David C.L. Nellor

Taxation of Mineral and Petroleum Resources

David C.L. Nellor

  • What factors make taxation of mining and petroleum enterprises different from taxation of other enterprises?

  • How does the choice of tax instruments influence the development of a country’s mineral and petroleum resources?

  • What are the different types of tax instruments that can be used in the mining and petroleum sector? What is role of each of these instruments?

Many developing countries and some economies in transition depend heavily on mineral and petroleum extraction for fiscal revenue and foreign exchange. Because these mineral and petroleum projects are often enclave activities with few direct links to the domestic economy, the public sector must be the principal agent for translating resource production into wider economic benefits.

The government has two fiscal roles with respect to the natural resources sector: it is the sovereign tax power and the resource owner. As the sovereign tax power, the government has the responsibility to ensure that the resource sector makes its due contribution to public revenues in the same manner as other industries. As the resource owner, the government must determine when to exploit its natural resources as well as ensure that it gets an appropriate price for its resources and distributes the benefits of resource exploitation so as to promote sustainable economic growth and intergenerational benefits.

Factors Shaping the Choice of Fiscal Instruments

At one level, there is a fundamental conflict between resource companies and the government over the division of the risk and reward of resource development. Both parties want to maximize rewards and shift as much risk as possible to the other party. At another level, resource agreements and the associated fiscal rules are a means of creating an identity of interest between the resource company and the government. The magnitude of revenues to be divided is maximized by designing fiscal arrangements that encourage a stable fiscal environment and efficient resource development.

Most mining and petroleum agreements are written for periods of 10 to 30 years, with the aim of defining a stable relationship between the investors (often multi-national companies) and the government. One reason resource projects are developed under long-term agreements is that the balance of power shifts over the life of a project. Before exploration begins or in its early stages, the power is with the resource companies because there is worldwide competition to attract potential investors. The power then shifts to the government, and political pressure for renegotiating the original agreement can become almost irresistible once a successful project has come on stream and is generating significant positive net cash flows. At the tail end of the project, when the resource deposit is almost depleted, the balance of power shifts again. The resource company can walk away from the project should it conclude that the government is making excessive demands.

Given the shifting balance of power over the life of a project, it is difficult to achieve an identity of interests. Further, both the division of the rewards from natural resource exploitation and the relative weights assigned to various fiscal instruments involve political judgment. A unique best policy cannot be proposed—there are unavoidable trade-offs between revenue, risk, and timing of the receipt of revenue. But, it is likely that multiple fiscal instruments will be needed to protect the interests of both parties over the life of the agreement. Product-based instruments can ensure that the government receives at least a minimum payment for the exploitation of natural resources of the country. Profit-based instruments reduce the likelihood of unplanned changes in resource contracts because they mean that the government shares in the returns from projects that turn out to be more profitable than expected.

Fiscal Instruments

This section reviews the instruments available to government both in its role as sovereign tax power and as owner of natural resources.

Sovereign tax power

Income taxation. The income tax is best suited for meeting the objectives of the government’s general tax power and should be levied on all resource and nonresource companies. It would be levied on resource sector companies whether or not the government owned the resources in the ground. Consequently, the decision by some governments to design income tax provisions specific to resource projects—such as accelerated capital deductions—is inappropriate. Rather, specific resource sector issues should be addressed by changing the price the government is charging for the use of its resource wealth. This price is levied by fiscal instruments, such as royalties, that are discussed below.

Similar to income taxation of other industries, income taxation of the resource sector involves matching of income and expenses. Most important, expenditures that produce a benefit over more than one accounting period should be capitalized and written off over their “useful life.” This ensures a rough matching of income with the expenses necessary to produce that income. Depletion of natural resources is simply a special case of capital recovery. Nevertheless, the particular features of the mineral and petroleum sectors mean that some income tax issues are more important than in other industries. Owing to the large initial capital outlays incurred in exploration and development of resource projects, defining capital deductions and the permissible debt equity ratio are important to maintaining the tax base and, in the latter case, to avoiding earnings stripping through artificially high debt-equity ratios.

The involvement of multinational companies increases the likelihood of transfer pricing as does the incentive created by the differential tax treatment of resource and nonresource companies within a country. Countries require provisions in their tax law enabling a price adjustment to be made where under- or over-pricing between associates has resulted in a lowering of taxable profit. To enforce such provisions, tax returns, at a minimum, should request details of domestic or international transactions with related parties.

In addition, an income tax is important in sectors, such as resources, in which foreign investment plays an important role. Many countries tax the worldwide income of their companies and allow a foreign tax credit. Investors from these countries—including Japan, the United Kingdom, and the United States—will want an income tax so as to have a creditable tax in their home country.

Import duties. Mineral and petroleum companies should be subject to import duties like other companies. Import duties are an mixed of the general tax powers of government rather than an instrument to secure a return on resource ownership. Thus, as a general rule, the tariffs applied to the resources sector should be those generally applicable in the economy. Resources sector companies rely heavily on imported capital equipment and intermediate inputs for their exploration, development, and operational activities. This makes import duties an important, timely, and relatively stable source of government revenue from the resources sector.

The government as resource owner

The fiscal arrangements with respect to natural resources need to take into account that the government is the landowner or the owner of mineral rights. If a valuable resource is going to be extracted, the government should receive a payment for this resource, separate from the regular income tax. The choice among fiscal instruments hinges on the timing of revenue, ease of administration, and risk-sharing. Hidden costs (or implicit taxes) such as mandated construction of schools, medical clinics, local roads, training, and localization requirements reduce what companies are willing or able to pay in direct taxes.

Traditional efficiency considerations would likely lead to the choice of one fiscal instrument that does not distort investment and production decisions or at least minimizes those distortions. There are, however, broader considerations of efficiency in reality. For example, the impact on investment decisions of the risk that future governments may change contractually agreed upon fiscal rules or even nationalize a mine. It may be in the interests of both parties (and efficiency) to design a combination of fiscal instruments to reduce this possibility.

The fiscal instruments chosen will influence the amount that the investor is willing to pay for the right to extract the resource. An auction of a mining lease, for example, provides revenue immediately, is efficient in a traditional sense, but may yield little revenue if there is a risk that the fiscal rules will be changed once the mining activity has commenced. Thus, the government may be faced with a trade-off if it values both the level of revenue and the receipt of revenues sooner rather than later.

The government also has an intertemporal production decision to make—it must assess whether resources should be exploited today or at some point in the future—this analysis of the opportunity cost of extraction should help it to define the minimum return that it should accept from an investor.

Finally, some governments want to extend the scope of resource taxes downstream to various manufacturing activities, such as refining or liquefaction of gas. At least conceptually, the analysis of mineral and petroleum taxation based on resource ownership is restricted to taxing the resource at the well or minehead.

Lease bonuses. Lease bonuses are up-front payments that could be determined by auction or at the government’s discretion. These payments are generally easy to administer. They mean that the investor bears the risk that the project will not be commercially viable because the return to government is fixed.

Reliance on a lease bonus offers little incentive for future governments to abide by the terms of natural resource lease arrangements. This raises a further mixed of risk—risk that a subsequent government will change either the fiscal arrangements or nationalize the mine or petroleum project. Thus, even if there was perfect foresight concerning the ex ante profitability of a prospect, an auction for a lease may yield little return to the government.

Royalties. Royalties are levied either on the volume or on the value of resources extracted. Royalties secure revenues as soon as production commences, are considerably easier to administer than most other fiscal instruments, and ensure that a minimum payment is made by the companies for the resources that they extract.

The conventional advice in the literature is to discourage the use of royalties. Proponents of this view note that royalties raise the marginal cost of extracting a resource and this may discourage development of otherwise marginal projects. In projects that do proceed, lower quality ores are left undeveloped. These propositions have been used to justify the use of profit-based fiscal instruments that are viewed as less distorting of investment and production decisions than royalties.

An alternative view is that a royalty is the price for natural resource extraction and, as such, is not necessarily distorting. The royalty, as a price for resource extraction, serves a role in determining whether investment should or should not proceed. The governmentowned ore or petroleum should be left in the ground for future development if companies are not prepared to meet this price reflecting the government’s opportunity cost of resource extraction. The government should determine what minimum payment it is willing to accept for the resource recognizing that it has given up its capital (i.e., the resource in the ground) once the resource is extracted. There is no reason to provide the resource to companies for nothing—however, under some “neutral” profit-based fiscal arrangements, the government is at risk of receiving little or nothing from resource ownership.

The case for use of royalties is reinforced by the significant administrative and monitoring advantages of royalties over other fiscal instruments. The royalty should be based on a transparent price formula agreed upon as part of the mining or petroleum agreement. Conceptually, the commodity price on which the royalty is based should be the mine or wellhead price, and the maximum royalty that could be imposed is then defined as the difference between the wellhead price and the cost of extraction. In some countries, the price used for determining the royalty is the export f.o.b. price. An overriding concern should be the use of an observable price and this could necessitate using the downstream price. In such cases, the rate of royalty would need to be adjusted in a simple and predictable way to reflect extraction and other intermediate costs. Royalties should be deductible for purposes of determining income tax liability because they are a cost of production.

Royalties can vary across projects depending on government perception of profitability (royalty rates could be auctioned) and higher royalty rates could be triggered by higher commodity prices according to an agreed formula based on transparent market prices for the commodity.

Resource rent tax. A resource rent tax (RRT) is similar to a cash-flow tax but is imposed only if the accumulated cash flow is positive. The net negative cash flow is accumulated at an interest rate that, in theory, is equal to the company’s cost of capital or discount rate. An RRT takes a share of returns once this rate of return has been earned by the company.

As envisaged by its designers, the RRT efficiently captures a share of natural-resource rent which is the return over and above the company’s cost of capital. Because the RRT only shares in returns in excess of the company’s opportunity cost of capital, it does not distort investment and is thus viewed as a superior fiscal instrument to royalties. Another advantage of the RRT is that it cannot incur losses for the government unlike other fiscal instruments, such as cash-flow taxes, or equity which can yield similar returns to the RRT. Further, the RRT may enhance contract stability because it automatically provides additional revenue in highly profitable projects.

Contrary to its theoretical attractiveness, the RRT can discourage exploration in practice. The RRT cannot be neutral with respect to the exploration decision because investors know that they will be taxed on highly successful projects whereas unsuccessful projects will be unaffected. Consequently, the company’s expected return from exploration is reduced by the RRT, and this distorts exploration decisions.

Also, excessive capital investment or a reduced rate of production will be encouraged if the RRT accumulation rate is set above the company’s discount rate, which will vary from company to company and can never be known with certainty. For example, assume a company’s discount rate is 15 percent and the RRT accumulation rate is 20 percent. Absent the RRT, the company would just be willing to invest 1 million today if it received a payback of 1.15 million a year from now. This investment would not be marginal if the company expects to be subject to the RRT because for RRT purposes, the 1 million outlay this year would be uplifted to 1.2 million next year, giving the company a 0.05 million loss that will reduce RRT taxable income in the future, providing an unintended tax benefit. In fact, if the accumulation rate is set too high, companies will have an incentive to stretch out development of a project.

The RRT is a high-risk measure for the government gaining a return on resource ownership; although revenue could be sizable in favorable circumstances, there is also a significant chance that resource development will yield little revenue. The RRT only provides a return to government on those projects yielding above normal rates of return. It is possible that the project may be seen to be earning high positive net cash flows but yield no revenue, creating political pressure for revision of the resource contract. Further, because the investor receives the threshold return before the government receives any revenue, the revenue stream, if any, is “back-ended.”

In summary, the RRT can play a role of capturing rents not collected by royalties and of enhancing contract stability by improving revenue buoyancy in relation to highly profitable projects, but it should not be relied on as the major fiscal instrument from which a return on resource ownership is gained.

Government equity. Government equity in mineral and petroleum projects is an important political symbol in many countries. Government equity gives a sense of participating in the development of the country. Beyond these arguments, however, there is a compelling case for the government not taking an equity interest in mineral and petroleum projects. Nevertheless, should the government decide to take an equity position in mineral and petroleum projects, it should use a carried interest.

There are a number of costs associated with public ownership. First, when the government takes an ownership position, it exposes itself to risk. At the time the government is required to exercise its equity option, it can never be known with certainty whether it is making a good investment. Though it may appear that a particular project will be highly successful, unexpected events, such as a fall in mineral prices, can turn a promising equity investment into a significant government liability. Second, taxation is more likely to maximize government revenue flow than an equity interest that looks to dividends that may never be paid. Third, equity requires the government to divert funds that otherwise could finance priority development projects. Moreover, a government equity interest could weaken the country’s external position. If the government borrows externally to pay for its equity interest, there will be years when the government is required to pay interest on its indebtedness even though it received no dividends from its investment. Fourth, there can be a conflict between the government’s role as a shareholder (or joint venturer) and its role as a regulator. As a shareholder, the government will want to maximize its return from its investment. As a regulator, the government will want to ensure that the mining project fully complies with all government regulations.

Fiscal instruments can be designed that yield the government the same return as equity but that are preferable because they eliminate some of the potential costs of equity. An RRT earns the government the same present value return as an equity interest that is purchased for cash, assuming the revenue streams are discounted at the RRT accumulation rate. But, the time profile and risk exposure of paid-up equity and RRT are quite different. With a paid-up equity stake, the government initially incurs substantial negative cash flows—its share of costs—and is subsequently compensated by revenues once production commences. In present value terms, the RRT yields the same revenue. Revenue, however, is only received once the project earns the accumulation rate of return. Should a project be unprofitable, the RRT prevents the government from incurring a loss whereas losses can be incurred when the government holds paid-up equity.

Should the government decide to take an equity interest, it should use a carried interest.1 This form of gaining equity is less risky than a working interest which requires up-front cash. An RRT, at a 35 percent rate for example, is equivalent to the government having a 35 percent carried interest if the accumulation rate for the RRT is the same as the interest rate charged for the carried interest. For example, assume the accumulation rate and interest rate are both 20 percent. In the case of the RRT, the government receives nothing until the project earns 20 percent. The government then receives 35 percent of additional profits. In the case of the carried interest, the government receives nothing until the project earns enough for the carried interest to be paid off. If the interest rate on the carry is 20 percent, this will not occur until the project earns 20 percent. Thus, an RRT provides the government all the benefits of a carried interest without the downside of ownership that the government would be exposed to after the carried interest crystallized.

Production sharing can be viewed as another form of government equity. In theory, the government and the private investors are partners. The government contributes capital to the project in the form of the ore body while the private investors contribute the exploration and development costs and operate the project. The government and the private investors agree to share production from the project, though the government often can require the private investors to market its share of the product. Production-sharing arrangements can take many forms and often are quite difficult to monitor and administer as the arrangements are complex and the parties can disagree on just how the arrangement should be interpreted. In a simple production-sharing arrangement, the government and the private investors only share production after the investors have recovered the original exploration costs, development costs, and operating costs in the form of product. A production-sharing agreement along these lines is essentially equivalent to the government having a carried interest, and thus is less risky than a working interest which requires the government to purchase its equity. Some production-sharing agreements limit the cost recovery in any one year to 30 or 40 percent of production, thus ensuring that the government receives some share production when the project first begins to produce.


The fiscal arrangements for resource projects involve political judgments regarding the trade-off between factors such as revenues, risk, and timing of revenues. A fiscal regime that is less reliant on income taxation and more on royalties will generate a relatively more stable and timely revenue stream. Thus, in many developing countries and economies in transition, the fiscal regime should comprise a broad range of instruments with emphasis on current revenue, lowering government risk exposure, and reducing tax and other administrative burdens. The following issues should be considered in establishing the fiscal regime for mineral and petroleum projects:

• Resource contract stability is likely to be enhanced by use of a variety of fiscal instruments.

• Mining and petroleum projects should be subject to the income tax like other activities in the economy. The various income tax provisions must be designed carefully particularly in relation to capital deductions, permissible debt-equity ratios, and transfer pricing. Transfer pricing is not only a problem at the international level but also domestically because of the different tax treatment of the resource versus nonresource sectors.

• Import duties can play an important role in providing revenue early in the life of a project because of the importance of imported capital equipment. In many countries, capital equipment is often exempt from duties so if revenue is a primary objective, a minimum tariff on capital equipment could be recommended.

• Royalties should play an important role—the rate of the royalty cannot be prescribed as a general rule—but will depend on perceptions of profitability and other aspects of the fiscal package. The rates of royalty may vary across mining leases and they could have stepped rates triggered by higher prices. A transparent price should be used for determining the royalty liability.

• A resource rent tax may be used as one mixed of the resource sector fiscal regime but should not be relied upon as the major part of the fiscal package.

• If government equity is to be used, consideration should be given to using a carried rather than a working interest.

Taxation of the Financial Sector

Janet Stotsky

  • What special issues does the financial sector raise under an income tax?

  • Why are financial services typically exempted from the VAT?

A well-functioning financial system is critical to a modern economic system. It is essential to directing capital to its best uses and to financing business and consumer activities. Financial institutions encompass a wide range of institutions, including commercial and savings banks, credit unions, insurance companies, pension funds, and brokerages. The taxation of financial institutions is in many respects similar to the taxation of other business sectors; nevertheless, it poses specific problems that require separate consideration. In some developing countries, financial institutions may be secondary to the unorganized financial sector. This sector includes moneylenders, cooperative and trade credit, pawnshops, and other arrangements, and it typically escapes taxation. This chapter first provides an overview of the types of financial institutions in the organized financial sector and the services they provide. It then examines issues relevant to income taxation and consumption taxation of financial institutions. Although financial institutions are also typically subject to a variety of franchise and stamp taxes, these are not discussed in this chapter.

Nature of Services

Financial institutions provide diverse services. They play a critical role in matching lenders and borrowers, by acting as an intermediary between them. Financial institutions generally earn a return on their services by charging a higher interest rate to borrowers than they pay to lenders. By aggregating the savings of many lenders and applying specialized knowledge and managerial skills, financial institutions can create financial securities that vary in several dimensions, including size, maturity, and riskiness. These securities are valuable to lenders and borrowers by satisfying their diverse preferences and needs. Thus, they serve the important role of channeling capital to its best uses and enhancing the efficiency of economic markets. Financial institutions could also play a more limited role by acting as a broker between participants in financial markets rather than by creating financial securities.

Another important function of financial institutions is to offer insurance against a large variety of risks. Insurance companies agree to accept obligations to pay for uncertain losses in return for the payment of fees or premiums. Since the payment of fees and compensation for losses is rarely simultaneous, insurance companies accumulate reserves to meet contingencies. In addition, some kinds of insurance companies, typically those providing life insurance, play an explicit role in saving. Policyholders accumulate money in their insurance policies; insurers invest this money, providing both policyholders with insurance and an explicit financial return on their investments.

Financial institutions also offer a wide variety of financial management services and consulting. Pension funds are an important part of the financial system, with current workers (or their employers) contributing substantial sums to these funds to manage their money and provide them with income in their retirement or in the event that they become disabled. Financial management services encompass a range of specialized services that businesses and individuals need for managing their assets or obtaining credit.

Income Taxation

In principle, the income of financial institutions should be measured the same way as the income of other businesses. Although financial institutions are subject to the usual corporate income tax, their income often receives special treatment in several respects.2

The timing of income

One issue that complicates the measurement of income of financial institutions relates to the timing of income. In general, businesses sell goods and services that involve no future obligation on their part, although, they typically make provisions for costs associated with warranties and bad debts. In contrast, financial institutions typically provide services that have an explicit time dimension. Banks, insurers, financial managers, and pensions provide services that may last over extended periods of time. Accounting for the possibility of future payments as a result of loans, insurance, and other financial activities is thus an integral part of the activities of financial institutions.

A corporate income tax typically applies to business income as it accrues. There has been vigorous debate over the years whether banks (and other lenders) should be given special deductions for bad debts.3 It is generally agreed that a bank should be allowed to deduct bad debt from income as it accrues. One controversial issue is the question of when bad debts accrue. Another controversial issue is whether the deductions are sufficient to compensate banks for these losses and whether banks deserve special treatment of reserves set aside for bad debts. An example will illustrate the nature of this problem. Consider a bank that makes loans to creditworthy borrowers at a rate of 10 percent. The bank’s income is measured as the 10 percent interest minus the costs of servicing the loans. Consider the same bank making loans to less creditworthy borrowers at a rate of 20 percent, with the 10 percent point differential compensating the bank for the greater probability that some of the less creditworthy borrowers will not repay their loans. Even if over some time, the bank would make the same return on its two loan portfolios, the pattern of income on the two portfolios is not likely to be the same since the bank receives more money early on from the riskier portfolio owing to the higher rate of interest. Suppose that in anticipation of expected future defaults, the bank holds some money in reserve for the riskier portfolio. If the bank is taxed on flows of income as they accrue, it only receives a deduction for the bad debt when they actually accrue. As a result, it earns more income initially on the riskier portfolio and hence pays more tax initially on the riskier portfolio. Even though the earnings may ultimately be the same on the two portfolios, in present value terms, the bank has paid more taxes on the riskier portfolio. For the bank to pay the same tax in present value terms on the two portfolios, banks could be allowed a deduction for the bad debt reserve at the time they allocate funds to this use. If the amount set in reserve evens out the flow of income relative to the less risky portfolio, the present value of taxes paid on the two portfolios would be the same, hence the bank would not be penalized for undertaking a riskier investment.

The main argument against allowing banks to deduct reserves for bad debt is that the ability to take these deductions could lead to abuses if banks attempt to shelter an excess amount of income in this fashion by overestimating potential losses, thus deferring income and reducing their taxes in present value terms. Thus, for a deduction for bad debt to be fair, the tax rules must allow a deduction that only evens out the flow of income. In light of the complexity and variety of loan portfolios a bank is likely to have, there would be many difficulties in administering fairly such a system.

There are two main alternatives under an income tax to allowing banks to take a deduction for bad debts. First, banks could receive no special treatment, taking deductions for bad debts when they are written off. One advantage of this method is that since expected future income is never taxed under an income tax, it offers equity in the treatment of expected future income and losses. Banks might wish to set aside money as reserves, but would receive no preferential tax treatment for this reserve. The main administrative issue in this case is determining when banks should be allowed to write off bad debts, which is a complicated issue. Without clear rules, banks can choose which year to recognize the loss. Second, banks could purchase insurance against bad debts on loans made on new loans. The costs of the insurance would be deductible against income at the time of purchase, thus income would be accurately measured. At the time the losses are realized, the deduction would be offset by insurance payments, resulting in no net tax liability. Similarly, banks could purchase insurance against bad debts on loans that are written off that year.

A similar problem arises with insurance companies since they must also set aside reserves for the purpose of meeting future payments on policies. Often, government regulations set required reserve levels for insurance companies to ensure that they are able to meet their financial obligations. These reserves may not bear an exact relation to the reserves that an insurance company would need to meet its obligations. Often, the reserves are a conservative estimate of insurance company needs.

Tax-favored investments

Income tax laws also provide tax preferences to financial institutions by allowing them to offer certain tax-favored investments that directly benefit investors by reducing their individual income taxes. These tax preferences may take several forms. Financial institutions may be able to offer some investments whose income is completely tax exempt while in other cases, the investment income is only tax deferred. Financial institutions may also be able to offer deposits or bonds whose interest income is tax exempt. Life insurance companies may be able to offer life insurance policies whose cumulative investment income on the death of the policy holder is exempt from inheritance taxes. Financial institutions may also be able to offer retirement savings accounts on which the income is tax deferred until it is withdrawn after retirement. Similarly, insurance companies may be able to offer life insurance policies on which the investment income, referred to as “inside build-up,” is tax deferred. In addition, in some cases, investments may be deducted from taxable income at the time they are undertaken. Depositors in pension funds and other retirement accounts are often able to exclude from tax their contributions to these funds. Often the investment income is not taxed until the income is withdrawn many years later, offering a significant advantage over taxable investments.

These tax preferences encourage investors to purchase tax-favored instruments, at the expense of other investments, which may be more productive. The application of the income tax rules to financial institutions is typically not uniform. Life insurance companies are frequently able to offer tax exempt or deferred investments, while these opportunities are more limited for banks. These tax preferences may thus give certain institutions an advantage in providing certain assets, leading to inefficiencies and inequities.

The intertwined nature of services

One problem arises from the intertwined nature of provision of investment services and other services by financial institutions. Taxpayers typically owe tax on investment income net of costs incurred in earning it while the costs of services not related to investment income are not deductible from income. It may be difficult for a financial institution to separate the two components of costs and hence, it may mix the costs of earning investment income with the costs of other services, leading to an understatement of taxable income. This tax advantage encourages taxpayers to purchase services from financial institutions that also perform investment services. This, however, could be largely a mirage. Thus, for instance, taxpayers may purchase “free” checking services because the cost of servicing the checking account is included in investment costs rather than paid explicitly, even though it is paid implicitly through forgone interest income.

Separating return on capital from return of capital

Another problem arises from the difficulties in separating return on capital from return of capital. At the corporate level, any income that accrues to creditors should not be taxable to the institution. At the individual level, only return on capital should be taxable while return of capital should not be taxable.4 The issues are somewhat different between banks and insurance companies. In the case of banks, their income is the difference between investment income and the costs associated with their operations. A deposit represents a liability to a bank and does not represent income. The withdrawal of a deposit represents a reduction of the liability and hence does not represent a cost. It is thus not that difficult to separate return on capital and return of capital in the case of banks. In the case of insurance companies, their income is derived from premium payments and investment income. Premium payments may represent both a payment for services and a “deposit” for investment. The only taxable part should be the payment for services, since the investment component must be returned. In practice, it may be difficult to separate these two components of premium payments.

A further complication arises when separating return on capital from return of capital, if income is allowed to accumulate without tax. If income is withdrawn from a tax deferred account, then it must either be treated as the investor’s return of capital or return on capital. If it is treated as return of capital, then it should be untaxed and if it is treated as return on capital, then it should be taxed. This requires a set of rules which classifies withdrawals as return of or on capital.

Other issues

There are many other issues that arise in the taxation of income of financial institutions, including issues related to income from currency fluctuations or to transfer pricing (see Chapter IV). Since it is often difficult to determine with financial services where the taxable transaction has taken place, apportioning income geographically may also present thorny tax issues. These issues also arise in the context of general corporate taxation.5

Consumption Taxation

There are several different commonly used consumption taxes. A widely applied consumption tax today is the value-added tax (VAT). The treatment of financial institutions under a VAT is a difficult issue.6 This discussion focuses on the treatment of financial institutions under a consumption-type VAT based on the invoice method and destination principle, since this is the most common form of VAT. Under a well-designed tax system, a VAT would apply to all forms of consumption, including financial services. It is, however, difficult in practice to apply a VAT to financial services primarily because of the difficulty in measuring value added associated with financial services.


In principle, it is possible to measure value added in the banking sector by adding profits, wages, rent, and interest or, alternatively, by taking the difference between investment income and the cost of funds (interest expense plus the cost of equity financing) and other costs of the bank. The application of the invoice system, however, requires that the VAT liability be attributed to each transaction. This is not possible in the banking sector because most financial services provided by banks do not have specific charges attached to them. Instead, charges for services result from differences in interest rates charged to borrowers and those paid to lenders. Even charges for some services, such as checking account activities, that could be separated from financial intermediation activities are often reflected in interest rates.

Insurance companies

With respect to insurance companies that provide casualty insurance (and other forms of noninvestment insurance), value added is measured by the loading charge, essentially the 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 re-distribution from one policyholder to another (e.g., when one policyholder makes a claim, there is a redistribution to that policyholder from other policyholders). For insurance with an investment component, the value added is again only properly measured by the loading charge, not the savings component. In either case, it is difficult to measure the loading charge, making it difficult to apply VAT to insurance activities.

Methods of taxing financial institutions under a VAT

There are three obvious ways of including financial institutions in a VAT. These institutions could be exempted, they could be zero-rated, or they could be made fully liable to VAT. Most countries with an invoice method VAT have chosen to exempt financial institutions. The advantage of exemption is that many financial services are provided to businesses that are taxable under a VAT, ensuring that these services are taxed, in effect, even if financial institutions are not subject to VAT. Since exemption does not allow firms to credit VAT paid on inputs, however, some cascading occurs with respect to financial services provided to businesses that are taxable under a VAT Exemption does allow services provided to households and businesses that are not taxable under a VAT to escape tax, although the inability to credit VAT paid on inputs results in some VAT burden. Exemption puts domestic financial institutions at a disadvantage relative to offshore institutions, if exports of financial services are zero-rated. Exemption may also encourage financial institutions to produce some intermediate goods themselves, rather than purchasing them, since they could not credit VAT on these purchases. Finally, if financial institutions are only partially exempt, this creates problems apportioning VAT paid on inputs to taxable and nontaxable items and could make the tax more vulnerable to tax evasion schemes.

The advantage of zero-rating is that it avoids many of the problems with exemption, but it has the disadvantage of generating less revenue and lowering the tax burden on financial services compared with other consumption activities.

The advantage of incorporating financial institutions into a VAT is that it enhances the tax base quite considerably and also results in an equal treatment of financial services and other business services. Nevertheless, taxing insurance poses additional administrative and conceptual problems in determining what proportion of insurance services provided to households is consumption and what proportion is investment.


Table VI. 1 summarizes the treatment of financial services under a VAT in OECD countries. All OECD countries exempt the intermediary functions of banks from VAT although a majority of them apply VAT to certain other activities, including check printing, rental of safe-deposit boxes, and foreign exchange transactions. All OECD countries exempt life insurance services fully from VAT while most countries also exempt fire and general insurance services. Finland and New Zealand subject these types of insurance to the standard VAT. Although financial services are typically exempted from VAT, exports of financial services are often zero-rated to prevent a loss of international competitiveness. Other consumption taxes typically apply to certain financial services.

Table VI.1.

VAT Treatment of Financial Services in OECD Countries

(January 1987)

article image
Source: This table is taken from Table 13.11 of Messere (1993).

Generally includes dealings in money, shares, stocks, bonds, or other securities, lending money or advancing credit (or the arranging thereof), and the operation of any current, deposit, or savings accounts.

Financial advice, safe-deposit boxes, debt collection, and the keeping of securities.

Only if exported outside the European Community (EC).

Supplier can choose to have such services taxed where they are supplied to a trader for use in his business. The exemption also applies to the management of credit and credit guarantees, and the management and safekeeping of securities.

Safekeeping of securities is, however, exempt.

Applies only to insurance on boats.

Does not apply to credit and reinsurance.

Fire insurance only.

Life insurance brokers and agents are exempt.

Israel has been the only country to attempt to apply a VAT to a comprehensive base of financial services using an addition-type VAT. Its VAT was applied to the sum of payroll and profits of financial institutions. The tax did not allow an offset for VAT paid by financial institutions on their purchases nor did it allow purchasers of financial services to credit the VAT against the VAT due on their sales. This form of VAT on financial institutions proved to be very unpopular and was dropped. It was replaced with a separate consumption tax applied to the same base.

Developing countries that have adopted a similar form of VAT to the OECD countries have also typically exempted financial services from the VAT. A number of developing countries, however, use selective consumption taxes on particular activities, such as insurance premia, which still raises the issue of separating consumption from investment components.

Intergovernmental Fiscal Relations

John Norregaard

  • Which are the main principles guiding decentralization of the public sector?

  • How can subnational governments best be financed?

  • What is the role of grant and equalization schemes?

  • Will decentralization pose a problem for macroeconomic control and stabilization?

The question of how to design an efficient fiscal system of multilevel government is of substantial importance in most countries. Although some basic principles, guidelines and criteria may be established, however, no ideal system exists as witnessed by the immense variety of multilevel governmental systems actually in operation in different countries.7

This section summarizes in a nontechnical way the core issues which arise when a country is about to re-form or review the existing or to establish a new fiscal system of several tiers of government. In accordance with the basic nature of this issue, the emphasis is on structural problems, although macroeconomic aspects are also touched upon. The scope of this chapter is broad and encompasses mainly normative issues such as expenditure assignments, criteria for what constitutes a good local tax, and the need for and design of grants and equalization systems. This section is to emphasize the basic thesis that the tax issues involved must necessarily be seen in the context of broader structural issues if the resulting tax system is to work efficiently. Being of a structural nature, reforms of intergovernmental fiscal relations must necessarily be implemented with a medium- to long-term time horizon. It is acknowledged that many countries, in particular economies in transition, have serious short-term economic stabilization problems, the solution of which must precede the types of structural reforms dealt with here.

This section encompasses the design of intergovernmental fiscal relations in federal as well as in unitary countries. The section focuses on the general issues that are common to the two constitutional regimes, disregarding problems of design and implementation which are specific to federal countries. Issues of supranational fiscal relations are excluded (such as convergence criteria and tax harmonization within the EU). The treatment here also excludes issues and policies related to development problems aiming at helping underdeveloped regions, not because these policies are not important, but because they are of a different nature from issues of intergovernmental fiscal relations.

The Rationale for Decentralization

Following the Tiebout/Musgrave tradition, the basic economic rationale for decentralization is 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. Thus, through decentralization, a “market” is established for local public goods, which—by forcing consumers to reveal their preferences (by “voting with their feet”)—entails a welfare gain compared to a situation with only one homogeneous level of services throughout the economy. This outcome, of course, assumes that the cost of providing the services can be internalized and thus fully borne by the receivers of the benefits: in accordance with the so-called decentralization theorem advanced by Oates (1972), each public service should be provided by the jurisdiction having control over the minimum geographic area that would internalize the costs and benefits of such provision.

Basically, this mechanism is based on three propositions: (a) that the benefits of decentralized public goods are spatially limited; (b) that consumer preferences vary across individuals as regards private versus public goods as well as for different types of public goods; and (c) that consumers are mobile and, in fact, respond to what has become known as fiscal localization factors such as differences in local service levels and tax rates.8 To the extent that this is not the case, the fundamental economic efficiency rationale for decentralization disappears.

It has also been held that—even in cases where the economic conditions for decentralization may not be evident—it will promote local democracy and thereby contribute to democratization of society. Finally, decentralization has been held to enhance macroeconomic performance by reducing growth of public expenditures and by mobilizing local financial resources, although—generally—evidence about the correlation between the degree of decentralization and the growth rate of public expenditure is mixed.9

The Functions of Lower Levels of Governments

From this general starting point, the first question that arises is what kind of tasks should be assigned to subordinate levels of governments (“local governments” in what follows) and which should be retained at the central level. The answer is closely related to the fact that regions and localities typically are very open economies.10 Therefore, following Musgrave’s classification of public tasks:

• Distributional policies should be assigned only to the central government because people bearing the burden of distributional policies might migrate and thus render local distributional policies inefficient;

• Likewise, local stabilization policies will be inefficient because of spillover effects and because local governments typically do not posses the necessary arsenal of policy measures (e.g., monetary policy instruments);

• The allocative function should be the core function of local governments, that is, their prime task is to provide public goods and services to the local population in accordance with the preferences of this population and with the financing burden as far as possible internalized and thus borne by the beneficiaries of the services (although, of course, a number of important allocative functions owing to their “nonlocal” nature will remain with the central government).

According to this very schematic picture, one of the main guiding criteria for the design of multilevel fiscal systems has been expressed by the term accountability which is to be interpreted in a much broader sense than usual (i.e., that local politicians should be held responsible for their decisions to local constituencies), namely that local governments should provide services, the benefits of which accrue to the local population which should also bear the major part of the economic burden associated with the provision of the services.

Looking at actual policies, however, it is evident that countries for varying reasons in practice assign tasks to lower levels of government which clearly have distributional- or stabilization-related objectives or both. This may in part be to exploit local expertise on, for example, local labor markets, but partly because in practice, it may be extremely difficult to distinguish between tasks which are clearly of an allocative nature and those which are not. Also, in some countries, central governments have “pushed downwards” stabilization measures in order to relieve financial pressures on the central government.

Closely linked to this discussion is the crucial question of how the different tasks of lower levels of governments—once identified and defined—should be assigned to different lower levels of government. This is probably one of the key issues in the policy discussion on this topic in most countries, and it also constitutes an area where lack of a clearly defined and communicated policy has led to inefficient systems in a number of countries. In other words, it is extremely important that the assignment of tasks to different levels of government is very carefully thought out, and once the allocation of tasks has been decided upon, that it be kept as stable as possible.

Following the principle of benefit areas indicated above, the theory of local finance “predicts” an optimal structure of local governments which consists of a multitude of layers with different “localities” corresponding to the different services provided. This model of “organized chaos” is seen by some as the actual situation in the United States. Again, this is an area where different solutions have been chosen by different countries, with perhaps the U.S. situation as the one extreme and with a much more rigid or “purist” approach with very clearly defined levels and allocation of tasks chosen, for example, by the Scandinavian countries. The French system with voluntary cooperation between a number of local governments in specific areas may be seen as an intermediate solution. The question of the optimal structure of lower levels of government raises many and very complex problems; among these problems are the optimal size of local governments, the number of levels of government, and the flexibility of the system with respect to borders and cooperation between entities at equal and different levels. Although literature has something to say about some of these issues, it is fairly safe to state that, in practice, the out-come is very much influenced by historical factors and current administrative procedures and capabilities.

Again, it is characteristic that different countries have chosen very different solutions, but experience strongly suggests that it is important to get the basic structure right, and that, once it is chosen, it may be difficult to change.

Financing of Lower Levels of Government

One of the reasons why the importance of structural issues has been emphasized so strongly is that, however efficient and well thought out is the financing system, it can never overcome systemic flaws in the basic structure of the different tiers of government.

Once these basic structural issues have been decided upon, the question of how to design an efficient financing system arises, including the question of the role of taxation. The guiding principle of accountability heavily supported by practical experience suggests that a prerequisite for an effectively functioning system of local government finance is that local governments be given at least one substantial revenue source over which they can decide and the burden of which cannot be exported to taxpayers outside the locality. In addition, the structure of financial sources has to be chosen. The general fiscal policy issues in this regard are (a) what should be the predominant source or sources of finance for local governments over which they can decide; (b) should they be able to decide on the tax base as well as on rates; (c) what should be the relative weight of other sources of finance; and (d) what should be the weight of central government grants and borrowing.

In most countries, lower levels of government have access to four main types of finance, namely, user charges, taxes, grants, and borrowing. There are many theoretical, conceptual, and practical problems related to the use of user charges as well as to borrowing. Most observers, however, seem to agree that user charges should be used wherever possible, that is, when local government provides marketable services, and that the charges should reflect current costs and the use of capital. As far as borrowing is concerned, most countries regulate heavily the use of this source of finance, which is typically allowed only for the financing of local government investment (and in many cases, only for specific purposes), in order to prevent fiscal indiscipline at the local level, and to reduce pressures on capital markets.

The principle of accountability is particularly difficult to apply in practical policies with regard to taxation and grant finance. The following observations are, therefore, focused on these areas. At the outset, it should be emphasized that taxation and grant-and-equalization systems necessarily must be seen as interrelated: without efficient equalization measures, even a well-designed tax system will not be able to achieve the objective of maximum accountability. There is almost universal agreement that local government taxes as far as possible should satisfy the same requirements as taxes in general, particularly the following requirements:

• Local taxes should—for a given level of revenue—be as neutral as possible and interfere as little as possible with the behavior of consumers and producers.

• They should be easy to administer.

• They should provide local governments with sufficient and stable revenue, given other available sources of finance.

• They should be perceived as fair.

• Finally, they should be as transparent as possible.

This last point is perhaps of special importance for local taxes, and touches upon the following additional criterion which is particular to local taxes:

• Local taxes should as far as possible be internalized, that is, be borne by the local population to which the benefits of the services provided accrue.

A good local tax, that is, a tax which among other criteria satisfies the accountability criteria, is one where local governments are not able to export the tax burden to the citizens of other localities and where the tax base is the least mobile. There is a consensus in the literature that land and property taxes are better than other taxes to satisfy these ideal requirements, though many local government experts also recommend local income taxes as a necessary or possible revenue raiser which to a sufficient degree satisfies the requirements for a good local tax. There is some disagreement as to the extent to which local governments should be allowed to make use of taxes on mobile factors/mobile tax bases, such as capital income taxes (e.g., corporate profits taxes) and consumption taxes (such as the VAT).11 This discussion coincides with the parallel discussion on the potential existence of beneficial impacts of tax competition between localities or regions and the need for tax harmonization, an issue which is not yet finally resolved.

To illustrate the very different tax structures chosen by different countries, Table VI.2 provides for selected developed and developing countries information about the allocation of the main taxes between different levels of government.

Table VI.2.

The Percentage Allocation of Tax Revenue Between Different Levels of Government

article image
Source: IMF, Government Finance Statistics.Note: na means not applicable.

Includes supernational authorities share of general government total tax revenue for Belgium (1.5 percent), France (0.7 percent), Germany (0.9 percent), the Netherlands (11.4 percent), and the United Kingdom (1.2 percent).

Data for general government do not include local government.

As regards the question of whether local governments should be allowed to determine the base as well as the tax rates, the general principles outlined at the outset seem to entail that only the tax rates should be allowed as policy parameters. The main argument for this limitation is that determining the base involves distributional considerations and consequences, but also that differences in the base across localities and regions would reduce transparency and thus accountability.12

Looking at actual tax structures in different countries, it is—once again—striking how different are the policies that countries have chosen. Some countries follow the Anglo-Saxon tradition of heavy reliance on property taxation, whereas other countries (e.g., the Scandinavian countries) use a spectrum of different taxes, but with heavy reliance on local income taxes. A general experience seems to be that property taxes, although they constitute an important revenue raiser, cannot be expected to raise more than approximately 10 percent of total revenue; that is, in countries with heavy or sole reliance on this tax source, central government grants typically play a substantial role.

In many countries, the revenue accruing from property taxes has been decreasing as a percentage of total revenues. One of the explanations put forward for this development is the unpopularity of the tax owing to its visibility and because it is hard to avoid—that is exactly the reason why it is considered to be a good local tax. Also, heavier reliance on property taxes might involve very high and thus distortive tax rates. Another important reason for its decreasing revenue importance is the perception of it being an unfair tax because, in many countries, the system of property evaluation is strongly deficient. This includes long lags in value updates, implying the need of relatively high tax rates to raise a given amount of revenue, and with haphazardous consequences for the distribution of the tax burden and for effective tax rates across taxpayers. This emphasizes the particular importance of a precise measurement of the tax base associated with this tax source. It goes without saying that these problems of valuation may be particularly complex in developing countries and in economies in transition in which well-functioning free markets for property may be limited. In such cases, simpler systems must be applied (e.g., different forms of square meter taxes).

In any case, in designing a local tax system, governments will have to make difficult choices concerning tax base(s), reliefs, exemptions, rates, etc. Practical experience suggests that the following important tax policy issues may arise: (a) to the extent that local governments should rely on local property taxes, to what extent should reliefs be provided for specific groups of taxpayers or specific sectors? (b) how can the tax base be measured in cases where no market prices exist? (c) to the extent that income taxes are preferred, to what extent should the base differ from that used by the central government? (d) to what extent should piggybacking be applied? (e) to what extent should local government taxes be based on tax-sharing arrangements with central government? and (f should local governments be given any influence on the determination of the tax share accruing to local governments?

Again, a wide variety of systems are applied in practice with respect to the use of different tax sources, differences in tax bases for the same type of taxes across levels of governments, and with respect to the degree to which tax-sharing arrangements are applied.

Tax Sharing and Grant Financing

On the question of tax sharing, the guidelines provided by the general principles outlined at the outset, in particular the principle of accountability, are in many respects similar to those guiding the use of government grants and are, therefore, dealt with together here (generally, these revenue sources represent lump-sum revenues on which local governments have no or only modest influence). In most OECD countries, the trend (or at least the political intentions) in recent decades has been to reduce the overall levels of grants to lower levels of governments and to convert remaining grants from specific to general grants, partly to improve local accountability and incentive structures, partly to improve central government finances.

Grants may be given for a variety of reasons, but generally they are provided for the following:

• To correct for vertical imbalances resulting from the assignment of expenditures and other revenue sources.

• To ensure that all local governments are able to provide a minimum of service, in general or in particular areas.

• To induce local governments to expand services in particular areas which are seen as beneficial by the central government (i.e., to correct for positive externalities, or to induce the provision of merit goods).

• To overcome lack of sufficient local revenue capacities (i.e., to correct for an insufficient local financing system), either in general or for underdeveloped or poor regions.

• To correct for unequal economic conditions in different localities or regions—or horizontal imbalances—due to differences in objective expenditure needs and fiscal capacities (normally referred to as equalization).

• To ensure a minimum level of redistribution in the system of taxation as a whole by raising more revenue via progressive central government taxes as opposed to the generally more proportional regional or local taxes.

Using the principle of accountability, two general statements may be made: first, unless grants clearly correct for externalities, their amounts should be minimized, and a correspondingly larger role be attributed to local governments’ own finance13 (which, in addition to the increased efficiency gain, may entail an improvement in the general government budgetary situation); second, a basic prerequisite for an efficiently working local income and/or property tax is that a minimum of equalization with respect to differences in expenditure needs and fiscal capabilities (e.g., as measured by the income tax base per capita) is taking place: the objective of equalization grants should be to ensure that each local government has the capacity to provide a standard level of service if it makes the same effort to raise revenue from its own sources and conducts its affairs with average efficiency. If this principle is followed, differences in average tax rates across localities should reflect differences in the politically determined service levels chosen by the individual governments concerned, and not external factors such as differences in the underlying cost of providing services arising out of, say, the age composition of the population, or differences in revenue capacities arising out of, for example, differences in the average levels of income in different local government regions.

Thus, to ensure accountability, the tax system in its interaction with the equalization system should as far as possible ensure that what might be called the “tax/service ratios” across localities are equal. It is, in this context, important to note that this criterion does not imply that tax rates should be equalized; it only implies that observed variations in tax rates reflect as closely as possible variations in service levels politically determined by local governments. Only if this is achieved will the efficiency gains of decentralization be fully realized, and only in this situation will local governments within a country be given reasonably equal opportunities, thereby enhancing the working of local democracy.

Also in this area, there is considerable variation across countries, some countries using only a minimum of equalization, typically achieved by fairly rough measures (e.g., per capita grants), whereas other countries apply sophisticated schemes including tax and expenditure needs equalization, with objective expenditure needs measured by econometric and other statistical methods.

Control of Local Government Expenditure

Looking at actual day-to-day policies toward subnational governments in, for example, most OECD countries as well as in many developing countries and countries in transition (such as China, Brazil, and the Russian Federation), it seems that the question of how to ensure—as an mixed of current stabilization policies—that aggregate expenditure levels of local governments are in conformity with macroeconomic policy objectives is a basic one. This issue is also a major source of disagreement between local and central governments.

Central governments may want to control local government expenditure first of all because of the important size of these expenditures in total domestic demand (in some countries, public consumption of lower levels of government is larger than that of the central government). Control instruments may take on many different forms, and may be enacted on the expenditure side (e.g., agreed or legislated growth ceilings) or on the revenue side (e.g., rate capping, tax rate ceilings). Second, as referred to earlier, most central governments control borrowing by lower levels of government which—in combination with limited revenue sources—force on local governments an indirect control of expenditures.

There are two schools of thought on this issue: one in favor of central government active intervention, the other against. The main argument against intervention is that, assuming that the system of local government finance is designed in accordance with the general principles described above, as given by the criterion of accountability, local governments will be “self-policing” in the sense that only expenditures demanded by local voters (and financed mainly by the same local populations) will be agreed upon, and intervention will therefore imply a loss of efficiency and welfare. In Musgrave’s terms, stabilization policies should be enacted on the basis of measures neutral with respect to allocation such as central government income taxes, and should not interfere with the (allocation) functions performed by local governments.

There seems to be some (albeit very limited) empirical evidence indicating that the stronger local governments and local democracy, and the more decentralized and self-financing the public sector, the slower the growth of public expenditures.14 In addition, in some countries (e.g., in Poland), decentralization of revenue and expenditure functions may be seen as an important way, on the one hand, to mobilize local financial resources, and on the other hand, to improve the budgetary position of central governments.

The main argument of the interventionist school seems to be that, generally, “self-policing” does not work, and that due to the relatively limited time horizon of local politicians and the existence of strong local interest groups, expenditure growth of local governments must necessarily be checked. These kinds of policies often go hand-in-hand with exporting of expenditure and revenue burdens from the central to local levels.

Also on this issue, a tremendous variation is observed in policy objectives and in the design of policy instruments across countries and over time.

Alternative Methods of Revenue Forecasting and Estimating

John R. King

  • What purposes are revenue forecasts and revenue estimates designed to serve?

  • What methodologies are available for forecasting revenues from different taxes?

  • How is the process of estimating the revenue effects of possible changes to the tax system related to revenue forecasting?

The Purposes of Revenue Forecasting and Estimating

Forecasts of government revenues from different taxes and nontax revenue sources are produced for a variety of purposes.

By far, the most important of these is government budgeting. In conjunction with expenditure estimates for the period covered by the budget, a forecast of total revenues is needed to indicate the prospective deficit that will have to be financed. Since the deficit is the difference between two much larger totals, accuracy in forecasting those totals is clearly of central importance: an error in one of them, other things being equal, will result in a much bigger proportional error in the fore-cast of the overall deficit or surplus.

It may be even more important, however, that the aggregate revenue forecast and the expenditure estimates be consistent with one another. For example, if both expenditures and revenues vary in rough proportion to changes in prices or wages in the economy during the budget period, errors in forecasting the underlying growth in wages and prices may have a relatively small effect on the deficit forecast—provided that both revenue and expenditure forecasts are based on the same assumptions. Similarly, errors in forecasting revenues from particular taxes may be relatively unimportant so long as those errors tend to offset one another, since it is the total from all sources that is most important in the budgeting context.

Closely related to their role in budget preparation is the use of tax revenue forecasts in monitoring budget outturns. To serve this purpose, revenue forecasts made at the time of the budget may be revised at several points during the budget period. In addition, Ministries of Finance often find it helpful to set operational revenue collection targets for the tax administration, based on forecasts of revenues from different taxes; and the central tax administration may set operational targets for local tax offices in a similar way.

“Revenue estimating” is the process of assessing the impact on revenues of tax law changes proposed at the time of the budget, or subsequently.15 It is a process closely related to revenue forecasting, but sufficiently different that in some countries (including the United States), it may be performed by different people. Fore-casts are required even when no change to the law is proposed; on the other hand, revenue estimates must often be made for proposals that are not subsequently adopted, and that therefore do not need to be taken into account in any revenue forecasts.

Revenue-Forecasting Methodologies

This section summarizes a number of different methodologies for revenue forecasting that are used in practice, in different contexts.16 A fundamental distinction is between revenue forecasts that are conditional on forecasts of other economic variables such as GDP, and those that are made unconditionally.


The most straightforward method of making an unconditional forecast of revenues from a particular tax is simply to extrapolate an established linear trend in receipts. More complex procedures using this general approach include the Box-Jenkins ARIMA procedure. When “univariate” procedures of this kind are used to derive a forecast of revenues from a particular tax i, in a particular period t, the revenue forecast Tti depends only on revenues observed in the past:

Tti=f(Tt-1i, Tt-2i,...)

Such a procedure does not make use of any knowledge the forecaster may have of the structure of the tax, and of probable relationships between the revenues that it will generate and other economic magnitudes. As a result, although univariate techniques are sometimes used as a standby when nothing better is available, they are not widely used by revenue forecasters.

Forecasting using elasticities

When a forecast of GDP (or GNP) is available, the simplest conditional approach to forecasting revenues from a particular tax (such as an income tax), or from a group of taxes, is to employ an estimate of the elasticity of revenue from the tax with respect to GDP, εi This elasticity is defined as:


where Y denotes GDP. Hence, if this elasticity may be assumed to be constant, a forecast of Ti in the forecast period may be derived straightforwardly from a forecast of Y in the same period, together with actual figures for both Ti and Y in some previous period.

To estimate the elasticity from time series observations on receipts from the tax, and on GDP, it is necessary to remove from the former time series the effects of any changes to the tax law that may have been made during the period. To do this, two main approaches have been adopted.

a. The most commonly used method makes use of the estimate that was made by the government, at the time of each change to the law, of what the revenue effect of that change would be. In such exercises, it is usually assumed that the change would have the same proportional effect on revenues in each subsequent year. A hypothetical time series Tti can then be constructed showing what revenues from the tax would have been, in past years for which data are available, if the tax law had been the same as it is in the current year. With sufficient past observations, a simple double-logarithmic regression of the hypothetical revenue series on GDP can then be used to derive a least-squares estimate of the elasticity from the equation:


where e is an error term that is assumed to be normally distributed about a mean of zero, and the estimated coefficient b corresponds to the elasticity εi. When fewer than about ten observations are available in the time series, cruder methods may have to be used.

b. An alternative approach may sometimes be possible, if there have been no more than one or two significant changes to the law over the period for which revenue observations are available. In this case, the elasticity may be estimated from a regression of the actual revenue series on GDP, in which “dummies” are also included as explanatory variables to capture the effects of those changes to the law:


where D is a (vector of) 0/1 dummy variables denoting different policy regimes.

An estimate of elasticity with respect to GDP can, in principle, be made for any tax. It is more natural, however, to relate certain taxes to other macroeconomic variables. For example, import duty revenues are likely to vary with the value of imports; revenues from a broad-based consumption tax such as VAT are likely to vary with aggregate private consumption expenditures; and so on. The elasticity of revenues from each tax, with respect to GDP, can then be seen as the product of two separate elasticities:

a. the elasticity of revenues with respect to the “tax base” (such as imports or private consumption); and

b. the elasticity of that tax base with respect to GDP.

Separate estimates of those two elasticities are useful, if forecasts of the components of GDP are already available to the revenue forecaster.

More complex macroeconomic models

The methods summarized above are based on the assumption that the relevant elasticities are constant. In principle, there is no reason why this should be a valid assumption. More generally, regression methods may be used to estimate functional relationships between revenues from particular taxes and a variety of macro-economic variables; these relationships need not be constrained to imply constant elasticities. Revenue equations for particular taxes were estimated along these lines in the macroeconomic models, which were developed for forecasting purposes in the major OECD countries from the late 1950s onwards.

An important advantage of this approach to revenue forecasting, using econometrically estimated tax revenue functions, is that the revenue forecasts are integrated with the corresponding macroeconomic forecasts. As a result, consistency between the two is guaranteed. There are, however, some serious limitations to the approach. First, many countries do not have sufficient (or sufficiently reliable) data from which detailed macroeconomic models, including tax revenue functions for individual taxes, can be estimated. Second, revenues from certain taxes (such as taxes on wealth, capital gains, and capital transfers) may not be closely related in practice to any variables that are included in conventional macroeconomic forecasting models.

Third and perhaps most important, the approach constrains the revenue forecast to depend on only a small number of macroeconomic variables—for instance, in the case of a corporate income tax, on forecasts of aggregate profits in different sectors, and relevant expenditures such as fixed investment and stock-building. Such macroeconomic relationships are likely to break down, however, if a substantial number of companies are subject to losses.

Structural models of individual taxes

Since the late 1960s, the major OECD economies (and individual states within the United States) have developed “microsimulation models” for their major taxes, particularly those on personal and corporate incomes.17 These models are constructed from samples of tax return data. Their focus is on the detailed application of the tax law to the structure of the tax base, at the level of individual taxpayer liabilities.

In most cases, the primary purpose of these models has been to assist in revenue estimating (as discussed in the next section). In many countries, however, they have also come to be used in the process of revenue forecasting. For this purpose, data in the sample are projected forward over the forecast period (on the basis of macroeconomic forecasts of the relevant variables); the microsimulation model is then used to estimate the tax liabilities that will arise; finally, adjustments are made for collection lags, to convert that estimate of liabilities arising into an estimate of tax receipts during the period.

Integrated forecasting systems

An obvious problem that arises in using such micro-simulation models as a basis for revenue forecasts is that of potential inconsistency between the macroeconomic forecasts (which are used as inputs in projecting the sample data over the forecast period) and the revenue forecasts themselves. In some countries (such as the United Kingdom), this has led to the development of iterative procedures, under which a macroeconomic forecast is first made and revenues are estimated using simple tax revenue functions (as in the section above); the microsimulation models are then used to adjust those revenue forecasts, using “residuals” in the macro tax revenue functions; and the process is repeated several times until convergence occurs.

With developments in computer technology, it should be possible to short-circuit these rather cumbersome procedures by linking the various models together in an integrated forecasting system. Such a system has recently been developed by private consultants for the U.S. State of Massachusetts,18 and similar systems seem likely to be adopted more widely in the coming years.

Estimating the Effect of Tax Changes

Conceptual issues

The basic objective of revenue estimating is to assess how possible changes to the tax law will affect tax revenues. Several problems arise in summarizing those changes in a single number.19

First, a particular change to the law can generally be expected to have revenue effects that vary over time. A change that is introduced at the beginning of a fiscal year will usually have a smaller effect on revenues during that year than it would in a “full year,” as a result of collection lags. Furthermore, some changes will have only a temporary effect on revenues (e.g., a change in the basis of business income taxation from a “preceding year” to a “current year” basis), while others will have effects that can be expected to build up gradually over several years.

Second, a change to one tax will often affect revenues from other taxes. For example, VAT is commonly levied on the value of sales, inclusive of any excise duties and import duties levied on the same product. In this case, an increase in the rate of excise duty on a particular product will have a direct effect on VAT revenues as well. It will also, of course, have indirect effects. As a result of an increase in excise duty (plus VAT), consumers can be expected to reduce the volume of that good which they consume. Depending on the elasticity of demand, their total expenditure on the good (including taxes) will either rise or fall. Either way, there is likely to be an effect on spending on other goods—and hence, on revenues from the taxes that are levied on those other goods. In addition to these substitution effects, the initial tax increase can be expected to have “second round,” macroeconomic effects which have implications for revenues from many other taxes.

The question then arises: Which of these different kinds of effects on receipts from other taxes should be taken into account in the revenue estimate for the initial change to the excise duty? The appropriate answer may well depend on the context in which the estimate is to be used. For example, if revenue estimates are being prepared to show the implications of alternative ways of reducing a fiscal deficit, it may be appropriate to neglect macroeconomic effects on revenues from other taxes—since they will be broadly similar for all the alternatives that are being considered. On the other hand, to judge how much of a tax increase is necessary to eliminate a deficit, it will be necessary to take account of those macroeconomic effects as well as of the first round, “direct” effects of the tax change.


It is harder to generalize about methods of revenue estimating than it is about methods of revenue fore-casting, since tax law changes for which revenue estimates may be needed vary widely. Changes to the rate of a proportional tax (for instance, on wage incomes or total consumer expenditures) are reasonably straightforward to deal with, once the appropriate conceptual basis for the estimate has been established. Changes to an existing tax allowance in a progressive income tax are more difficult, since the revenue estimate requires information about the present allowances and the marginal tax rates of those that receive them. It was to answer questions of this sort that the sample-based microsimulation models (referred to in the section on more complex macroeconomic models above) were initially developed.

But there are many possible tax changes for which even such complex tools provide no assistance to revenue estimators. For example, a proposal to extend a personal income tax to income that is presently nontaxable (such as social security benefits) cannot usually be assessed on the basis of a sample of existing tax returns—since those returns will typically contain no information on exempt income. In these circumstances, revenue estimators may need to use a wide variety of information sources, and techniques, to estimate the effect of the proposed change. Revenue estimating is, in practice, as much a creative art as a science.

Evaluating Revenue Forecasts and Estimates

Revenue forecasts that are made unconditionally can easily be evaluated in retrospect, when the difference between the forecast and the actual outturn is known. But revenue forecasts that are conditional on particular macroeconomic assumptions are harder to evaluate, since the forecast error may result from inappropriate assumptions as well as errors in the conditional forecast, and the two sources of error can often be difficult to disentangle. Most revenue estimates are even harder to evaluate in retrospect, since such an evaluation usually requires a hypothetical comparison of actual revenues with what those revenues would have been, if the specific change for which the revenue estimate was made had not been implemented.

Despite these difficulties, retrospective evaluations are an important component of an effective revenue estimating and forecasting system because they provide the estimators with an opportunity, and also with a motive, to learn from past mistakes and to improve their estimating methods over time. Revenue forecasters and estimators should not, therefore, devote all their efforts to the future: to the extent possible, systematic ex-post evaluation of past forecasts and estimates should be a part of their work program.

Presumptive Taxation

Kenan Bulutoglu

  • Why do countries levy presumptive taxes?

  • What are the major design issues in a presumptive tax?

  • What are the administrative advantages of a presumptive tax?

Many developing countries have weaknesses in their tax administration that make it difficult to levy effective taxes. Presumptive taxation is one means to overcome these administrative weaknesses. It may also lead to certain efficiency and equity gains. Although presumptive taxation has been practiced for years in many countries as a pragmatic approach to increasing tax revenues, the analysis of presumptive taxes is relatively scant.20 The purpose of this chapter is to describe presumptive taxation and its advantages and disadvantages.

Presumptive taxation is used mostly as a proxy for an income tax on small businesses. It may also be used as a substitute for an income tax on small farmers, owners of rental property, professionals, independent contractors, and other hard-to-tax groups. Presumptive taxation is occasionally used as a proxy for indirect taxes. For instance, in Turkey, Tunisia, and Morocco, it is used as a substitute for a VAT on small businesses, and in Pakistan, it is used as a substitute for excise taxes.

If used judiciously, presumptive taxation may broaden the tax base by increasing the number of taxpayers and their tax payments, and it may reduce tax evasion—all at a relatively low administrative cost. Although its revenue per taxpayer is generally low, it may have substantial spillover benefits in facilitating the movement of small taxpayers from the informal to the formal sector and as a source of information to reduce evasion.

The main virtue of presumptive taxation is that it may be the only effective way to tax small businesses in developing countries. Since small businesses represent the vast majority of enterprises, this may lead to a substantial increase in the number of taxpayers. In fact, in countries where there is a solid presumptive tax tradition, the number of presumptive taxpayers may be ten to twenty times higher than the number of those subject to self-assessment on recorded transactions, although this ratio declines with the level of economic development. Presumptive methods may also be effective in cutting audit time and cost, particularly in countries where accounting illiteracy is widespread. Nevertheless, enforcement without safeguards may lead to harassment of and extortion from taxpayers by un-scrupulous tax officials. Although, the participation of citizens in presumptive tax commissions may reduce the opportunities for abuse by corrupt officials.

Some countries deliberately charge punitive presumptive tax rates in order to push small taxpayers into the option of self-assessment on an actual income basis. Experience shows that this can backfire, by causing taxpayers to go underground. This shift of taxpayers to the underground economy can outweigh the benefits of adding some taxpayers to the declarative tax regime. Instead of forcing small businesses into a declarative tax regime, the tax administration should keep them in the presumptive tax regime until they are sufficiently large and sophisticated to participate in the regular tax system.

Presumptive taxes may also enhance the efficiency and equity of the tax system.21 Presumptive taxes generally take the form of a tax on average or “normal” income. Hence, the marginal tax rate on income above this average income is zero, avoiding the negative incentives associated with high marginal tax rates. In addition, by facilitating more effective taxation of hard-totax groups, presumptive taxation may lead to greater horizontal equity in the tax system.

Determination of the Presumptive Tax

There are several different ways in which presumptive taxes on income or profits may be levied. One frequently used approach is for the tax administration to determine a method for estimating income and then to apply it to each taxpayer. A second approach is to apply an assets tax. This approach has become more popular in recent years. A third approach is to apply a turnover or gross receipts tax. A fourth approach is to base the tax on external indicators of income. The first approach is typically applied in developing countries to small businesses or establishments. The second and third approaches generally apply to all businesses and are not used exclusively for presumptive tax purposes but may also serve as minimum taxes or as an additional source of revenues. Several Latin American countries use presumptive methods based on wealth. Many francophone African countries use presumptive methods based on gross receipts or turnover. These two approaches are discussed in Chapter IV.

The level of economic sophistication in a country influences the choice of presumptive tax methods. Countries with the least sophistication typically apply very simple presumptive tax methods, while those with more sophistication may apply highly developed methods. Presumptive taxes may apply generally to entire classes of taxpayers or only to taxpayers who fail to file a standard tax declaration.

The first approach to presumptive taxation may make use of very simple techniques for estimating income or more complicated techniques. The simplest approach is just to levy a lump sum on all businesses, although clearly this approach has serious drawbacks. Under a more sophisticated presumptive tax, applying it to business income generally follows a sequence of stages. First, the tax administration conducts a census of taxpayers, under which it registers and records the annual sales volume of each business, if this information is available. Second, the tax administration, sometimes along with a tax commission, determines profit margins for each business activity. To account for differences in average profit margins, activities may be classified by business type, such as trade, craft, and service; by location; and by size. Business types may be further broken down by activity. The tax administration then applies certain rules to determine profit margins by activity. For instance, one rule of thumb may be that the higher the inventory turnover, the lower the profit margin.

If information on turnover is not available, then the tax administration must also devise methods for determining turnover. In manufacturing, the tax administration may collect information on purchased materials and infer production volume from the use of these materials. It might then determine profits as a percentage of sales. In service industries, material purchases are less significant as an indicator of business volume; instead, the tax administration might use the service capacity of the establishment (such as the number of tables, seats, and employees in restaurants, theaters, and barber shops); the amount of cargo space in motor vehicles; and so on, as an indicator of turnover or profits.

Several countries have adopted the tachshiv (assessment guidelines) approach originally developed in Israel, which emphasizes the use of objective factors to estimate the income of taxpayers who fail to keep adequate books and records.22 Physical inputs and factors such as the number of employees are critical to the determination of an enterprise’s income. The tachshiv contain instructions for estimating the enterprise’s income according to the kinds of services it provides, its equipment, its location, work schedules of employees, and other criteria. Each tachshiv is prepared with detailed research and visits to a representative sample of businesses. This determination results in measures of the average profitability of the sector and the relation-ship of specific characteristics of the enterprise to profitability.

This approach has been criticized. One criticism is that taxpayers whose incomes are above average for their line of business may intentionally fail to keep adequate records for a declarative tax so that they fall under the presumptive tax. In addition, reliance on precise factors to determine income may transform the tax into one on those factors, rather than a general income tax, thus distorting optimal use of factors of production and other business decisions.

Presumptive tax systems generally provide for the public to contest the government’s determination of profit margins and costs. Individual taxpayers generally cannot contest the decisions that establish profit margins and costs. But some countries allow both taxpayers’ representative bodies, such as legal trade unions, chambers of commerce, and the tax administration, to contest regulatory decisions. In practice, however, grievances are usually settled through political decision.

The tax administration may determine the presumptive tax collectively or individually. The collective assessment method has the advantage of administrative simplicity while the individual assessment method has the advantage of greater accuracy. In France and in some of its former colonies, the collective presumptive tax has traditionally been applied to the smallest businesses and to some selected sectors, and individual assessment to businesses intermediate in size to the small taxpayers and the self-assessed. Some countries, such as Turkey, have only collective assessment.

Under collective assessment, the tax administration establishes presumptive income for groups of taxpayers, corresponding to different activities. The tax administration assigns each taxpayer to a group, which determines the taxpayer’s tax liability. The tax liability and taxpayer’s classification may be revised periodically. Finally, taxpayers are given the opportunity to contest the classification of their business activities. In some countries, taxpayers can resort to court action if they are unsatisfied with their tax treatment.

Under individual assessment, taxpayers are not expected to remain passive as in collective assessment; they may be required to submit certain information annually so that the administration can assess their net income by applying cost-profit ratios. The tax administration assesses the taxable income of each taxpayer, by mimicking the declaration procedure. Finally, under individual assessment, the taxpayer has the opportunity to negotiate with the tax administration or to appeal through the judicial system.

In individual assessment, the tax officer may tutor the taxpayer in the self-assessment procedure, as is found in the British system of assessment. Such assessment, however, is costly and a potential source of corruption and taxpayer harassment. When facing a tax officer invested with discretionary powers, the taxpayer is in a weak position; a scared and/or astute taxpayer may seek collusion with the tax officer. This danger is particularly strong if the taxpayer is not protected by a right to court action. Even when tax officers are honest, a negotiated assessment can create a feeling of fiscal harassment, even inquisition on the part of the taxpayer. A zealous application of individual forfait in France in the 1950s caused a tax revolt that culminated into a strong political party (Poujadism).

Under both collective and individual assessment, the tax administration may update records only every few years to lower administrative costs. Multiyear applicability of the presumptive tax is advantageous to businesses in expansionary or inflationary times, but it becomes more burdensome when profits decline.

Promotion of taxpayers from individual assessment to a self-assessment system has several benefits. First, the tax administration may reduce the costs of individual assessment. Second, small businesses get invoices to deduct their costs and register their transactions, boasting self-enforcement in the business tax. Third, malpractice stemming from contacts between taxpayers and tax officers is largely eliminated.

Half-Presumptive Methods in Self-Assessment: Presumptive Cost

Halfway between the presumptive tax and the self-assessment tax is the method of presumptive cost deduction, under which the taxpayer declares revenues but not costs. Costs are then estimated on a presumptive basis. When costs are deductible presumptively, most countries also allow taxpayers the option to deduct the actual costs. A generous presumptive cost deduction erodes taxable incomes. When deduction of actual cost is an option, there is no justification for a generous presumptive cost deduction. Taxpayers experiencing high costs can opt for actual cost deduction. When there is an option, however, the taxpayer can erode the tax base by concentrating the costs in a year of actual cost deduction, then shifting to presumptive cost deduction in subsequent years. A minimum time period, say three years, should be required to prevent such abusive shifts from one method to another. One common application occurs in developing countries, and increasingly in the former centrally planned economies, where many people obtain rental income without setting up a business. A bookkeeping obligation for cost deductions would be cumbersome for most of these property owners. Presumptive cost deduction alleviates their compliance costs.

Presumptive Taxation as a Shortcut to Audit

In the self-assessed income tax, cost or profit ratios are used to reduce tax evasion and audit time and cost. A deviation of taxpayer declarations from such norms can be sanctioned in two ways: either the presumptive norms are non-negotiable and applied automatically, or the taxpayer has the right to challenge the presumptive ratios, while still bearing the burden of proof. The former is cost effective but may be inequitable for some taxpayers. The latter is prevalent in advanced income tax systems, and its effectiveness depends on the existence of a fair and speedy judicial procedure.

The very existence of cost norms can boost self-enforcement, provided that they are applied randomly—even if sporadically. They can induce taxpayers to declare plausible costs and profit margins and to keep evidence of their transactions. In developing countries, profit, cost, and output standards are seldom used in computerized programs to identify the suspicious returns and to speed up the audit; therefore, self-enforcement is either weak or absent. The use of presumptive cost and profit norms without any possibility of taxpayer’s contestation is a common practice. Any declaration that reduces the tax base by deviation from official cost and profit ratios is rejected. Modern income tax laws, however, allow taxpayers to prove the authenticity of their declarations. But such declarations can trigger an audit and eventually a costly litigation. To avoid this outcome, some countries give the tax administration power to negotiate with the taxpayer. Unless rules of negotiation are clear and discretionary powers minimal, however, corruption can be rampant.

Presumptive Taxation Based on Income Indicators

The presumptive tax models described so far determine the tax base by estimating sales, costs, or asset values. Another form of presumptive taxation departs further from actual transactions and determines the tax base by using external indicators of income, such as personal expenditures and wealth accumulation.

The tax administration may use external income indicators to identify nonregistered taxpayers and to verify the income of existing taxpayers. When taxpayers do not comply with their bookkeeping obligations, make no income declaration, or file a declaration that is rejected due to bookkeeping irregularities, many countries allow the use of indirect indices—such as living standards and annual increases in net wealth—to assess the taxable income. Indirect or external income indicators that are inconsistent with the declared income can be used for assessment even if the taxpayer’s books show no irregularities. But a tax based on external income indicators is exceptional, and used only after the taxpayers are given a chance to prove their nontaxable sources of income and wealth.

Taxation based on income indicators is usually complementary to self-assessed income taxation. In most countries, its use is restricted to sustained criminal or underground activities, such as drug dealing, smuggling, and gambling. Wealth accumulation and standard of living of a taxpayer can be investigated by the tax administration as part of an audit work. But some countries require taxpayers to make annual declarations of their living standards and total wealth. If accumulated over several years, such information can render the discrepancies between actual and declared income increasingly obvious and the taxpayer vulnerable to audit.

In developing countries, indicator-based taxation is a speedy way to audit. It reduces the evasion of small, hard-to-tax taxpayers. It can capture incomes from underground activities. Adequate safeguards, however, must be devised to protect honest taxpayers and prevent corruption. To that effect, a special committee can be invested with the power to determine the presumptive tax based on wealth and expenditure indicators. Taxation on living standards or wealth accumulation may increase compliance. In some countries, wealth indicators are by themselves a base for a minimum income tax.

Incentive Effects of Presumptive Taxation

Presumptive taxation may produce a more efficient outcome than regular taxation by offering no discouragement to economic activity on the margin;23 rather, it may encourage production because producers keep all profits on the margin. This incentive, however, tapers off the closer the presumptive tax approximates the actual tax liability.

The presumptive tax may be an entry barrier to small business compared to the absence of any tax. When the formal sector or large enterprises are taxed, however, the exemption of the informal sector or small enterprises would distort economic decisionmaking. A minimum contribution to tax revenues may enhance the risk of business creation and encourage enterprises to go underground. These potential negative effects of presumptive taxation on small enterprises argue for a moderate presumptive tax and taxes that vary according to potential incomes.

The most important problem with presumptive taxation is that it may lead to erosion of the self-assessed business tax base through the use of fake invoices. Experience shows that enterprises under the declarative regime use presumptively taxed enterprises as their suppliers of fake invoices to boost their costs. To limit such practices, the tax administration has to monitor the invoices issued by the presumptively taxed businesses. This, in turn, reduces the cost savings of the presumptive tax.


Experience suggests that the shortcomings of presumptive taxes stem mostly from a lack of clear objectives. When overstretched for revenue purposes, a presumptive tax may lead to corruption and oppression of taxpayers. When used for administrative convenience and low compliance costs, it may degenerate into outright exemption as a result of insufficient monitoring of small businesses. A successful presumptive tax should serve primarily to prepare small businesses for self-assessment and secondarily to enhance revenues.

The stages of presumptive taxation are outlined in Table VI.3.

Table VI.3.

The Stages of Presumptive Taxation

article image

Minimum Taxes

Janet Stotsky

  • Why do some countries levy a minimum tax?

  • What are the main issues in the design of a minimum tax?

  • Should the minimum tax be levied on businesses or individuals?

Minimum taxes overlap to some extent with both assets taxes and presumptive taxes, but they may also be neither. Minimum taxes are generally based on assets, turnover, or redefined income and they apply to both large and small enterprises. Assets taxes may serve as minimum taxes but may also serve as nonminimum taxes that apply to all enterprises. Presumptive taxes may have the same form as minimum taxes but they generally apply only to small firms that are incapable of being incorporated into the regular tax. Since Chapter IV discusses assets taxes and the previous section discusses presumptive taxes, this section focuses on the distinct role that minimum taxes can play in a tax system.

Definition and Use

Many countries have introduced some sort of minimum tax to supplement either the regular business income tax or individual income tax. The purpose of a minimum tax is largely to ensure that businesses or individuals with economic income do not regularly avoid paying income tax. Many developing countries have business minimum taxes while individual minimum taxes are rare. In developing countries, the business minimum tax may apply to both corporate and noncorporate enterprises. A few industrialized countries also have explicit minimum taxes. The United States has both corporate and individual minimum taxes, where they play a complementary role. Canada has only an individual minimum tax. Norway and Denmark also levy a supplementary tax, which is a form of minimum tax. This discussion will examine the rationale for minimum taxes and important issues in their design.

Rationale for Business Minimum Taxes

A business minimum tax may be used to enhance the equity (or perceived equity) and efficiency of the business income tax. Inequities in a business income tax may arise because of differences in tax compliance across businesses, because of differences in the ability to use tax preferences, and other reasons. Under an equitable business income tax, businesses with the same taxable income would pay the same taxes. In many developing countries, the ability to administer the business income tax is limited, leading to widespread tax evasion. As a consequence, businesses with equal incomes may pay very different taxes, simply because of differences in tax compliance. A business minimum tax may reduce this inequity by taxing businesses on some notion of income that is independent of their declared income for tax purposes. A business minimum tax may also facilitate tax collection from taxpayers with business income. This was one rationale for the adoption of the business assets tax in Mexico.24

In many industrialized and developing countries, the business income tax is characterized by a proliferation of tax preferences, granted in the form of tax exclusions and deferrals. The aggregate business income tax base may be significantly eroded by these tax preferences. A minimum tax offers two ways of increasing business tax revenues from businesses that have benefited excessively from these tax preferences. First, it forces all businesses to pay either the regular business income tax or the minimum tax. Second, it may discourage businesses from taking advantage of tax preferences that would lower their regular tax liability so much that they end up incurring a minimum tax liability. By avoiding excessive tax preferences, they increase their regular business income tax liability. Ambivalence about the role of tax preferences was a main argument in favor of the adoption of the corporate alternative minimum tax in the United States.

In economies beset by high inflation rates, the favorable treatment accorded debt finance through the deduction of nominal interest payments supports this rationale for a minimum tax. With high inflation, businesses with significant amounts of debt may get such large deductions for nominal interest payments that they routinely have tax losses and are able to avoid paying the business income tax for many years, even though they are earning substantial economic income. In Argentina and Mexico, the erosion of the tax base from high inflation was one important factor that led to the adoption of the business assets tax.

On efficiency grounds, there are arguments for and against a minimum tax. Sadka and Tanzi (1993) argue that businesses may use capital inefficiently because socially inefficient uses are intangible and thus untaxable while socially efficient uses are taxable. They argue that a tax on “normal” or presumptive income (which may be seen as a minimum tax) may enhance efficiency because it may have fewer distortions at the margin than the regular income tax. This would encourage businesses to use capital in socially efficient ways. In the U.S. context, Graetz and Sunley (1988) argue that by equalizing the marginal tax rate across industries, a minimum tax may enhance efficiency. On the other hand, they argue that by penalizing businesses that may have some advantage in tax-preferred activities, the minimum tax may reduce efficiency across businesses in the same industry. The ultimate efficiency effects of a minimum tax are quite complicated.

A minimum tax can also be justified as a form of business license tax. Business license taxes have not precluded the adoption of a minimum tax in many countries including France, which levies both taxes. Nevertheless, the two taxes must be distinguished. The minimum tax is always credited against the regular business income tax. But there is no refund if the minimum tax turns out to be higher than the regular tax. Business license taxes are generally allowed among deductible charges in the business income tax.

Forms of Business Minimum Taxes

The business minimum tax may take many forms. The appropriate form of the tax depends on the specific objectives that it is intended to accomplish. In its simplest form, the tax may require the payment of a fixed nominal amount from each taxpayer. Although such a tax is efficient in that it essentially functions as a lump-sum tax (as long as it applies equally to all forms of businesses), it is inequitable in that it is not based on any proxy for income. It may also be used in combination with a business minimum tax where the information needed for applying a minimum tax is lacking. The amount has to be set carefully to insure that businesses do not simply pay this nominal sum rather than the minimum tax.

An alternative is to levy the tax as a relatively low percentage of the turnover (or gross receipts) of the business. This form of business minimum tax is frequently found in countries influenced by French tax practice. France introduced a minimum business tax in 1973. Initially, this minimum was a fixed amount of FF 1,000, but now the rate varies according to the size of turnover, and it currently ranges from 0.5 percent for corporations with a turnover of FF 1 million or less to 0.215 percent for those with a turnover of FF 10 million or more. Many countries in francophone Africa also use a minimum business tax based on turnover.25

The advantage of a turnover-based business minimum tax is that turnover is likely to be the most easily measured financial variable for a business and must be computed for the payment of other taxes, such as the VAT. Thus, turnover is likely to be more readily available to tax authorities. In addition, in an economy where prices are not regulated, prices and hence turnover adjust with inflation. The tax is thus not distorted by inflation. The disadvantage of a turnover-based business minimum tax is that turnover bears no necessary relationship to any measure of income. A business with a large turnover might have negligible profits or losses, but would still be required to pay a substantial minimum tax, while a business with a small turnover might have substantial profits and would pay a small minimum tax. Thus, a turnover-based business minimum tax does not provide a good proxy for an income tax. In addition, a turnover tax is deficient compared to a VAT because it cascades from one level of production to the next. Thus, the rationale for a turnover-based business minimum tax rests mainly on practical considerations.

Another alternative is to levy the tax as a relatively low percentage of the assets of the business, applying to gross, net, or fixed assets.26 The use of an assets tax as a business minimum tax is found in Latin America. An assets-based business minimum tax has a stronger theoretical appeal than one based on turnover in that economic income could be expected to bear some systematic relationship to assets. Typically, it is assumed that in a well-functioning capital market with capital mobility, the rate of return on capital is equalized across investments. Hence, a business assets tax could be a reasonable, though imperfect, proxy for an income tax. This form of tax is perhaps most appropriate for most developing countries.

Mexico has applied a business assets tax since 1989. The tax is levied on business assets at a 2 percent rate. The purpose of the assets tax is to facilitate collection of the income tax on business income, but it also functions as a minimum tax. The assets tax liability is designed to be roughly equal to a taxpayer’s income tax liability. For instance, 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 are allowed to credit their Mexico income tax payments against their Mexico assets tax liability. Thus, if the income tax liability is greater than or equal to the assets tax liability, no additional assets tax is levied.27

Another alternative is to levy the tax on some redefined notion of business income, as in the United States. As part of the 1986 reform of the U.S. income tax code, a corporate alternative minimum tax was added to the regular corporate income tax. This tax replaced a previous minimum tax, which was levied as a surcharge on the corporate income tax for the use of certain preferences. Under the reconstituted minimum tax, businesses are required to compute taxable liabilities under both the regular corporate income tax and the alternative minimum tax, and then they must pay the larger of the two tax liabilities. The difference between the alternative minimum tax payment and the regular tax liability may, with some exclusions, be credited against future tax liabilities, but may not reduce the regular tax liability below the alternative minimum tax liability. The alternative minimum tax is computed by making certain adjustments and adding certain tax preference items to income. Gerardi, Milner, and Silverstein (1992) discuss the major adjustments and preferences in the U.S. corporate alternative minimum tax.

This form of minimum tax is perhaps appropriate for the United States in that the minimum tax is motivated by an ambivalence about the use of legal tax preferences rather than by concern about administrative weaknesses or the distorting effects of inflation on business income. Thus, it gets directly to the heart of the problem. This form of minimum tax, however, may be less appropriate in countries where administrative weaknesses or high inflation are the principal problems since it is really not intended to solve these problems.

Determination of the Tax Base

The appropriate application of the minimum business tax requires careful determination of the tax base: either turnover, assets, or redefined income. Turnover is the most readily measured of these tax bases. Nevertheless, in many developing countries, it may be difficult for the tax authorities to get accurate measures of turnover for some businesses, particularly retail establishments and farmers. It is easier to get accurate measures of turnover for manufacturing businesses. Even in industrialized countries, it may be difficult to get accurate measures of turnover for small businesses, particularly those providing services, such as tailors and hairdressers. To overcome obstacles in measuring turnover, the French developed the “forfait” system, essentially a presumptive tax, in which the tax is based on some objective characteristics of the business.

The assets tax is generally imposed on a taxpayer’s gross business assets. Assets include cash and securities, receivables, inventories, land, and other fixed assets at depreciated value, and intangible assets at amortized value. Alternatively, it is possible to impose the tax on fixed assets (land, plant, and equipment) or on net assets (gross assets net of debt-financed liabilities) alone.

Various considerations enter into the choice of which asset measure to use for the tax base. If the purpose of the tax is to provide a proxy for a broad-based income tax that does not favor debt-financed assets, then levying the tax on gross assets is appropriate. If instead, the purpose is to provide a proxy for a business income tax that allows a deduction for interest payments on debt but not for dividends, then levying the tax on net assets is appropriate. Levying the tax on fixed assets has merit on practical grounds. A tax on assets may be difficult to administer fairly because of the inability to measure some components of assets accurately because of transitory variations in asset value, illiquidity, and inflation. Current assets have a larger transitory mixed, but are more liquid and less distorted by inflation. Ideally, fixed assets and intangible assets should be valued at fair market value, but usually the only available measure of value is historic costs reduced by depreciation or amortization, respectively. Land, although it is not a depreciable asset, poses a similar problem. It is possible to adjust balance sheet items through indexation to account for inflation.

Double taxation of certain assets is potentially a problem in certain instances where taxable businesses own financial interests in each other or own each other exclusively. It is possible to solve this problem, however, by selectively exempting these assets from the tax base where these conflicts arise.

The determination of the base of the U.S. alternative minimum tax reflects the objectives of reducing the benefits of tax preferences and enhancing the equity of the business income tax. Thus, the additions to income are derived from parts of the tax code which create large differences between economic and taxable income. Graetz and Sunley (1988) and Lyon (1991) discuss these issues in detail.

Determination of the Tax Rate

The tax rate that is chosen should reflect the revenue and incentive objectives of the business minimum tax. A key point is that the minimum tax rate should not be set so high that it results in hardship for a business with true economic losses, but the rate should be set high enough that the business tax’s equity and efficiency goals are achieved. Thus, the appropriate rate is likely to vary, depending on revenue needs, the nature of the minimum tax, and the structure and rates of the regular income taxes that they supplement.

Other Issues

Another design issue relates to international coordination of income tax regulations to allow businesses to credit their minimum tax liabilities against their domestic business income tax liabilities. It is important to develop rules so that any minimum tax qualifies for the foreign tax credit. A problem may arise with minimum taxes based on turnover or assets in that some countries only allow income tax credit for taxes that are explicitly levied as income taxes. This problem arose for U.S. firms under the initial version of the Mexico assets tax since U.S. law does not allow a U.S. taxpayer to credit an assets tax payment against the U.S. income tax liability. To allow U.S. businesses to qualify for the foreign income tax credit, the Mexico income tax was modified by allowing businesses to credit their income tax payment against their assets tax liability rather than allowing businesses to credit their assets tax payment against their income tax liability.

Another important design issue relates to the ability to credit minimum tax payments against the regular income tax liability in the future. Under a typical business minimum tax, businesses are able to credit in part or in full the difference between the minimum tax and regular income tax liabilities against future business income tax liabilities. If this credit is not indexed, it falls in present value terms the longer the business must defer the use of the credit. If the credit is indexed, this problem does not arise. Under the Mexico assets tax, the credit is indexed, thus preserving its value under conditions of high inflation. Under the U.S. corporate alternative minimum tax, the credit is unindexed, though the relatively low rate of inflation in the United States makes this less of a problem than it would be in countries with higher inflation rates.

Individual Minimum Taxes

There are also different ways to levy an individual minimum tax. One alternative is to levy a tax on the assets of self-employed business owners or on personal property, which allows its extension to all individual taxpayers. The United States and Canada have explicit individual minimum taxes. The U.S. individual minimum tax has the same basic structure as the U.S. corporate alternative minimum tax, with the two complementing each other. The Canadian minimum tax is similar to the U.S. tax; there is, however, no corresponding business minimum tax. The objectives are the same: to enhance the equity and efficiency of the individual income tax.

Some countries, such as Norway and Denmark, use supplemental taxes on high-income taxpayers to en-sure that they pay their fair share of tax. These taxes are similar in nature to minimum taxes, though with somewhat different application. Norway levies an 8.5 percent supplementary tax on gross income, applicable over a high threshold. Denmark levies a 12 percent supplementary tax on “personal income” (which is roughly gross income minus income from capital), applicable over a high threshold. Shome (1993) discusses other ways to levy supplemental taxes.

Under a fully integrated individual and business income tax system, there would be no need for a separate business minimum tax, since all income would be taxed at the individual level. Even without full integration, it is possible to question the need for a separate business minimum tax, if the ultimate incidence of the business tax falls on owners of capital, a matter which remains open, however. As a practical matter, the business minimum tax is an essential counterpart to the individual minimum tax and is more relevant than the latter for most countries.

The Interrelationship Between Tax Policy and Tax Administration

Angelo G.A. Faria and M. Zɒhtɒ Yɒcelik

  • What are the interrelated roles of tax policy and tax administration in designing and operating a tax system?

  • How do tax withholding schemes contribute to improved compliance and enforcement and alleviate the burden of tax administration?

  • Should schedular taxes be preferred over a global income tax in developing countries with weak tax administration?

  • How do exemption thresholds, presumptive taxation, and minimum taxes help improve revenue productivity and remedy problems in enforcement?

While there may be some disagreement on the extent to which government should be involved in attaining economic objectives, its overall desirability is generally accepted to be crucial. The government’s role requires both that it is able to finance its activities in a noninflationary way through compulsory extraction of resources from households as well as that the resultant distortions constituted by taxes as wedges that influence relative prices are minimized. Herein lies the primacy of tax policy in helping to attain economic policy objectives.28 More importantly, in the same way that tax policy should have a dynamic orientation to respond to changing economic circumstances, tax administration must itself evolve an internal dynamic to promote the effective application of tax policy.

Thus, tax policy and tax administration are inextricably related.29 For the design of tax policy to be successful, it must also pay due attention to administrative constraints; and, measures to improve tax administration should help to make the implementation of designed tax policies more effective. Put simply, idealistic tax policy can complicate tax administration, while ineffective tax administration can undermine tax policy.30 Hence, failure to coordinate these activities is likely to affect adversely the pace and sustainability of the tax reform process.

Tax revenue yield is influenced by both tax policy and tax administration.31 The concern of tax policy is to ensure the elasticity or responsiveness of potential revenue to overall economic growth and this depends on how the tax bases, and the tax rates applied to them, are established. For example, if tax bases are eroded because of tax exemptions and deductions for achieving various social and political welfare objectives, tax rates would have to be raised on taxable items, in principle, to secure a given level of potential revenues. There is, however, a limit to which tax rates can be raised before they begin to affect adversely the behavior of economic agents, in particular, as regards their decisions to work, consume, save, and invest.

Tax administration assumes that tax bases and rates are appropriately established and seeks to secure as much as possible of the resulting potential tax revenue effectively and efficiently. The more complex it becomes to administer taxes because of the need to ensure the correct application of numerous exemptions and deductions and multiple tax rates, the less effectively will the tax administration collect the potential tax. At the same time, its efficiency will also be diminished because the costs of such collection will rise.

In sum, an appropriate strategy for tax reform would first involve studying the tax structure and setting appropriate policy goals, and then modifying these in the short term by taking cognizance of the associated administrative problems. If the ordering were reversed and administrative considerations became the binding constraint in a tax reform, which by its very nature is a longer-term process, the tax system is likely to play only a very limited role in achieving economic policy objectives.

Inasmuch as the characteristics of tax policy have been explored in detail in the other sections of this Handbook, the remainder of this section is devoted to an examination of the various aspects of a tax system that have a bearing on the interplay between policy and administrative concerns.

Statutory Versus Effective Tax System

A tax system may be described in two ways: (1) its statutory provisions covering tax rates and bases, methods of payments, and so on as prescribed in tax laws; and (2) its effective impact reflecting actual implementation of tax laws.32 Several factors are responsible for the possible divergence between the statutory and effective aspects of a tax system:

• Tax evasion resulting partly from complex tax laws reducing the compliance level; and largely from the inability of tax administration in enforcing tax laws;

• Poor level of accounting and bookkeeping practices in the country precluding reliable records for determination of taxable bases, which is a crucial mixed of modern income and sales taxes;

• Lack, or poor application, of sanctions against tax offenders. For example, low- and middle-income developing countries, such as Côte d’Ivoire, Ghana, Lesotho, Mauritius, Sierra Leone, Togo, and Zaire collected on average about 37 percent of their tax revenue during 1986–92 from foreign-trade based, convenient tax handles, notwithstanding the existence of a broad range of domestic taxes, reflecting their administrative limits.33 This happens even though excessive dependence on taxation of foreign trade compromises efficiency and equity objectives.

Simplification of the Tax System

A tax system with several rates and various deductions and exemptions can substantially reduce its enforceability by the tax department and compliance by taxpayers; hence the simplification of the system is essential in tax reforms. A tax administration with limited resources of staff and means cannot effectively monitor, in situations involving a large number of taxpayers, sales tax with multiple rates or an income tax with various deductions and exemptions.34 In industrial countries, a tax administration may have sufficient number of well-trained staff and material resources to do this task, but both compliance and enforcement costs may be significantly higher than economically justifiable. The experiences in Argentina, Bolivia, Colombia, and the United States point in the direction of simplification. In Bolivia, the 1986 tax reform replaced enterprise income tax by a net worth tax. A simple tax based on gross sales revenue replaced all taxes (in-come taxes and VAT) on small enterprises. A broad-based VAT was adopted with a single rate and few exemptions, which replaced excise taxes with multiple rates.35 The 1989 tax reform in Argentina abolished several taxes because either their revenue yield was very little or they created significant inefficiencies in the economy.36

The 1986 tax reform in Colombia simplified the income tax by eliminating personal exemptions, income splitting, and most itemized deductions and hence the need for filing tax returns by most taxpayers.37

In the United States, the 1986 tax reform simplified the tax system to a large extent by eliminating itemized deductions. Compliance cost was also reduced significantly. According to simulations run by Slemrod (1989), payments for professional assistance by taxpayers were likely to decline ranging from 28 percent to 37 percent.

In Israel, on the other hand, both the 1975 and 1987 tax reforms proposed complex tax policies to ensure economic efficiency and tax equity without taking into account the constraints and opportunities of tax administration and therefore serious gaps were observed between goals and actual results. Several aspects of the reform were “diluted” because of administrative limits in applying requirements of the tax reforms. The tax administration resisted or simply ignored several aspects of the reforms.38

Current Payment Systems and Tax Policy

Current payment systems are used mainly for income taxation purposes to make income tax more effective as a fiscal policy instrument. They enhance the automatic stabilizer role of a progressive income tax.

Current payment systems play a crucial role in maintaining the real tax revenue in economies with high inflation; without them, the income tax might have a destabilizing effect in such economies by failing to absorb excess demand. Current payment of income tax is also convenient for taxpayers because it is transferred to the Treasury during the year in which the income has been earned, so that taxpayers will budget their taxes accordingly and avoid payment of large sums in one installment at the end of the taxable period. In inflationary conditions, collection of current payment installments becomes more crucial to prevent the erosion of real revenue.39

Current payment systems may use two techniques: (1) withholding on wages, salaries, and interest and dividends; and (2) estimated tax payments for self-employment incomes.40

Tax withholding

Withholding on wages and salaries represents a major source of current payment revenue in many countries. The idea of tax withholding was first introduced in the United States in 1943 to accelerate collection of income tax during World War II. It was adopted gradually by other countries. The withholding system is now applied in two major forms: (1) noncumulative withholding; and (2) pay-as-you-earn (PAYE).41

Noncumulative withholding may have proportional or progressive rates but it requires an annual adjustment of tax liability for the taxable year as a whole. Under the PAYE system, withholding of the tax is made in a cumulative manner from one pay period to the next. The employer should determine cumulative totals of wages paid and of tax withheld for each employee in each pay period. The difference between the tax due on total wages paid-to-date and total tax withheld-to-date gives the amount of withholding required; total withholding for the year gets very close to the actual liability of employees. The PAYE system has been used in many countries as a final tax and thus freeing taxpayers from filing an annual tax return and the tax administration from processing a large number of returns with little revenue prospects.

Under both systems, definition of the withholding base is important. A comprehensive withholding scheme would require inclusion of all payments, that is, bonuses, premiums, overtime pay; and monetary values of benefits-in-kind such as housing, vehicle, food, and servants provided by the employer.

A major problem in withholding schemes may arise from retention of taxes withheld by employees. The tax administration should monitor closely transfer of taxes to the Treasury and apply sanctions on delays. Such retentions constitute interest-free loans for employers and in particular, partial nonpayment in inflationary conditions. Monitoring of employers would be relatively easy because the number of employers would be much less than that of employees.42

Withholding is also used for taxation of interest and dividends as it constitutes a convenient tax handle applied at the time of payment of such incomes by payers. Withholding rates are proportional and the payee may be required to file a tax return at the end of the taxable year and include such income in their total income. They may claim tax credit for tax amounts previously withheld. In many countries, tax withholding represents final taxation for interest and dividends and their inclusion in total income is not required, mainly for administrative convenience. Otherwise, the tax administration would need lists of payments for such incomes by payers, indicating amounts paid and tax withheld by payee; and it would also need to cross-check these data against tax returns filed by payees, which may create a large drain on administrative resources.

Estimated tax payments

This type of current payment is applied to companies and self-employed taxpayers, for example, professionals, traders, craftsmen, and farmers. In the absence of such practice, equity and revenue objectives of taxation would be compromised in view of tax withholding on wages and salaries of employees.43

As withholding on self-employment income would be impossible, current payment would require estimation of current income by taxpayers.44 Since taxpayers cannot estimate easily their income for the current year, the preceding year’s declared income and tax payments are usually taken as reference. Taxpayers are required to pay the same amount of tax for preceding year minus total tax withholding. Taxpayers should be authorized to amend their current income levels in case of fluctuations during the year compared with those initially declared; and underestimations exceeding a set percentage, say 10 percent to 25 percent, should be penalized to deter willful reduction of estimated tax payments.

Schedular Income Taxes Versus Global Income Tax

Serious doubts have been raised about global income tax in countries with limited administrative resources despite a worldwide movement toward globalization in the 1970s and 1980s. A global income tax may be desirable for equity purposes but not viable for various reasons.45

First, there may be political obstacles to globalization especially from members of the emerging middle class in many developing countries. Fears are sometimes expressed also on account of heavy taxation of investment income leading to the possibility of capital flight.

Second, determination of global income by the taxpayer constitutes a major problem in developing countries. Professional, rental, business, and capital incomes cannot be easily determined because of the poor state of accounting practices, the high level of illiteracy among taxpayers, and limited use of financial institutions in commercial transactions. All these conditions hinder production of reliable data for the tax administration to determine global income.

Third, the tax administration has limited resources of staff and materials to satisfy requirements of a universal filing. As a result, a global income tax in a country with weak enforcement may be transformed in practice into a schedular tax applied only to employment income.46

An interesting experience took place in the Philippines in 1987 as schedular taxes replaced a global and comprehensive income tax which did not achieve its equity and efficiency objectives; moreover, it caused hardships on both taxpayers and the tax administration. Despite its progressive rates, the global income tax base was eroded by large optional deductions. The present income tax system comprises:

  • a tax on wages and salaries with progressive rates reaching a maximum of 35 percent and allowing deductions only for dependents;

  • a uniform final withholding tax of 20 percent on capital income;

  • a progressive tax on business income, also reaching a maximum of 35 percent; and

  • a company tax of 35 percent.47

Exemption Thresholds, Presumptive Taxation, and Minimum Taxes

Although a comprehensive tax base may be desirable for equity and efficiency, tax policy experts have to agree to limit the coverage of each tax to the number of taxpayers with which the tax administration can effectively cope through high exemption thresholds.48 Administrative improvements feasible in the short term may allow expansion of coverage to some extent. With experience, this coverage may be further expanded by the administration over the medium term. If a large coverage is desirable on equity and/or efficiency grounds, the administration should be prepared to assume additional tasks provided that an adequate lead time is allowed for administrative changes.49

Concentration of administrative efforts is justifiable on revenue productivity grounds. In many developing countries, a small percentage, say 5 percent to 10 percent, of potential taxpayers account for about 80 percent to 90 percent of total revenue collected from major taxes, that is, business income taxes and sales taxes. Concentration of administrative resources on large taxpayers would be productive,50 while the diversion of staff time and materials beyond a reasonable level on small taxpayers and hard-to-tax groups would yield little revenue. Efforts to tax these taxpayers, however, may still be justified on equity grounds, as well as on the possibility of obtaining information related to large taxpayers for auditing purposes. But it must be recognized that this represents a short-term expedient that should not become a permanent feature of the tax system. In the short term, as it is often difficult to ascertain the taxable bases of small taxpayers, presumptive taxation can be used as a tax policy instrument to bring them into the tax net. Nevertheless, the objective should be to bring them into the universe of regular taxpayers over time.

Minimum business taxes can also be levied on business profits of medium- and large-size enterprises effectively in developing countries to remedy the following three major problems of net profit taxation: (1) low level of compliance by taxpayers; (2) poor state of enforcement by tax administration; and (3) large-scale use of tax incentives by new enterprises as opposed to those which do not benefit from such schemes. Minimum taxes are even used in developed countries such as Canada, Denmark, Norway, and the United States to enhance the equity and efficiency of income taxes. They are also convenient for assessment and collection purposes by the tax administration. Minimum taxes are payable on a current basis, and are credited against the final tax levied on net profits, but no refund is allowed if the former is higher than the latter.51


  • Abdel-Rahman, A.M., “The Egyptian Tax System: Recent Developments of Tax Policy and Administration,” in Tax Systems in North Africa and European Countries, ed. by Luigi Bernardi and Jeffrey Owens (Boston: Kluwer Law and Taxation Publishers, 1994).

    • Search Google Scholar
    • Export Citation
  • Bahl, Roy W., and Jorge Martinez-Vasquez, “The Nexus of Tax Administration and Tax Policy in Jamaica and Guatemala,” in Improving Tax Administration in Developing Countries, ed. by Richard M. Bird and Milka Casanegra de Jantscher (Washington: International Monetary Fund, 1992), pp. 66123.

    • Search Google Scholar
    • Export Citation
  • Bird, Richard M., “Tax Administration and Tax Reform: Reflections on Experience,” in Tax Policy in Developing Countries, ed. by J. Khalilzadeh-Shirazi and Anwar Shah (Washington: World Bank, 1991), pp. 3856.

    • Search Google Scholar
    • Export Citation
  • Casanegra de Jantscher, Milka, “Administering the VAT,” in Value Added Taxation in Developing Countries: A World Bank Symposium, ed. by Malcolm Gillis, Carl S. Shoup, Gerardo P. Sicat (Washington: The World Bank, 1990), pp. 17179.

    • Search Google Scholar
    • Export Citation
  • Casanegra de Jantscher, Milka, and Richard M. Bird, “The Reform of Tax Administration,” in Improving Tax Administration in Developing Countries, ed. by Richard M. Bird and Milka Casanegra de Jantscher (Washington: International Monetary Fund, 1992), pp. 115.

    • Search Google Scholar
    • Export Citation
  • Chand, Sheetal K., Some Procedures for Forecasting Tax Revenue in Developing Countries, International Monetary Fund, Departmental Memorandum No. DM/75/91 (October 1975).

    • Search Google Scholar
    • Export Citation
  • Federation of Tax Administrators, “Revenue Forecasting and Estimation: How It’s Done, State by State,” State Tax Notes, Vol. 4 (Arlington, Virginia, May 3, 1993), pp. 103863.

    • Search Google Scholar
    • Export Citation
  • Gerardi, Geraldine, Hudson Milner and Gerald Silverstein “Temporal Aspects of the Corporate Alternative Minimum Tax: Results from the Corporate Panel Data for 1987–1990,” Proceedings of the 85th Annual Conference, National Tax Association—Tax Institute of America (1992) pp. 11727.

    • Search Google Scholar
    • Export Citation
  • Gillis, Malcolm, “The VAT and Financial Services,” in Value Added Taxation in Developing Countries, ed. by Malcolm Gillis, Carl S. Shoup, and Gerardo P. Sicat (Washington: World Bank, 1990), pp. 8394.

    • Search Google Scholar
    • Export Citation
  • Goode, Richard B., “Some Economic Aspects of Tax Administration,” Staff Papers, International Monetary Fund (Washington), Vol. 28 (June 1981), pp. 24974.

    • Search Google Scholar
    • Export Citation
  • Graetz, Michael J., and Emil M. Sunley, “Minimum Taxes and Comprehensive Tax Reform,” in Uneasy Compromise: Problems of a Hybrid Income-Consumption Tax, ed. by Henry J. Aaron Harvey Galper and Joseph A. Pechman (Washington: Brookings Institution, 1988) pp. 385429.

    • Search Google Scholar
    • Export Citation
  • Griffith, Peter S., “Current Payment of Taxes on Income,” in Readings on Income Tax Administration, ed. by Patrick L. Kelly and Oliver Oldman (Mineola, New York: Foundation Press, 1973), pp. 345 –68; pp. 38187.

    • Search Google Scholar
    • Export Citation
  • Hudder, John J., “Use of Models in Tax Policy and Revenue Analysis: A Great Leap Forward,” State Tax Notes, Vol. 4 (Arlington, Virginia, May 17, 1993), pp. 118185.

    • Search Google Scholar
    • Export Citation
  • Jenkins, Glenn P., Tax Reform: Lessons Learned (Cambridge, Massachusetts: Harvard Institute for International Development, 1989).

  • Johnson, J. Walker, “Common Trends in the Taxation of Banks, Thrifts and Insurance Companies,” Taxes-The Tax Magazine, Vol. 67 (August 1989), pp. 485501.

    • Search Google Scholar
    • Export Citation
  • King, David N., Fiscal Tiers: The Economics of Multi-Level Government (Boston: Allen and Unwin, 1984).

  • King, John R., “On the Revenue Effects of Tax Changes,” Fiscal Studies, Vol. 7 (February 1986), pp. 5160.

  • Lyon, Andrew B., “The Alternative Minimum Tax: Equity, Efficiency, and Incentive Effects,” in Economic Effects of the Corporate Alternative Minimum Tax, American Council for Capital Formation, Center for Policy Research (Washington, 1991), pp. 5182.

    • Search Google Scholar
    • Export Citation
  • Mansfield, Charles Y., “Tax Administration in Developing Countries: An Economic Perspective,” Staff Papers, International Monetary Fund (Washington), Vol. 35 (March 1988), pp. 18197.

    • Search Google Scholar
    • Export Citation
  • Mansfield, Charles Y., “Tax Reform in Developing Countries: The Administrative Dimension,” Bulletin for International Fiscal Documentation, Vol. 44 (March 1990), pp. 13743.

    • Search Google Scholar
    • Export Citation
  • McLees, John A., “Fine Tuning the Mexico Assets Tax,” Tax Notes International, Vol. 3 (February 1991), pp. 11720.

  • McLure, Charles E., Jr. and Others “Net Wealth and Presumptive Taxation in Colombia,” in Taxation in Developing Countries, ed. by Richard M. Bird and Oliver Oldman (Baltimore: Johns Hopkins University Press, 4th ed., 1990), pp. 28895.

    • Search Google Scholar
    • Export Citation
  • Messere, Kenneth C, Tax Policy in OECD Countries: Choices and Conflicts (Amsterdam: International Bureau of Fiscal Documentation Publications BV, 1993).

    • Search Google Scholar
    • Export Citation
  • Musgrave, Richard A., “Who Should Tax, Where and What?,“In Tax Assignment in Federal Countries, ed. by Charles E. McLure Jr. (Canberra, Australia: Centre for Research on Federal Financial Relations, 1983), pp. 219.

    • Search Google Scholar
    • Export Citation
  • Musgrave, Richard A., “Tax Reform in Developing Countries,” in The Theory of Taxation in Developing Countries, ed. by David M. Newbery and Nicholas H. Stern (New York: Oxford University Press, 1987), pp. 24263.

    • Search Google Scholar
    • Export Citation
  • Musgrave, Richard A., “Income Taxation of the Hard-to-Tax Groups,” in Taxation in Developing Countries, ed. by Richard M. Bird and Oliver Oldman (Baltimore: Johns Hopkins University Press, 4th ed. 1990), pp. 299309.

    • Search Google Scholar
    • Export Citation
  • Mutén, Leif, et al. Morocco: Review of the Tax Reform, Fiscal Affairs Department (Washington: International Monetary Fund, 1988).

  • Naert, Patrick, La Reforme des Impots Directs en Tunisie, World Bank Study (Washington: World Bank, 1989).

  • Neubig, Thomas, and C. Eugene Steuerle The Taxation of Income Flowing Through Financial Institutions: General Framework and Summary of Tax Issues, OTA No. Paper 52 (Washington: U.S. Treasury Department, 1983).

    • Search Google Scholar
    • Export Citation
  • Oates, Wallace E., Fiscal Federalism source (New York: Harcourt Brace Jovanovich), 1972.

  • Oates, Wallace E., “Fiscal Federalism: An Overview,” in Public Finance with Several Levels of Government, ed. by Remy Prud’ homme (Brussels: International Institute of Public Finance, 1990).

    • Search Google Scholar
    • Export Citation
  • Oldman, Oliver, and Richard ML Bird, “The Transition to a Global Income Tax: A Comparative Analysis,” in Bulletin for International Fiscal Documentation, Vol.-31 (October 1977), pp. 43954.

    • Search Google Scholar
    • Export Citation
  • Organzation for Economic Cooperation and Development, A Comparative Study of Personal Income Tax Models: A Report (Paris: OECD, 1988).

  • Radian, Alex, and Ira Sharkansky, “Tax Reform in Israel: Partial Implementation of Ambitious Goals,” in Policy Analysis, Vol. 5 (Summer 1979), pp. 35166.

    • Search Google Scholar
    • Export Citation
  • Sadka, Efraim, and Vito Tanzi, “A Tax on Gross Assets of Enterprises as a Form of Presumptive Taxation,” Bulletin for International Fiscal Documentation, Vol. 47 (February 1993), pp. 6673.

    • Search Google Scholar
    • Export Citation
  • Schroeder, Larry, and Michael Wasylenko, “Public Sector Forecasting in the Third World,” International Journal of Forecasting, Vol. 5 No. 3 (1989), pp. 33345.

    • Search Google Scholar
    • Export Citation
  • Shome, Parthasarathi, The Taxation of High Income Earners, IMF Paper on Policy Analysis and Assessment No. PPAA/93/19 (Washington: International Monetary Fund, 1993).

    • Search Google Scholar
    • Export Citation
  • Shome, Parthasarathi, and Julio Escolano, The State of Tax Policy in the Central Asian and Transcaucasian Newly Independent States, IMF Paper on Policy Analysis and Assessment No. PPAA/93/8 (Washington: International Monetary Fund, 1993).

    • Search Google Scholar
    • Export Citation
  • Sicat, Gerardo P., Tax Reform in the Philippines: Paper Presented at the Conference on Tax Policy and Economic Development among Pacific Asian Countries (Taipei, 1990).

    • Search Google Scholar
    • Export Citation
  • Sunley, Emil M., and Randall D. Weiss, “The Revenue Estimating Process,” Tax Notes, Vol. 51 (June 10, 1991), pp. 1299314.

  • Tait, Alan A., Value Added Tax: International Practice and Problems (Washington: International Monetary Fund, 1988).

  • Tanzi, Vito, “Inflation, Lags in Collection and the Real Value of Tax Revenue,” Staff Papers, International Monetary Fund (Washington) Vol. 24 (March 1977), pp. 15467.

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, “Quantitative Characteristics of the Tax Systems of Developing Countries,” in The Theory of Taxation for Developing Countries, ed. by David M. Newbery and Nicholas H. Stern (New York: Oxford University Press, 1987), pp. 20541.

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, The IMF and Tax Reform, IMF Working Paper No. WP/90/39 (Washington: International Monetary Fund, 1990).

  • Tanzi, Vito, and Milka Casanegra de Jantscher, Presumptive Income Taxation: Administrative, Efficiency, and Equity Aspects, IMF Working Paper No. WP/87/54 (Washington: International Monetary Fund, 1987).

    • Search Google Scholar
    • Export Citation
  • Zemrani, Anas Bensalah, Fiscalite Face au Developpement Economique et Social du Maroc (Paris: LGJD, 1982).

  • Zins, Barry L., “Current Developments in International Taxation of Financial Services and Products,” Tax Planning International Review, Vol. 15 (February 1988), pp. 313.

    • Search Google Scholar
    • Export Citation

Under a carried interest, funds are deemed to be loaned to the government by the project investors. Interest is charged on the govern-ment’s carried interest at a prescribed rate and the loan is repayable out of the government’s share of profits from the project. The govern-ment’s equity interest only crystallizes when the “loan” is paid off. In essence, under a carried interest, government purchases equity by means of a nonrecourse loan provided by the project investors.


See Chapter IV for further discussion of this issue.


Return on capital refers to the income earned from a capital investment, while return of capital refers to the return of the original investment.


For a general treatment of the main principles involved, see King (1984), Musgrave (1983), and Oates (1972 and 1990).


Which seems to be supported empirically in the sense that, in some countries, these factors have been shown to be reflected in property values.


It could be added that macroeconomic performance is influenced at least as much by the quality of public expenditures as by its overall level.


Disregarding states or provinces in very large federal countries such as Canada and the United States.


Again, it must be emphasized that countries have chosen very different solutions with respect to tax assignment to different levels of government. Thus, a large number of countries have assigned profits taxes and/or consumption taxes to subordinate levels of govern-ment—either exclusively or under tax-sharing arrangements.


Again, in practice, many countries allow some discretion of sub-ordinate levels of government over the tax base, particularly for the provinces or states in larger federal countries (such as Australia).


Assuming that any resulting vertical gaps can be closed by higher local revenue sources. Against this change, it has been held that good candidates for local taxes (i.e., taxes on immobile resources) may be undesirable for equity or neutrality reasons.


Generally, however, available empirical analyses have not provided a clear answer to this question.


This is the general use of the term in the United States; see, for example, Sunley and Weiss (1991). In the United Kingdom, the same process is commonly referred to as “tax costing.”


For a fuller account of these methodologies, see, for example, Chand (1975); Schroeder and Wasylenko (1989); and Federation of Tax Administrators (1993).


See, for example, OECD (1988).


For a fuller discussion of these problems in the context of the United Kingdom, see King (1986).


See Chapter IV for further discussion of assets taxes.


This linkage may formally be expressed as: Tax revenue/GDP = (Tax base/GDP) x (Tax collected/Tax base). The maximization of the first term on the right-hand side of the equation represents the concern of tax policy, while that of the second term represents the concern of tax administration. This ordering brings out the key insight that the tax base must first be clearly defined through tax policy and then fully captured for revenue purposes through tax administration.


See Tables 4, 5, and 6 in the Appendix to Chapter VII.






This is generally known as the Tanzi effect. See Tanzi (1977).


Only a few countries (Benin, Mali, Niger, and Senegal) use withholding for turnover tax purposes. It is applied as a percentage of imports of the informal sector and constitutes a final tax.


For a technical description of current payments systems, see Griffith (1973).




Excluding interest and dividends subject to tax withholding.


Experience has shown this resulted in several countries, namely Argentina, Morocco, Senegal, Tunisia, and Turkey.


Exemption of the agricultural sector in VAT in most countries, including Western Europe, is an indication of this trend.


For a more detailed discussion of presumptive taxation and minimum taxes, see earlier sections of this chapter.