Taxation in Latin America: Structural Trends and Impact of Administration

Contributor Notes

Author’s E-Mail Address:

From the mid-1980s to early 1990s, Latin American tax policy provided rich lessons for other reforming countries. Meaningful innovations led also to perceptible revenue gains. Later in the 1990s, tax policies began to drift. Shining examples of fundamental reform seemed to lose their luster. Revenue in terms of GDP also stagnated, partly reflecting over-reliance on consumption taxes and neglect of taxable capacity on incomes. The stagnation has been exacerbated by excessively simplified administrative practices. Based on these developments and on the limited taxability of internationally mobile capital, the paper anticipates a likely tax structure for the new century.


From the mid-1980s to early 1990s, Latin American tax policy provided rich lessons for other reforming countries. Meaningful innovations led also to perceptible revenue gains. Later in the 1990s, tax policies began to drift. Shining examples of fundamental reform seemed to lose their luster. Revenue in terms of GDP also stagnated, partly reflecting over-reliance on consumption taxes and neglect of taxable capacity on incomes. The stagnation has been exacerbated by excessively simplified administrative practices. Based on these developments and on the limited taxability of internationally mobile capital, the paper anticipates a likely tax structure for the new century.

I. Introduction

During the 1980s and 1990s, Latin American tax policy changes have held the interest of tax experts and researchers without ceasing to arouse scrutinization and study, making the Latin American experience rich in lessons to be drawn by other reforming countries. This is because Latin American economies have often taken bold steps towards fundamental tax reform, for example, Bolivia after its 1985 hyperinflation, or Argentina, Colombia, and Peru in the early 1990s, through a series of meaningful changes in the tax structure and in tax administration. Many have demonstrated a capacity for innovation and experimentation in particular tax concepts and constructs such as Brazil with an inter-governmental VAT, or Chile with a single customs duty, or Mexico with its minimum corporate tax based on gross assets.

At the same time, however, some countries have experienced periods of perceptible slowdown in their tax reform effort or in the optimal design of taxes, for example, Argentina with its tax on bank checks with other countries such as Brazil debating over, and then introducing, similar tax concepts such as its prevailing tax on financial transactions, or an inability in many to design or collect appropriate property taxes. On the whole, the rich experience of Latin America, therefore, provides many examples of what policies to undertake and what to avoid in structuring a tax system and in its application (CEPAL, 1998). In this context, given their recent experiences, an area of interest is how they themselves are likely to go forward in further modernizing their tax structures as they go into the twenty-first century.

Nevertheless, a particular area of increasing concern that seems to have emerged pari passu with tax reform is how the tax structure is effectively administered in the field i.e., how closely does the administered system resemble the legal structure. This concern pertains not only to Latin America, but also to Asia and Africa, together with many European countries. The paper takes up this issue with particular reference to Latin America, since it should have important implications for the modernization of their tax systems when seen in their totality, i.e., tax policy in combination with tax administration.

Among the various cross-country studies on Latin America may be mentioned Bird (1992), González-Cano (1996), Perry and Herrera (1994), Rodríguez (1993), Shome (1992, 1995a), Turro (1993), and others. There are also many helpful individual country studies including Morisset and Izquierdo (1993) and Durán and Gómez-Sabaini (1994) for Argentina, Harberger (1988) and Mann (1990) for Bolivia, Canto (1989) and Shome and Spahn (1997) for Brazil, Boylan (1992) and Toro (1994) for Chile, Shome (1995b) and Weizman (1994) for Colombia, Aguirre and Shome (1988) and McLees (1991) for Mexico, and various individual country studies that appeared in CIEDLA (1995). These studies comprise a testament to the breadth of tax policy experiences in these countries.

This author’s 1992 paper, originally delivered at an annual conference of the Center for Inter-American Tax Administrators (CIAT), examined the sweep of Latin American tax reform over the 1980s and identified common traits as well as an agenda for future reform. His 1995 paper, delivered at an International Monetary Fund (IMF) sponsored conference with the Brazilian authorities on Latin American economic reforms, took stock of further progress in tax reform. He has been asked now by the U.N. Commission for Economic Policy in Latin America (CEPAL) to look towards the future for Latin American tax reform. That is the objective of this paper.

In what follows, Section II summarizes the changes in the structure of selected taxes and attempts to draw conclusions regarding the major trends. Section III addresses the issue of administrative simplification of the legal tax structure, with important ramifications to economic efficiency and distributional equity. This, in turn, raises the question of what is an advanced tax system and which Latin American countries would fit such a categorization as they enter the next century. Section IV deals in some detail with issues that are likely to preoccupy the tax authorities in the evolution of tax structures as they move into the twenty-first century. Section V makes selected concluding remarks on tax policy improvements and cautions against oversimplification in tax administration.

II. Structures of Selected Taxes

Many of the earlier trends in the structure of taxes illustrated in Shome (1992, 1995a) have continued to hold in the 1990s. However, looking at the income tax, these changes are now occurring at a somewhat diminished pace for the personal income tax, but more rapidly for the corporate income tax, while the primacy of consumption taxes, and the value added tax (VAT) in particular, continues. This is illustrated in the accompanying tables. Using recent data, they indicate the changes in the rates of the personal income tax, the corporate income tax and the VAT, the level of personal exemption in proportion of per capita GDP, the treatment of various capital-based taxes, and the tax burdens in terms of GDP, together with the weight of various taxes in the overall structure.

A. Income and Capital Taxes

First, it is readily seen that the top personal income tax rate has continued to fall further: from an average of 50 percent in 1985–86 to 38 percent in 1991 and 34 percent in 1997, a rate of decline considerably more rapid than in the OECD, where the same overall trend has occurred (Table 1).2 This continuing decline has been experienced in the 1991–97 period in Chile, El Salvador, Guatemala, Mexico, Nicaragua, Panama, Peru, and Venezuela. Only Argentina experienced a small increase in the top rate during this period, with Paraguay and Uruguay initiating the tax in a limited, scheduler form. The net result has been a significant decline in the average top marginal rate. However, the lowest marginal rate which had declined from an average rate of 7 percent in 1985–86 to 6 percent in 1991, went up again to 9 percent in 1997.

Table 1.

Personal Income Tax Rates, 1985 or 1986, 1991 and 1997

(Percent of taxable Income)

article image
Sources: Secondary published sources such as publications of tax summaries by Price Waterhouse, Coopers and Lybrand, International Bureau of Fiscal Documentation, International Financial Statistics (IFS) of the IMF; and other similar sources.

The average shown is a joint average of the two years.

Second, while the top marginal tax rate has been brought down during the 1990s, this top rate has begun to be applied at a lower multiple of per-capita GDP as might be expected (Table 2). At the same time, the raising of the lowest marginal rate during the 1990s, is also consistent with the personal exemption level in terms of per capita GDP being steadily raised in many countries, from 1.29 in 1991, and 1.36 in 1997 (Table 2).3 This leaves a greater number of potential taxpayers out of the tax net and, therefore, begins the coverage of the first bracket at a higher marginal rate. There is an important ramification of this structural change if carried too far, however, as seems to have occurred in some of the countries. Thus, the universe of personal income tax payers has decreased in many countries, calling for an emergent need to expand the number of taxpayers, as will be elaborated upon in Section III. In turn, this seems to have contributed to continuing low reliance on the personal income tax for revenue in most countries.

Table 2.

Personal Income Tax: Exemption Level and Upper Income Bracket, 1985 or 1986, 1991 and 1997

(Multiples of per capita GDP)

article image
Sources: Secondary, published sources such as publications of tax summaries by Price Waterhouse, Coopers and Lybrand, International Bureau of Fiscal Documentation, Governament Finance Statistics (GFS) of the IMF; and other similar sources.

Average are taken over the set of countries for which data for 1997, 1991 and 1985 or 1986 are available.

Allowanc equals 12 months minimum wage in zone of residence (13 months with Christmas bonus). The data provided correspond to the Federal District.

However, other structural reasons also result in low revenue from the personal income tax. For example, a major factor explaining this phenomenon appears to be the inability of the authorities to remove existing personal allowances, deductions, and incentives that erode the tax base, even as tax rates have been reduced. In addition, poor structural links between the personal and corporate income tax such as in Ecuador, or the lack of clarity in the definition of income such as in Argentina, could also exacerbate the problem.

Third, the rate of the corporate income tax also continued to fall on average, falling from 43.3 percent in 1986 to 36.5 percent in 1992, and to 27.6 percent in 1997 (Table 3). Its rate of decline between 1992 and 1997 (especially after 1994)4 has been quite significant, thereby moving the top personal income tax rate considerably above that of the corporate income tax rate on average. This had been true even in the 1980s, but the difference has been getting more marked in the 1990s. These trends could be expected especially in the context of globalization and the commensurate difficulty in taxing factors of production that are mobile across international boundaries, such as capital or high-income professionals who are likely to fall in the highest tax bracket, a matter addressed further in Section IV.

Table 3.

Corporate Income Tax Rates, 1986, 1992 and 1997

(Percent of taxable profits)

article image
Sources: Secondary published sources such as publications of tax summaries by Price Waterhouse, Coopers and Lybrand, International Bureau of Fiscal Documentation, International Financial Statistics (IFS) of the IMF’; and other similar sources.

Other charges and contributions effectively make the tax rate higher.

Fourth, the trend towards easing the tax burden on capital is revealed through changing structures in other forms of taxation of capital as well. For example, capital gains tend to be either taxed at normal income tax rates, or are exempt, or taxed at lower-than-normal rates (Table 4). Similarly, withholding tax rates on remittances of dividends, interest, and royalties from capital have also continued to decline (Table 5). This trend of lower taxation of capital is also visible in the form of fewer countries applying capital taxes on net worth or assets in the 1990s than in the 1980s (Table 6). Mexico has become almost unique in the late 1990s in its continued successful use of gross assets as a base for a minimum tax on capital. Even that has often faced a rocky terrain, being tested through the justice system and having to restructure the tax to accommodate the concerns of foreign investors. Other countries such as Argentina had implemented it earlier, but subsequently repealed it, mainly on political grounds and, though a few are considering its introduction again, the overall prospects for the minimum corporate tax based on gross assets, seem to be bleak for the foreseeable future.

Table 4.

Treatment of Capital Gains, 1980, 1991 and 1997

(Percent of capital gains)

article image
Sources: Secondary published sources such as publications of tax summaries by Price Waterhouse, Coopers and Lybrand, International Bureau of Fiscal Documentation, International Financial Statistics (IFS) of the IMF; and other similar sources.

Less than normal corporate tax rate.

“Normal” throughout the table indicates that the prevailing income tax rate is applicable.

Table 5.

Withholding Taxes on Foreign Remittances, 1986, 1992, and 1997

article image
Sources: Secondary, published sources such as publications of tax summaries by Price Waterhouse.

Simple average of figures presented.

Pertains to dividends in cash or kind, other than stock dividends. The beneficiary must be identified: otherwise, the rate is 22.25 percent. Dividends and remittances of branch profits in excess of 12 percent of registered investment are subject to a special remittance tax ranging from 15 percent t 25 percent.

Services derived from agreements ruled by the Foreign Technology Law: (a) technical assistance, technology, and engineering—27 percent (45 percent on assumed profit of 60 percent); (b) cession of right or licenses for inventions, patents, exploitation, and others—36 percent (45 percent on assumed profit of 80 percent); and (c) nonregistered agreements—45 percent (profit of 100 percent is assumed).

These rates have been established as of January 1, 1996. Treaty rates in excess of those of in force for nontreaty countries are automatically reduced.

Taxes on dividends are withheld at the basic tax rate with surcharges. If the dividends are paid from undistributed profits of prior years, credit is allowed for the tax already paid on such profits by the company.

No withholding required on interest remitted or credited abroad on loans. A special tax of 0.5 percent to 2 percent on the portion of the loans payable up to two years is levied (only once) at the time loans are registered at the Central Bank of Ecuador. If the loan is due after two years, the special tax is not payable.

Interest on cash foreign-source loans brought into the country is not subject to withholding taxes.

The withholding taxes are an average of different interest and royalties rates.

Interest payments to nonresidents are exempt of Mexican income tax in the case of (a) loans to the federal government; (b) fixed-rate loans for five or more years, by duly registered financial institutions; and (c) certain securities and bank acceptances issued in foreign currency.

Interest payments to nonresidents are exempt from Mexican income tax for (a) loans to the federal government; and (b) loans for three or more years by duly registered financial entities that promote exports by special financing; (c) these gains are taxable as interest; (d) when royalties are paid for the use of patents in connection with the technical assistance required for their use under the same contract, both the licensing fee and amounts paid for the technical assistance will be subject to the lower 15 percent rate; (e) the nonresident taxpayer may elect to pay at the regular corporate tax rate on net profit if he has a resident representative and advises the customer accordingly. The latter, then, makes no withholding.

Interest payments to nonresidents are exempt from Mexican income tax in the case of: (a) loans to the federal government; and (b) loans for three or more years by duly registered financial entities that promote exports by special financing; (c) preferential loans granted or guaranteed by foreign financial entities to institutions authorized to receive tax-deductible donations in Mexico, provided that these institutions are properly registered and use the funds for purposes consistent with their status.

The election if available only if the payee is a resident taxpayer of a country that has signed an income tax treaty with Mexico and the treaty is in force.

35 percent income tax must be withheld on payments made to foreign persons or entities located in low-tax jurisdictions.

When these payments involve items on which tax must be withheld at either the 15 percent of the 35 percent rate, the tax must be calculated by applying the applicable rates to the payments made for each of the corresponding items; when no distinction can be made, the 35 percent rate must be applied to the total payment.

The alienation (even as a capital contribution) of drawings, models, plans, formulas, or procedures is treated as an authorization for their use; accordingly, the corresponding amount is taxed at the 15 percent rate.

Under certain circumstances, exemptions are granted.

Taxable income is determined as gross rentals less depreciation computed as provided by law.

Payments for transfer of technology or for information regarding commercial, industrial, or scientific knowledge are deemed to be royalties.

Table 6.

Net Worth or Assets Tax, 1986, 1992 and 1997

(In percent)

article image
Sources: Secondary published sources such as publications of tax summaries by Price Waterhouse, Coopers and Lybrand, International Bureau of Fiscal Documentation, International Financial Statistics (IFS) of the IMF; and other similar sources.

Minimum corporate income tax; can be credited against normal corporate tax. In Mexico, the income tax can be credited against the gross assets tax in order to avoid the foreign investors’ problem of crediting against tax liability in the home country.

The base is real estate. The tax, however, is conceived not as a property tax but as an additional corporate tax.

This tax has the form of a license to do business. The maximum tax amount is US$ 20,000 per year.

B. Consumption Taxes

In general, taxes on the domestic consumption of goods and services comprised the most important component of tax revenue in Latin American countries during the 1980s as many of them introduced the VAT which replaced cascading turnover taxes and most simplified prevailing complex income tax structures by reducing rates and raising personal exemption levels. This characteristic has continued into the 1990s: looking at consumption taxation and, in particular, the VAT, its standard rate in most countries has gone up, the few exceptions being Chile, Paraguay and Peru (Table 7). Certainly in such countries, recent improvements in tax administration may have had an impact on the feasibility to maintain or even decrease VAT rates. On the whole, however, it is important to note that VAT productivity, defined as VAT revenue as a percent of GDP divided by VAT tax rate, cannot be said to have increased significantly over the decade. Chile’s VAT productivity has been the highest at about 50 percent; but in other countries, it has mainly hovered between 30 percent–40 percent, reflecting a mix of causes including complex structures represented by base erosion due to exemptions, tax evasion, and gaps in administration. It is clear that the VAT is exhibiting signs of aging in many countries, and needs thorough reform to reduce exemptions and improve administration. Nevertheless, it is obvious that, the objective of tax policy has been to continue to rely more heavily for revenue on the VAT than on income taxes, a trend that had been pointed out earlier, in Shome (1992, 1995a).

Table 7.

Cross-Country Comparisons: Value-Added Tax Rates 1/

(In percent)

article image
Sources: Fiscal Affairs Department, IMF.

Rates shown in bold type are so-called effective standard rates (tax exclusive) applied to goods and services not covered by other especially high or low rates. Some countries use a zero rate for a few goods, and tax exports.

Supplementary VAT rates of 8 percent and 9 percent on noncapital goods imports; through “catch-up,” these can revert to 18 percent retail.

Effective rate (legislated tax-inclusive rate is 13 percent).

On interstate transactions depending on region.

On intrastate transactions.

Venezuela was the last country to introduce a VAT in October 1993, had removed it by March 1994, but reintroduced it soon thereafter.

The tendency to rely heavily on consumption taxes compared to income taxes (two exceptions being Mexico and Panama) is also illustrated by a cross-country comparison of tax revenue in proportion to GDP (Table 8).5 On average during the 1990s, revenue from the VAT and similar taxes, in terms of GDP, has been significantly higher than either personal income tax or corporate income tax and, when combined, revenue from taxation of goods and services has been significantly higher than income tax revenue. While, initially, a focus on the VAT represented an improvement in overall efficiency, today’s underutilization of the income tax could be said to reflect a combination of factors, including structural deficiencies such as a preponderance of exemptions, deductions, and allowances that erode the tax base, as well as an over-simplification on administrative grounds, a point taken up in Section III. Economically less advanced countries such as Bolivia may use it less as a revenue source than as an instrument to cross-check and control the VAT. But even economically more advanced countries such as Argentina and Brazil continue to possess greater potential than they actually collect from this tax, a practice that contradicts the general virtue of diversification in revenue sources.

Table 8.

Tax to GDP Ratios, 1992 and 1996

article image
Source: International published documents and Fund staff estimates.

In some countries, the social security institution operates outside central government. For comparability, it is shown under central government, here.

Includes state-level VAT.

Includes miscellaneous taxes: for Uruguay, on goods and services, financial assets, and sale of foreign exchange; for Argentina, other indirect taxes; for Bolivia, transactions tax mainly; for Colombia, stamp and some other taxes; for Mexico, fiscal fines and other taxes; for Panama, stamp and other taxes; for Guatemala, stamp and other taxes; for Paraguay, stamp and other taxes; for Peru, it includes the combined effect of other taxes and tax credit.

Includes VAT and transactions tax paid by state oil company.

Includes excise and VAT paid by oil company.

Includes income and excise tax paid by oil company.

C. Drift in Tax Policy

It had been pointed out in Shome (1992) that, many countries that had undertaken tax reform in the 1980s, experienced an increase in their tax-GDP ratios in the range of 2-4 percent of GDP, within about five years of undertaking reform. Some countries could have had revenue generation as a specific objective of reform, but in others, the objective might not necessarily have been revenue enhancement. The revenue increase could be attributed, then, to the likelihood that tax reform was carried out in an overall reform environment in the economy, with improvements taking place at the same time in many sectors, including the fiscal sector. This could have generated the higher incomes necessary to yield revenues in proportions higher than income growth.

Table 8 indicates that a similar experience for the 1990s is difficult to perceive.6 As in the earlier case, Latin America can still conveniently be divided into three groups—high tax, medium tax, and low tax—of countries reflecting their tax-GDP ratios in the early 1990s.7 While it is true that some countries have had a significant increase in the tax-GDP ratio, such as Bolivia, Paraguay, and Venezuela, they have been traditionally low-tax countries attempting to reduce their dependence on petroleum revenue or simply to catch up with others, or Peru, which has been in a recuperative phase since its revenue collapse during the heterodox period of the late 1980s. Colombia has gained too, though this not only reflects reform in fiscal federalism combined with complementary measures in the VAT and social security taxation, but also revenue pressures in the face of rapidly increasing expenditure. Other countries that have experienced a small growth in revenue are Brazil, Costa Rica, Guatemala, and Nicaragua, but remaining well within their groups, and not necessarily reflective of any major reform effort. Still others have experienced no growth or even a decline in their tax-GDP ratios, including Argentina, Chile, Ecuador, Mexico, Panama, and Uruguay.

Thus, around the mid-1990s, the broad experience seems to be one in which countries that have traditionally lingered behind in their revenue effort, have been attempting to catch up, combined with a drift in tax policy in others, without many shining examples of fundamental or continuing reforms that improve the efficiency or equity of the overall tax structure. Given a traditionally strong demonstration effect across Latin America, it is not surprising, therefore, that the late 1990s appear to be a period experiencing the justification and introduction of awkward or inefficient taxes, and continuing tax-base erosion, rather than one of fundamental improvement. It is important, therefore, for a few important countries to address tax reform as a renewed challenge and embrace the necessary systemic changes, for other countries to follow. Of course, even without the appearance of leading cases, there is no reason why smaller countries cannot begin fundamental reform themselves; and, indeed, some of them such as Bolivia and Paraguay have initiated technical studies with that objective.

Among the issues to be raised in any revitalized reform process are, first, whether it is advisable to continue to rely so heavily on goods and services taxes, second, which among the income taxes might be focussed upon in the future for enhanced revenue generation and, third, if a country does not have a broad-based personal income tax, whether it should be introduced. This is especially so since it is clear that, in some of the sample countries, the VAT is beginning to exhibit signs of increased complexity in structure in the form of an increase in the number of rates and erosion of the base as well as a deterioration in administration that is often the counterface of a complex structure resulting in difficulty or manipulation in interpreting the tax code. To add to this, the personal income tax may be covering less and less potential taxpayers in relation to the growing incomes being experienced by economic agents in many of these countries and, where not adequately applied, an important tax instrument is being ignored. This issue, with particular reference to administrative mechanisms, is discussed further in Section III below.

III. Impact of Tax Administration, on Tax Structure

The lack of primacy of the income tax has another facet in many parts of the world and, in particular, in Latin America perhaps, comprising a blend of practices for the expressed purpose of facilitating its administration. When the conception of a tax structure is enacted into law, the tax administrator takes over, as it were, going full steam ahead with its implementation. In a sense, the framework of taxation rests there, for it is the tax administrator who holds the reins over what part of the tax law is actually applied, what part is modified for the sake of practical simplicity or in order to achieve revenue goals under the revenue department’s budget constraint, and what part is ignored since it is considered to be too impractical for field application. It is perhaps in this context that Bird and Casanegra (1992) wrote, “In developing countries, tax administration is tax policy.”

Increasingly, however, it seems to become obvious that the obverse is also true, i.e. there has to be greater responsibility and onus on tax administrators regarding, to what extent, they attempt to implement the tax structure without sacrificing the law for the sake of simplification. One example is working mainly with a set of large taxpayers while assigning low priority to potential taxpayers such as nonfilers in growing sectors like services, in effect operating without a perceptible strategy for expanding the universe of taxpayers to reduce the population of nonfilers. Targeting mainly large taxpayers is an efficient strategy in the short run but, in the long run, it comprises an insufficient approach. Another example is the use of various withholding schemes that may even be enacted under the law to facilitate tax administration, but that effectively negates or modifies the intended base and rate structure under a global income tax. Thus, it often boils down to a matter of trade-offs between appropriate structure and effective administration, and the solution is to arrive at an optimal combination of the two.

In this vein, the adage that tax administration is tax policy in developing countries is, indeed, not just true but incomplete. The statement could be further generalized to all countries by virtue of the fact that, in any scenario, it should not take too much imagination to conclude that it is the administration of a tax structure that is ultimately faced and perceived by the taxpayer. If there is little difference between tax policy and tax administration, then the taxpayer’s burden reflects the tax structure as laid out in the law. As this difference increases, the taxpayer’s burden begins to increasingly reflect the way the tax is administered and away from the framework of the tax structure.

This relatively stronger frontline position of the tax administrator over that of the tax policy designer puts the former in a position of advantage. But this should add a degree of responsibility in ensuring that its application resembles, as closely as possible, the tax structure as represented in the tax code. Indeed, not rarely, in many a Latin American country the tax code tends to become dotted with executive orders for administrative convenience, with actual insertions into the text of single leaflets that announce these orders. Upon inspection, it is remarkable how far the day-to-day interpretation and application sometimes are from the original design and intent behind the law.

Tax policy designers, on their part, tend to retreat to a passive role. Usually, a tax reform is carried out by a committee of experts comprising tax policy technicians, perhaps buttressed by tax lawyers (and influenced by lobbyists representing various sectors of economic activity), but rarely including experienced tax administrators. The latter, in effect, take over at a subsequent stage after the tax reform is passed, holding sway over how, what, when and where the new tax configuration is enforced. Most interestingly, yet surprisingly, the tax policy experts tend to distance themselves from this next stage—field application of their recommendations now configured in the form of a tax code—with scarce continuing links with the tax administration.

It is not as though the need for designing an implementable tax structure has not been examined in the literature before (Faria and Yucelik, 1995), and major tax reforms in the late 1980s and early 1990s in countries such as Argentina, Bolivia, Colombia, and Peru were indeed based primarily on the objective of simplification (Shome 1995a; 1995b). Therefore, once the tax structure is thus simplified, it should be obvious that tax administrators would be justified to further simplify practices for administrative convenience only under extraordinary circumstances. Otherwise, further simplification would not only affect efficiency and equity underlying the tax structure, but could actually lead to a lower revenue collection than intended. In turn, government expenditure, financed by revenue, would get curtailed. The alternative of a higher fiscal deficit cannot be sustained without detrimental macroeconomic effects. A simple framework is presented in Appendix I to demonstrate the link that, without proper tax administration practices, tax policy objectives cannot be realized.8

The separation between tax policy and tax administration serves the purpose of providing comfort to both parties who do not have to share ultimate responsibility for the successful delivery of a well structured and efficiently administered system of taxation. The policy expert can always blame the administrator for his inability to implement ideas (he tends to ignore or undervalue the task of mere administration) and the tax administrator can likewise blame the policy specialist for designing an impractical set of tax laws (he tends to set aside the background principles for a meaningful set of laws based on economic efficiency, equity, and stabilization considerations, being overwhelmingly influenced by the revenue objective). This can be expected, given the quintessential pressures for revenue generation from a finance minister’s office.9

Specific examples in the context of a shifting economic environment should help illustrate the point. In many Latin American countries, the relative shares in GDP of various economic—agriculture, industry, services—sectors are changing over time with differential rates of economic growth among them. With globalization, these shifts have speeded up, as many economies move away from heavily regulated structures of economic activity towards unregulated forms of production. At the same time, the services sector has tended to expand more rapidly compared to the industrial sector and, in some countries, international comparative advantage has also encouraged agricultural activity.10

The framework for tax policy, in general, should not differentiate between sources of income whether they be from industry or services, or from the organized or unorganized sector (though agricultural income often enjoys special tax treatment on practical or political grounds). This makes tax administration difficult because, understandably, services are typically more difficult to control, or the unorganized sector often includes small taxpayers who are difficult to identify especially in a changing world. Thus the relative shift from industry towards services and agriculture and, within industry, towards unorganized forms of activity, should underline the need for a clear strategy for expanding the taxpayer base through successfully including an increasing number of potential contributors--both large and small--from the growing sectors. These, and other phenomena, lead to particular concerns regarding the disconnectedness between policy and administration, as elaborated below.

A. Large Taxpayer Units

Tax administrators are prone to develop strategies to control large taxpayers. Though they tend to agree, at least conceptually, that all taxpayers should be targeted, in practice and under pressure to generate revenue, they are more likely to implement a strategy to control and expedite the flow of revenue from large taxpayers, than to develop or implement a clearly defined strategy for expanding the overall universe of taxpayers. Their universe, in effect, tends to collapse to that of large taxpayers and an expansion of that universe concerns the expansion in the number of large taxpayers. The telling indicator is the setting up of large taxpayer units in contrast to the lack of allocation of comparable resources for, or the assignment of specific responsibility for revenue generation from, the taxpayer population as a whole. This is not to minimize the role of large taxpayer units especially in the short run but, rather, to point towards the need for broadening the overall taxpayer base in the medium to long run.

Though large taxpayer units are set up as a short term measure with the assertion that, in the medium term, all potential taxpayers would be targeted, an examination of a country’s taxpayer profile, even over decades, is likely to demonstrate that the large taxpayer unit has effectively transcended from short-term to long-term policy, and that the share of small taxpayers in the taxpayer list has declined or remained static. A few countries do come to mind, however, where the tax administrations have set up departments to reduce the number of non-filers, but such cases are hard to identify in Latin America.

If tax administration activities are broken up into registration, collection, assessment, and audit, the resources allocated to large taxpayers under each activity would immediately demonstrate that a heavy proportion goes towards the control of large taxpayers. For example, registration of new taxpayers is focussed on large ones. Collection from large taxpayers is considered more important. The incidence of audit is higher for them (as is to be expected, of course), as is arrears collection, with the collection of small arrears being an activity that may be relegated to private collection agencies. Justification is usually offered in terms of the overwhelming share of revenue actually collected from the large ones.

It is to be admitted that the fact that large taxpayers account for such a high proportion of revenue is because they are large. However, at least three counterpoints are often overlooked. First, large taxpayers—rather than the small, potential taxpayers—are the ones on whom administrative resources are overwhelmingly spent. Second, the fact that large taxpayers account for such a high proportion of revenue should be taken as a (hopefully) temporary shortcoming in tax administration, rather than as a rationale for the tax administration to pour more resources to that activity. Third, calculations often disprove the contention of administrators that even theoretically, potential revenue collection from small taxpayers does not justify spending resources on them. Often the small taxpayers do possess significant revenue contribution potential, not uncommonly around one-third of total potential, depending of course on where the line is drawn between small and large.

This bias regarding large taxpayers, unless seen as a short-run strategic phenomenon, would contradict the concept of equity in taxation. That vertical inequity may and can manifest itself in the reverse direction when small taxpayers are ignored and when most improvements in tax administration are focussed on how to control large taxpayers, is often ignored. Even horizontal equity in taxation suffers when a wage earner in a factory is subjected to a tax that is not collected from a wage earner in a restaurant, or a self-employed worker with the same income. From a tax policy point of view, differential effective treatments of this nature can be acceptable at most temporarily for, otherwise, the design of a tax structure based on principles of equity would become meaningless. A temporary departure from the basic principles may be justified only to the extent that it is a third-best solution, with clear indications that these departures would be bridged within a specified period of time. That could happen only with the design and implementation of a strategy devoted towards closing the gap. In general, tax administrations do not employ such strategies and tax policy experts do not deem it necessary, or to their interest, to be concerned with the gap.

B. Simplification Schemes

Tax administrators in general prefer simplification of administrative mechanisms. One example is the tendency to introduce simplification regimes whereby small taxpayers, below a threshold that is defined by the tax administration, are allowed to contribute all taxes—income, excises, sales or the value added tax (VAT)—based on a common denominator such as turnover or sales. This could be a once-for-all contribution, the success of the measure being the inclusion of the taxpayer in the taxpayer list, irrespective of the legal requirement for their revenue contribution. Again, three points are missed here. First, there is a built-in acceptance that small taxpayers should pay little tax—or certainly less than the stipulation under that implied by the tax structure—on administrative grounds. However, unless the applied tax rate is allowed to vary according to presumed differential retail margins of different economic activities, acute horizontal inequity could result. Second, reverse vertical inequity could occur since any loss of revenue has to necessarily be made up by those who are not under the threshold, leading to their tax burden being higher than what it would be in the absence of the simplification regime. Third, there is an incentive for the typically large self-employed sector around the margins of the threshold to continue to underdeclare their incomes even in the long run to circumvent graduation from the regime.

A second example is the practice of self assessment by taxpayers which is used as a hallmark of modern tax administration. In this context, self-assessment is assumed to imply that a taxpayer files his tax return on his own, with the probability of a face-to-face encounter with a tax official being reduced only to the occasion of a detailed audit. This is supposed to reduce the incidence of corruption. However, what is rarely asked is if self assessment is appropriate in every environment. Self assessment presumes a mature taxpaying habit that, first, may not exist in every country at every moment in time and, second, it may take significant time for habits to change sufficiently to make self assessment work successfully. If self-assessment is imposed in an environment in which tax evasion and avoidance are likely to appear, then various pretexts can be created under which taxpayer-tax official encounters can occur. There it should be better to use more rudimentary administration mechanisms based partially on physical checks and controls in order to ensure equity and efficiency from a tax policy viewpoint. In such environments, often tax administrations have neither adequate resources nor the professional wherewithal to follow up with techniques for adequate taxpayer control including assessment and selective and incisive audit. External assistance based on experience from much advanced countries is unlikely to take hold and to continue after the departure of foreign experts. The result is that effective burdens of taxation get distributed unequally and the resultant allocation of economic resources is inefficient. Typically, however, tax policy experts are wary of wetting their fingers with preoccupations of this sort, remaining oblivious of these gaps between practice and theory.

Third, withholding mechanisms are another common example. It is true that withholding requirements are often written into law, to facilitate collection at an early stage (in a manner similar to the usefulness of a VAT for collecting revenue in steps, rather than only at the final retail stage at which the likelihood of tax evasion is higher). However, it is obvious that if withholding were final, for example on interest income, then the administration of the system moves away from a global income tax (again, in a manner comparable to truncating the VAT at an earlier stage such as that of the manufacturer or wholesaler). Under the latter, all sources of income should be taxed under the same rate structure, which is achieved by pooling all incomes together to arrive at a concept of global taxable income. Only if withholding were treated as a collection mechanism prior to the eventual summing up of global income, would withholding serve the objective of tax policy.

The above examples are but a few that illustrate how practice can obfuscate the principles of tax policy that underlie tax design. It is also noteworthy that such practices tend to exacerbate the low reliance on the income tax compared to consumption taxes. This is not to deny that there are indeed instances in which even the most modern tax administrations do need to simplify operations. For example, it is not possible to audit every taxpayer. Nor can it be expected that all potential taxpayers be made to pay taxes under the general regime. However, there do exist various in-between administrative practices that should bridge the policy-administration gap more successfully. What is being emphasized here is that these gaps are likely to be large rather than small in practice and much greater vigilance is needed to reverse this trend if, first, the right underlying principles of taxation are to be reflected in the field, second, if post-tax behavior of taxpayers is to have any semblance to that designed and predicted by theory and, third, if greater diversification in the sources of revenue is to be encouraged. As globalization and factor mobility spread across international borders, necessitating lower tax rates for many sources of income, the need for tax administration to carry out the original intent of the income tax structure, as fully as possible, becomes all the more imperative. The future of taxation in an environment of expanding globalization is addressed next.

IV. Emerging Issues in Taxation

With the advent of globalization as Latin American countries liberalize, they become the beneficiaries of increased capital flows. At the same time, they, as other liberalizing countries across the world, have become prone to easy capital flight. This is making them understandably and increasingly cautious regarding limits to the taxation of capital and, in general, of incomes (Tanzi, 1995). This will represent the main consideration spearheading taxation issues in the next century.

A. Effects of International Tax Developments

International tax developments have given rise to the need for an examination of the taxability of factors of production, or income tax. These factors are capital, labor, and land. Today, international capital—in particular, financial—flows are like swarms of bees moving from coast to coast, instantaneously changing direction from shore to shore, triggered by tax structure changes and non-tax factors such as speculative attacks on currencies with underlying weaknesses. Even without necessarily being able to identify whether tax or non-tax determinants are more important, countries have already become sensitive to the fact that international financial capital cannot be taxed at internationally non-comparable rates. In turn, this also has a direct impact on the taxation of nonfinancial capital. This is already happening in this century and is likely to accelerate in the next.

Indeed, in Latin American countries, this has manifested itself in the form of reduction in not only corporate income taxes, but also in withholding taxes for foreign remittances of dividends, interest and royalties. This occurred during the 1980s and seems to have accelerated in the 1990s. Thus the decade 1986-97 has witnessed a decrease, on the withholding tax rates on dividends from 24 percent to 16 percent, on interest from 23 percent to 20 percent, and on royalties from 31 percent to 30 percent (Table 5). Similarly, as mentioned above (Tables 4 and 6), other forms of taxation of capital have also tended to lower their burdens during the same time period. It is clear that, in the future, the capital tax base is likely to shrink. The same is true for the personal income tax, as far as increasingly mobile professionals are concerned, though not for others. This, in turn, would produce a “fiscal squeeze” (Grunberg, 1998) and, consequently, countries would need to firm up tax administration, collecting the income tax from all potential taxpayers, rather than mainly targeting the already existing large taxpayers, or merely attempting to expand the list of large taxpayers.

In the scheme of things to come, therefore, labor taxation should remain important because, though labor could also move like capital, labor movement could be envisaged to be slower than that of capital. And its taxation is likely to take the form of withholding taxes, but this should not take the form of final withholding taxes, for that would mean more scheduler taxation, i.e., each source of income being subjected to a different schedule of tax rates. Such a movement away from a global basis of taxation could augur an era of inefficient and inequitable taxation. Schedular taxation could be avoided only if great care is taken to make withholding an initial collection mechanism within a global structure of taxation.

It will be necessary to expand the universe of personal income taxpayers by, for example, progressively reducing the personal exemption level and bringing in more taxpayers from below. The personal exemption levels in terms of per capita GDP have continued to be higher in Latin American countries than in any other country-group (Shome 1995a). For example, in Argentina a representative taxpayer with a spouse and two children with an annual income of up to three times per capita GDP remains exempted. Many other countries in Latin America have comparable exemption levels. These must be brought down in the future in order to improve the reliance on income taxes relative to that of consumption taxes and since, with increasing incomes, there is ample justification to reverse the prevailing bias. This would be in line with the practice in many OECD countries.

Another issue that appears from the above discussion is the possibility of taxing other non-mobile bases of potential revenue. This begs the question of taxability of land and property. Available experience reveals a poor record not only in developing countries, but it seems also in many European, and East-Asian advanced countries. For example, recently, even countries such as New Zealand and Ireland have had rather poor experiences with respect to land taxation. Thus it is difficult to be too optimistic regarding land taxation even though land is a non-mobile factor. This does not mean that attempts at better land or property taxation should not be made, especially since, even in some of the middle-income Latin American countries such as Mexico (as opposed to say, Argentina or Uruguay), property taxation has remained negligible in terms of GDP.

Clearly, despite the anticipated movement towards greater reliance on the personal income tax, the taxation of consumption may be expected to continue to be a very important revenue source. It could take on even larger proportions since it can target domestic consumption of goods and services in those countries in which services are so far not taxed adequately. Some Latin American countries have already shown what can be done. With an 18 percent VAT rate, Chile gets 9 percent of GDP in VAT revenue. Other countries also generate significant revenues from the VAT. Indeed, a broad variety of indirect taxes may have to be explored further as the usability of selected direct taxes diminishes. Even in OECD countries, apart from the VAT, individual excises exist. In Latin American countries, excises also play an important role but VAT systems, with a functioning credit mechanism for input taxes paid earlier, have tended to replace other systems that taxed turnover. In some, the role of traditional excises has diminished. However, it may be expected that, in search of revenue, dependence on domestic excises could increase. And, in fiscally federal countries such as Argentina or Brazil, harmonized VATs at the level of states or provinces—which do not suffer from tax base erosion through tax competition—are difficult to operate where they already exist, or are difficult to formulate where they do not exist. In such an environment, sales taxes other than the VAT (or some contorted form of VAT) could prevail, requiring the tax administration to operate custom-tailored control mechanisms, including financial or physical methods of assessment and audit. This would depend on the need for the use of particular administrative instruments to ensure that the intention of the stipulated tax structure is fully carried out.

Administrative improvements could reduce the compliance gap in many countries, with a salutary effect on revenue. Especially in Latin American countries, there should also be continuing attempts to cover labor income from the unorganized sector in the tax net, with presumptive methods of taxation being used for administrative convenience. However, success depends not only on the efficacy of tax administration, but on the transparency of public policy and governance. Many countries lack these in general and need to improve their stature in these matters (Tanzi, 1997). If they are unable to improve overall governance at the higher levels, tax compliance by the average citizen is unlikely to follow.

Finally, as the next century approaches, discussions over global taxes become commonplace. Some global taxes should be avoided—for example, an international tax on financial transactions—which would be adverse for both developing countries and developed countries, because it will distort trade and financial flows (Shome and Stotsky, 1996). A second candidate is a global carbon tax (Shome, 1996). But there are difficulties associated with this tax too. The track of economic treaty negotiations reveals a lack of preparation or full knowledge of developed countries’ strategies among developing countries. Recent GATT rules on agriculture or patent laws are cases in point. Also, selected Latin American countries have historically been rather lukewarm towards bilateral tax treaties, and quite understandably so. In turn, fear that they might loose relatively in a negotiated solution could lead towards qualified cooperation and, therefore, to continuing lack of international cohesion, even if there are global reasons or imperatives for greater cohesion.

B. Next Century Tax Model

Looking into the next century, what would be the nature of a tax model that it is likely to emerge? Economies are unlikely to achieve or be in restful equilibrium. Rather, they might form complex patterns at the edge of chaotic economic movements. The theory of complexity, in effect, says that equilibrium does not exist. Instead, simply, patterns have to be identified within or at the edge of ever-changing phenomena. Thus, economics and its understanding will change. The understanding of economic processes is likely to undergo changes unfathomable by today’s standards (King, 1996). Issues such as economic growth, international mobility of factors of production, and taxation will no longer be driven by the objective or hope of generating them through economic equilibria or stability but, rather, through extracting them from ephemeral patterns around ever-changing disequilibrium formations.

Though such developments and their ramifications on the future structure of taxes could only be speculated upon, a model of Latin American taxation anticipated in such an environment, might comprise the following elements.

  1. The VAT would continue to be the primary revenue source perhaps, but with a declining share in total tax revenue. In federal countries excises and provincial sales taxes will be important, as harmonization within such countries tends to become more challenging, with particular regions emerging as more prosperous and powerful than others.

  2. The share of the personal income tax should increase, though the structure of taxes on professional labor is likely to be scaled back further. Hopefully, exemptions will be reduced to widen the income tax base. There is likely to be heavier use of withholding taxes. Hopefully, withholding will remain a collection mechanism rather than being converted to a schedular form of final taxation.

  3. Corporate income tax rates will remain low and, in continuation of prevailing trends, less than the top personal income tax rates, as capital continues to be internationally mobile.

  4. Other forms of taxation of capital, such as on capital gains, assets, or remittances of interest, dividend, and royalties, are also likely to decline further, in keeping with current trends, with their rates remaining internationally competitive.

  5. The use of customs duties will tend to diminish even further, and they will be based on low and a declining number of rates, provided WTO rules reflect equitably the interests of developing and developed countries alike.

  6. There will be attempts to increase the property and land taxes, though much success in such an effort is doubtful, given their limited success in general.

  7. More green taxes will emerge, as the environmental tax movement and the idea of double dividend from green taxes gain force. It is not unlikely for them to be considered on a global scale or as a global tax.

  8. Tax administration improvements, as a reflection of a conscious effort at closing the gap between tax administration practices and tax policy intentions, may be expected to be stepped up, as the discussion of the issue becomes more prominent. The development and monitoring of indices such as: (i) increase in the universe of taxpayers; and (ii) compliance gap or a measure of tax evasion, should occur more commonly, provided tax administrators become more aware and accepting of fundamental economic principles behind tax policy.

IV. Conclusions

Though Latin American countries experienced overall trends in tax structures in the early 1990s similar to those in the 1980s (with reductions such as in corporate taxation gaining speed while others such as in personal income tax continue, though less rapidly and, the central role of the VAT in revenue generation continues as well), a general drift in tax policy, without a clear vision or example, seems to have set in during the late 1990s. Though there seems to be an obvious need for reviving the role of income taxation from an exclusive focus on consumption taxation, the forces of globalization, however, are likely to make this difficult. All the more, therefore, renewed vigor and vigilance are needed in this area.

Globalization is also likely to lead to significant changes in other areas of tax systems and arrangements in the twenty-first century, particular features being: (1) international cooperation despite the sacrifice of some fiscal sovereignty; (2) the introduction of global taxes such as an environment tax, rather than one on financial transactions which would adversely affect financial flows; and (3) an international tax organization focussed on international tax policy and related tax administration issues (Tanzi, 1998) which would reduce the hold of tax havens, increase international exchange of tax information, develop multinational conventions, and reduce the prevalence of bilateral treaties.

Modern tax administration is moving forward in leaps and bounds and, as globalization spreads, tax policy and tax administration become even more inter-connected. In this context, the responsiveness of tax administration to ensure that administrative practices closely resemble the underlying principles of taxation, becomes a fundamentally important issue. On their part, tax structures also have to, by necessity, reflect what is administrable, minimizing the number of tax rates and tax concessions, ridding the system of nuisance taxes such as small excises, and stopping short of changing the tax statute too often. To conclude, the one-to-one correspondence between tax policy and tax administration cannot be overemphasized, the hallmark of a developed tax system being, on the one hand, how closely the administration of a tax replicates its original policy objective and, on the other, how cognizant the design of the tax is to make its implementation feasible.