Parthasarathi Shome


Parthasarathi Shome

The scope of the Tax Policy Handbook includes a consideration of the theoretical underpinnings of taxation but encompasses the broader perspective of taxation issues of more practical interest. This introduction, Chapter I, attempts to conceptualize the driving force behind selected topics and, in the process, draws attention to important prevailing concerns in taxation and summarizes the coverage of the various chapters.

Tax Policy, Tax Revenue, and Tax Reform

The justification for taxes lies in a government’s need for resources to carry out its essential functions. Nevertheless, there are many costs to taxation. These costs reflect the direct cost of tax collection as well as the effects on the efficiency of resource allocation and on the equity of income distribution. Estimates of the welfare cost of taxation—defined as the excess cost to society of collecting $1 of tax revenue—seem to be about $0.50, according to some studies for the United States.1 It is surprising, therefore, how governments tend to forget or seem to ignore taxation costs. For example, to reduce the fiscal deficit, before curtailing expenditures, often governments tend to immediately identify new tax measures. The counterargument that reducing expenditures could impose costs on society as well could, of course, also be made.

Governments seem to regard fiscal policy as “their” policy within the mix of macro policies. This is because fiscal policy is more amenable to direct action, its results are often more clearly visible, and its impact is felt more quickly. Further, within fiscal policy, a tax revenue increase, though not always easily introduced, is generally easier to implement than the politically more difficult expenditure cuts. This is especially so, as increasing proportions of the expenditure budget reflect debt service and other costs, limiting the possibility to maneuver. These factors explain country authorities’ heavy focus and concerns on taxation issues.

Therefore, in the post-Second World War era, there has been a pervasive presence of tax reforms more often than not brought about in the context of the need for increased revenue. Examination of actual country cases would reveal that most countries that carried out tax reform with the stated objectives of improving the efficiency, equity, neutrality, and administrative feasibility (simplicity) of their tax systems also experienced a perceptible increase in their tax revenue to GDP ratios. Thus, even though tax reform need not necessarily be linked to revenue increase, in effect, this seems to have been commonly the case.

The fascination with tax revenue increase is understandable in cases where governments that undertake reform have been preoccupied with a downward slide in tax revenue attributable to a prevailing cumbersome tax structure that they wished to correct. However, where the tax structure is broad-based and relatively neutral, tax revenues have been stable and opportunities for expenditure reduction exist, the focus of fiscal reform has to be on the expenditure side because of the remarkably high costs associated with taxation.

In what follows, Chapter II presents the conceptual ramifications of taxes by surveying how taxation affects efficiency, equity, intertemporal decision making, and the like. Subsequent chapters consider more technical issues in the context of particular taxes. Chapter III focuses on the taxation of consumption and production, describing not only the broad differences among major taxes such as the sales tax, value-added tax (VAT), and excises, but also draws attention to the details of various technical complications that arise in the way they function. Chapter IV examines income, property, and social security taxes or, more precisely, those issues that most commonly arise in the context of providing technical assistance to policymakers. Chapter V deals with policy issues in international trade taxation which have been reflected in tariff reforms implemented in recent years in many parts of the globe. Chapter VI selects topics that do not comprise an integral part of the preceding chapters but are nevertheless important in their own right. Chapter VII summarizes the main trends in global tax reform and sketches the nature of IMF tax policy advice. An appendix reports recent tax revenue data in terms of GDP and total tax revenue across groups of countries between 1975 and 1992.

General Issues and Concepts: Theoretical Underpinnings

This Handbook explains the conceptual basis for taxation in Chapter II.

Efficiency, Equity, and Incidence

The concepts and measurement of the efficiency cost of a tax were developed in many stages over many years. Efficiency cost refers to the reduction in a consumer’s welfare over that which can be related solely to his lost income as a result of the tax. It is, therefore, varyingly called “deadweight loss” or “excess burden” of a tax. It arises from the change in the relative prices between taxed and untaxed commodities in the post-tax situation. The commonly used measure of excess burden is the “Harberger triangle”2 which uses the concept of the loss in consumer surplus as a result of the tax. It varies positively with the pretax price elasticity of demand and the tax rate. The measure has its detractors, however. This is because it can be different from the monetary compensation that a taxed consumer would require to make him as well off as he was before the tax. These compensation measures—“equivalent variation” and its later variant, “compensating variation”—were initiated by Hicks3 and later developed by others. Zee elaborates on these concepts.

The equity considerations of taxation comprise preoccupations with whether taxpayers with similar incomes are treated equally under the tax system—“horizontal equity”—and with ensuring that taxpayers with unequal incomes are treated differentially—“vertical equity”—often through the income tax side of the tax system. The vogue of the post-War era was to design tax systems that considered both of these aspects of equity. The result often was a rather complicated tax structure, too cumbersome to administer and not revenue productive. Zee also examines these aspects.

Also, the true incidence of such structures was difficult to derive, possibly because the “nominal” and “effective” tax rates were vastly different. Two distinct movements were developed that tried to measure a tax or a tax system’s true incidence, one by Harberger4 that used a rudimentary general equilibrium framework, and another by Musgrave5 that also used a rudimentary approach by preassigning incidence of various taxes in a tax system across different consumer groups. Even though both approaches have been much extended, finessed, and applied to examine many country tax structures, important difficulties remain with both approaches since both are dependent on a number of strong assumptions. These assumptions include elasticities of demand and supply as well as of factor substitution, the ability of factors to move across sectors, the existence of risk and uncertainty in production, and the realism in preassigning of tax incidence. Krelove considers the various aspects of incidence. The indeterminate nature of the ramifications of important elements of taxation seem to emerge from these considerations.

Static Versus Intertemporal Effects of Taxation

The efficiency and equity effects of taxation are usually cast in a static context. Many effects of taxation, however, are essentially intertemporal, such as on interest rates, savings, capital accumulation, or economic growth. Also, it is sometimes important to track the path of relevant economic aggregates as a result of a tax to fully appreciate its effects. A simple example demonstrating this issue is an inefficient tax that can raise revenue in the short run but leads to a slowdown in economic growth and, ultimately, in revenue. Further, the development of endogenous growth models has made possible the linking of long-run growth to present and future tax and fiscal measures and to expectations of future events.6 This suggests that the growth effects of taxation could be permanent, rather than merely transitory, as previously believed.

An important element of taxation of savings has been associated with David Ricardo whose view, in effect, was that a current tax cut financed by government debt has no real effects. It does not change the economy’s path—Ricardian equivalence—because private agents are rational, anticipate higher tax liabilities in the future, and adjust their behavior accordingly. This equivalence has been picked up by contemporary economists and, even though its empirical verification remains unproven, it could approximate the actual behavior of the economy depending on the circumstances. For example, first, if the planning horizon of the economic agent is his life cycle and he saves for retirement, then the equivalence is likely to break down if the likelihood of taxes is seen to be remote during his lifetime. On the other hand, if a bequest is the motive for saving, the equivalence is likely to be valid. Second, if households face liquidity constraints, a tax cut would increase their consumption as though in effect the government has borrowed on their behalf, in opposition to the equivalence. On the other hand, without a liquidity constraint, the equivalence is more likely to hold.

Intertemporal substitution effects play a role in the area of capital income taxation. The tax treatment of capital income affects capital accumulation and growth since savings have been found not to be inelastic with respect to the interest rate.7 Empirical evidence as well as analytical considerations point to a larger welfare cost of capital taxation than had previously been anticipated. The postulation of altruistic links between successive generations tends to exacerbate these findings.

The classical theory of economic growth considered technological change, human capital accumulation, and similar growth-inducing factors to be generally outside the realm of economic policies. The new theory of growth recognizes such factors and, although it is still early for conclusive results, it sheds light on the effect of taxation on the long-term growth prospects of an economy. Escolano analyzes the dynamic or inter-temporal effects of taxation, especially in instances where static analysis might lead to erroneous conclusions.

Taxing Expenditure or Income

If the income expenditure equation of an economy is considered, meaningful questions that should be posed would include the following: which side of the equation should be taxed, which side is easier to tax, or whether there is a rationale for taxing both sides. There are pros and cons to taxing consumption expenditures as an alternative to income taxation.8 Even John Stuart Mill was concerned with this issue. Among the pros, first is the exemption of savings under the expenditure tax in contrast to the double taxation of investment returns under the income tax. Second is the possibility of designing even the expenditure tax with a progressive structure. Nevertheless, the cons comprise doubts that have always been raised regarding this matter. First, proponents of the income tax rely on the Haig-Simons definition of income as accretion of power to consume and, therefore, think of income as the proper equity criterion for taxation. Second, the expenditure tax is seen as a payroll tax from a lifetime perspective excluding bequests and inheritances and, therefore, causes undue burdens on wage earners.

In India and Sri Lanka, where the expenditure tax was applied, the definition of the tax base foundered on its conceptual basis, and it was revoked. In practice, the income tax has been more commonly implemented, while the expenditure tax has been only temporarily tried in a handful of countries. Nevertheless, when expenditure taxation was tried within the narrower range of consumption, it became easier to conceptualize as well as to implement in particular forms, such as the VAT, because of its cross-control feature. Many countries have now introduced the VAT which exists side by side with the income tax. Particular groups of countries, for example in Latin America, have spent much of their tax administration resources on the VAT, which has become their main vehicle for generating revenue while others, such as in Asia and Europe, have continued to use the income tax as the main revenue generator. This is not to say that the scope of the VAT is not increasing even in the latter countries.

On a purely conceptual basis, taxing both the income and the expenditure side of the equation may smack of double taxation. As was just described, however, in practice, it has not stopped tax practitioners short of using taxation of both. This is perhaps because what is conceptually clear and correct may not always be easily implemented. The practice of taxation reflects a combination of what is implemented with the intermittent incorporation, in the form of tax reform, of concepts developed through the progress of tax theory. Thus, it has always been common to tax both income and expenditure, using all revenue sources alike. Only a couple of countries have actually confined themselves to taxing either income or expenditure and this could turn out to be temporary. Escolano considers the conceptual issues related to the taxation of income or expenditure.

Taxation and Risk Taking, Imperfect Markets, and the Second Best

Profits are in part a compensation for bearing risk. A tax on the return from an undertaking taxes risk taking. Thus, there has long been a concern that capital income taxation leads to a reduction in risk taking. Nevertheless, the new literature in this area demonstrates that, on the one hand, if the tax system shares sufficiently in the risk of an investment as well as in the expected return, when loss offsets are partial, it is not impossible that taxing the return on risky assets can actually increase risk taking. On the other hand, even when there is effective full loss offset for “income risk,” tax systems, as they are usually designed, imply that a government rarely shares completely in “capital risk.” When risky returns are taxed so that a government is sharing in risk, tax revenue is uncertain. In sum, the recognition of investment as a risky business modifies many of the conventional results of capital taxation and the revenue it can generate.

Another situation in which conventional wisdom on the effects of taxation fails is the existence of an imperfect market. It is often assumed that a monopolist can just pass on a tax to consumers. It can be shown, for example, that an excise tax would tend to be borne partially by a monopolist and that the form of the tax— ad valorem or specific—would determine its extent. This is because, for any given revenue raised by the tax, the monopolist’s output will be higher with an ad valorem tax than with a specific tax. Similarly, in the case of a partial tax on a factor of production, such as the corporation income tax, or a more general tax, such as an income tax, the ultimate incidence will depend on the market form.

An area that has attracted much attention in the development of taxation theory is how taxation affects efficiency in an environment in which efficiency did not initially exist. This is the theory of the “second best.” By nature, second-best policy involves an investigation of the interactions among market distortions. Hence, it is inherently general equilibrium. Rationalization and formulation of algebraic constructs have led to some interesting conclusions. First, if a tax distortion exists in one market, adding another tax can be beneficial in improving efficiency. Second, if several taxes exist, removing one distortion may not be beneficial. Third, policies that would not be desirable in a distortion-free world may have a role to play in a second-best environment. Fourth, policy design in a second-best world is complex and not at all obvious. These facets of tax theory, based on the seemingly unusual or extraordinary, yet representing an everyday occurrence in a representative economy faced by the policymaker are discussed by Krelove.

Domestic Consumption and Production Taxes

Chapter III, on domestic consumption and production taxes, covers a wide range of issues reflecting methodological advances as well as particular details that have emerged in the course of IMF technical assistance in the context of consumption and production taxes.

Optimal Taxation

A segment of taxation literature—“optimal taxation”—has emphasized that the minimization of efficiency cost of collecting revenue should be the main criterion for determining the structure of tax rates. Optimal tax rules based on elasticities in a mix of producer prices—fixed or variable—have been developed. As expected, the lower the elasticities of demand, the higher the optimal tax rate on a commodity—the “in-verse elasticity rule,” which is based on special assumptions. Those who believe in optimal tax rules do not consider the need for multiple rates to be a forbidding enough property that the rules themselves should be discarded. The critique that the rules may lead to some regressivity led even to the incorporation of income maintenance elements in the rules. They have recommended tax structures based on optimal taxation theory in selected developing countries. Thus, optimal taxation comprises an interesting chapter of tax policy that must be examined. Zee provides a critical examination of the theory of optimal commodity taxation.

General Consumption Taxes

The VAT ideally taxes the value added in every stage of production and distribution which implies that the taxpayer would have to be given credit for the taxes he pays on his purchases. Many countries have moved from sales taxes levied at a single stage in the production-distribution chain to a VAT mainly for reasons of tax administration.9

Many countries that introduced the VAT, however, preferred to opt initially for a “uproduction”;-type rather than a “consumption”-type VAT mainly for revenue considerations. The production-type VAT typically does not allow credit to be given for capital goods purchases. This leads to “cascading,” that is, some value added is taxed more than once because the tax base calculation at every stage of production fails to exclude all the value added that had been taxed before. The extent of cascading itself depends on the elasticities of demand and supply, the ratio of taxed to untaxed inputs, the number of stages in the production-distribution chain, and other factors. Zee delineates the concept and measurement of cascading.

There are other complications in the design of a VAT. Multiple rates make the tax difficult to administer, robbing it of the simplicity of cross controls. Important differences occur if a good is “zero-rated”—credit is given for taxes paid on inputs even though tax on output is zero, as opposed to if it is “exempted”—the output is not taxed but neither is credit allowed for taxes on inputs. The calculation of the net tax liability based on the “credit method”—gross tax liability minus taxes already paid on purchases, as opposed to the “subtraction method”—using income and cost for the calculation of net tax—would lead to differing effects of exemption versus zero-rating at different points in the production-distribution chain. A narrow tax base reduces the neutrality of the VAT and renders its administration difficult. Very often, countries have had to deal with declining bases reflecting political reasons. A whole methodology for the estimation of the “theoretical” VAT base compared with actual revenue collected has developed to estimate VAT evasion. McMorran and Zee discuss issues that bear on general consumption taxes.


The inverse elasticity rule derived from optimal tax theory is conveniently applicable in the case of selective excises. For example, traditionally, selective excises have been favored in addition to a broad-based sales tax or VAT to tax the consumption of “demerit” goods such as tobacco, alcohol, and gasoline. Their elasticities of demand are low at least in the short run. Thus, high tax rates could be used to generate revenue without diminishing consumption significantly. It should be obvious that this type of tax should be used selectively in order not to burden the consumers of other goods with low elasticities of demand, such as necessities. Indeed, the other face of selective excises is the possibility, or even the desirability, of imposing them on luxuries that may have high elasticities of demand but nevertheless may not be adequately taxed under a low-rated general consumption tax. Various practical issues arise in the design of excise taxes—for example, should they be levied on production or on consumption, on an ad valorem or specific basis—as well as in the consequences of optimal taxation—for example, are there any simple tax structures that may be developed. These are the subjects that McCarten and Stotsky address in this chapter.

Environmental Taxes and User Charges

Environmental taxes apply wherever there are environmental objectives. They are termed “Pigouvian taxes” after Pigou, the classical economist who first enunciated the design of taxes that would internalize the externalities—for example, pollution—associated with economic activities, by equating their private and social costs. A Pigouvian tax would reduce and, in an extreme case, eliminate the level of pollution depending on whether society regards the benefit to be gained from its elimination to be worth the cost of doing so. Some environmental taxes “earmark” their revenue to help clean up the environment. In addition, modern environmental tax designs recognize political economy concerns such as potential conflicts with equity objectives. Nellor analyzes various aspects of environmental taxes.

In particular cases such as road maintenance and other selected public services, it has been found that charging users according to the extent of use is efficient—the “benefit principle”—without sacrificing considerations of equity. Common issues, such as in the event of the provision of technical assistance, include methods to adequately identify the user, measure the intensity of use, safeguard fairness, and develop a topology of appropriate user charges. Bell covers considerations relating to user charges.

Income and Wealth Taxes

The longest surviving discussions in the area of taxation policy have perhaps been on how best to conceptualize and design corporate and personal income taxes. Chapter IV is divided into many sections, which attempt to define the concept of income, deal with various aspects of income taxes, and cover other variants of income and wealth taxes.

Income, Taxable Income, and Optimal Taxation

An issue that affects income taxation considerably is the concept or definition of income during a period. The most widely accepted theoretical basis was developed in early writings by G. von Schanz (1896), R.M. Haig (1921), and H.C. Simons (1938). Schanz-Haig-Simons income is the sum of the market value of rights exercised in consumption and the change in the value of the store of property rights between the beginning and the end of the period in question. Thus, this definition of “comprehensive” income equals consumption plus net wealth accumulated during the period.10 Alternative definitions of income exist but are not so commonly used. King examines issues regarding the concept of income.

Optimal income taxation examines the trade-off between efficiency and equity. Economic theory yields insights into the optimal degrees of income tax progressivity under different theories of distributive justice. Models rely on assumptions about the distribution of pretax income, the government’s objective to maximize social welfare using income tax, and the disincentive effects or efficiency costs of individual work effort from the tax system. Zee examines the application of these models.

The normative recommendations of optimal income taxation are reflected in the design of many personal income tax systems. Nevertheless, the design of personal income tax often reflects the outcome of decisions reflecting many other goals beyond efficiency and equity. Personal income taxes may be global or schedular in their basic structure. The concept of taxable income, the treatment of particular types of income, allowable deductions, exemptions, credits, the rate structure and number of brackets, and the neutrality to inflation all differ across personal income tax systems. Stotsky covers the basic structure of the personal income tax.

Design Issues

In the design of the personal income tax, there are, in principle, four different features that determine tax liability: the choice of taxable unit, sources of income subject to tax, tax preferences, and the tax schedule. Thus, issues that matter include the right choice of the taxable unit—household or individual—and its interaction with other features of the tax system such as income subject to tax, income allowances and tax credits, and tax rate structure. Each combination has a different implication for the equity and efficiency of the tax system. Stotsky examines these considerations.

Again, while progressivity is generally accepted as a favorable feature to achieve a minimum of equity, the definition and determinants of progressivity itself have varied. Indeed, sometimes they may yield conflicting results. Thus, one measure, focusing on the distribution of taxes, may yield high progressivity as long as all taxes fall on a few, say the richest decile of taxpayers, even if the overall tax burden is low, say 1 percent of GDP. Another measure, focusing on the after-tax distribution of income, may conclude that the same tax system reflects low progressivity. Expectedly, the appropriate progressivity measure has been a matter of debate. Its particular design will have significant consequences for equity and efficiency. Norregaard covers concepts of progressivity and actual practices in selected developed and developing countries.

The reliance on the concept of comprehensive income does not mean that its measurement is straightforward. Measurement of income can be complex in the context of a business or corporation. First, correct valuation of assets and liabilities is not obvious and it may be based on different criteria such as economic value, original cost, market value, value to the owner, replacement cost, or other bases. Second, assets and liabilities may be designated in monetary or real terms and inflation will affect their values differentially. An improper inflation adjustment will mismeasure profits. Third, timing of income generation may be viewed differently, either when it accrues or when it is actually received in cash. If a loss occurs, the appropriate length of time for which it may be carried forward or backward needs to be determined so that risk-taking enterprises are not penalized relative to risk-averse enterprises by the tax system. Problem areas in the measurement of taxable income include the specification of rules regarding how to depreciate different assets over their useful lives (typical methods being straight line, declining balance, and accelerated depreciation), how to value inventories (typical methods being last in first out, first in first out, and period average methods), or how to treat exchange rate changes when a business has assets designated in a foreign currency. A particular selection will have implications for both the business’s balance sheet and its income statement. King and Chua focus on aspects related to the appropriate measurement of income and taxable income.

An aspect of corporate taxation that has assumed importance in recent years is the tax treatment—as opposed to regulatory or book treatment—of provisions for possible loan loss, in particular by banks. When financial assets are transacted, their current value is easily ascertainable for tax purposes. Such is not the case for loans, consumer credit, or other financial claims. Some of these may become worthless before maturity if they are obviously uncollectible. It may be advisable to allow some tax deductions for setting aside provisions for loan loss, albeit with specific and transparent rules. Essentially, two methods have been used. The first is the charge-off method, which expenses bad debt only as it becomes wholly or partially worthless. The second is the reserve method, in which a reserve account is set up as an allowance against the eventuality of a bad debt, while all receivables are recorded at their face value until they become worthless. Receipts associated with the risk of making loans are part of interest receipts and take place throughout the life of a loan, while loan losses may be concentrated in particular periods. It becomes apparent that both the charge-off method and the reserve method tend to favor loan portfolios with early loan losses and disadvantage those with late loan losses. Escolano looks at the advisability of tax deductions and of selecting alternative methods, and surveys selected country practices.

Integration of Income Taxes

A factor that is crucial in the design of income taxes is the “integration” of the personal income tax with the corporate income tax. The problem arises when some sources of income, such as dividends, are taxed once at the corporate level before they are distributed to individuals, and then a second time as individual income. The case for integration is built on many premises, including that corporations have no ability to pay and are simply a conduit through which income flows to individuals. As such, they should not be taxed at all, except as a withholding mechanism for the individual income tax.11 To avoid the double taxation that is caused by taxing both corporate and individual incomes, full or partial tax relief may be given at either the corporate or individual level.12 The “classical” system, however, prefers not to attempt to integrate the two taxes. This is based on several grounds, including possible revenue loss as a result of integration, the lack of definitive empirical proof that integration would lead to less bankruptcies, as well as the relative ease of administering a classical system. Country experiences reveal the adoption of partially integrated or classical systems. Various forms may be devised essentially to eliminate or reduce the extent of double taxation. Chua and King weigh the pros and cons of integration and present possible models describing the mechanics of integration.

Taxation of Capital Gains, Interest, and Dividends

The definition of comprehensive income should clearly include capital gains on real property or financial assets such as company shares. Nevertheless, reflecting concerns regarding the impact of capital gains taxation on investment, and as the measurement of accrued capital gains is impracticable in many circumstances, income tax systems have tended to view capital gains as a source of income that deserves specific provisions, often exempting them from tax or taxing them at lower rates. For example, owner-occupied residences and government securities are often exempted. On the other hand, if the gains are realized on depreciable assets, tax systems attempt to “recapture” the depreciation allowances by taxing such gains as ordinary income. Differential taxation of income and capital gains is based on the premise that taxable income is a flow from capital sources, distinct from any changes to the value of those sources themselves. If the tax schedule for capital gains is quite different from that of income, then the interpretation as income or capital gains of a particular increase in the businesses value assumes importance. Thus, for example, if money made through exchange gains is not taxed until realized, should it be ordinary income or capital gains for tax purposes? Other variations also appear. Thus, sometimes capital gains tax provisions differ depending on whether the gains are made by individuals or companies. All this leads to structures for the capital gains taxes independent from income taxes as reflected in tax rates, thresholds, inflation adjustments, and appropriate holding periods of the assets to minimize the potential for “locking in” investments and necessitating rollover provisions. King surveys these structural aspects together with selected country practices.

The tax treatment of interest and dividends poses related problems. In corporate tax systems, typically interest payments to the company’s creditors are deductible from taxable profits. The treatment of dividend payments varies more widely. A proportion of dividends may be deductible. Or, as in the majority of cases, the payment of dividends may not directly affect the calculation of taxable profits. This implies that corporate taxation has a bias toward debt financing. In turn, this increases companies’ incentives to disguise the payment of a return to its owners as “interest.” This problem of “thin capitalization” has become increasingly important in the case of foreign subsidiaries, necessitating the introduction of provisions on the part of tax authorities to limit this form of tax avoidance, for example, by introducing a ceiling on the debt/equity ratio. Recently, a new view has emerged, however, that the so-called bias toward debt financing may not be so clear as long as profits are reinvested and the realization of capital gains is indefinitely postponed. Alternative approaches to treat debt and equity financing in a neutral way include integration, the imputation of a deductible interest charge to equity financing, or a corporate cash-flow tax that would render both dividends and interest nondeductible.13 King examines the various ramifications of different tax treatments of interest and dividends and also provides a few illustrative country examples.

Other Income Tax Issues

One question that assumed considerable importance in the 1980s was up to what point would a profit maximizing corporation invest to replace depreciated capital. It was postulated that a firm invests to the point where the incremental unit of capital provides, in present value terms, a stream of real returns that is just enough to cover all costs, including taxes, associated with that investment; or where the marginal benefit of a dollar’s worth of capital per period is equal to the cost of holding a dollar of capital for that period. The cost associated with the holding of a dollar of capital per period is known as the cost of capital. This cost has two components: the cost of loan or equity financing and the cost of capital consumption reflecting depreciation. A capital income tax will affect both costs, possibly differently, affecting marginal investments. The marginal effective tax rate (METR) comprises the sum of the distortions created by the tax. The METR is also defined as the difference between the before-tax rate of return on a marginal investment and the after tax rate of return on the savings that is used to finance that same marginal unit of investment. Many tax factors affect the METR. These include tax codes that affect the financing operations—debt versus equity—of a firm; whether tax deductions for depreciation compare with true economic depreciation; and the after-tax returns to individual savers who are the suppliers of loanable funds. Inflation is a nontax factor that undermines the ability of a firm to recover the real economic cost of using its capital, while the real value of debt will decline. Thus, without indexation, inflation can either increase or decrease the true economic cost of investment. The direction in which inflation will affect the METR is not possible to deduce theoretically, given the two opposite forces at work. Analyzing the determinants of the METR, Chua examines available empirical evidence suggesting that METRs tend to vary widely across capital assets and sectors reflecting, on the one hand, more favorable tax treatment for investment in agriculture, forestry, and fishery and, on the other hand, more generous depreciation allowances in services and manufacturing industries.

Many countries use or have used tax incentives of one form or another to stimulate investment—generated from within and augmented by foreign inflows— in preferred economic activities. Various types of tax incentives have been put in place including tax rate reductions for priority sectors, tax holidays for a certain number of years, and allowances with fast write-off of investment expenditure such as accelerated depreciation or investment tax credits that allow an investment expense to reduce tax liability. Nevertheless, empirical investigations in both developed and developing countries have overwhelmingly demonstrated that nontax factors such as economic and political stability supported by adequate infrastructure and a well-trained or trainable labor force, as well as the existence of natural resources, are likely to be more important in determining the level of investment. Among tax factors, the dependability regarding the continuation of a tax structure and a low overall level of taxation are more attractive to investment than differential tax incentives. Additionally, tax incentives have efficiency costs since they distort the allocation of resources. They tend to erode the tax base, create an opportunity for “tax planning,” and tend to benefit short-term investments and those that are already profitable. They also make monitoring and, therefore, tax administration more difficult. A common experience in developing countries that use tax incentives widely is that they become an instrument for tax avoidance and tax evasion as well as for corruption by the bureaucracy. Chua critically examines the role of tax incentives and their costs and describes the experience of Canada with tax incentives in a representative case in the province of Nova Scotia.

Cash-Flow Tax

Some economists have criticized the corporate income tax arguing that for all its complexity in design, in practice, it is not a tax on income at all but rather on some base residual of various ad hoc exemptions and deductions. They have recommended a simpler tax base to be defined as the cash flow of businesses. Three variants of the corporate cash-flow tax (CCFT) are as follows: The first variant is the R—or real—base CCFT in which the tax base is net real transactions (the difference between sales and purchases of real goods and services). As opposed to a corporate income tax (CIT), the RCCFT allows immediate expensing of capital outlays but not the deduction of interest payments. Interest received is not taxable. The second variant is the RF—or real plus financial—base CCFT, and in addition, includes in its tax base nonequity financial transactions (the difference between borrowing and lending). Interest and retirement of debt are deductible, while borrowing and interest received are taxable. The third variant is the S—or shareholder—base CCFT, which taxes the net flow from the corporation to shareholders (dividends paid plus purchases of shares minus issues of new shares) and conforms closely to the interpretation that the CCFT is a “silent partnership” of the government in any investment.

The CCFT’s advantages lie primarily in the theoretical clarity of the tax base insofar as it does away with the problems of defining true economic depreciation, measuring capital gains, costing inventories, and accounting for inflation (although not in all variants of the tax). The CCFT, however, can give rise to problems—for example, tax-base erosion through avoidance and evasion. This could be contained by carefully designing the tax code and by selecting an RF-base over the R-base CCFT, thereby including the financial sector. On the other hand, an important advantage of the R-base CCFT—nondeductibility of interest, which eliminates incentives for debt over equity financing and obviates any need for inflation adjustments for the calculation of real interest—is not shared by the RF-base variant. The S-base CCFT, while sometimes favored because it has been perceived to be administratively simpler, could lead to a tax rate of over 100 percent because of the definition of the S base. Thus, the choice among variants of the CCFT is not at all clear. In addition, international considerations turn out to be extremely important in any future implementation of the CCFT because of the unresolved treatment of foreign tax credits under a CCFT. The CCFT remains a theoretically attractive option with accompanying practical difficulties. The CCFT may prove particularly difficult to implement for a single—especially developing—country in an environment that may not necessarily accommodate its smooth and effective operation. Shome and Schutte examine theoretical aspects as well as the practical ramifications of the CCFT.

Payroll Taxes

An income tax that has become an important revenue source in both developed and developing country tax structures is one that is levied on payrolls. Payroll taxes are generally applied as a flat percentage of an employee’s gross wages up to a limit. They are the principal means of supporting social insurance programs including social security, health insurance, unemployment insurance, and disability insurance programs. Payroll taxes are unusual in that their revenues are generally earmarked for specific purposes. While earmarking introduces rigidities into the budget by requiring revenues to be spent regardless of need, taxpayers may be more willing to pay a tax when they see a tangible return from the payment. In any event, the nature of expenditures that payroll taxes finance is found to be sufficiently socially important to safeguard it through earmarking.

Social insurance systems are financed either on a pay-as-you-go or funded basis. Under the former, the taxes that current workers pay cover expenditures for current beneficiaries. In contrast, the latter accumulates the taxes paid by individual taxpayers in a reserve fund to cover their future benefits. In recent years, social insurance programs based on the pay-as-you-go principle have faced the problem of rapid growth in current and future liabilities as a result of aging populations or expansion of benefits or both. At the same time, a reduction in the productivity growth rate has often led to a stagnating tax base. As a result, there has been some tendency to move toward funded systems. In terms of their design, payroll taxes typically adhere to simple structures, with no exemptions from gross wages, single, though different, rates for employee or employer contributions, and mostly ad valorem rates. Different rates are typically used for different program categories, however. While both employers and employees must make statutory payments, the burden of the payroll tax is likely to fall on the labor market since wages are likely to ultimately reflect labor productivity which is not altered by the payroll tax. Assuming that the incidence of the tax is on the employees, the tax is regressive. This may lead to important ramifications for the composition of labor. Stotsky surveys various considerations under the payroll tax and practices in selected developed and developing countries.

Taxes on Assets, Property, Inheritance, and Gifts

In many countries, business assets or receipts are taxed, sometimes as a supplement, replacement, or minimum contribution to the income tax. A low tax rate is applied to a typically large base. Even with reasonable capital mobility, it is generally found that the average return on capital is dissimilar across industries. Thus, a tax based on an average rate of return on business assets is likely to be an imperfect proxy for the income tax. An assets tax is therefore better used as a supplemental or minimum tax. Many Latin American countries levy assets taxes using a variety of bases such as gross assets, net worth, fixed assets, and others. The gross assets tax emerges as the best tax, given theoretical and practical considerations. A gross—rather than net—assets base is appropriate if the objective is not to favor debt financed assets. A fixed assets base is simpler but tends to discourage investment in fixed assets as opposed to other asset forms. Under a gross assets tax, the incentive to choose output and input prices strategically is diminished, thus minimizing transfer pricing issues. A gross receipts tax base may include only product sales or also asset and equity sales, as well as receipts from issuance of new debt. Nevertheless, a gross assets tax also needs to be designed carefully. Valuation becomes complex, as the base includes varying proportions of long-term and current assets requiring average asset values to be determined for tax purposes, and as the impact of inflation becomes difficult to track during the life of a long-term asset. The timing of valuation is also important since a business’s asset holdings fluctuate over the year. Other design issues can be quite important, including double taxation when taxable businesses own financial interests in one another. Also, the design of a gross assets tax must neutralize liquidity constraints that might otherwise arise from complying with tax liability. Krelove and Stotsky examine the pros and cons of the various taxes in this category.

Taxes on land and property are among the oldest forms of taxation. They have been justified on grounds of both the benefit and the ability to pay principles since government provides benefits in the form of law and order enabling the maintenance of property rights, while the ownership of property indicates ability to pay. Forms of property taxation include taxes based on rental value, capital value, or land value. But there are many practical problems associated with the tax. First, property is a heterogeneous good. Second, an accurate assessment of “fair market rental” or “fair market value” is often difficult, since they have to be based on elusive concepts such as income generation capacity or, if owner occupied, on comparable property value. Third, sometimes assessments are set low for social considerations. Fourth, while a flat tax rate would be better in terms of tax compliance and is easier to administer, the continuance of low assessments often results in high, progressive and, in turn, unimplementable tax rates. With respect to agricultural land taxation, while it may have comprised the oldest property tax, its importance has declined in terms of revenue even though some countries maintain records on them. An important reason is that taxes that have been applied, such as those based on land area, may not reflect land value and are inelastic in inflationary times. Thus, one of the major reform areas remaining to be implemented is agricultural—income and property—taxation. Stotsky and Yücelik examine various aspects of land and property taxes.

Another vintage tax is that on property transferred at death, reflecting a social as well as an economic philosophy. The objectives of the tax include limiting one’s right to dispose of one’s wealth at death or to acquire wealth through bequests without “own effort,” establishing a final point to capture lifetime capital income while reducing disincentives for saving and investment, and redistributing wealth. However, these and gift taxes typically generate little revenue. Their bases are often eroded by large exemptions and under-valuations. Tax avoidance is practiced by “generation skipping” whereby a bequest is left to grandchildren to minimize the number of times an estate changes hands. While rates are progressive, usually with three or four rates as evidenced in a large number of countries, they tend to lead to parcelization of estates. Valuation problems relate mainly to the heterogeneity of properties such as life interests, annuities, personal effects, and businesses and shares for which cadastral values may not be used. Another issue is the timing of valuation, which has to be determined as the fair market value at the date of transfer or the date of death. Yiicelik examines these and related issues pertaining to the taxation of bequests, inheritance, and gifts.

Taxation and the Open Economy

The traditional concerns pertain to the design and effects of international trade taxes. As links through international trade and finance have made the world more economically integrated, however, so have the complexities in the method of taxation, for example, whether to be based in the country where the “source” of income lies or where taxpayers are “resident.” Chapter V focuses on various aspects of taxation in the context of an open economy.

Free Trade, Protectionism, and Nominal and Effective Protection

The case for free trade goes back to Adam Smith and David Ricardo. It is “static” in nature. The idea is based on the “comparative advantage” of trading partners in their production activities, transforming exports into imports according to their international relative prices. Also, international trade expands the set of goods and services that consumers can afford by reducing their prices or by making new commodities available. A country will achieve a higher welfare with free trade than with protection, which distorts prices and restricts market exchange. Recent theory postulates that “dynamic” economic growth is the result of various factors including human capital accumulation, investment in research and development, and that open trade policies allow a country to profit from growth-enhancing factors. Closer economic links increase the transmission of new technologies. For example, imports of intermediate products may embody technology developed abroad. Greater competition leads to greater assimilation of new technologies, improved products, and a wider diversification of output. Protectionist policies have the opposite effects.

Arguments in favor of protectionism assume that governments can “pick winners” or “infant industries” to determine which industries to support. In practice, however, trade policies tend to be influenced by special interest groups, hampering the selection of both winners and infant industries. Beneficiaries of protectionist trade policies devote considerable resources to “rent-seeking” activities aimed to maintain and extend existing protection, further magnifying distortions and inefficiencies in resource allocation. The rationale for tariffs, therefore, has to emerge from elsewhere.

Unlike nontariff barriers—quotas, voluntary export restraints, and subsidies—trade taxes yield revenue, sometimes constituting its main and most stable source. Notwithstanding, import and export tariffs are not optimal instruments to raise revenue. A combination of domestic taxes levied equally on domestic and imported products along with revenue neutrality with respect to tariffs will cause a lower efficiency loss. The inward-oriented bias caused by tariffs can produce large inefficiencies and hamper growth. Domestic taxation of consumption or income can meet the revenue target with lower rates, broader bases, and without a protectionist bias. If domestic tax sources are not easily available, as in rudimentary economies, tariffs are the available option. An argument in favor of tariffs is their lower administrative cost. Countries with weak tax administrations and a lack of accounting sophistication on the part of taxpayers tend to rely on them. Sometimes tariffs or across-the-board import surcharges are applied to contain an imbalance in the external sector and thereby temporarily avoid necessary domestic adjustment. Countries that favor exemptions within the general structure of tariffs may include a minimum tariff to capture all potential contributors. In an increasingly complex trading environment, some countries resort to tariffs as an “antidumping” measure.

In principle, a tariff may be beneficial for a country that can affect the international price of its imports or exports. For example, an import tariff can result in an optimal pricing strategy when the country has a monopsony in import markets. The “optimal tariff is determined in this context by setting the benefits— government revenue and private profits—against the costs—domestic distortions and misallocation of resources. Similarly, those countries that have an oligopoly in a natural resource can improve their terms of trade by restricting or taxing exports. The rest of the world, however, would be worse off. Small trading economies, such as most developing countries, however, cannot expect to improve their own welfare through optimum tariff policies. Further, a “new trade theory” postulates that economies of scale and externalities cause decreasing average costs and thus, larger producers tend to have an advantage over smaller ones. Producers that gain initial control of a large share of the market are able to drive competitors out. This in turn results in retaliatory tariffs. The design of strategic trade policies, however, requires extensive knowledge of many details of world markets that are often unavailable to small trading economies.

How much protection is obtained for an importable item by imposing an import tariff cannot be gauged directly from the statutory tax structure. “Effective protection” may be quite different from “nominal protection.” A tariff on an import raises its domestic price and shelters it from international competition. The amount of protection is usually expressed as a percentage of the international price. If the tariff is an ad valorem tax proportional to the value of imports, the tariff rate measures the nominal rate of protection. If the tariff is specific, the nominal rate of protection is given by the tariff divided by the price net of the tariff. The nominal rate of protection, however, is not always a good indicator of true protection since it only looks at output price. If a tariff applies to intermediate goods, it will increase domestic input prices, thereby bringing down the true or effective protection enjoyed by the domestic producer. It therefore depends on the magnitude of the value added. Indeed, the effective rate of protection can be defined as the amount by which the value added in a sector at domestic prices exceeds the value added in the sector at international prices expressed as a percentage of the latter. Effective and nominal rates of protection for a sector may diverge widely. Neutrality of a trade regime implies equal rates of protection across industries of tradable goods, both exports and imports. Even high tariffs could be devised to maintain neutrality even though they would be very distortive. Thus, neutrality of a trade regime is not an indicator of the efficiency loss from protectionist policies. Escolano examines these conceptual issues in some detail.

Tariff Reform: Import and Export Duties

In recent years, policy advice on trade reform, notably in IMF-and World Bank-supported adjustment programs, has emphasized the need to lower average tariffs and to curtail the dispersion of existing rates. Reform aims at a low, uniform tariff and at the removal of non-tariff barriers. The rationale is not purely theoretical; rather, it is based on a wide scope of considerations ranging from revenue and administration costs to the suboptimality of protectionism. The case for a low tariff rests also on its small economic cost which is the value of the output forgone due to losses in economic efficiency. Minimizing the dispersion of rates is also based on several grounds. Multiple rates substantially increase the cost of administering the tariff. A uniform rate implies uniform effective protection of all domestic industries. Rate dispersion encourages special interest groups to devote resources to increase the amount of protection granted to them. Distortions prompted by a tariff grow more than proportionately with the tax rate.

Nevertheless, the literature on optimal taxation lends favor to greater dispersion of rates. One result of this theory is that final goods should be taxed at higher rates than inputs. This is because tax-induced changes in the relative prices of inputs can lead to the choice of inefficient technologies since some inefficient technologies may become more profitable to the extent of overtaking efficient technologies. Extending that argument, the theory of optimal taxation recommends higher tariff rates on final consumption goods than on intermediate products. Among final goods, those with a more inelastic demand should have higher rates. The theory of optimal taxation, however, has found little application in practice because of the vast amount of information required to construct an optimal tariff structure. Independent of the directives of optimal taxation theory, in many cases, policymaking needs to operate within the parameters of a given level of protection. The optimal tariff structure under that constraint will depend on the reasons for protection. For instance, if the objective is to redirect resources from industry to agriculture, a uniform tariff will not help achieve that goal. As mentioned previously, however, a case for a single rate tariff structure might still be made on efficiency, administrative, or political grounds.

Despite the success of outward-oriented trade strategies, many developing countries maintain high levels of protection. Often the objective of import substituting policies has been to protect domestic producers of final consumption goods through high tariffs or quantitative restrictions. Protectionist policies have also led to an anti-export bias. Since most developing countries are small traders and cannot expect to increase the international prices of their exports, high cost and limited availability of inputs result in a negative effective protection rate for those industries that employ importtables to produce export tables. In contrast, many of the world’s manufactured exports come from places where exporters face relatively low trade barriers with regard to the taxation and availability of inputs. One way in which some successful exporting countries have provided exporters with inputs at world prices has been by applying a policy of zero tariffs on all inputs. Other methods that have been practiced to insulate exporters from taxation are duty drawbacks that allow a rebate to exporters on all duties and indirect taxes paid, duty waivers and exemptions from nontariff barriers, and the installation of bonded factories and duty-free export processing zones, which are essentially islands of production solely for export. Little analytical work exists on methods of isolating exporters from inward biases created by protectionist policies.

Export duties have played an important role in the tax structures of many developing countries. In 1980, 67 countries imposed export duties, mostly on one or two commodities that accounted for a large share of the traditional exports of the country. They have been defended on various grounds: they improve terms of trade, substitute for income taxation, stabilize export revenue and tax away windfall gains. Empirical studies, however, show that exports have been overtaxed, leading to excessive reduction in supply and loss in foreign currency earnings. They have also in general failed to stabilize producers’ incomes. In the 1980s and 1990s, many developing countries dismantled them. Sometimes the export sector has either been taxed implicitly through foreign exchange surrender requirements at artificial exchange rates, or it has had to contribute to export stabilization funds or face prices designated by state marketing boards. These have occurred to stabilize domestic prices and incomes rather than to generate revenue. In practice, however, accumulated fund surpluses have often not been sufficient or have not been used to compensate for losses incurred in other periods. The operations of state marketing boards tend to become complicated and fall captive to special interest groups. Escolano examines matters pertaining to import and export tariff reform in some detail.

Source and Residence Principles, Tax Treaties, and Tax Harmonization

Taxation of productive factors in an economy— whether directly or indirectly—involves balancing national against international considerations. While national objectives are concerned with revenue, allocative efficiency, and equity effects of taxation, such tax systems may have fundamental ramifications on the volume and allocation of productive resources internationally. This is because factor returns and the underlying tax bases straddle national boundaries. To deal with possible conflicts, adjustments to domestic taxation are needed. One set relating to direct taxation is represented by the residence and source principles. The analogous set for consumption taxes such as the VAT is the origin and destination principles.

For income taxation, the residence principle asserts that individuals are taxable in the country or tax jurisdiction in which they establish residence regardless of the source of income. For legal entities, establishing residence is less clear cut, though it is increasingly tied to the location where its business activities are registered. The source principle asserts the prior, and even sole, claim of the country in which the income arises to natural or legal persons, to tax such income without reference to other criteria. In practice, countries have tended not to stay with the pure application of a single principle but to apply a mix—residence for nationals residing in the country, and source for income earned within the country by nonresidents or nonnatural persons or both. The nature of mixes has depended on a country’s objectives regarding foreign investment, revenue, administrative capabilities, and degree of cooperation with competing jurisdictions. Since the various mixes are not uniform, double taxation might result, leading to the need for tax treaties.

For consumption taxation, for example, the VAT, when exports are zero-rated and imports are taxed, a destination-based VAT results. Most countries apply the destination principle, although within multicountry trading blocs, such as the European Union, the elimination of national boundaries would result in the use of the origin or source principle. This needs to be supplemented by clearing-house arrangements to compensate for revenue gains or losses.

The inherent conflict between the residence and source principles is often resolved through tax treaties initially contracted on a bilateral basis between developed countries with a rough balance in the exchange of income flows. Over time, however, the treaties have diverged on concepts, structure, and operating rules. Nevertheless, they cover complex and conflicting concerns of the contracting parties about sharing income taxing jurisdictions. They address concepts and definitions for tax purposes of a permanent business establishment, tax treatment of dividends, interest, royalties, and capital gains as well as arrangements such as “tax sparing,” which allows any tax benefit accruing to a business in a capital importing source country to be spared from being recouped in a capital exporting residence country. Tax treaties also contain provisions for the exchange of information with tax implications between national tax jurisdictions.

If taxes across international borders were fully “harmonized,” there would be little need for tax treaties. At another level, however, tax competition through market pressure may be desirable because it would exert a downward pressure on taxes, on public expenditure, and on greater efficiency in the allocation and use of world public resources. But tax competition has destabilizing short-term macroeconomic spillover effects which could interfere with the efficient functioning of global trade and capital markets. Thus, tax harmonization does play a salutary role in supporting interjurisdictional equity, locational neutrality, and taxpayer equity. In this chapter, Faria focuses on an analysis of the various principles, considerations in the design and content of tax treaties and tax harmonization, as well as a brief look at their development.

A major problem for tax authorities has been to monitor how and to what extent multinational enterprises allocate their global income among fiscal jurisdictions in order to minimize their overall tax liability. In the literature, this issue is referred to as “transfer pricing,” where transfer price is the price for the internal sale of a good or service in intrafirm trade, that is, among affiliates in different countries. Transfer price manipulation has increased in importance as intrafirm trade has multiplied over the last two decades. For example, in 1989, 86 percent of U.S. parent company imports were from foreign affiliates and 89 percent of U.S. parent company exports were destined for foreign affiliates. National tax authorities are faced with the task of establishing and enforcing rules for the setting of such prices while minimizing conflicts with other jurisdictions. Fortunately, a consensus has developed that a single guideline should be adhered to, generally known as the “arm’s-length” criterion, which stipulates the price that would have been negotiated if the parties were unrelated. The OECD, for example, has developed several guidelines for determining arm’s-length price, based on a known comparable uncontrolled price or a price at which a multinational sells to an unrelated party. Nevertheless, developing country tax authorities in particular continue to be ill-equipped to cope with transfer-pricing issues. McCarten investigates these matters, together with experiences of developed and developing countries.

Relatedly, the tax treatment of branches and subsidiaries of multinationals has many facets.14 In general, attributable profits for tax purposes must adhere to arm’s-length conditions. But there are specific advantages and disadvantages of operating as a branch or subsidiary. For example, a corporation can usually deduct fully the losses incurred abroad by a branch against its home country tax liability. Further, a branch can repatriate after-tax profits to the parent company without further taxation, such as a withholding tax. Usually, parent companies can transfer property to a branch without incurring taxation in the home country. Tax incentives are often carried over to a branch but not to a subsidiary. There are also many disadvantages of a branch operation. For example, a parent company cannot defer home country taxation on branch income that is not remitted to the home country. Host countries typically grant more generous tax options to subsidiaries than to branches, such as for loss carried forward or backward, or deductions, to limit transfer pricing abuses. McCarten examines these and related issues in this chapter.

Selected Issues in Taxation

Particular issues in tax policy exist that cannot be clearly categorized under the above sections. In other cases, even though it would not be impossible to consider them as a part of those topics, they are unique enough to be best considered independently. These special issues are grouped together in Chapter VI.

Taxation of Mineral Resources

The taxation of petroleum resources deserves special consideration. One way of considering the issue is to treat it as a tax on the return on capital assets such as mineral resources. Government is cast to have a dual role: it is a sovereign taxing power as well as a resource owner. As the sovereign tax authority, it will impose the same tax arrangements that apply to economic activities in general. As the resource owner, it will seek a return on its resource assets. It can use various tax instruments such as income tax, royalties, resource rent tax, and others. Questions often arise as to the appropriate tax instruments under those circumstances. The taxation of mineral and petroleum resources, especially as they have arisen in practical queries and design of tax systems, is the subject of Nellor’s examination in this chapter.

Taxation of the Financial Sector

The financial system that directs capital to its various uses is critical to a modern economic system. It includes commercial and savings banks, credit unions, insurance companies, pension funds, brokerages, and other institutions. In some developing countries, the informal financial sector comprising money lenders, cooperative and trade credit, pawnshops, and various funds, may be more important. The financial sector increasingly provides a variety of services, acting as intermediary between borrower and lender, offering insurance against risk, financial management services, and consulting. The taxation of this sector—in the form of income or consumption taxation—demands separate consideration.

In the case of income taxation, the measurement of income of financial institutions relates to the timing of income. They provide services that have an explicit time dimension. The possibility of future receipts or payments is a part of their business activities. A corporate income tax typically applies to business income as it accrues. Yet banks have to set aside funds for bad loans and insurance companies for future payments on policies. Governments often require them to set aside minimum funds which may not necessarily match, and will probably fall below, their actual payment needs. Thus, the extent of tax deducibility becomes an issue. Another problem arises from the intertwined nature of investment and other services provided by financial institutions. Taxpayers typically owe tax on investment income net of costs while costs of services not related to investment income are not deductible. To the extent that financial institutions are not able to separate the two cost components, there may be an understatement of taxable income. Another problem relates to the separation of return on capital from return of capital. Essentially, the fact that a deposit represents liability and not income needs to be recognized in the calculation of taxable income. An insurance company’s income is derived from premium payments and investment income. Premium payments may represent both a payment for services and a “deposit” for investment. Only the former should comprise income since the other component must be returned. In practice, it may be difficult to separate the two. In the case of banks, however, deposits can be identified more clearly.

In the case of consumption taxation, for example the VAT, its base should include financial services in principle. It is possible, in general, to measure value added by adding profits, wages, rent, and interest or, alternatively, by taking the difference between investment income and the cost of funds. The use of invoices under the VAT requires that VAT liability be attributed to each transaction. This is not possible for the banking sector, however, since financial services provided by banks do not have specific charges attached to them. Charges result from differences in interest rates charged and paid. In the insurance sector, value added in non-investment insurance is measured by the “loading charge”—earnings of the insurer over and above payments of claims. Value added is not properly measured by the value of premiums or claims since this includes the component of premiums that is a redistribution from one policy holder to another. Again, value added has to be measured as the loading charge, eliminating the savings component. It is, therefore, not easy to tax financial institutions under the VAT. They are, therefore, often exempted. For example, OECD countries exempt their intermediary functions while taxing selected activities such as check printing, safe deposit rental, and foreign exchange transactions. Israel is one country to have attempted to apply a comprehensive base deriving it by adding different value added components. Stotsky examines various issues pertaining to the taxation of the financial sector.

Fiscal Federalism and Tax Assignment

Governments comprise many levels. Different taxes are assigned to different levels. Revenue from the same tax is also shared across levels. Many questions arise as to their best division. While valuable literature has developed over the years on these issues, only recently are their ramifications on macroeconomic stabilization being questioned. The rationalization for decentralization of fiscal responsibility to various levels of government rests on the potential efficiency gain which follows from the possible differentiation of the provision of public goods and services in different locations in accordance with the tastes of the local population. From this starting point, governmental functions are assigned to various levels. The consensus that has emerged is that distributional policies should be assigned to the central government since otherwise, people bearing the burden of those policies would migrate. Likewise, local stabilization policies will be inefficient because of spillover effects. Thus, the core function of local governments should be allocation. The financing of lower levels of government may come from user charges, taxes, central government grants, and borrowing. Of these, user charges seem to be the most practical and commonly used. The ability to borrow relatively freely has led to impediments in fiscal stabilization and to limits on this power. The assignment of taxes to lower levels of government has not necessarily led to their determined effort to generate revenue, particularly in developing countries. Thus, grants and revenue sharing from central government taxes have played an important role in financing lower level governments.

A good local tax is one whose burden cannot be exported by the local jurisdiction elsewhere. Though some experts recommend local income taxes, land and property taxes seem to satisfy the criterion best. However, “piggyback” local taxes on a central income tax could be devised. On the question of designing the tax, while tax rates can be left to local taxing powers, the tax base should be determined at the central level since determining the base involves distributional considerations and consequences and differences in base across localities would reduce transparency and accountability. Apart from tax assignment, governments may use revenue sharing and grants to achieve particular objectives, such as income redistribution across regions, to overcome insufficient local revenue generation capacity and to induce local governments to expand particular services. The objective, however, should be to equalize, across localities, the ratio of tax revenue to services provided. Thus, unless grants clearly correct for externalities, their amounts should be minimized. Nevertheless, actual country policies vary considerably in terms of the equalization criterion. Norregaard considers in a nutshell various issues pertaining to fiscal federalism, focusing on the revenue side.

Revenue Forecasting and Estimating

Revenue estimating is the process of assessing the impact on revenues of tax law changes proposed at the time the budget is presented. On the other hand, revenue forecasting takes place even when no change in the law is proposed, just for the regular budgetary process. A number of different methodologies are employed for revenue forecasting. First, one method of making an unconditional forecast is to extrapolate an established linear trend in receipts from a particular tax. Alternatively, to make a revenue forecast when a GDP forecast is available, an estimate of the elasticity of tax revenue with respect to GDP may be employed. Of course, the elasticity itself will have to be re-estimated from time to time since the procedure assumes that elasticity values are constant. This is done by removing from the time series on tax revenue the effects of any changes to the tax law that may have been made during the period. Particular procedures have been developed for this purpose under different assumptions regarding tax revenue behavior over time resulting from a tax law change. Second, dummy variables may be used to capture the revenue effect of a tax law change. Third, more complex macroeconomic models are also used for revenue forecasting. Generally, regression methods are used to estimate functional relationships between revenues from particular taxes and a variety of macroeconomic variables. The relationships need not be constrained to imply constant elasticities. An advantage of this procedure is that revenue forecasts are integrated with the corresponding macroeconomic forecasts. Fourth, structural “microsimulation” models of major taxes have also been developed based on tax return data. Many OECD countries use this method. They apply the tax law to the structure of the tax base at the level of individual taxpayer liabilities. This method, however, suffers from a potential inconsistency between macroeconomic forecasts and revenue forecasts.

Revenue estimating is the assessment of how possible changes to the tax law will affect tax revenue. Several factors need to be accounted for to obtain the right revenue estimate. First, the effect will vary over time. Second, collection lags will affect revenue. Third, some changes will only have temporary revenue effects. Fourth, change in one tax law may affect revenues from other taxes. It is harder to generalize about revenue estimating methods than revenue forecasting methods since the tax law changes are varied. For example, changes to the rate of a proportional tax, say on consumption or wage income, may be straightforward to deal with, while changes to an existing tax allowance in a progressive income tax may be more difficult. Originally, microsimulation models were developed to tackle these types of issues. But even these models may not suffice in revenue estimating. For example, estimating the revenue impact of extending the personal income tax to cover social security benefits may not be feasible from a sample of tax returns since the latter would not contain information on currently exempt income. Thus, revenue estimating remains relatively complex. King focuses on the different procedures and methodologies of revenue forecasting and revenue estimating as well as their pros and cons.

Presumptive Taxation

Many developing countries use presumptive methods of taxation to counter deficiencies in tax administration. Often they are used to tax small businesses, which may represent the majority of enterprises. But they may apply to entire classes of taxpayers or only to taxpayers who fail to file a standard tax declaration. If used judiciously, presumptive taxation may broaden the tax base by increasing the number of taxpayers and their tax payments. Even if the revenue per taxpayer is low, it may have spillover benefits in facilitating the movement of small taxpayers from the informal to the formal sector and as a source of information to combat evasion. Presumptive methods may help reduce audit time and cost. Since they generally comprise a tax on average or “normal” income, the marginal tax rate on income above this average income is zero. Therefore, they avoid the negative incentives associated with high marginal tax rates. Nevertheless, they carry the danger of harassment and extortion of taxpayers by unscrupulous tax officials. The use of punitive presumptive tax rates to push small taxpayers into self assessment may backfire by causing taxpayers to go underground.

There are several different ways to levy a presumptive tax. Estimated income, assets, turnover, or external indicators of income are all alternatives. The level of economic sophistication influences the choice of presumptive tax methods. A simple approach is to levy a lump sum on all businesses. A more sophisticated approach involves a census of taxpayers and a determination of average profit margins, using objective factors differentiated by activity, based on which presumptive schemes may be developed. Nevertheless, the detailed approach has been criticized because using precise factors to determine income may transform the tax into a tax on that factor rather than approximating a general tax on income. Assessment may be either on an individual or a collective basis. Under the latter, individual taxpayers cannot contest the levy, and tax administration is simplified. Under the former, taxpayers may be required to submit certain annual information so that the administration can assess their net income by applying cost-profit ratios. At the same time, the taxpayer can negotiate tax liability and appeal through the judicial system. Bulutoglu analyzes the many properties of presumptive taxation including a shortcut to auditing, the use of income indicators, the pros and cons of individual versus collective assessment, and their incentive effects.

Minimum Taxes

Another method for generating revenue when tax administrations are not perfect is to require taxpayers to make minimum contributions to selected taxes. By virtue of their uses, presumptive taxes overlap with minimum taxes, but this may not always be the case. Similar is the comparison between assets taxes and minimum taxes since increasingly assets-based taxes are being used as minimum contributions toward business income taxes. The purpose of a minimum tax is to ensure that businesses or individuals with economic income do not regularly avoid paying tax on it. The United States, Canada, Denmark, and Norway have a minimum income tax based on a broader concept of income (with less deductions). Among Latin American countries, Argentina, Ecuador, Mexico, and Peru have minimum business income taxes based on gross assets.15 Several other countries are considering minimum income taxation. A business minimum tax may reduce the inequity of the business income tax which may arise because of differences in tax compliance across businesses. Also, as tax preferences such as deferrals and exclusions proliferate, the resultant narrowing of tax bases is partially restored by minimum taxes. In an inflationary environment, the gain made by debt-financed firms in reducing tax liability is also checked. A business minimum tax has also been justified as a business license tax. Nevertheless, the two are different in that a business minimum tax is always creditable against the regular income tax, but a minimum license tax is not.

There are various forms of a business minimum tax. In a simple form, it may comprise a requirement for each taxpayer to pay fixed nominal amounts. This would function as a lump-sum tax and hence would be efficient. But it would be inequitable since it would not be a proxy for an income tax. Using turnover as the base provides certain advantages since turnover is the most easily measured financial variable for a business and most easily available to tax authorities. An assets based minimum tax has a theoretical appeal in that economic income could be expected to bear a systematic relationship to assets. It has to be designed carefully Its base is gross business assets including cash and securities, receivables, inventories, land and other fixed assets at depreciated value, and intangible assets at amortized value. Alternatively, it is also possible to impose the tax on fixed assets—land, plant, and equipment—but this discriminates against particular asset forms, or on net assets—gross assets net of debt-financed liabilities—but this does not remove the incentive to reduce the tax base through increased borrowing. Mexico’s minimum assets tax has a rate of 2 percent based on gross assets. The assets tax liability is designed to be roughly equal to a taxpayer’s income tax liability. If the taxpayer is assumed to earn a 6 percent return on assets and the business income tax rate is 35 percent, then a 2 percent tax on assets is roughly equivalent. Taxpayers can credit their income tax liability against their assets tax liability. Another alternative is to levy the tax on some redefined notion of business income, as in the U.S. corporate alternative minimum tax, which is computed by making certain adjustments and adding certain tax preference items to income. Businesses are required to compute taxable liabilities under the regular and alternative systems and pay the higher of the two. Stotsky analyzes aspects that must be kept in mind in designing minimum taxes for both businesses and individuals, such as appropriate bases and rate structures.

Tax Policy and Tax Administration

Tax policy and tax administration are inextricably related. The government’s role requires that it be able to finance its activities in a noninflationary way through compulsory extraction of resources from households while minimizing distortions. Herein lies the primacy of tax policy in helping to attain economic policy objectives. Tax administration must therefore evolve an internal dynamic to promote the effective application of tax policy.

In nascent economies, tax administration may have to focus on large taxpayers, while rudimentary presumptive methods may have to be used for contributions from smaller taxpayers. As economies mature, however, consideration should be given to the extent to which the administered tax system resembles the legislated tax structure. If, to simplify tax administration, potential taxpayers have had to be ignored or actually left out of the functioning ambit of a tax such as the VAT or income tax, consideration must be given to the extent to which the universe of taxpayers can be increased over time to include more of the “minor” taxpayers. Relatedly, while the role of “large” taxpayer units in improving auditing and revenue performance should not be minimized, the rate at which the coverage of these units can be expanded—if not merged with the general taxpayer population—should be considered an important criterion for measuring the maturity of a tax system. Faria and Yücelik consider the interrelationships between tax policy and tax administration.

Tax Reform Experience and IMF Tax Policy Advice

Policymakers often ask for the nature of cross-country experiences in tax reform, and to view the extent of common trends and divergences. Especially as the transition economies of Eastern Europe and the former Soviet Union undertake reform, it is interesting to compare their experiences with tax reform patterns in market economies. In this light, it is also interesting to examine the content of IMF tax policy advice. These comprise the focus of Chapter VII. This chapter is supported by an appendix containing selected tables on tax revenue data (1975–92), their relation to GDP, and their shares in total tax revenue across different country groups.

Recent Tax Reform Experiences

In market economies, recent tax reform experiences form certain patterns. Their patterns change with time; for example, the 1960s and 1970s were quite different from the 1980s. The patterns also differ according to geographical area; for example, trends of Latin America may be distinguished from those of Europe and Asia. In the 1990s, the concerns in carrying out tax reform in ex-Socialist or transition economies are again quite different. Overall, among market economies, it can be said that the VAT is now universally accepted as an economically efficient and administrate tax. The bases have been broadened and there is a recognition that few rates are better. As dependence on the tax has grown, however, so have the rates. The number of selected excises has diminished, the main ones being on tobacco, alcoholic beverages, and petroleum products, even though other excises continue, especially with tariff reform. Customs tariffs have been scaled down and export duties have mostly been eliminated. Income tax rates and rate dispersion have decreased. Top marginal rates apply at lower levels in terms of per capita GDP. Rates of taxation affecting international flow of capital have also come down. While there is a consensus that the income tax base should be broadened and made more transparent, there has been less success in implementing this. Many developing countries are introducing presumptive taxes, minimum contribution requirements, and withholding taxes to buttress tax administration.

Patterns are less easily identified among transition economies. Even as modern tax policy is taking form in those economies and, while recognizing that they may have come a long way from the system of transfers typical of a command economy, it would be pertinent to note that transition economy tax systems are rapidly acquiring complex, distortionary features. Many of the taxes conform neither to broad-based tax structure design nor to a simple—though distortionary—interim tax structure focused on closing the fiscal deficit. Needless to say, nascent tax administrations could hardly be expected to implement these tax structures efficiently. As a result, tax systems need much improvement.

To list the causes behind this situation, first, transition economies were operating their own complex tax/transfer mechanisms—reflecting social considerations as well as priority activities—before the period of change. Some of the current practices—even though modified—reflect old centralized operations in favor of intervention and differentiation. Effort has to be made to steadily reduce such methods. Second, many transition economies continue to cast their tax policies in the model of (or as a reaction to) the Russian Federation. Third, transition economies are receiving technical assistance from diverse sources. Their tax policies reflect a mix of such advice. For example, European thinking on the VAT with the accommodation of multiple rates is different from that based on wider experience, inclusive of Latin America and Asia, which tends to indicate that multirated VATs are difficult to administer. Or, some experts may think that a modified cash-flow tax may be ideal in a fresh environment, while others may think that it may be quite complex as a starter (Tait, 1992). Fourth, experience in both Western and Eastern Europe reveals that tax re-form is a complex process and may be expected to be the same in transition economies. Long preparation is needed for major tax reform even in industrial countries in terms of tax administrator and taxpayer education, as well as in terms of the fruition of the process—from tax policymaking, to giving it legal form, to putting it into action. In Eastern Europe, such as in the Czech Republic, tax packages have had to be withdrawn after they were announced and published, to be ratified later in a modified form, through a referendum.16 Similarly, in transition economies, already, tax packages proposed by the government’s executive branch—reflecting, for example, technical assistance recommendations of multilateral institutions—have not been accepted by the legislative branch. This simply implies that tax reform will be a slow process, as experienced in many countries in Latin America or East Asia, where fundamental tax reform is taking hold after decades of experimentation. Faria surveys the various experiences with tax reform.

IMF Tax Policy Advice

IMF tax policy advice has emphasized the need to enhance the neutrality of tax systems and, in line with international reform trends, to improve the administrability of the tax code. In general, Fiscal Affairs Department missions have recommended two courses of action: (1) to simplify the structures of existing taxes (by reducing the number of rates, broadening the bases, and eliminating preferential treatment of particular economic agents or activities); and (2) to introduce new and simple taxes (such as a single-rate VAT) to replace old and complicated ones (such as a multi-rate turnover tax).

When countries have sought IMF advice for identifying policy options to mobilize additional budgetary resources, technical assistance missions have been uniformly guided by the principle of designing measures that would generate adequate revenue to meet the countries’ budgetary needs in as economically neutral a manner as possible. As circumstances warrant, however, they have also suggested interim measures that deviate in varying degrees from the long-term goals of tax reform.

In general, IMF tax policy advice has had a discernible impact on the course of tax reform in many countries. By type of tax, reform recommendations on the domestic consumption and international trade taxes have met the greatest success in terms of the extent to which they have subsequently been implemented. By geographical location, technical assistance advice has had the greatest influence on Western Hemisphere economies and selected economies in transition. As expected, timely and concrete policy actions have materialized more frequently when technical assistance advice was given in the context of IMF-supported programs. But even when IMF resources were not involved, the missions’ analytical work supporting their recommendations often provided significant assistance to authorities in their policy deliberations. Stotsky carries out a review of IMF tax policy advice.


  • Ballard, Charles L., John B. Shoven, and John Whalley (1985a), “General Equilibrium Computations of the Marginal Welfare Costs of Taxes in the United States,” American Economic Review, Vol. 75 (March 1985), pp. 12838.

    • Search Google Scholar
    • Export Citation
  • Ballard, Charles L., John B. Shoven, and John Whalley (1985b), “The Total Welfare Cost of the United States Tax System: A General Equilibrium Approach,” National Tax Journal, Vol. 38 (June 1985), pp. 12540.

    • Search Google Scholar
    • Export Citation
  • Gordon, Roger H., “Fiscal Policy During the Transition in Eastern Europe,” Paper Presented at the Conference on the Transition in Eastern Europe, National Bureau of Economic Research, February 2629, 1992(unpublished).

    • Search Google Scholar
    • Export Citation
  • Hansson, Ingemar, and Charles Stuart, “The Welfare Costs of Deficit Finance,” Economic Inquiry, Vol. 25 (July 1987), pp. 47996.

  • Harberger, Arnold C., “The Incidence of the Corporation Income Tax,” Journal of Political Economy, Vol. 70 (June 1962), pp. 21540.

    • Search Google Scholar
    • Export Citation
  • Harberger, Arnold C., Taxation and Welfare (Boston: Little, Brown, 1974).

  • Hicks, John Richard, Sir, Value and Capital; An Inquiry into Some Fundamental Principles of Economic Theory (Oxford: Clarendon Press, 2nd ed., 1946).

    • Search Google Scholar
    • Export Citation
  • Hicks, John Richard, Sir, A Revision of Demand Theory (Oxford: Clarendon Press, 1956).

  • International Monetary Fund, Technical Assistance on Tax Policy: A Review, IMF Working Paper No. WP/93/65 (Washington: International Monetary Fund, 1993).

    • Search Google Scholar
    • Export Citation
  • Kopits, George, “Hungary: A Case of Gradual Fiscal Reform,” in Transition to Market: Studies in Fiscal Reform, ed. by Vito Tanzi (Washington: International Monetary Fund, 1993), pp. 6590.

    • Search Google Scholar
    • Export Citation
  • Musgrave, Richard A., K. Case, H. Leonard, “The Distribution of 0,” Public Finance Quarterly, Vol. 2 (July 1974), pp. 259311.

  • Shome, Parthasarathi, “Corporate Tax Reform and Capital Formation,” Proceedings of the 72nd Annual Conference, National Tax Association—Tax Institute of America (1979), pp. 14146.

    • Search Google Scholar
    • Export Citation
  • Simons, Henry Calvert, Personal Income Taxation: The Definition of Income as a Problem of Fiscal Policy (Chicago: University of Chicago Press, 1938).

    • Search Google Scholar
    • Export Citation
  • Tait, Alan A., “A Not-So Simple Alternative to the Income Tax for Socialist Economies in Transition: A Comment on Mc Lure,” World Bank Research Observer, Vol. 7 (July 1992), pp. 23948.

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, Tax Reform in Economies in Transition? A Brief Introduction to the Main Issues, IMF Working Paper No. WP/91/23 (Washington: International Monetary Fund, 1991).

    • Search Google Scholar
    • Export Citation
  • Xu, Bin, Tax Policy Implications in Endogenous Growth Models, IMF Working Paper No. WP/94/38 (Washington: International Monetary Fund, 1994).

    • Search Google Scholar
    • Export Citation

See Xu (1994) for a survey of the literature on tax policy in the context of endogenous growth models.


Changes in the after-tax interest rate, prompted by capital income taxation, could have important dynamic effects and savings could not be represented as a stable function of the contemporaneous return on capital. See Tanzi (1991), for example.


This could also be cast as a debate between the efficacy of taxing personal cash flow versus taxing personal income.


The sales tax seems to be left in operation at the provincial level, while the VAT operates often at the federal level.


The inclusion of net wealth lends the phrase “comprehensive” to the definition.


Having introduced above the concept of cascading in the context of consumption taxation, note that double taxation under the income tax could be thought of as a form of cascading.


The imputation of a deductible interest charge to equity financing has been tried in, at most, a handful of countries, including Egypt. The corporate cash-flow tax, however, has not been tried.


A foreign enterprise that operates in a host country but does not incorporate is a “branch.” A “subsidiary” operates under incorporation. Both forms may set up a “permanent establishment” of operation and thus come within the purview of taxation of the host country.


Argentina has just repealed it.


Also, see Gordon (1992) on Poland, and Kopits (1993) on Hungary.