gratuitous transfers. Second, because of the difficulties of estimation, most accretions to wealth are ordinarily not included in an individual’s taxable income until they are “realized”—that is, converted into cash or some easily valued form. Finally, and for much the same reason, most countries have chosen not to include in taxable income such forms of imputed income as the rental value of owner-occupied homes.

In some countries the individual income tax is imposed on the total income of an individual or family unit, whereas in others income from different sources is taxed under separate rules and often at somewhat different rates. The use of multiple schedules is questionable on grounds of both neutrality and horizontal equity (persons with the same income, under like circumstances, paying the same amount of tax), and countries with schedular taxes frequently supplement them with a progressive rate scale applicable to total income. These schedular income taxes are today found in some South American and African countries. In most industrialized countries, such as Great Britain, personal income has to be reported on one of a number of separate schedules, but assessable income is then lumped and only one tax is imposed. This kind of individual income tax is not usually regarded as a schedular tax. By comparison, the Scandinavian countries have recently adopted “dual” systems in which labour income is subject to graduated rates, but capital income is subject to flat rates. The United States has adopted antishelter provisions that have the effect of converting a nominally global income tax into one having schedular features.

Before a tax on personal income can be considered to be a completely fair tax, it has to meet the tests of horizontal and vertical equity. Pivotal to the first test is the definition of “like circumstances” when considering taxes imposed on individuals with the same income. Clearly, two families with the same income would not be equally able to pay taxes if one consisted of husband and wife and the other of husband, wife, and four dependent children. On the other hand, if neither family had any children but in one the entire income was earned by the husband whereas in the other both husband and wife worked, would horizontal equity require that they pay the same or different taxes? Similar questions have been raised concerning families whose equal incomes take the form of wages and salaries in one case and dividends and interest in another or whose income has to be used to pay personal debts (such as medical expenses) or to pay state and local taxes to a greater extent in one case than in the other. In order to compensate for those differences in the sources and uses of income that are thought to affect an individual’s ability to pay income tax, most countries allow a wide variety of deductions from statutory personal income before the tax is imposed.

The concept of vertical equity relates to the taxes paid by individuals at different income levels. Clearly, if income is a good index of ability to pay, the taxes for these individuals should not be the same, but how different should taxes be at different income levels? If a single rate of tax is applied to all individual income in excess of the allowed exclusions, exemptions, and deductions, the tax will be proportionate to taxable income (although it may be progressive when compared with total income). If, however, different tax rates are applicable to different blocks or brackets of income, and if these rates rise as one moves from the lowest bracket to successively higher ones, the tax will be progressive. Those countries that tax total individual income today almost always use graduated or progressive rates; those with schedular income taxes may or may not do so.

Many attempts have been made to develop a theory that would not only justify the principle of progression but also result in a mathematically exact scale of equitable taxation. Some theorists, accepting the notion that the taxes a person pays ought to bear some close relation to the benefits the taxpayer enjoys from the operation of government, have tried to show that, at some levels of income, benefits increase more rapidly than income. But their efforts have served to do little more than reveal the shortcomings of “benefit theory.” Others, starting with the premise that an equitable tax is one that imposes equal sacrifices on individuals at different income levels and accepting the view that the utility of any given unit of money becomes less the more money one has, have tried to demonstrate that progression is needed if the sacrifices imposed on the wealthy are not to be less than those imposed on persons less-well-off. But it is debatable whether a dollar has less utility for a very rich person than for a moderately rich one or whether it is scientifically possible to make the sort of interpersonal comparisons that the “sacrifice theories” call for. Ultimately, the progressivity of a nation’s tax system depends on each society’s view of what is fair, as expressed through the political process.

So long as prices are stable and the tax is basically a tax on realized income and does not require an assessment to be made of accrued but unrealized capital gains and losses, the income tax is generally held to be easier to administer than either an expenditure tax (a tax on spending) or a wealth tax (a tax on one’s worth—as opposed to a tax on one’s earnings). An income tax fails, however, to calculate the effects of inflation and timing issues in the measurement of income. Inflation erodes the real value of interest income and of deductions for interest expenses, depreciation, inventories, and the cost of capital assets sold by the taxpayer. Furthermore, it is not always clear when income is earned and when taxes are incurred; a direct tax on consumer spending would require the subtraction of net saving (or exemption of capital income in the case of a flat tax, which imposes the same level of tax on all taxpayers) from realized income, and balance sheets would be required in order to prove that saving was correctly reported. Some favour the direct consumption tax and the flat tax because they are based on cash flows, which means that these taxes eliminate the need to adjust for inflation. They also overcome problems in the measurement of income and questions of timing. The administration of a wealth tax would be far more complicated, requiring, for example, a complete accounting for assets and liabilities.

The enforcement of the income tax in many countries, such as the United States, has been made easier by the practice of withholding (retaining) the tax from wage and salary payments. The same approach has not been extended to interest and dividends in the United States, although it has in other countries. Compliance is undoubtedly incomplete, and complex provisions increase costs for both taxpayers and the fiscal authorities, but, in general, the income tax raises revenue efficiently and at low out-of-pocket cost to the government, if not to taxpayers.