artwork from the cover of this publication

NO.  27

Other Titles in this Series

Also available
Tax Policy for Developing Countries
Vito Tanzi, Howell Zee

© 2001 International Monetary Fund
March 2001

[Preface]  [Tax Policy for Developing Countries
[Level of Tax Revenue]  [Composition of Tax Revenue?
[Selecting the Right Tax System
[Tax Policy Challenges Facing Developing Countries]  [Author Information]


The Economic Issues series aims to make available to a broad readership of nonspecialists some of the economic research being produced on topical issues by IMF staff. The series draws mainly from IMF Working Papers, which are technical papers produced by IMF staff members and visiting scholars, as well as from policy-related research papers.

This Economic Issue is based on IMF Working Paper 00/35 "Tax Policy for Emerging Markets," by Vito Tanzi and Howell Zee. Citations for the research referred to in this shortened version are provided in the original paper which readers can purchase (at $10.00 a copy) from the IMF Publication Services or download from David Driscoll prepared the text for this pamphlet.

Tax Policy for Developing Countries

Why do we have taxes? The simple answer is that, until someone comes up with a better idea, taxation is the only practical means of raising the revenue to finance government spending on the goods and services that most of us demand. Setting up an efficient and fair tax system is, however, far from simple, particularly for developing countries that want to become integrated in the international economy. The ideal tax system in these countries should raise essential revenue without excessive government borrowing, and should do so without discouraging economic activity and without deviating too much from tax systems in other countries.

Developing countries face formidable challenges when they attempt to establish efficient tax systems. First, most workers in these countries are typically employed in agriculture or in small, informal enterprises. As they are seldom paid a regular, fixed wage, their earnings fluctuate, and many are paid in cash, "off the books." The base for an income tax is therefore hard to calculate. Nor do workers in these countries typically spend their earnings in large stores that keep accurate records of sales and inventories. As a result, modern means of raising revenue, such as income taxes and consumer taxes, play a diminished role in these economies, and the possibility that the government will achieve high tax levels is virtually excluded.

Second, it is difficult to create an efficient tax administration without a well-educated and well-trained staff, when money is lacking to pay good wages to tax officials and to computerize the operation (or even to provide efficient telephone and mail services), and when taxpayers have limited ability to keep accounts. As a result, governments often take the path of least resistance, developing tax systems that allow them to exploit whatever options are available rather than establishing rational, modern, and efficient tax systems.

Third, because of the informal structure of the economy in many developing countries and because of financial limitations, statistical and tax offices have difficulty in generating reliable statistics. This lack of data prevents policymakers from assessing the potential impact of major changes to the tax system. As a result, marginal changes are often preferred over major structural changes, even when the latter are clearly preferable. This perpetuates inefficient tax structures.

Fourth, income tends to be unevenly distributed within developing countries. Although raising high tax revenues in this situation ideally calls for the rich to be taxed more heavily than the poor, the economic and political power of rich taxpayers often allows them to prevent fiscal reforms that would increase their tax burdens. This explains in part why many developing countries have not fully exploited personal income and property taxes and why their tax systems rarely achieve satisfactory progressivity (in other words, where the rich pay proportionately more taxes).

In conclusion, in developing countries, tax policy is often the art of the possible rather than the pursuit of the optimal. It is therefore not surprising that economic theory and especially optimal taxation literature have had relatively little impact on the design of tax systems in these countries. In discussing tax policy issues facing many developing countries today, the authors of this pamphlet consequently draw on extensive practical, first-hand experience with the IMF's provision of tax policy advice to those countries. They consider these issues from both the macroeconomic (the level and composition of tax revenue) and microeconomic (design aspects of specific taxes) perspective.

Level of Tax Revenue

What level of public spending is desirable for a developing country at a given level of national income? Should the government spend one-tenth of national income? A third? Half? Only when this question has been answered can the next question be addressed of where to set the ideal level of tax revenue; determining the optimal tax level is conceptually equivalent to determining the optimal level of government spending. Unfortunately, the vast literature on optimal tax theory provides little practical guidance on how to integrate the optimal level of tax revenue with the optimal level of government expenditure.

Nevertheless, an alternative, statistically based approach to assessing whether the overall tax level in a developing country is appropriate consists of comparing the tax level in a specific country to the average tax burden of a representative group of both developing and industrial countries, taking into account some of these countries' similarities and dissimilarities. This comparison indicates only whether the country's tax level, relative to other countries and taking into account various characteristics, is above or below the average. This statistical approach has no theoretical basis and does not indicate the "optimal" tax level for any country. The most recent data show that the tax level in major industrialized countries (members of the Organization for Economic Cooperation and Development or OECD) is about double the tax level in a representative sample of developing countries (38 percent of GDP compared with 18 percent).

Economic development will often generate additional needs for tax revenue to finance a rise in public spending, but at the same time it increases the countries' ability to raise revenue to meet these needs. More important than the level of taxation per se is how revenue is used. Given the complexity of the development process, it is doubtful that the concept of an optimal level of taxation robustly linked to different stages of economic development could ever be meaningfully derived for any country.

Composition of Tax Revenue

Turning to the composition of tax revenue, we find ourselves in an area of conflicting theories. The issues involve the taxation of income relative to that of consumption and under consumption, the taxation of imports versus the taxation of domestic consumption. Both efficiency (whether the tax enhances or diminishes the overall welfare of those who are taxed) and equity (whether the tax is fair to everybody) are central to the analysis.

The conventional belief that taxing income entails a higher welfare (efficiency) cost than taxing consumption is based in part on the fact that income tax, which contains elements of both a labor tax and a capital tax, reduces the taxpayer's ability to save. Doubt has been cast on this belief, however, by considerations of the crucial role of the length of the taxpayer's planning horizon and the cost of human and physical capital accumulation. The upshot of these theoretical considerations renders the relative welfare costs of the two taxes (income and consumption) uncertain.

Another concern in the choice between taxing income and taxing consumption involves their relative impact on equity. Taxing consumption has traditionally been thought to be inherently more regressive (that is, harder on the poor than the rich) than taxing income. Doubt has been cast on this belief as well. Theoretical and practical considerations suggest that the equity concerns about the traditional form of taxing consumption are probably overstated and that, for developing countries, attempts to address these concerns by such initiatives as graduated consumption taxes would be ineffective and administratively impractical.

With regard to taxes on imports, lowering these taxes will lead to more competition from foreign enterprises. While reducing protection of domestic industries from this foreign competition is an inevitable consequence, or even the objective, of a trade liberalization program, reduced budgetary revenue would be an unwelcome by-product of the program. Feasible compensatory revenue measures under the circumstances almost always involve increasing domestic consumption taxes. Rarely would increasing income taxes be considered a viable option on the grounds of both policy (because of their perceived negative impact on investment) and administration (because their revenue yield is less certain and less timely than that from consumption tax changes).

Data from industrial and developing countries show that the ratio of income to consumption taxes in industrial countries has consistently remained more than double the ratio in developing countries. (That is, compared with developing countries, industrial countries derive proportionally twice as much revenue from income tax than from consumption tax.) The data also reveal a notable difference in the ratio of corporate income tax to personal income tax. Industrial countries raise about four times as much from personal income tax than from corporate income tax. Differences between the two country groups in wage income, in the sophistication of the tax administration, and in the political power of the richest segment of the population are the primary contributors to this disparity. On the other hand, revenue from trade taxes is significantly higher in developing countries than in industrial countries.

While it is difficult to draw clear-cut normative policy prescriptions from international comparisons as regards the income-consumption tax mix, a compelling implication revealed by the comparison is that economic development tends to lead to a relative shift in the composition of revenue from consumption to personal income taxes. At any given point of time, however, the important tax policy issue for developing countries is not so much to determine the optimal tax mix as to spell out clearly the objectives to be achieved by any contemplated shift in the mix, to assess the economic consequences (for efficiency and equity) of such a shift, and to implement compensatory measures if the poor are made worse off by the shift.

Selecting the Right Tax System

In developing countries where market forces are increasingly important in allocating resources, the design of the tax system should be as neutral as possible so as to minimize interference in the allocation process. The system should also have simple and transparent administrative procedures so that it is clear if the system is not being enforced as designed.

Personal Income Tax

Any discussion of personal income tax in developing countries must start with the observation that this tax has yielded relatively little revenue in most of these countries and that the number of individuals subject to this tax (especially at the highest marginal rate) is small. The rate structure of the personal income tax is the most visible policy instrument available to most governments in developing countries to underscore their commitment to social justice and hence to gain political support for their policies. Countries frequently attach great importance to maintaining some degree of nominal progressivity in this tax by applying many rate brackets, and they are reluctant to adopt reforms that will reduce the number of these brackets.

More often than not, however, the effectiveness of rate progressivity is severely undercut by high personal exemptions and the plethora of other exemptions and deductions that benefit those with high incomes (for example, the exemption of capital gains from tax, generous deductions for medical and educational expenses, the low taxation of financial income). Tax relief through deductions is particularly egregious because these deductions typically increase in the higher tax brackets. Experience compellingly suggests that effective rate progressivity could be improved by reducing the degree of nominal rate progressivity and the number of brackets and reducing exemptions and deductions. Indeed, any reasonable equity objective would require no more than a few nominal rate brackets in the personal income tax structure. If political constraints prevent a meaningful restructuring of rates, a substantial improvement in equity could still be achieved by replacing deductions with tax credits, which could deliver the same benefits to taxpayers in all tax brackets.

The effectiveness of a high marginal tax rate is also much reduced by its often being applied at such high levels of income (expressed in shares of per capita GDP) that little income is subject to these rates. In some developing countries, a taxpayer's income must be hundreds of times the per capita income before it enters the highest rate bracket.

Moreover, in some countries the top marginal personal income tax rate exceeds the corporate income tax by a significant margin, providing strong incentives for taxpayers to choose the corporate form of doing business for purely tax reasons. Professionals and small entrepreneurs can easily siphon off profits through expense deductions over time and escape the highest personal income tax permanently. A tax delayed is a tax evaded. Good tax policy, therefore, ensures that the top marginal personal income tax rate does not differ materially from the corporate income tax rate.

In addition to the problem of exemptions and deductions tending to narrow the tax base and to negate effective progressivity, the personal income tax structure in many developing countries is riddled with serious violations of the two basic principles of good tax policy: symmetry and inclusiveness. (It goes without saying, of course, that tax policy should also be guided by the general principles of neutrality, equity, and simplicity.) The symmetry principle refers to the identical treatment for tax purposes of gains and losses of any given source of income. If the gains are taxable, then the losses should be deductible. The inclusiveness principle relates to capturing an income stream in the tax net at some point along the path of that stream. For example, if a payment is exempt from tax for a payee, then it should not be a deductible expense for the payer. Violating these principles generally leads to distortions and inequities.

The tax treatment of financial income is problematic in all countries. Two issues dealing with the taxation of interest and dividends in developing countries are relevant:

  • In many developing countries, interest income, if taxed at all, is taxed as a final withholding tax at a rate substantially below both the top marginal personal and corporate income tax rate. For taxpayers with mainly wage income, this is an acceptable compromise between theoretical correctness and practical feasibility. For those with business income, however, the low tax rate on interest income coupled with full deductibility of interest expenditure implies that significant tax savings could be realized through fairly straightforward arbitrage transactions. Hence it is important to target carefully the application of final withholding on interest income: final withholding should not be applied if the taxpayer has business income.
  • The tax treatment of dividends raises the well-known double taxation issue. For administrative simplicity, most developing countries would be well advised either to exempt dividends from the personal income tax altogether, or to tax them at a relatively low rate, perhaps through a final withholding tax at the same rate as that imposed on interest income.

Corporate Income Tax

Tax policy issues relating to corporate income tax are numerous and complex, but particularly relevant for developing countries are the issues of multiple rates based on sectoral differentiation and the incoherent design of the depreciation system. Developing countries are more prone to having multiple rates along sectoral lines (including the complete exemption from tax of certain sectors, especially the parastatal sector) than industrial countries, possibly as a legacy of past economic regimes that emphasized the state's role in resource allocation. Such practices, however, are clearly detrimental to the proper functioning of market forces (that is, the sectoral allocation of resources is distorted by differences in tax rates). They are indefensible if a government's commitment to a market economy is real. Unifying multiple corporate income tax rates should thus be a priority.

Allowable depreciation of physical assets for tax purposes is an important structural element in determining the cost of capital and the profitability of investment. The most common shortcomings found in the depreciation systems in developing countries include too many asset categories and depreciation rates, excessively low depreciation rates, and a structure of depreciation rates that is not in accordance with the relative obsolescence rates of different asset categories. Rectifying these shortcomings should also receive a high priority in tax policy deliberations in these countries.

In restructuring their depreciation systems, developing countries could well benefit from certain guidelines:

  • Classifying assets into three or four categories should be more than sufficient—for example, grouping assets that last a long time, such as buildings, at one end, and fast-depreciating assets, such as computers, at the other with one or two categories of machinery and equipment in between.
  • Only one depreciation rate should be assigned to each category.
  • Depreciation rates should generally be set higher than the actual physical lives of the underlying assets to compensate for the lack of a comprehensive inflation-compensating mechanism in most tax systems.
  • On administrative grounds, the declining-balance method should be preferred to the straight-line method. The declining-balance method allows the pooling of all assets in the same asset category and automatically accounts for capital gains and losses from asset disposals, thus substantially simplifying bookkeeping requirements.

Value-Added Tax, Excises, and Import Tariffs

While VAT has been adopted in most developing countries, it frequently suffers from being incomplete in one aspect or another. Many important sectors, most notably services and the wholesale and retail sector, have been left out of the VAT net, or the credit mechanism is excessively restrictive (that is, there are denials or delays in providing proper credits for VAT on inputs), especially when it comes to capital goods. As these features allow a substantial degree of cascading (increasing the tax burden for the final user), they reduce the benefits from introducing the VAT in the first place. Rectifying such limitations in the VAT design and administration should be given priority in developing countries.

Many developing countries (like many OECD countries) have adopted two or more VAT rates. Multiple rates are politically attractive because they ostensibly—though not necessarily effectively—serve an equity objective, but the administrative price for addressing equity concerns through multiple VAT rates may be higher in developing than in industrial countries. The cost of a multiple-rate system should be carefully scrutinized.

The most notable shortcoming of the excise systems found in many developing countries is their inappropriately broad coverage of
products—often for revenue reasons. As is well known, the economic rationale for imposing excises is very different from that for imposing a general consumption tax. While the latter should be broadly based to maximize revenue with minimum distortion, the former should be highly selective, narrowly targeting a few goods mainly on the grounds that their consumption entails negative externalities on society (in other words, society at large pays a price for their use by individuals). The goods typically deemed to be excisable (tobacco, alcohol, petroleum products, and motor vehicles, for example) are few and usually inelastic in demand. A good excise system is invariably one that generates revenue (as a by-product) from a narrow base and with relatively low administrative costs.

Reducing import tariffs as part of an overall program of trade liberalization is a major policy challenge currently facing many developing countries. Two concerns should be carefully addressed. First, tariff reduction should not lead to unintended changes in the relative rates of effective protection across sectors. One simple way of ensuring that unintended consequences do not occur would be to reduce all nominal tariff rates by the same proportion whenever such rates need to be changed. Second, nominal tariff reductions are likely to entail short-term revenue loss. This loss can be avoided through a clear-cut strategy in which separate compensatory measures are considered in sequence: first reducing the scope of tariff exemptions in the existing system, then compensating for the tariff reductions on excisable imports by a commensurate increase in their excise rates, and finally adjusting the rate of the general consumption tax (such as the VAT) to meet remaining revenue needs.

Tax Incentives

While granting tax incentives to promote investment is common in countries around the world, evidence suggests that their effectiveness in attracting incremental investments—above and beyond the level that would have been reached had no incentives been granted—is often questionable. As tax incentives can be abused by existing enterprises disguised as new ones through nominal reorganization, their revenue costs can be high. Moreover, foreign investors, the primary target of most tax incentives, base their decision to enter a country on a whole host of factors (such as natural resources, political stability, transparent regulatory systems, infrastructure, a skilled workforce), of which tax incentives are frequently far from being the most important one. Tax incentives could also be of questionable value to a foreign investor because the true beneficiary of the incentives may not be the investor, but rather the treasury of his home country. This can come about when any income spared from taxation in the host country is taxed by the investor's home country.

Tax incentives can be justified if they address some form of market failure, most notably those involving externalities (economic consequences beyond the specific beneficiary of the tax incentive). For example, incentives targeted to promote high-technology industries that promise to confer significant positive externalities on the rest of the economy are usually legitimate. By far the most compelling case for granting targeted incentives is for meeting regional development needs of these countries. Nevertheless, not all incentives are equally suited for achieving such objectives and some are less cost-effective than others. Unfortunately, the most prevalent forms of incentives found in developing countries tend to be the least meritorious.

Tax Holidays

Of all the forms of tax incentives, tax holidays (exemptions from paying tax for a certain period of time) are the most popular among developing countries. Though simple to administer, they have numerous shortcomings. First, by exempting profits irrespective of their amount, tax holidays tend to benefit an investor who expects high profits and would have made the investment even if this incentive were not offered. Second, tax holidays provide a strong incentive for tax avoidance, as taxed enterprises can enter into economic relationships with exempt ones to shift their profits through transfer pricing (for example, overpaying for goods from the other enterprise and receiving a kickback). Third, the duration of the tax holiday is prone to abuse and extension by investors through creative redesignation of existing investment as new investment (for example, closing down and restarting the same project under a different name but with the same ownership). Fourth, time-bound tax holidays tend to attract short-run projects, which are typically not so beneficial to the economy as longer-term ones. Fifth, the revenue cost of the tax holiday to the budget is seldom transparent, unless enterprises enjoying the holiday are required to file tax forms. In this case, the government must spend resources on tax administration that yields no revenue and the enterprise loses the advantage of not having to deal with tax authorities.

Tax Credits and Investment Allowances

Compared with tax holidays, tax credits and investment allowances have a number of advantages. They are much better targeted than tax holidays for promoting particular types of investment and their revenue cost is much more transparent and easier to control. A simple and effective way of administering a tax credit system is to determine the amount of the credit to a qualified enterprise and to "deposit" this amount into a special tax account in the form of a bookkeeping entry. In all other respects the enterprise will be treated like an ordinary taxpayer, subject to all applicable tax regulations, including the obligation to file tax returns. The only difference would be that its income tax liabilities would be paid from credits "withdrawn" from its tax account. In this way information is always available on the budget revenue forgone and on the amount of tax credits still available to the enterprise. A system of investment allowances could be administered in much the same way as tax credits, achieving similar results.

There are two notable weaknesses associated with tax credits and investment allowances. First, these incentives tend to distort choice in favor of short-lived capital assets since further credit or allowance becomes available each time an asset is replaced. Second, qualified enterprises may attempt to abuse the system by selling and purchasing the same assets to claim multiple credits or allowances or by acting as a purchasing agent for enterprises not qualified to receive the incentive. Safeguards must be built into the system to minimize these dangers.

Accelerated Depreciation

Providing tax incentives in the form of accelerated depreciation has the least of the shortcomings associated with tax holidays and all of the virtues of tax credits and investment allowances—and overcomes the latter's weakness to boot. Since merely accelerating the depreciation of an asset does not increase the depreciation of the asset beyond its original cost, little distortion in favor of short-term assets is generated. Moreover, accelerated depreciation has two additional merits. First, it is generally least costly, as the forgone revenue (relative to no acceleration) in the early years is at least partially recovered in subsequent years of the asset's life. Second, if the acceleration is made available only temporarily, it could induce a significant short-run surge in investment.

Investment Subsidies

While investment subsidies (providing public funds for private investments) have the advantage of easy targeting, they are generally quite problematic. They involve out-of-pocket expenditure by the government up front and they benefit nonviable investments as much as profitable ones. Hence, the use of investment subsidies is seldom advisable.

Indirect Tax Incentives

Indirect tax incentives, such as exempting raw materials and capital goods from the VAT, are prone to abuse and are of doubtful utility. Exempting from import tariffs raw materials and capital goods used to produce exports is somewhat more justifiable. The difficulty with this exemption lies, of course, in ensuring that the exempted purchases will in fact be used as intended by the incentive. Establishing export production zones whose perimeters are secured by customs controls is a useful, though not entirely foolproof, remedy for this abuse.

Triggering Mechanisms

The mechanism by which tax incentives can be triggered can be either automatic or discretionary. An automatic triggering mechanism allows the investment to receive the incentives automatically once it satisfies clearly specified objective qualifying criteria, such as a minimum amount of investment in certain sectors of the economy. The relevant authorities have merely to ensure that the qualifying criteria are met. A discretionary triggering mechanism involves approving or denying an application for incentives on the basis of subjective value judgment by the incentive-granting authorities, without formally stated qualifying criteria. A discretionary triggering mechanism may be seen by the authorities as preferable to an automatic one because it provides them with more flexibility. This advantage is likely to be outweighed, however, by a variety of problems associated with discretion, most notably a lack of transparency in the decision-making process, which could in turn encourage corruption and rent-seeking activities. If the concern about having an automatic triggering mechanism is the loss of discretion in handling exceptional cases, the preferred safeguard would be to formulate the qualifying criteria in as narrow and specific a fashion as possible, so that incentives are granted only to investments meeting the highest objective and quantifiable standard of merit. On balance, it is advisable to minimize the discretionary element in the incentive-granting process.

Summing Up

The cost-effectiveness of providing tax incentives to promote investment is generally questionable. The best strategy for sustained investment promotion is to provide a stable and transparent legal and regulatory framework and to put in place a tax system in line with international norms. Some objectives, such as those that encourage regional development, are more justifiable than others as a basis for granting tax incentives. Not all tax incentives are equally effective. Accelerated depreciation has the most comparative merits, followed by investment allowances or tax credits. Tax holidays and investment subsidies are among the least meritorious. As a general rule, indirect tax incentives should be avoided, and discretion in granting incentives should be minimized.

Tax Policy Challenges Facing Developing Countries

Developing countries attempting to become fully integrated in the world economy will probably need a higher tax level if they are to pursue a government role closer to that of industrial countries, which, on average, enjoy twice the tax revenue. Developing countries will need to reduce sharply their reliance on foreign trade taxes, without at the same time creating economic disincentives, especially in raising more revenue from personal income tax. To meet these challenges, policymakers in these countries will have to get their policy priorities right and have the political will to implement the necessary reforms. Tax administrations must be strengthened to accompany the needed policy changes.

As trade barriers come down and capital becomes more mobile, the formulation of sound tax policy poses significant challenges for developing countries. The need to replace foreign trade taxes with domestic taxes will be accompanied by growing concerns about profit diversion by foreign investors, which weak provisions against tax abuse in the tax laws as well as inadequate technical training of tax auditors in many developing countries are currently unable to deter. A concerted effort to eliminate these deficiencies is therefore of the utmost urgency.

Tax competition is another policy challenge in a world of liberalized capital movement. The effectiveness of tax incentives—in the absence of other necessary fundamentals—is highly questionable. A tax system that is riddled with such incentives will inevitably provide fertile grounds for rent-seeking activities. To allow their emerging markets to take proper root, developing countries would be well advised to refrain from reliance on poorly targeted tax incentives as the main vehicle for investment promotion.

Finally, personal income taxes have been contributing very little to total tax revenue in many developing countries. Apart from structural, policy, and administrative considerations, the ease with which income received by individuals can be invested abroad significantly contributes to this outcome. Taxing this income is therefore a daunting challenge for developing countries. This has been particularly problematic in several Latin American countries that have largely stopped taxing financial income to encourage financial capital to remain in the country.


Author Information


Vito Tanzi was the Director of the Fiscal Affairs Department of the IMF from 1981 to 2000. He retired from the IMF on December 1, 2000. He holds a Ph.D. from Harvard University and is the author of many books and articles in professional journals.

Vito Tanzi
Howell Zee is the Chief of the Tax Policy Division in the IMF's Fiscal Affairs Department. He holds a Ph.D. from the University of Maryland (College Park) and is the author of many articles in professional journals. Howell Zee