In the long run, the business unit or source will yield more revenue to the public treasury than the individual; and the place where the income is earned will derive larger revenues than the jurisdiction of the person.

In the long run, the business unit or source will yield more revenue to the public treasury than the individual; and the place where the income is earned will derive larger revenues than the jurisdiction of the person.

—T.S. Adams

I. Introduction

This chapter examines the details of international income tax as an aid to understanding and drafting the parts of the income tax law dealing with international issues. Given the large literature on basic policy issues in international taxation, I deal with general policy matters only in passing.1 The chapter accepts the general parameters of international income tax law as it is now established without questioning whether the structure provides the best solution to international tax problems.2 Within that structure, it seeks to provide a detailed discussion of policy, design, and drafting issues. Although the chapter draws on the experience of industrial countries with international taxation, the special concerns of developing and transition countries are emphasized throughout.

The major difference between international income tax law and the remainder of the income tax lies in the pervasive importance of treaties.3 Most countries have entered into one or more bilateral tax treaties that supplement and sometimes replace the income tax law, but only as regards the parties to the tax treaty in question. This chapter gives considerable emphasis to tax treaties and to the work of the Organization for Economic Cooperation and Development (OECD) and the United Nations (UN) in this area.

II. The International Dimension of Taxation

In the development of a country’s tax laws, the international dimension plays an increasingly important role that significantly restricts the rules that might be adopted if regard were had only to domestic considerations. The increasing role of international factors is mainly attributable to the globalization of the world economy.

A. Importance of International Taxation

International trade has existed since the birth of nations, but there has been an accelerating growth not only in trade but also in finance and investment since the end of World War II. This growth has far outstripped the general growth in the world economy. One important cause has been the gradual removal of barriers to international trade through the various negotiating rounds of the General Agreement on Tariffs and Trade (the GATT, which as of 1995 is administered by the World Trade Organization, or WTO). For finance, the removal of exchange controls in most industrial countries, commencing from the floating of exchange rates in the early 1970s, has been a notable factor leading to the globalization of world capital and financial markets. The international organizations most involved here have been the IMF and the Bank for International Settlements.

In relation to investment, the main multilateral push is yet to come. In recent years, the foreign direct investment laws of investee countries and the investment rules for various institutional investors in investor countries have been liberalized, and bilateral investment treaties have grown. The Multilateral Agreement on Investment is currently under negotiation in the OECD. When this treaty is concluded in the near future, it is proposed to extend its regime worldwide through the cooperative efforts of the OECD and the WTO, which will see further global relaxation of investment controls. In addition, the end of the cold war has freed up the international transfer of technology, and labor is also becoming more mobile, especially for high-cost services (such as professional, management, and consulting services) and within trade blocs.

Overlaying all these developments and substantially contributing to many of them are the great advances in international communications and computer technology.

It is a corollary of this growth in international transactions that international tax laws (along with international trade, finance, and commercial laws) have become more significant to each country’s legal system. Moreover, as restrictions in other areas are reduced or removed, taxation is brought increasingly into focus, but there is a significant difference in the tax case. Whereas it may be possible to liberalize or abolish rules in other areas affecting international transactions, taxation needs to be retained in some form for the financing of governments. The international challenge for taxation is the development of a system that does not act as an undue impediment to international transactions while protecting the revenue of each state.

Although this challenge is present for all kinds of taxes, this chapter deals with the income tax.4 The income tax is usually the major source of revenue and the most complex tax in industrial countries. For both these reasons, the tax causes the most problems in the international arena. In developing and transition countries, the income tax may not be the most important tax in terms of revenue, but it is looked to as serving that role in the future and it will also generally be the tax of greatest concern to foreign investors and expatriate personnel.

B. The Challenge for International Taxation

There are two main categories of case that international tax rules have to deal with. First, there is the taxation of persons from outside a country who work, enter into transactions, or have property or income in the country. Second, there is taxation of persons who belong to a country and work, enter into transactions, or have property or income abroad. The usual term used in international taxation to denote the concept of a person’s belonging to a country is “residence” (“resident” and “nonresident” being used to indicate whether a particular person belongs to a country or not); similarly the usual term for income arising in a particular place is “source” (“domestic” and “foreign” being used to indicate whether particular income is sourced inside or outside a country).

The two categories arise in virtually all areas and types of taxation. For the income tax, the issues are the taxation of domestic income of nonresidents and the taxation of foreign income of residents. In both categories of case, the main problem is the potential for double taxation or double nontaxation of the income. That is, more than one country may seek to tax without reference to tax levied in another country, or no country may tax (usually on the assumption that another country is taxing, although often it will be the result of the increased opportunities for tax planning or tax cheating on the part of taxpayers that international transactions offer). Double taxation is likely to act as a barrier to international transactions, and the nations of the world are generally agreed on the desirability of removing such barriers as a means of increasing global welfare.

By similar reasoning, double nontaxation of international transactions will create a bias in favor of international over domestic transactions, leading to a loss of global (and national) welfare, not to mention tax revenue. While, however, there is general agreement among taxpayers and governments on the undesirability of double taxation, double nontaxation is obviously desired by taxpayers and to some extent tolerated or even encouraged by governments. Developing countries often express the view that any increase in global welfare arising from the removal of international barriers accrues mainly to industrial countries. International agreements sometimes contain special regimes to deal with these concerns of developing countries, such as the generalized system of preferences in the GATT, which allows industrial countries to confer tariff privileges on developing countries without being obliged to extend them to all GATT members.

In the income tax field, this developing country view finds expression in the desire to offer tax incentives to international investors in order to attract capital and to ensure that the tax systems of industrial countries do not negate the effect of the incentives by collecting the tax that the developing countries have given up. The desired result of developing countries is generally achieved by tax-sparing provisions in bilateral tax treaties, which effectively sanction double nontaxation and hence create a bias in favor of international investment in developing countries. This particular policy in favor of double nontaxation is dealt with elsewhere in this volume.5 In this chapter, the general premise is that the basic goal of the international income tax system is to avoid double taxation and double nontaxation.

C. Consensus on International Tax Rules

As the importance of the international dimension of income taxation has grown, an international consensus has emerged about the structure of the international income tax regime. The income tax is typically levied by a country on (1) the domestic and foreign income of its residents and (2) the domestic income of nonresidents. These basic rules are referred to respectively as the residence and source principles of taxation. The tax legislation of a country should in succinct terms state in some suitably conspicuous place (either the general provision levying the income tax, or the beginning of the group of provisions dealing with international issues, or both) whether and to what extent it has adopted these rules.

If a resident of one country earns income from a source in another country, double taxation is likely to result because one country will tax that income on a source basis and the other country on a residence basis. In this case, the internationally accepted regime is that the source country has the prior right to tax (although this right may be limited by treaty), and the residence country is responsible for relieving any double taxation that results. Such relief is generally achieved through one of two systems—the exemption system whereby the foreign income is exempted from tax in the residence country, and the foreign tax credit system whereby the tax of the residence country on the foreign income is reduced by the amount of source country tax on the income. Most countries employ some combination of the two systems.

The details of the rules necessary to implement these apparently simple concepts and their interaction with tax treaties will take up the remainder of this chapter. Before embarking on these rules, I will explore briefly the structure, purpose, and effect of tax treaties.

III. Tax Treaties

Tax treaties (also often referred to as double taxation conventions or double tax agreements) are international agreements entered into by countries and hence subject to general international law on treaties as codified in the Vienna Convention on the Law of Treaties.6 Most tax treaties are bilateral, that is, involve two countries only, and cover income and capital taxes, though there are some examples of multilateral tax treaties. There are well in excess of 1,000 tax treaties and the number is growing rapidly.7

A. Structure of Tax Treaties

The history of tax treaties can be traced to the League of Nations, which was pressed to deal with the problem of double taxation after income taxes became important during the First World War and which developed a number of models for use in negotiation of bilateral tax treaties.8 The major modern successor to these models is the OECD Model Tax Convention on Income and on Capital (the OECD Model), which itself has gone through various versions.9 Of especial interest to developing and transition countries is the 1980 UN Model Double Taxation Convention (the UN Model), which was based on the 1977 OECD Model but designed to take into account the special interests of developing countries.10

The typical structure of tax treaties is most easily seen from the chapter and article headings of the OECD Model as follows:

article image
article image

This structure (and even the numbering) is followed with only a few variations in nearly all existing tax treaties. The treaties apply to income and capital taxes11 levied on residents of either of the countries that are parties to the treaty. Chapter III sets out the major substantive rules of the model treaty; they operate by dividing income into classes and setting out rules for each of the classes. These rules generally give the residence country an unlimited right to tax the income and at the same time limit or eliminate the source country’s right to tax, with the source-country rights the greatest with respect to active income (business, professions, and employment) and income from immovable property, and the least with respect to passive income from intangibles. The treaty recognizes the source country’s prior right to tax by requiring the residence country to relieve double taxation of its residents for taxes levied by the source country in accordance with the treaty. Chapter VI deals with administrative matters, to ensure that the treaty is effective in practice, and with the important issue of nondiscrimination.

On the basis of these models and its own particular policies, each country generally develops its own model that serves as the starting point in negotiations to conclude a tax treaty with another country.12 A bilateral tax treaty takes about two years on average to negotiate and bring into force. In view of this long period of gestation, most treaties fix a minimum time period for their operation (generally about five years), but the expected life of a treaty before replacement by an updated version will usually be of the order of 10-30 years. This long life dictates both that the treaty be expressed in general terms so that it is flexible enough to handle the inevitable changes in the domestic tax laws of the treaty partners that will occur during the life of the treaty, and that the treaty contain mechanisms to deal with issues that arise during its life (primarily through each party keeping the other informed of changes in tax laws and through the consultative mechanisms provided by the mutual agreement procedure).

B. Purpose of Tax Treaties

The purpose of bilateral tax treaties is typically expressed in their preamble to be “the avoidance of double taxation and the prevention of fiscal evasion.”13 As most countries contain within their domestic law provisions to prevent double taxation of their residents in the most common case (where another country taxes the same income on a source basis), the main operation of tax treaties in this respect is for other types of double taxation that can arise as elaborated below. The prevention of fiscal evasion primarily refers to cases where taxpayers fraudulently conceal income in an international setting and rely on the inability of tax administrations to obtain information from abroad. The exchange of information article in tax treaties is the major provision dealing with this problem. Because of the capital flight experienced by many developing and transition countries, exchange of information is important, but in practice there are some considerable hurdles to successful exchange for reasons developed below.

From the perspective of developing and transition countries, there are a number of other purposes of tax treaties that are usually unstated but in many cases are more important. First, there is the division of tax revenues to be derived from income involving the two countries that are parties to the treaty. Where flows of income from business and investment are balanced between two countries, or even among a group of countries, it often does not make a large difference if each country agrees to significantly curtail its source jurisdiction to tax, as its residence taxation of income sourced in the other country is correspondingly increased. Where the flows are substantially unbalanced, the conclusion of a treaty under which each country gives up some of its source jurisdiction to tax generally has the effect of transferring revenue from one country to the other. Typically, developing and transition countries (and many smaller industrial countries) will be in the position vis-á-vis industrial countries of substantial net capital importers and hence will want to preserve source-country tax rights.

Second, developing and transition countries nowadays generally desire to encourage capital inflows from capital-exporting countries. Tax treaties may facilitate this process in a number of ways. In a very general sense, entering into tax treaties acts as a signal that a country is willing to adopt the international norms. This symbolic function is reinforced by the nondiscrimination article of tax treaties, by which the country undertakes not to discriminate under its tax laws against residents of treaty partners. Many potential investors attach great importance to the nondiscrimination article, in light of the historical antipathy that many developing and transition countries have in the past exhibited to inward investment. It is no coincidence that many tax treaties with transition countries are negotiated alongside investment protection treaties.

In the past, many developing countries took the view that they did not need tax treaties.14 The countries very often adopted a policy that growth of their economies could best be achieved through domestic production by domestically (often state) owned firms of goods and services for domestic consumption. Hence, foreign investment was not needed and economic policy bolstered the natural human emotional response against ownership by foreigners. As tax treaties involved giving up part of the revenues from source taxation, there seemed little to be gained from them. Likewise, it was a consequence of the domestic focus that investment abroad by residents was not encouraged (a policy often enforced through very strict exchange controls). This situation has now changed, as demonstrated by the rapidly expanding tax treaty networks of many developing countries. Partly, the new attitude is due to a policy shift that accepts the benefits that flow from international trade and, in particular, from export-led growth in the model of the newly industrialized economies of Asia. Another factor has been the practical impossibility of making exchange and investment controls work effectively in a global economy.

Transition economies did enter into tax treaties in the past, but these were mainly political gestures given that there were no significant capital flows from the West.15 The provisions of the old treaties were often inappropriate for the new situation and they therefore had to be speedily replaced (a phenomenon particularly noticeable in the case of the Russian Federation). The need to do so, along with the large needs for capital, has spurred many transition countries to develop their treaty networks in recent years. The tax laws of transition countries are often not sufficiently developed or clear to enable the tax administration to utilize treaty rules. For example, domestic legislation may lack rules for adjusting transfer prices between related parties. This is another matter that the countries are generally remedying.

The remainder of the discussion in this chapter therefore proceeds on the assumption that most developing and transition countries will be actively pursuing the development of a tax treaty network and that, in the case of the transition countries, changes will be made to domestic law to remove the elements that form impediments to this development. What effect does this assumption have on domestic law?

C. Relationship of Tax Treaties and Domestic Law

It is not necessary to incorporate into domestic law the contents of treaties that operate only between states and do not directly affect private persons. A tax treaty, however, is intended to confer enforceable rights on taxpayers against the countries that are parties to the treaty. How this occurs is a matter for the constitutional law of each state, but in many cases it is necessary for each country to carry out some formal law-making process, such as approval of the tax treaty by parliament.

Further, the provisions of tax treaties are intended to have precedence over any inconsistent provisions of domestic tax law. Again, how this is effected is a matter for the constitutional law of the countries concerned. A common practice is to insert such a provision either into the law giving effect to the treaty or into the domestic tax law itself.16 The usual result of such a provision under the law of most countries is that, apart from the administrative treaty provisions on the mutual agreement procedure and the exchange of information, a treaty sets limits on the operation of domestic law but does not expand its operation.

Thus, if a country taxes business profits arising from sales to residents of the country by a resident of another country without reference to a permanent establishment concept, the business profits article of a tax treaty will usually prohibit such taxation, unless those profits are attributable to a permanent establishment in the country. The outcome is the same if the domestic law uses a permanent establishment concept, but the concept is wider than that used in a relevant treaty. Similarly, if the tax applied under domestic law to dividends and interest paid to a resident of the other treaty country exceeds the maximum rates permitted in the treaty, the source state is obliged to reduce its taxation accordingly.

If, however, a country levies no tax on dividends or interest paid to nonresidents, then the fact that a treaty allows such taxation up to a specified limit does not mean that such dividends and interest are taxable. It is possible, however, for domestic law to provide that if a treaty permits taxation that does not otherwise occur under domestic law, then the treaty rule will become the domestic rule for this case. This is the position in France17 (and many Francophone African countries under their tax legislation) and in Australia with respect to source rules contained in treaties under legislation giving force to tax treaties.18 Such a result is fairly uncommon, however.

By contrast, the administrative provisions of tax treaties (which may include articles on mutual agreement, exchange of information, and assistance in collection) by their very terms expand domestic law in the sense of giving powers that generally do not exist under domestic law. Thus, the mutual agreement procedure as contained in article 25 of the OECD Model gives an avenue of recourse to challenge assessment to tax in certain cases that does not exist under domestic law and overrides domestic limitation periods. Article 26 gives power to exchange information that does not usually exist under domestic law and modifies the secrecy provisions of domestic law accordingly.

The consequence of this relationship between tax treaties and domestic law suggests an important guideline for drafting the domestic tax rules themselves. If the domestic rules by and large follow the rules typically found in tax treaties, this will simplify the question of the relationship between tax treaties and domestic law and provide transparency to foreign investors as well as indicating (even in the absence of an extensive tax treaty network) the intention of the country to adopt internationally accepted standards.19 This approach also gives instant access to a substantial body of commentary that is accepted by international consensus as elaborating and explaining the wording in question. The consequences of following—or not following—this guideline will be explored below. Because an international consensus exists on the structure and content of tax treaties, no one country, except perhaps the United States, is able to depart substantially from international norms. Accordingly, having a country tax treaty model that departs radically from the existing international models and following that model in domestic law generally is not a viable option for developing or transition countries.20 Moreover, no country can sensibly adopt a policy of residence taxation only (i.e., excluding the source principle). Neither would it make sense for developing and transition countries to adopt a policy of source-only taxation.

IV. Definition of Residence

Residence is almost invariably a central concept in the international tax rules of the domestic tax law of a country, with residents taxed on their worldwide income (or at least some categories of income).21 It is difficult to enter into tax treaties without a concept of residence in domestic tax law because, by the first article of the international models, the tax treaty applies to the residents of each country that is a party to the treaty, and the definition of resident in the treaty refers to a resident under the domestic tax law of the countries. The basic idea behind the residence concept is that a person is a resident of a country if the person has close economic and personal ties to the country. It is possible for a person to be a resident of more than one country.

A. Individuals

Applying this basic policy idea in the case of an individual usually leads in domestic tax legislation to the adoption of one or more of three approaches. First, there is a facts-and-circumstances approach where no criterion is definitive but all the facts are weighed to determine residence. In many countries, this approach is not specifically defined by statute, and it is left to the courts or tax administration to give content to the concept. Tax treaties in the article defining residence give an indication of the factors that are most often used for this purpose: permanent home, personal and economic relations, and habitual abode. The problem with this approach is its uncertainty, which can be ameliorated by combining it with one of the following tests.

Second, the tax legislation may adopt rules for residence that are used for other purposes in the civil law of the country concerned (such as entitlement to work or remain in the country indefinitely under immigration laws, domicile, or citizenship). Many European countries use domicile. The United States is the only major country that uses citizenship as a residence-type test and, in view of the very liberal nationality laws of many countries, the citizenship criterion does not seem appropriate in most cases.22 The problem with civil law tests is that the policy underlying a test devised for other purposes may not be appropriate for tax purposes, but the advantage is that they are more certain than the facts-and-circumstances approach, unless the civil law concept itself is vague.

Third, a rule of thumb based on the number of days that a person spends in the country during either the tax year or a moving 12-month period may be employed, the usual period being half a year (expressed as 183 days or more). Under this test, physical presence in the country for any part of a day usually counts as one day except when the person is in transit between other destinations and does not pass the customs or immigration barrier.23 The advantage of using the tax year for this purpose compared with any moving 12-month period that ends or begins in the tax year is that a person can determine residence in relation to a particular tax year at the time of filing the tax return for that year. For example, if the calendar year is the tax year and the due date for filing declarations is March 31, a person who arrived in the country on October 1 will not know until the following October 1 whether the person was a resident from the time of arrival under a moving 12-month test. The disadvantage of a rule that looks solely to the number of days of presence during the tax year is that it effectively allows a person to remain in the country for up to 364 days consecutively spread over two tax years without becoming a resident. An intermediate rule that avoids these problems would look to presence in any consecutive 12-month period ending in the tax year in question.

In either form, the test can be criticized as unfair because it is mechanical—one individual can be treated as a resident despite very short periods of stay in the country (e.g., where a person drives to and from work through a neighboring country each work day),24 while others can manipulate their period of stay to avoid crossing the 183-day threshold even when they are substantially connected to the country. Most countries use some variation of the 183-day test but, because of its problems, often adopt a more substantive test of residence in addition and enact other measures to ameliorate its arbitrary nature, as discussed below.

Often, there is a special residence rule deeming specified government employees stationed abroad to be residents. The main purpose is to ensure that the diplomatic or other government staff of a country who may spend most of their working lives outside the country are nonetheless resident and therefore taxable on their salaries by the country (as they will often not be taxable in other countries, either by virtue of their diplomatic status or by virtue of the government service article in tax treaties).

Under all tests, questions arise as to whether a person can be a resident for tax purposes for part of the tax year and nonresident for part of the year. Most countries permit this possibility mainly to cover the case of migration where a person is moving permanently from one country to another. Where an individual is a resident for only part of a tax year, tax allowances tied to residence are often apportioned.

Some language encompassing these possibilities (other than when reliance is placed on other features of domestic civil law) follows.

  • (1) Subject to (2) and (3), an individual is a resident of X for the entire tax year if that individual

    • (a) has closer social and economic relationships with X during the tax year than to any other country;

    • (b) is present in X for 183 days or more in any consecutive period of 12 months ending in the tax year; or

    • (c) is an official of the state service of X posted overseas during the tax year.

  • (2) An individual who was not a resident in the preceding tax year shall not be treated as a resident for the period preceding the day the individual was first present in X during the tax year.

  • (3) An individual who is not a resident in the following tax year shall not be treated as a resident for the period following the last day on which the individual was present in X during the tax year if during that period the individual had a closer social and economic connection to a foreign country than to X.

  • (4) For the purposes of (1)(b),

    • (a) presence in X for part of a day is counted as a full day, and

    • (b) presence in X without immigration clearance in transit between other countries is disregarded.

Some countries distinguish varying degrees of residence, such as residence and permanent residence, for different purposes under domestic tax law.25 This approach may create confusion in the operation of the law unless the different terms are used with care in drafting. It can also cause problems in the application of the tax treaty article defining residence. Countries with a number of residence concepts need to review their model tax treaty to ensure that it is in harmony with the domestic tax law. Generally, it is better to avoid the use of differing residence concepts in the law and to deal with the concerns that give rise to them in other ways (such as special rules for expatriates; see section VI(B) of this chapter).

Because countries use different tests of residence, individuals with dual residence are not uncommon. In fact, even if all countries adopted the most common definition of residence—namely, the 183-day test—it would still be possible for the same individual to be resident in more than one country under each country’s tax law at the same time. An example is the frontier worker who lives in one country but works in another and crosses the border between the two countries each work day. Dual residence creates problems of double taxation where each country taxes the worldwide income of its residents. The mechanisms for giving relief for double taxation arising from combined source and residence taxation of the same income are not able to solve this problem (sometimes called residence-residence double taxation). It is difficult for a country to solve this problem on its own, and so tax treaties provide a tiebreaker mechanism to allocate the residence of the individual to one country alone for the purposes of the treaty. This allocation is achieved through a hierarchy of tests involving the individual’s permanent home, center of personal and economic relations, habitual abode, and nationality.26

For frontier workers, this mechanism may not solve the practical difficulties that average people face in being subject to two tax jurisdictions (either because a taxpayer is resident in one country and receives income sourced in the other country where the taxpayer’s employment is conducted, or because the taxpayer is regarded as a resident of both countries). Accordingly, tax treaties between contiguous countries often contain provisions to ensure that frontier workers are taxed on their wages in one of the countries alone.27

B. Legal Entities

The residence of other taxpayers besides individuals, that is, corporations and other entities taxed as separate taxpayers, involves similar problems. From a policy perspective, legal entities are ultimately owned by individuals and the residence of the owners should determine the residence of the entity. This is not a practical test for a number of reasons: it may be necessary to trace ownership through many tiers of entities, which is not administratively feasible, and in any event the ultimate owners may themselves be resident in different countries. Hence, a number of other tests are used. The first test is the country under whose laws the entity came into existence, commonly referred to as the place-of-incorporation test. Even more than the 183-day rule, this test is quite mechanistic and susceptible to manipulation. Therefore, additional tests and other safeguarding mechanisms are often provided.28

The second test is usually based on the place of management of the legal entity. In Anglo-Saxon countries, this is often expressed in the phrase “central management and control,” which basically means where the board of directors meets. European countries look to the location of the head office of the legal entity.29 These tests are based in part on a facts-and-circumstances approach to residence and so are not quite as mechanistic as the place-of-incorporation test, but they are susceptible to manipulation nevertheless.

Tax treaties seek to deal with the problem of dual residence of legal entities as for individuals, but are much less successful in this area mainly because there is no real international consensus on the appropriate tiebreaker, even though the OECD Model uses the place of effective management.30 Moreover, dual-resident companies can give rise to problems that are not adequately addressed in tax treaties, especially the double claiming of deductions on a residence basis. Hence, a number of countries have enacted rules denying deductions to dual-resident companies in certain cases.31 For developing and transition countries, it may be better to rely on general antiavoidance rules to deal with this kind of tax planning, as discussed below.32

These problems by no means exhaust the issues regarding residence of entities. Most countries have a variety of legal entities, not all of which can be easily fitted into the category of company or corporation for domestic tax law and tax treaty purposes. As exotic entities are being used increasingly in international tax planning,33 countries should consider the need for special tax residence rules for various kinds of entities. Further, if it is not regarded as clear from definitions based on the above criteria that governments (central, regional, or local) and other public bodies of a country are resident in the country, then provision may be made to that effect.

It is common to define both individuals and legal entities that are not resident under the definitions in the domestic law as nonresidents for the purposes of the law,34 although this may simply be stating the obvious.

V. Definition of Source of Income

Residence establishes a relationship between a country and the taxpayer deriving the income, whereas source concerns the connection between the income itself and a country. The basic policy idea is that income should be sourced in the country with which it has a substantial economic connection. Obviously, income may often have substantial connections with more than one country, in which case it may be appropriate to determine source by apportioning the income between the countries. Source rules have traditionally used differing concepts for active and passive income. In broad terms, active income is usually sourced by a place-of-taxpayer-activity test, while passive income (where the taxpayer often engages in no significant activity in deriving the income) is sourced by the place of activity of the person paying the income. To the extent that a clear distinction can be drawn between active and passive income, the growth of international trade in services raises questions as to whether the place-of-taxpayer-activity test is always appropriate for active income.

A. Geographical Extent of Country

As source involves a geographic connection, it is necessary to define the geographical area in question. For landlocked countries, this definition question does not present a real problem because the land area of the country is the relevant area. For countries with a maritime boundary, the territorial sea is treated under international law as part of the country and the country’s jurisdiction also extends to the natural resources of the sea and seabed of the continental shelf. It is customary to extend source tax jurisdiction to the continental shelf. This extension may be effected in a way that reflects the limited rights that a country can exercise over the continental shelf (i.e., the country taxes only those continental-shelf activities over which it has sovereignty) or may be more general and cover all activities on the continental shelf. The resulting difference in tax jurisdiction over the continental shelf is shown by the example of a floating hotel owned by a nonresident and moored on the continental shelf; if the tax jurisdiction of a country is limited by reference to its sovereign powers under international law, the country cannot tax the profits of the hotel, whereas it can tax the profits if the broader formulation is adopted.

It is common to include a similar provision in tax treaties in the definition article. Given the importance of potential oil or gas resources in the continental shelf, oil-producing countries commonly include in tax treaties special provisions on this topic that preserve source-country jurisdiction as far as possible. In the case of other resources of the continental shelf, such as fisheries, some developing countries levy license fees in lieu of income tax, although in either case there are significant enforcement problems.

There is no agreement in international law that countries must limit their taxing jurisdiction for nonresidents to income sourced in the country. Some international lawyers consider that a country can assert the right to tax everybody in the world on their worldwide income,35 but it will never be able to enforce such a claim and may attract various forms of retaliation from other countries. In other words, the adoption of the residence and source principles of taxation has been very much guided by practical considerations of enforcement and reciprocity. In marginal cases, such as floating hotels moored on the continental shelf, an assertion of tax jurisdiction is not likely to cause any problems practically or in international relations.

B. Structure of Source Rules

Many industrial countries do not have elaborate source rules in their domestic tax laws,36 instead relying on such general expressions as income arising (from activities) in the country to express the source concept. In these countries, there will usually be a well-developed body of practice as to the detailed application of the general principle,37 and in any event, there will be an extensive network of tax treaties in place containing explicit and implicit source rules for virtually all types of income. In the past, it was possible for developing countries to elaborate their domestic tax laws without detailed source rules, both because international income tax was not as important for the reasons outlined above and because the countries could usually rely on the body of practice in industrial countries because their tax laws would usually be modeled on the law of one or another industrial country. Transition countries are in all cases actively encouraging foreign investment. However, there is no tax tradition and in most cases no tax treaty network on which they can call to fill in the gaps in their laws on sourcing rules. For both developing and transition countries, fairly detailed source rules will give comfort to foreign investors as to when their income will or will not be taxed.38

Tax treaties contain only a few source rules explicitly identified as such, for example, article 11(5) of the OECD Model dealing with interest. Nonetheless, for most kinds of income, there are implicit source rules. The source rule is implied by the way in which a country is given jurisdiction to tax income derived by residents of the other treaty country; for example, in the case of business income, under article 7 a country can tax income only if attributable to a permanent establishment in that country of a resident of the other treaty country. Some countries include a provision in their tax treaties to make clear that these implicit rules are effectively the source rules under the treaty.39

Countries can appropriately take these implicit treaty rules as the basic guideline for their source rules, subject to some caveats. First, to give the country negotiating room in the tax treaty process, the source rules in the domestic law should generally be more expansive than those found in treaties. Second (a related point), as the treaty rules operate to divide revenues between source and residence country, the source country will usually want in its domestic law to take full advantage of its taxing powers and have broader rules than those found in treaties. Third, the rules in tax treaties are to some extent shaped by practical considerations of tax administration, with a country giving up taxing rights not because income cannot be regarded as sourced there but because it is simpler for taxpayer and tax administration not to attempt to tax the income. However, it is very helpful if the domestic law generally follows the categorization of income that occurs in tax treaties because this makes the interaction of domestic law and tax treaties easier to understand. It also allows an easy connection between the type of income, and the method of taxation and collection of tax, as outlined below.

Just as it is possible to have residence-source and residence-residence double taxation, so source-source double taxation can arise when more than one country asserts that the same income is sourced in each country.40 Again, it is difficult for any one country to solve this problem unilaterally, and tax treaties are the usual mechanism for resolving it. The method adopted in treaties is to specify expressly or impliedly for a single source rule to apply between the parties to the treaty for particular categories of income.41 In turn, this method creates some impulse for countries to adopt similar rules in their domestic laws as informal harmonization on the same approach will generally overcome source-source double taxation even without tax treaties. Against this background, the various categories of income are now considered basically in the same order as found in tax treaties.

C. Income from Immovable Property

For income from immovable property, such as the rental of buildings or mineral royalties, the income is sourced in the country where the property is situated, whether it is derived as part of a business or otherwise. Under tax treaties, the provisions based on article 6 of the OECD Model include income from agriculture and forestry in this category and have a fairly extensive definition of immovable property that includes reference to the domestic law concept of immovable property. These features can be incorporated in domestic law, although it is probably simpler to omit them. Their effect in practice is not significant.

D. Business Income

For business income, tax treaties start with the permanent-establishment concept, which refers to a relatively enduring presence in a country through either location (e.g., an office) or personnel. The definition article of this term is quite lengthy and can be simplified in domestic law by removing some of the qualifications that limit the concept. Further, some extensions of the concept found in the UN Model may be added, especially as they were designed to increase the taxing reach of developing countries and add negotiating room in the tax treaty process. Special rules on oil and mineral exploration activities may also be appropriate for some countries. A provision with these features could take the following form:

  • (1) A permanent establishment is a fixed place of business through which the business of a person is wholly or partly carried on.

  • (2) A permanent establishment also includes

    • (a) a building site or construction, installation, or assembly project in the country, or supervisory activities connected therewith; and

    • (b) an installation or structure used for the exploration or exploitation of natural resources in the country, or supervisory activities connected therewith.

  • (3) Where another person is acting on behalf of the person and has, and habitually exercises, in a country an authority to conclude contracts in the name of the person, that person shall be deemed to have a permanent establishment in that country in respect of any activities which that other person undertakes for the person. This paragraph does not apply to an independent agent acting in the ordinary course of business.

The primary sourcing rule for taxing business income will then be through association with a permanent establishment. In addition to the OECD Model and the UN Model tests of connection, many countries also tax technical, administrative, and management fees paid to a nonresident by an enterprise that is resident in the country or that constitutes a permanent establishment of a nonresident in the country. Such a rule deals with cases where persons use deductible service fees to reduce the tax base in the country of the paying enterprise without corresponding taxation by that country of the fees received by the nonresident (which will often be a company that is related to the payer).42 Alternatively, management and service fees may be taxed as royalties, which will usually be the preferable course. A suggested provision incorporating the UN features, but not technical or other fees, follows:

Business income is sourced in country X if

  • (1) it is attributable to a permanent establishment of the taxpayer in X;

  • (2) it arises from sales by the taxpayer in X where the taxpayer has a permanent establishment through which goods of the same or similar kind are sold; or

  • (3) it arises from other business activities carried on by the taxpayer in X where the taxpayer has a permanent establishment through which activities of the same or similar kind are carried on.

This provision will not exhaust the taxation of business income. First, there will often be special provisions for specific types of business income that take precedence over this general rule. Second, income of certain types that may or may not be income of a business depending on the circumstances (especially passive income, such as dividends, interest, royalties, or capital gains) will generally be taxable if it falls into either the business income rule or into the specific rules for the type of income in question (although the method of taxation will vary for each case as explained below).

Thus, tax treaties have special rules for international transport income, independent professional services, and income from entertainment and sporting activities. Many countries also add income from international communications and insurance. The OECD tax treaty approach for international transport income is premised on the view that the income will be equally balanced between the two countries, so that it is simpler from an administrative point of view to confine taxation to the country of residence of the company carrying out the international transport.43 In the case of air transport, this assumption will generally be correct because of the restrictions in international airline agreements entered into by governments, which try to share revenues between the airlines of each country, while for shipping very few countries nowadays have substantial shipping industries because of the way that business is organized internationally. While the tax treaty approach thus does little harm, some countries find it easier to use a simple 50/50 rule that divides the income equally between the start and end points of the international transport, an approach also used for international telecommunications income (not separately covered in tax treaties partly because of its recent development and partly because international agreements between countries often share the income between companies in each country).

Professional services income nowadays is generally regarded as the same as business income, and the existence of separate articles in tax treaties is mainly to be explained historically.44 As the outcome under such articles is similar to that for business income generally, special sourcing rules for such income are not often included in domestic law, except where it is intended to include a time threshold, which is discussed below in relation to employment income.

The taxation of insurance is a very specialized topic. Because of the difficulties involved in calculating the profit of an insurance company, some countries simply levy tax on a percentage of the premium income, either generally or specifically for certain types of insurance or in the international area. The basic sourcing rule adopted is the insuring of risks located in a country.

Business (and employment) income from entertainment and sporting activities is sourced in a country when the activity is carried out there; this is because very high incomes can be earned in short periods within a country that may not be captured under the general business income rules.

When special sourcing rules are adopted for particular types of business income in domestic law, they override the general business income source rule. In turn, tax treaties will generally overturn the special rules for insurance and telecommunication income and adopt the general business income rule unless the special rules are preserved by provisions inserted for that purpose (which does occur in bilateral treaties and the UN Model for insurance but not for telecommunications). Dividends, interest, and royalties are often regarded as passive income but may be received in a business context, in which event the rules for taxing business income generally apply.

E. Dividends, Interest, and Royalties

Dividends are usually sourced under domestic law, and tax treaties by the residence of the company paying them. Interest under tax treaties also uses as a criterion the basic residence of the payer,45 but where the interest is borne by the permanent establishment in connection with which the indebtedness is incurred, the interest is sourced by the location of the permanent establishment. Taken together, these rules on interest mean effectively that it is the place where the economic activity giving rise to the payment of the interest occurs that is its source.46 Interest source rules under domestic laws show some variation from this pattern, most commonly adding the case where the interest relates to a loan that is secured by property situated in the country, but tax treaties generally override this rule. The tax treaty rule for the source of interest differs in one respect from the rule suggested in the text, apparently because of the bilateral nature of tax treaties. Sourcing by a branch rather than by the residence of the debtor occurs only where the branch is in one of the treaty countries; otherwise, the residence of the debtor prevails. This treaty rule can give rise to difficulties and is thus not followed by some countries in their treaties.47

Royalties do not have a detailed source rule in the OECD Model, given that taxation is exclusively reserved to the residence country, but almost half of the OECD countries and the UN Model do not follow this pattern. Rather, they replicate the interest-source rule for royalties, that is, residence of the payer with the permanent-establishment qualification. The United States has a sourcing rule of where the property giving rise to the royalties is used48 and can usually have this accepted in its treaties, but less powerful countries may find it more difficult to go their own way. Certainly, domestic law should contain a clear rule for sourcing royalties, as they are one of the most important forms of income internationally—especially so in a world that is coming to be dominated by trade in technological innovation and services rather than goods.

One of the most important aspects of the source rules for dividends, interest, and royalties is the definition of the terms. Most domestic tax laws will have a definition of dividend in relation to the general rules for taxing distributions by legal entities,49 and tax treaties effectively adopt this definition. The reliance on the domestic definition of dividend under tax treaties can cause difficulties, as countries have widely differing definitions, which can lead to the consequence that one country regards a payment as a dividend whereas another country regards it as something else. For example, one country may treat a payment on the liquidation of a company to its shareholders, in whole or in part, as a dividend, whereas another country may treat it as a disposal of the shares (and so covered by the capital gain article in tax treaties). Tax treaties do not usually provide any clear resolution of this “conflict-of-qualifications” problem, except the possibility of the mutual agreement procedure. It follows that whatever definition of dividend is adopted for domestic purposes, problem cases can arise internationally under tax treaties. No simple solution is available.

By contrast, the definitions of “interest” and “royalties” in tax treaties do not rely on domestic definitions. The definition of interest in the OECD Model is income from a debt claim (but excluding penalty charges for late payment). While this definition operates clearly in many cases, financial innovation in recent decades has given rise to many instruments that are effectively loans but that do not relate to a debt claim and are therefore outside the definition (e.g., foreign exchange contracts and swaps can be structured to produce interest equivalents). Increasingly, countries are moving in their domestic laws to ensure that such instruments are taxed consistently with interest, but the rules required for such a regime are likely to be very complex. The result is that what is assimilated to interest under domestic laws varies greatly among countries and the definition used will depend on a number of fundamental policy choices in the taxation of interest.50

The definition of royalties is more straightforward. The essence of the definition in tax treaties, which is followed in the domestic law of many countries, is a payment for the use of intellectual property, including copyrights, patents, know-how, and secret processes. (The term “royalty” is also commonly used for payments to the owner of land or to the state for the right to extract natural resources, but these are income from immovable property and have already been dealt with above.) The OECD Model before 1992 covered equipment rentals in the definition of royalties by including payments for the right to use industrial, commercial, or scientific equipment; the deletion of this item in a bilateral treaty means that equipment rentals come within the business profits article under tax treaties. While the usage covering equipment leasing probably extends beyond the normal understanding of the term royalties, many countries still include equipment rentals in their domestic law definition of royalty.51 As long as positive tax rates are specified for interest and royalties in tax treaties (not the case in the OECD Model but not uncommon in practice), one justification for this inclusion is that interest can be converted into rental income through the device of the financial lease (i.e., a lease that is the equivalent of a loan), but the treatment under the domestic law and tax treaties will effectively be the same.52

One problem of the royalty definition in the OECD Model is the reference to payments “for the use of, or the right to use” patents, etc. This language apparently does not cover disposals of intellectual property and, if so, the royalty definition can be simply avoided as transactions for use can easily be converted into disposal transactions because of the flexibility of patent and copyright law in most countries. For example, a person could be given the right to use a patent in a particular country for a specific period of time in return for payments related to the number of items produced using the patented process, or the patent could be disposed of to the person in respect of that country and time period on the same payment terms. For this reason some countries provide that where some proportion of the payments in relation to intellectual property are contingent on use, then they will be treated as royalties even though the transaction takes the form of a disposal.53

In some countries, technical fees are assimilated into the definition of royalties or are taxed similarly to royalties.54 In the context of tax treaties, similar issues arise. Payments for technical services and the like may be incorporated into the royalties article or subject to a separate but similar article.55 If no such provision is made, then the domestic rules for taxing such income will be overridden by tax treaties.

F. Capital Gains

Capital gains are another area where variation in domestic laws can give rise to problems in their international treatment. Some countries (especially common law countries) have a general conception of capital gains as any gain on an asset other than inventory (and similar property) of a business and personal use assets of an individual (such as consumer durables). Within this group, a number of countries do not tax such capital gains while many others have beneficial rules and tax rates for them. Other countries, especially those based on civil law, have either a much narrower concept of capital gains or no such concept—business profits are taxed with no tax distinction drawn between gains on disposition of inventory and other assets, and individuals are simply taxed on gains on a list of assets without invoking any rubric of capital gain in either case.56 Hence, the use of the term “capital gains” can cause some confusion in an international setting, and it can be argued that it is better avoided even though it is used in the OECD Model.57

Following the tax treaty rules, gains on business assets are generally sourced at the permanent establishment to which the gain is attributable; gains on immovable property are sourced where the property is situated; and gains on other property are sourced where the person disposing of it is resident. A number of countries include special rules in their domestic law and tax treaties for sourcing gains on shares in resident companies in one or more of the following categories: companies whose major assets are immovable property, direct investment interests in companies (usually defined as a certain proportion of the shares, such as 10 percent or 25 percent), and, more rarely, any interest in a closely held company.58 The first two of these are intended to buttress the rules on taxing gains on business assets and immovable property. A taxpayer can easily avoid those rules by holding the relevant assets in a company and then selling the shares in the company.

While the purpose of the rules on companies is understandable, in practice it is not possible to prevent nonresidents from using variations on the same stratagem to avoid these rules. Rather than selling the shares in the resident company directly holding the relevant assets, a taxpayer can hold the assets through several tiers of companies (usually located in tax havens); it is then possible for one higher-tier nonresident company to sell the shares in the nonresident company below it in the tier and so effectively dispose of assets that may be several tiers below. While domestic law can have rules referring to disposal of shares in companies that amount indirectly to disposal of the relevant assets,59 such rules will be almost impossible to enforce and will usually be overridden to a greater or lesser degree by tax treaties.

G. Employment, Services, and Pension Income

1. Employment Income

Employment income is usually sourced by the place where the employment is carried out (and if it is carried out in several places, the income is apportioned between those places). This is followed in tax treaties, with the exception that the OECD Model contains a 183-day presence threshold before a nonresident employee is taxable, if employed by a nonresident employer that does not deduct the relevant salary as part of the expenses of a permanent establishment in the country. Some short time threshold, such as 30, 60, or 90 days, subject to the same conditions, is a sensible rule for domestic law, as no country can successfully tax such employees who are in the country for very short periods. Especially in the context of developing and transition countries that are seeking to attract foreign investment, this kind of rule allows the important exploratory visits to take place before investment decisions are made without tax impediments so far as the employees of the potential investor are concerned. A monetary threshold can also be used as an addition to the time threshold to eliminate small amounts for ease of administration, or as an alternative to the time threshold to try to capture very high amounts of income earned in a short period.60 Tax treaties contain time limits for employment income, but not usually monetary limits.

2. Fringe Benefits Tax

Following the lead of Australia and New Zealand, a few developing and transition countries have adopted fringe benefits taxes to deal with the problems of taxing benefits in kind provided by an employer to an employee.61 The tax is levied directly on the employer at a flat rate and the benefit is then tax exempt in the hands of the employee. Even from a domestic viewpoint, the technical problems of this approach to the fringe benefits problem indicate that the tax should be adopted only when it is politically the only possible way to ensure that the benefits are taxed.62 Otherwise, the more straightforward method of treating fringe benefits as the equivalent of cash wages is to be preferred.

From an international perspective, fringe benefits taxes cause significant problems. First, the application of the residence and source principles to the tax is unclear. Does the residence of the employer or the employee count? Is the sourcing rule the same as for wages? If so, one of the claimed advantages of the tax, the avoidance of allocating benefits to individual employees, is lost. Second, how is relief from double taxation effected in domestic law (especially as other countries may not be using the tax, but taxing employees instead)? At the moment, fringe benefits taxes often lack mechanisms to avoid double taxation. Third, no satisfactory tax treaty mechanism has yet been found for dealing with such taxes.63 Where the traditional approach of taxing fringe benefits to the employee is adopted, tax treaties experience little difficulty because the matter is dealt with by the employment income article.

For developing and transition countries, this fringe benefits tax problem is more than theoretical. As already noted, the taxation of the salaries and benefits of expatriate employees of foreign investors can be a significant factor in investment decisions. If a fringe benefits tax is adopted, it will not be relieved in the country of the expatriate employees’ residence if that country applies a foreign tax credit, with resulting double taxation. The foreign investor, rather than the employees, in practice will absorb the fringe benefits tax so that it is simply an additional cost of—and disincentive to—the investment. Given that fringe benefits for good reasons often figure importantly in the remuneration packages of expatriate employees, the cost can become significant. Indeed, even under the traditional approach to fringe benefits, there is an argument for special rules to deal with such employees. Carrying these rules over into a fringe benefits tax will ameliorate but not solve the problem that the fringe benefits tax causes in the international context.

3. Services Income

The employment income source rule is often extended to all forms of services income. This has two effects. First, not only is the employee taxable but also the employer, where the services of the employee are part of the rendering of services of the employer to a third party. Second, in the case of professional services and services with a high value added where no employment is involved, the person rendering the services is taxable without the need for some permanent presence, as is generally true for business income. Because of the increasing significance of high-cost services in international trade, it is sensible for countries to seek to tax such services. They can do this either by adopting a general rule for services based on the place of performance or by including the rendering of services other than as an employee in the definition of permanent establishment. If either is done, it would be sensible to include a short time threshold, for similar reasons as in relation to employment income. Tax treaties based on the OECD Model would eliminate the tax in such a case but the UN Model would allow it, subject to a time threshold.64 The addition of a monetary threshold in addition to or in lieu of a time threshold raises similar considerations as for employment income. The UN (but not the OECD) Model includes such a threshold for independent personal (i.e., professional) services but not for other services. Consistency across employment, professional, and other high-cost services makes sense from a policy viewpoint, but tax treaties will generally not produce this outcome.

Under domestic law, it is usually necessary (e.g., in relation to withholding on wage income) to draw distinctions between employment and business income.65 Employers and employees may gain some advantage (in relation not simply to the income tax, but also to payroll-based taxes and even labor law) in converting what is essentially an employment relationship into a business one. One way to achieve this is for the employee to form a company that then contracts the services of the former employee to the former employer (the person is now an employee of the company the person owns but has control over how much income is received from the company in wages and how much in other forms). Domestic tax laws often deal with this problem by expansively defining employment to include such cases or by extending withholding to certain types of business income. As far as the former route is adopted, the rules will generally flow over into tax treaties (tax treaties have a special rule for directors of companies, sourcing directors’ fees by the residence of the company, although under the domestic laws of most countries these fees are treated as employment income).

Arrangements designed to convert employment into business income have given rise to particular problems in international situations through manipulation of the time limits for taxing employment income under tax treaties. The OECD has accordingly developed rules for treaty purposes that seek to determine whether there is a genuine employment. These rules can be considered for use in domestic law. The rules address where the responsibility, risk, and authority to give instructions lie, where the work is carried out, the method of calculation of remuneration, who provides facilities, and the methods for the conduct of the work.66

One form of high-value service that is usually the subject of special rules concerns entertainers and athletes. Their income can be structured, as they desire, as business or employment income (in the latter case through the use of “star” companies similar to the situation just dealt with). Whether the income is employment or business, it is sourced under treaties by reference to the place of performance of the services without time thresholds.67 Monetary limits may be used to segregate highly paid pop stars from the lower-paid members of, say, a visiting symphony orchestra, although tax treaties usually employ other methods to make this kind of separation (usually based on exceptions for official cultural exchange programs). Tax treaties also usually contain special provisions to look through star companies to the entertainer or athlete and to attribute all the income to that person. A similar rule may be useful in domestic law.

Even with some or all of this panoply of rules to cater to the problems of taxing high-cost services, the growing importance of services in the world economy is going to increase pressure on both source and residence country taxation. A successful computer software company, for example, could locate its programming and management staff in some suitably pleasant tax haven and market its products through mail order solicited by advertisements in computer magazines or on the Internet. Taxing the profits of such a company and the salaries of its employees in the countries where its products are sold is almost impossible, is not provided for in the domestic laws of most countries, and may be prevented by tax treaties. Similarly, much of the income of entertainers and athletes comes from sources not directly related to actual performance, such as video and sound recordings and endorsements. Capturing this indirect income by the country of place of performance entails the same kinds of problems.68

As already noted in the discussion of business income and royalties, some countries are responding to this problem by employing a definition of source based on the residence of the payer.69 As yet, such a shift does not have general international acceptance. This may, however, be a case where it is wise for domestic law in developing and transition countries to depart from the norms implicit in existing tax treaties and to seek to change their treaty practice accordingly. Considerable resistance will be encountered in tax treaty negotiations with industrial countries if a developing or transition country adopts this course.

4. Pension Income

A form of income often closely related to services income is pension income. Where the pension has been financed by contributions out of services income that have received favorable tax treatment in the country of performance (by exclusion of the contribution from income or a deduction for the contribution), a rule based on the place of performance of the services may be thought suitable for sourcing the pension. This approach will not be practical when the services have been rendered in many countries over a period of many years. For this reason and because pensions can take other forms (such as government benefits, distributions from social security schemes, and purchased annuities), they are often sourced by the residence of the recipient of the pension or by the residence of the payer of the pension. The OECD Model adopts the former while many tax treaties in practice adopt some form of the latter, especially in regard to social security and government benefits. The UN Model adds, in one of its variants, a permanent-establishment sourcing rule as a gloss on the residence-of-the-payer rule. Pensions and similar payments also give rise to some more general problems under international taxation, which are taken up below.

In developing countries, pensions of all kinds are much less common than in industrial countries, whereas they are widespread in transition countries. In both developing and transition countries, pensions tend to be small in amount (especially as a result of recent inflation in transition countries) and are often not taxable either because of an express exemption in the domestic tax law or because they fall entirely within the tax-free zone established by the tax rate scale or by personal allowances. Some developing and transition countries have already experienced immigration of pensioners from industrial countries in part to take advantage of a lower cost of living. It is not advisable therefore to be dogmatic on a source rule for pensions.

As with residence rules, there may be special sourcing-type rules for government employees. Although these rules are not often found in domestic tax laws, tax treaties generally limit taxation of the employee’s wages to the government employing the person, except for local employees. A variant of this rule is extended by tax treaties to pensions paid by the government to its former employees.

H. Other Income

For other income,70 the OECD Model basically adopts a residence-only tax rule. The UN Model allows the country of source to tax the income in accordance with its own source rules without defining such rules. The domestic law of a transition or developing country can sensibly adopt this approach with some generally expressed source rule as a residual.

VI. Taxation of Residents

A. Rate Scale and Personal Allowances

The main reasons for taxing residents on their worldwide income have to do with the fairness of the tax system. When a country adopts a progressive income tax rate scale for individuals, it is usually motivated by the idea that it is fair for higher-income individuals to pay proportionally more of their income as tax. However, unless the individual is taxed on worldwide income, this goal may not be achieved for an individual with income from more than one country. If the progressive tax rates are the same in each country and each taxes only on a source basis, an individual receiving income from each country will pay less tax in total to both countries than an individual who receives the same total amount of income from only one of the countries. This is doubly unfair; not only are two like individuals taxed differently, but individuals are obviously encouraged to split their income between the countries, an avenue that is more likely to be availed of by a high-income taxpayer. There is also a lack of neutrality in such a system because of the splitting incentive that it creates.

Given that it is not practical for a country to tax all individuals in the world on a worldwide basis, the general policy that has been adopted is to tax only residents of a country in this way. A country can generally enforce its tax claims against residents (i.e., persons who have substantial personal contacts with the country), whereas a single-source country is unlikely to know the total income of a nonresident taxpayer and will face enforcement problems in relation to income arising outside the country. From a policy viewpoint, it also seems appropriate for the country taxing on the basis of personal allegiance of the taxpayer to be the one that takes account of the taxpayer’s personal attributes. This concept relates not just to the progressive rate scale but also to tax allowances, such as those relating to a zone of tax-free income (which is closely related to progressivity), family size and composition, medical costs, and subsidies for home ownership.

Conversely, for nonresidents, this approach implies flat-rate taxation of income sourced in a country and no tax allowances for personal attributes. If residence is changed part way through the tax year, then the taxpayer should change from one regime to the other and allowances should be adjusted to account for the fact that the entitlement is for only part of the year.

In practice, this approach to taxation of residents and nonresidents is often not fully realized. While dividends, interest, and royalties received by nonresidents are generally taxed on a flat-rate basis, the progressive rate scale is often applied to many other forms of income of nonresidents (although the zone of tax-free income is often not applied). Personal allowances (especially those applied by a developing or transition country to an expatriate from an industrial country) are often not significant in revenue terms in relation to nonresidents because of the small number of taxpayers affected. Hence, it is easier from an administrative perspective to apply them to all individual taxpayers and not just to residents or at least to apply them on a whole-year basis to any individual who is resident for part or all of the tax year.

Confining personal reliefs to residents of a country does not infringe on the nondiscrimination rules of tax treaties, which generally seek to ensure that residents and nonresidents are treated alike under the tax law of a country. The reliefs are recognized as part of the residence jurisdiction of the taxing country, so that residents and nonresidents are not treated as being in the same circumstances, which is a threshold condition for the application of the nondiscrimination principle.71 As tax treaties otherwise do not deal with personal reliefs, the tiebreaker rule in the tax treaty that addresses dual residence will not carry over into domestic law for this purpose. Hence, a dual resident will be entitled to the personal reliefs in more than one country, and a special rule in domestic law limiting entitlement to reliefs in such cases is necessary if it is desired to track the tax treaty rule. For developing and transition countries, this qualification seems an unnecessary refinement.

B. Expatriates

There are, however, a number of refinements that need to be considered by developing and transition countries in the taxation of expatriate employees who become residents of the country for a limited time. Expatriate employees will usually be brought into a country where the skills necessary for a particular job are lacking in the country and hence they will usually be very highly paid—especially in comparison with the general level of wages in the country. They can be employed either by foreign investors or by local employers. While in the past when colonial attitudes prevailed, foreign investors may have been inclined to use expatriate employees for all senior positions whether or not local skills were available, this position has now generally changed for cost reasons and out of greater sensitivity to national sentiment. It will be assumed in what follows that the expatriate employee is providing skills that are in short supply in a country and that the country wishes to encourage—or at least not discourage—the importation of the skills.

For the purposes of the discussion we can distinguish several different situations in which expatriate employees will be used in a country. First, there is the person who comes in to do a specific task and leaves when the task is complete; the stay is very short term. Such persons will generally not become residents under the domestic law of the country visited and, in the case of employment by a foreign resident investor, will often not be taxable on their employment income by reason of either tax treaties or provisions in domestic law that set time limits on source taxation of employees. Second, there is the person who comes for a more extended stay, say, six months to two years, but who leaves all or part of the person’s family and a permanent home behind in the home country. This person will generally become a resident of the country where the work is performed under a 183-day rule while remaining resident in his or her home country, and under tax treaties residence will usually be allocated to the home country by the tiebreaker rule. Third, there is the person who comes for a yet more extended stay, but always with the intention of returning to the home country (as evidenced by the ownership of property there and the limited period of the assignment). This person may cease to be a resident in the home country for the period of the assignment, but, if not, residence will usually be allocated by tax treaties to the country where the work is being performed. Finally, there is the person who, at the outset, or more usually after an initial period in the country, decides to remain in the country and “go local.” This person will usually cease to be a resident of the home country entirely.72

Because of the high costs involved with expatriate employees, employers will usually require them to go local after two or three years in the case of placement in an industrial country and after a longer period, say, three to five years, in developing or transition countries; that is, they are thereafter treated in the same way as local employees and do not receive special expatriate allowances. The basic structure of the remuneration of employees in the second and third categories above (which are the most common problem cases) will be to provide them with salary and benefits designed to keep their after-tax salary before the assignment intact, compensate them for the additional costs incurred as a result of the assignment, and provide them with a bonus for undertaking the assignment, which will often be viewed as having an element of hardship (such as separation from family, personal security and general living conditions in the country of assignment, and complication of personal affairs).

Typical benefits will thus include free or subsidized accommodation in the country of assignment, payment of private education fees for children, free airfares between the home country and the country of assignment on a regular basis, tax supplements to remove additional tax burdens and free access to specialist tax advice, special pension scheme arrangements, special medical insurance, free car and driver, and general security arrangements, plus a bonus of, say, 25-50 percent of salary.

1. Rate Scale

If an expatriate has become a resident of a developing or transition country under its law, taxation of worldwide income under the progressive rate scale will occur. The appropriateness of the rate scale to the expatriate thus becomes an issue. Generally, it will have been enacted with local incomes in mind. This means in many cases that the maximum tax rate is reached very quickly in comparison with industrial country tax rate scales, because of the generally lower level of local incomes. The result is a greater tax burden on the expatriate than in the home country, even if the maximum tax rates in the countries are the same. There is also a tendency for maximum tax rates in developing and transition countries to be higher than in some major industrial countries.73 The employer thus will pay a tax supplement to the employee to eliminate the additional taxation, and, because the tax supplement is really just additional salary, it should also be taxable and grossed up for the additional tax accordingly.

To obviate this problem, some countries provide in effect special tax rate scales for expatriates (by giving special additional personal allowances or by stretching the tax brackets) and do not impose tax on tax supplements. Both of these measures represent generous treatment of expatriates. They may be risky in a political sense as favoring wealthy foreigners, but they may send a very positive signal to both foreign investors and to potential expatriate employees. The amount of revenue at stake in terms of overall revenues will usually be small in view of the small numbers of employees involved.

2. Fringe Benefits

The case of fringe benefits given to expatriates is clearer, however. Benefits that are viewed as simply part of a person’s working conditions are not generally taxable as fringe benefits (such as pleasant office accommodation, access to labor-saving technology, or payment of costs of work-related travel). It can be argued that although many of the benefits received by expatriates would amount to taxable fringe benefits if received by local employees, in the case of expatriates they are simply part of the conditions of work in the country. Thus, free accommodation in the country of work when the expatriate has left family in a residence in the home country and airfares to return home on a regular basis are little different from payment of the cost of work-related travel. Arrangements to ensure personal security may also be regarded as part of the work conditions.

In addition, taxation of fringe benefits in many countries takes account of disadvantageous work conditions (such as working in remote locations), and, again, the expatriate situation can be assimilated into this thinking. For example, while the expatriate may have sent school-age children to a public school back home, the only realistic option for language and cultural reasons may be to send them to a private school in the country of work to get a comparable education that will allow the children to be absorbed back into the public schools on return home. Free provision of health care up to the standard in the home country can be justified in the same way.

Hence, provided the rules are carefully framed and judiciously enforced by selective audits of expatriates to prevent abuses, nontaxation of such benefits can be justified on the basis of the special position of the expatriate. Indeed, this approach can be generalized for the converse case where residents of the country become expatriates in another country, although in practice legislation on this topic will be much rarer. A provision to this end follows:

  • (1) A foreign service allowance paid in respect of the additional expenses incurred by reason of employment in X is exempt income [in an amount not exceeding x percent of income (apart from this exemption)].

  • (2) Paragraph 1 does not apply to any allowance in respect of income tax payable in X. Regulations may further limit the exemption provided under paragraph 1.

  • (3) This article applies to a taxpayer if

    • (a) the taxpayer was a resident of another country under its tax law immediately before undertaking the employment in respect of which the allowance is paid;

    • (b) the taxpayer became a resident of X for tax purposes solely as a result of carrying out the duties of the employment; and

    • (c) the employment in X lasts no longer than three years.

As indicated by the material in the article in square brackets, the exemption may be limited to a percentage of the taxpayer’s income before the exemption is applied, which is a method to limit abuse (e.g., to prevent an employer from paying such an employee a relatively low salary and a substantial foreign service allowance). A limitation is imposed by paragraph 2, which provides that the exemption does not extend to allowances in respect of income tax (tax supplements) that the individual may have to pay in X. Paragraph 3 provides a special test to determine whether the person is entitled to the allowance, based on where the person was resident before moving as a result of the person’s employment, and limits eligibility to the allowance to employment in the country for a maximum of three years.

3. Foreign-Source Income

The next problem is the treatment of income of the expatriate derived outside the country where the work is carried out. As already noted, expatriates in the second category referred to above will normally be treated as residents of the home country under the tiebreaker rule in tax treaties, whereas those in the third category will usually be treated as residents of the country where the work is performed. Depending on the precise terms of the treaty, the effect on the second category may be to eliminate taxation by the country where the work is conducted on income derived by the expatriate from sources outside that country and to limit or exclude taxation by that country on dividends, interest, royalties, and other kinds of income derived from sources in that country. For the third category of expatriate, the effect of tax treaties will usually be to permit unlimited taxation of their worldwide income by the country where the work is performed and to limit or eliminate taxation in the home country of income derived by the expatriate from sources in the country of work, the home country, or other countries. Where there is no tax treaty, there will frequently be unrelieved residence-residence double taxation for both the second and third categories. The overall tax position is thus complex and very likely to lead to excessive tax burdens.

If the country where the work is to be done wishes to attract the skills of expatriates, it may seek to deal with the problem in its domestic law. As with fringe benefits, the simplest mechanism is to exempt for a limited time income (other than employment income) derived by the expatriate from sources outside the country. An example of possible statutory language is as follows:

The foreign-source income of a resident of X is exempt income if

  • (a) it is not employment income;

  • (b) it does not benefit from a tax reduction under a tax treaty entered into by X;

  • (c) the taxpayer became a resident solely as a result of employment exercised in X; and

  • (d) the employment in X lasts no longer than three years.

4. Pensions and Social Security

It is common in many industrial countries for higher-paid employees with special skills to become members of private pension schemes. Under the tax law of industrial countries, the contributions to, income of, and distributions from the pension scheme will usually be subject to favorable tax treatment as a means of encouraging the employee to save for retirement and so not require support from the state in old age. When such a person comes to a developing or transition country as an expatriate employee, it will often be found that the country has no similar provisions in its laws (because of a lack of pension arrangements for old age in developing countries or because of full state provision of pensions in transition countries), or that such provisions as do exist do not apply to foreign pension schemes, or that ceilings on tax-favored contributions to local pension schemes are low by international standards. As entitlements under private pension schemes are often not portable between schemes within a country, let alone across international borders, the expatriate usually has no option but to remain a member of the pension scheme in the home country. The result is an increased tax burden on the employee and employer simply to maintain the existing pension entitlements of the employee, which will not come into effect until many years after the employee leaves the developing or transition country.

Although tax treaties increasingly are seeking to deal with this problem, most existing treaties do not.74 Hence, countries may wish in their domestic law to recognize the position of foreign pension schemes and to seek to remove the tax problems they and their members currently experience. How this is to be achieved depends on the existing arrangements for domestic pension schemes in the country where the work occurs. If the country has schemes similar to those used in most industrial countries, then it is possible to extend the same preferential treatment to foreign schemes on the basis of reciprocity. Alternatively, if there are no such schemes or if reciprocity is difficult to achieve, a deduction may be provided for contributions from expatriate employees and their employers to pension schemes in the home country of the employee limited by reference to some proportion of salary and employer contributions (say, 10 or 15 percent).

Social security taxes present similar problems, especially in transition countries, where they can amount to up to 50 percent of payroll before tax. Although such taxes are separate from income taxes and are not covered by tax treaties, they are intimately related as far as the employee is concerned, especially as regards provision for retirement. An expatriate employee (or the employer, depending on where the tax is formally levied) will often find that social security contributions must be paid in respect of the employee’s salary in both the home country and the country where the employee is working on the basis of residence in each country (the definition of residence under the social security tax law being in question here, but with similar issues to the income tax).75 There will generally be no relief from this double taxation, and in addition the expatriate employee or employer will often be insuring privately in respect of some of the matters that may be covered by the social security system (such as medical treatment) because of the difficulties of extracting adequate benefits from the systems of developing or transition countries in such cases. Even if there is no double tax, the local tax may be quite high.

For expatriate employees, the most relevant social security system is that of the home country, because they will avail themselves of very few or no benefits under the system of the country where the work occurs. Hence, it is sensible for the country where the work is carried out not to levy its social security tax on expatriate employees and at the same time to deny benefits under its system. This leaves the matter to be dealt with under the system of the home country or private insurance. Social security totalization agreements are nowadays being entered into between countries to deal with these kinds of problems, but the development of such treaties lags far behind tax treaties.76

C. Relief from Double Taxation

In industrial countries, the major residence country tax issue is generally seen as the relief of double taxation on income that has been taxed at source in another country. For developing and transition countries, this issue is less of a problem because residents will derive much less income from foreign sources. So far as there is foreign income, it will frequently be the result of (often illegal) capital flight to low-tax jurisdictions, in which event the problem for the residence country is detection and taxation of the income, not the relief of double taxation. Hence, the discussion of this issue will be fairly abbreviated and will not delve into all the well-known intricacies of credit and exemption systems of industrial countries.

It is necessary to distinguish among four basic methods in this area. The first is for a country not to assert jurisdiction to tax foreign-source income of residents (either at all or for selected types of income). This territorial approach to taxation (taxing only income sourced in the country) means that the country is not following the usual international norm of worldwide taxation of residents and so is not strictly a method for relieving double taxation as residence-source double taxation will simply not arise for its residents.77

The second method is the exemption system, under which foreign-source income is exempted in the country of residence. If the exemption is unconditional and the exempted income does not affect in any way the taxation of other income, then in substance the result is the same as a purely territorial system. Most exemption systems are not of this kind and so are to be distinguished from territorial systems. Most countries using an exemption system adopt exemption with progression, under which the total tax on all income of a resident is calculated, and then the average rate of tax is applied to the income that does not enjoy the exemption.78 Exemption systems are also increasingly subject to various conditions to ensure satisfaction of the assumption underlying the system (that the income has been taxed in the source country at its ordinary rates).79 These conditions can consist of subject-to-tax tests (including the specification of tax rates) or selective application of exemption to foreign countries under domestic law or tax treaties.80 In particular, the exemption is usually not given where the source tax has been reduced or eliminated by a tax treaty. The result is that there are no countries asserting jurisdiction to tax worldwide income that give an exemption for all kinds of foreign income; where a country is referred to as an exemption country, this generally means that it provides some form of exemption to business income, dividends received from direct investments in foreign companies, and often employment income, with a credit being used in other cases.

The third system is the foreign tax credit system under which a credit against total tax on worldwide income is given for foreign taxes paid on foreign income by a resident up to the amount of domestic tax on that income. This limit is designed to ensure that foreign taxes do not reduce the tax on the domestic income of residents and is calculated by applying the average rate of tax on the worldwide income before the credit to the foreign-source income. In its simplest form, this limit is applied to foreign income in its entirety, without distinguishing the type of income and the country where it is sourced.

The fourth system is to give a deduction for foreign income taxes in the calculation of taxable income. While this system is used in some countries, often as a fallback from a foreign tax credit where the credit may not be of use to the taxpayer,81 it is not widely accepted as a method for use on its own and, more specifically, is not used in tax treaties.

It can be argued that relief of double taxation in either credit or exemption form involves a number of complexities that are best avoided by developing or transition countries. Pure territorial taxation, however, simply invites tax avoidance through the moving of income offshore, and once qualifications on the pure territorial principle are admitted, such as limiting it to certain kinds of income, it is hard to see that any great simplicity is achieved as problems of characterization of income arise, as well as incentives to convert income from one form to another. Similar difficulties arise when a conditional exemption system is used. For this reason, a simple foreign tax credit system is probably suitable for most such countries—it asserts the worldwide jurisdiction to tax income of residents and does not require significant refinements of calculation. It leaves open the greatest scope for elaboration of the system by domestic law and tax treaties in the future without having to repeal or modify any exemption (often a difficult process politically because of entrenched interests). Given that tax treaties are premised on an item-by-item foreign tax credit limit, rather than on a worldwide limit aggregating all foreign income of the taxpayer, the item-by-item limit is probably easiest to use in domestic law.82

Whichever double tax relief system is adopted, some method of apportioning deductions between domestic and foreign income will be necessary. Where deductions allocated to foreign income exceed that income, the loss should not be available for use against domestic income. In practice, most credit countries do end up with some cases of effective exemption for foreign income.83 One possible example in this context is in relation to the foreign income of expatriates discussed above.

Tax treaties invariably contain an article for relief of residence-source double taxation (they are built on the assumption that each country will assert jurisdiction to tax the worldwide income of residents, which is another reason for asserting this jurisdiction in domestic law). The only methods specified in tax treaties are exemption and credit, but there is no need for the treaty method to follow that used in domestic law. Some countries have no relief method under domestic law, so that the only relief is under treaties,84 while some countries have the credit method in domestic law but use the exemption method for selected kinds of income in treaties. Where the country has the credit method in its treaties, this is not generally regarded as preventing it from using the exemption method in domestic law, as exemption is seen as more generous than the credit method and therefore not inconsistent with the treaty obligation. Where the exemption method is adopted by tax treaties, the exemption-with-progression system is usually expressly authorized.

Special double tax relief rules are often provided for foreign direct investment. As already noted, the exemption system is often targeted to foreign-source business income and dividends received by a resident company from a direct investment in a nonresident company. Direct investment in a foreign company is equated with business income to ensure that no bias is created as to the business form used. If an exemption is granted for the business income of a branch in a foreign country, then it should make no difference that the business income is generated by a subsidiary in that country and then repatriated as dividends. By parity of reasoning under a credit system, a resident company should get a foreign tax credit not only for foreign tax paid by a branch but also for foreign tax paid by a subsidiary. This credit, referred to as an underlying or indirect foreign tax credit, in practice involves a number of complexities that most developing or transition countries would do well to avoid. It needs to be recognized, however, that failure to grant an indirect credit creates a bias against investment abroad by residents of the country in the form of subsidiaries. If such investment becomes important to the country, the indirect foreign tax credit issue should be addressed either in tax treaties or in domestic law, or both.

The article in tax treaties on relief from double taxation may also contain special rules for direct investment in a developing or a transition country by a foreign investor to preserve the effect of tax incentives granted by the developing or transition country. This topic is discussed in relation to tax sparing in chapter 23.

D. Capital Flight

The more important residence tax problem for developing and transition countries is capital flight. Many residents, especially those with the greatest wealth, will seek to send their wealth abroad. They may be concerned about devaluation of their own currency and wish to hold foreign currency, which may not be legally possible in their countries; they may be afraid of confiscation (by the state or criminal gangs) or civil unrest; and they may seek not to pay tax on the income produced by the wealth, which itself may have been obtained by illegal means or may represent income that was not declared for tax purposes. Whatever the reasons in any given case, it is clear that capital flight from developing and especially transition countries is a major problem; the need of these countries, on the contrary, is to retain domestic capital for productive local investment.

Most of this fleeing capital finds its way to tax havens, which may be defined for this purpose as low-tax jurisdictions that have bank and other secrecy laws that allow the ownership of assets to be concealed. For transition countries, it is well known that Cyprus is a major destination of nervous capital. For developing countries, there are any number of other tax havens only too willing to assist. Indeed, so lucrative does the business seem that many developing and transition countries actively consider turning themselves into tax havens.

If the money simply finds its way into an anonymous bank account and the income earned thereon is not declared for income tax purposes, then assuming that the residence country asserts jurisdiction to tax the income, this is a case of tax fraud (deliberate nondisclosure). The problem here is one of detection and tax administration. More sophisticated taxpayers may wish to ensure that no tax liability arises in respect of the wealth, and there are a number of stratagems that they can employ. The simplest form is to invest in shares in a tax haven company that in turn simply invests in a very safe form (such as U.S.-dollar-denominated bonds with a high credit rating) and accumulates the interest income for further similar investment. If the shareholder desires the return of the original investment and the income that has accrued in the company, an associate simply buys the shares at a price based on their asset backing. The company is not taxed on the interest that accrues on the bonds (or is taxed at a very low rate) because it is located in a tax haven (from the point of view of the residence country of the investor, it is foreign-source income of a nonresident) and the investor is not taxable on the interest because it accrues to the company and not to the investor.85 The investor will be taxable in the residence country, if at all, only on the profit on the sale of the shares, but can postpone this tax for many years by not selling. In any event, many developing and transition countries do not tax gains on the sale of shares.

To counter this kind of activity, special rules are required in the domestic tax law of the residence country, in effect to look through the company and tax the resident investor on the underlying income. A number of industrial countries have such laws but they are usually very complex. For developing and transition countries, a simpler provision can be inserted in the tax law to give a discretion to tax and thus to send a signal that such cases will be pursued when detected.86 A provision may be drafted along the following lines:

  • (1) Where a resident of X has entered into a transaction that converts income into foreign-source income derived from a tax haven by another person, the tax administration may adjust the income and foreign tax credit position of the resident to reverse the tax effect of the transaction.

  • (2) The tax administration may treat a foreign country as a tax haven if that country has

    • (a) effective tax rates significantly lower than those of X; or

    • (b) laws providing for the secrecy of financial or corporate information that facilitate the concealment of the identity of the real owner of any asset or income.

This provision is not generally regarded as breaching tax treaty obligations in the unlikely event that there is a treaty with the tax haven. There will still be an information problem if such a provision is inserted into the law, and the investor will no doubt be relying on lack of information as much as the interposing of the company to avoid tax. To overcome this problem, it is necessary to have a question in the tax return or tax declaration that requires the taxpayer to disclose investments in nonresident entities, which will prompt the tax administration to inquire further. If the resident investor deliberately answers this question incorrectly, as is likely, the taxpayer’s position is back to tax fraud and problems of detection.

The information problem is almost impossible to solve. The tax haven will usually not enter into tax treaties, or if it does, it will change the exchange-of-information article so as not to require disclosure in relation to banks’ tax haven operations. As tax treaties generally provide the only way for tax administrations in different countries to exchange information, cooperation in the disclosure of the information from the tax haven will not be forthcoming. For this reason and many others, developing and transition countries should be wary of entering into tax treaties with tax havens. The best that developing and transition countries can do for now to deal with capital flight to tax havens is to try to remove the conditions that give rise to the flight in the first instance and to apply severe penalties in relation to tax fraud involving tax havens.

E. Change of Residence for Tax Reasons

One other residence country tax problem can be noted in conclusion. Some residents who anticipate deriving a substantial amount of foreign-source income may be tempted to change their residence before the income is received so that it becomes foreign-source income of a nonresident from the point of view of the former residence country. Obviously the change would be made to a country that would not tax the income (possibly a tax haven) and would occur only if there were no substantial source country tax on the income (because otherwise the residual residence country taxation is likely to be minor). Some industrial countries have special rules to deal with this problem, but they may be regarded as unnecessary for developing or transition countries. The main problem for such countries is in fact likely to be the other way around, that is, the rules of industrial countries in this area may create problems for expatriate taxpayers who become residents. The rules outlined above to deal with the tax problems of expatriates may also assist in overcoming this problem.87

VII. Taxation of Nonresidents

As already noted, general principles suggest that the income of nonresidents should be taxed on a flat-rate basis, as progression is a matter for the residence country. In practice, some taxes on nonresidents are collected on a flat-rate basis, but more for administrative convenience than principle. Because of the general rule found in most legal systems that one country will not assist another in enforcing its tax laws and because of the general administrative difficulties of dealing with persons and assets outside a country, the source country will be well advised to enforce its tax claim on the payer of the income before the payment leaves the country in cases where the recipient does not have any substantial connection with the country, such as a permanent establishment. Hence, it has become accepted as a general principle of international taxation that taxation of passive income unconnected with a business in a country is enforced by flat-rate final withholding taxes, whereas tax on business income arising from a permanent establishment is levied on net income and is collected by the normal assessment system applied to businesses of residents (which may also include some elements of withholding and payment of tax by installments).

For other forms of income, there is less consistency in practice between flat-rate withholding and tax by assessment, although where assessment is used it is normally in accordance with the rate scale applicable to residents, rather than with a special flat-rate scale for nonresidents (although personal allowances including a tax-free amount are often confined to residents). The discussion of taxation of nonresidents will thus start with the related issues of tax rates, method of collection, the use or not of assessments, and the effect of tax treaties, taking the categories of income in turn as for the source area. It will then turn to a number of other issues affecting nonresidents of concern to developing and transition countries.

A. Income from Immovable Property

Income of nonresidents from immovable property is taxed by some countries on a flat-rate final withholding basis on gross rent and by others on an assessment basis. Some countries provide an option to nonresident taxpayers as to the method of taxation88 since, although final withholding is simple, it can prove very rough and ready because of the wide variation that occurs in the amount of deductions relating to income from immovable property (e.g., the full amount to purchase the property, or none of it, may have been borrowed, leading to very different amounts of interest deductions). As enforcement in this case is not generally a problem (assuming that the tax administration can execute against the immovable property for unpaid tax), tax by assessment on a net basis seems the fairer approach, and requiring private residential tenants to withhold on rental payments is unlikely to be enforced effectively. Tax treaties do not generally constrain domestic law in this case.

B. Business Income

In the case of business income of a nonresident sourced in a country, income attributable to a permanent establishment (or otherwise associated with a permanent establishment and sourced in the country) is generally taxed on a net assessment basis. Tax treaties usually require this approach in the case of income subject to the business profits article but, because of their convoluted drafting, the actual extent of this obligation is not obvious at first sight. The business profits article is usually expressed to be subject to other articles of the treaty, but then other articles either refer the matter back to the business profits article in respect of profits attributable to a permanent establishment (dividends, interest, royalties, and other income) or adopt in effect the same rule as the business profits article (capital gains and implicitly, at least according to the OECD Commentary,89 income from independent personal services).

Articles that may involve business profits and that override the tax treaty requirement of taxation on a net basis concern income from immovable property (above), international transport, and entertainment and sporting activities. In the case of international transport, source taxation is generally excluded (although the UN Model has a little-used variant for shipping) and in the case of entertainment and sporting activities, taxation on a gross withholding basis is permitted. Taxation by withholding is usually permitted for dividends, interest, and royalties that are not attributable to a permanent establishment.

To the extent that the domestic law provides for taxation, on a net or a withholding basis, of technical fees paid to nonresidents, tax treaties will usually override and prevent the tax levy if the fees are not attributable to a permanent establishment while requiring as a result of the nondiscrimination article that a deduction be given to the permanent establishment or resident company that incurred the expenses, subject to the amount being at arm’s length in the case of related parties. Nonresident companies may try to exploit this situation, but depending on the circumstances, it may be possible to find means within the tax treaty to levy tax on both the technical fees and on the salaries of the personnel providing the services.90 It was noted above that most tax treaties do not deal separately with insurance and telecommunication income, so that the permanent establishment requirement applies, with the result that the profits from these activities in a country are often not taxable. A number of countries nonetheless apply (relatively low, say, 5 percent) flat-rate withholding taxes on insurance premiums, either in the international area specifically or more generally and seek to protect this levy in their tax treaties.

One particular problem that some transition countries experience in the area of business income is the treatment of deductions. In a number of countries, the tax laws, for the purpose of wage control, have denied deductions for wages in excess of a very low threshold. Deductions for other expenses, such as advertising and interest, may also be limited. There has been some debate on the extent to which the requirements of tax treaties that permanent establishments be taxed on a net basis override provisions of domestic law that deny deductions that affect the determination of profit. While it is unlikely that tax treaties will be interpreted to override the denial of deductions in marginal areas where denial is quite common under domestic laws (e.g., entertainment deductions), it is another matter where a fundamental matter of profit determination such as the treatment of wages is concerned.

A number of industrial countries have inserted special provisions in their recent tax treaties with transition countries to attempt to clarify the matter for permanent establishments and to ensure that subsidiaries of direct investors from their countries also get deductions for their full wage costs (because the only tax treaty rule that potentially covers the subsidiary case, the nondiscrimination article, is unlikely to be of assistance).91 Some transition countries have modified domestic law so that the denial of wage deductions does not apply to branches and subsidiaries of foreign direct investors, and others have repealed the wage deduction denial entirely. For some industrial countries, these rules in the transition country tax systems have raised the more fundamental question of whether their “profit” taxes are income taxes at all in the generally understood sense and have consequently slowed down the development of tax treaty networks.

C. Dividends, Interest, and Royalties

In the case of dividends, interest, and royalties paid to nonresidents, domestic law usually provides for flat-rate final withholding tax on the gross amount if they are sourced in the country and not attributable to a permanent establishment. The tax rate is typically set at 20-30 percent in developing and transition countries and then is often reduced to 10—20 percent in tax treaties. The rates are set at this level in domestic law to leave negotiating room in the tax treaty process but usually to be below the normal company tax rate in recognition of the fact that the tax is gross and does not take account of expenses. In tax treaty negotiations, developing and transition countries will come under considerable pressure from industrial countries to reduce withholding tax rates on interest and royalties to zero or near zero (special considerations applicable to dividends are discussed further below). The argument used by industrial countries is that the gross tax often wipes out the entire profit, with the result that the price charged to the resident or permanent establishment in the country is increased (i.e., the tax is passed back to the payer) with adverse consequences for the import of capital and technology.

While gross-up for withholding taxes (usually by increase in the interest or royalty rate) undoubtedly occurs and is detrimental to developing and transition countries, reduction of tax rates to zero or near zero likewise produces problems and the appropriate course to take is a matter of judgment. If the treaty tax rate on interest is 10 percent, then banks that lend to residents of the country will find it difficult to make a profit. For example, if the cost of funds of the bank is 9 percent and its lending rate is 10 percent, then on a loan of $1,000 it will make $10 before tax and other expenses besides interest, but the withholding tax will be $10 and so wipe out the profit, forcing the bank to increase the interest rate (assuming that it cannot use the excess foreign tax as a credit against other domestic tax in its residence country). If the OECD Commentary’s suggestion to deal with this problem is followed and loans from banks are exempted from tax,92 this opens the way for simple back-to-back transactions, which will mean that the exemption will be effectively extended to nonbank lenders. If a nonbank nonresident lender deposits money in a nonresident bank and the bank then makes a corresponding loan to a resident (less a small fee), what is effectively a loan from a nonbank becomes for treaty purposes a loan from a bank and is protected accordingly. Some of the problems of this kind can be dealt with better by provisions in domestic law that remove the withholding tax on interest for borrowings in the international capital markets where the debt is widely held (often referred to as Eurocurrency loans). The widely held requirement substantially removes the problem of back-to-back transactions. Many industrial countries have such provisions in their laws.93 Nonetheless, in a few cases, reduction of interest withholding to zero under treaties is common for developing and transition countries, especially for concessional loans made by development banks. A general lowering of the interest withholding rate to zero also worsens the thin capitalization problem described below.

Similar considerations apply to royalties, which are also particularly associated with the problem of treaty-shopping discussed below. Hence, there is a good argument for developing and transition countries to have a reasonable positive tax rate on interest and royalties under tax treaties (say, 10-15 percent). If royalties include equipment leasing rentals, there is also a strong argument for uniform tax rates under tax treaties on interest and royalties; indeed, the possibilities for conversion from interest to royalties or vice versa, especially in the case of related parties, extend beyond this area so that equivalence should be a goal in any event. Perhaps more important, because of the problem of treaty shopping, it is imperative to have the rates similar or the same across tax treaties with other countries in the case of interest and royalties. The industrial countries generally (but reluctantly) accept this position in their tax treaties with developing and transition countries; however, they often negotiate most-favored-nation clauses in protocols to the tax treaties in such cases, so that if the developing or transition country grants a more favorable rate or treatment to another industrial country (often defined in terms of membership of the OECD), then either the more favorable treatment is automatically extended to that country or an obligation to renegotiate that tax treaty arises.94

D. Capital Gains

Capital gains of nonresidents present a more difficult problem for withholding. While it is possible to have flat-rate withholding based on the sale price either generally or specifically in the case of nonresidents, the gain part of the sale price can vary considerably, and so an option for net taxation should be provided for in domestic law with appropriate administrative safeguards.95 Enforcement of such withholding is likely to be feasible only in the case of land (because land transactions are usually registered in some way and the collection of tax can be tied in with this procedure) or of a permanent establishment (with the gain taxed on a net basis like most other business profits). Many countries do without withholding in such cases, as it is possible with appropriate administrative mechanisms to deal with the capital gains.96 Attempts to levy capital gains in other cases will generally be overridden by tax treaties and any attempt to protect the power to levy tax on gains on shares in resident companies is likely to be futile for reasons already explained.

E. Employment, Services, and Pension Income

Employment income of nonresidents is usually subject to the normal wage withholding and not to any special final withholding, despite the policy arguments that flat-rate withholding is the appropriate method for nonresidents. There are special collection problems where the employer is a nonresident, but tax treaties will usually protect the employee from taxation by the country where the work is performed in this event through the 183-day rule unless a permanent establishment bears the wages (in which event enforcement will not usually be difficult). If the employee is present for 183 days or more, residence will usually arise and the more permanent connection with the country will facilitate withholding, although it is easy for temporarily present employees to slip through the net unless attention is given to this issue by the tax administration. Powers in the domestic law for the tax administration to prevent a person from leaving a country unless taxes are paid can provide some assistance to tax collection depending on how easy or difficult it is to exit the country.

Some transition countries find it difficult to cope with withholding on wages of expatriates because their wages are paid into bank accounts in foreign countries. This is partly a function of some wage taxation laws applying only to wages paid in a country (which should be rectified if necessary, making clear that the law applies to wages sourced in the country, whatever the place of payment) and partly a surrender to the difficulties that the international border creates. Most employers, however, will not use such a device to avoid tax as the penalties on employers for failing to withhold are typically and appropriately severe. Moreover, this is one area where information exchange under tax treaties with the country of the employer can be effective in assisting the tax administration.

Although wage withholding often is not formally final, the way in which obligations to file tax returns are expressed in many developing and transition countries means that many employees are taxed through withholding only, so that in effect the withholding is final.97 In the case of nonresident taxpayers, returns are not usually required or forthcoming so that the withholding is final in fact. For expatriate taxpayers, adoption of any of the special rules set out above may mean that special attention has to be given by the tax administration to withholding on wages and filing of returns in their case to prevent abuse of the rules.

Some countries extend withholding beyond the employment area (including deemed employments discussed above) to certain services rendered in a business context. As already noted, such income is required by tax treaties generally to be taxed on a net basis, but this obligation can be satisfied by permitting such taxpayers to file returns and to have the withholding credited against the tax liability (with refunds where necessary). The language of tax treaties (although not perhaps the OECD Commentary)98 suggests that final withholding on professional income is permitted where there is a fixed base (or a presence time limit is exceeded if included in the treaty).

For entertainment and sports-related income, flat-rate final withholding is clearly permitted under tax treaties and provides a simple and effective method of collecting tax via the promoter of the event. Provision for some form of withholding on this income at a reasonably substantial rate, such as 30 percent, should be provided in the domestic law and should apply whether the income accrues to the entertainer or athlete directly, which is very rare, or to some intermediary; that is, the law should permit the tax authorities to look through the intermediaries to the entertainer or athlete.

In the case of pensions, withholding in accordance with the rate scale for individuals is often provided for in domestic law in a similar way as for wage income. Tax treaties may override any tax depending on the source rule adopted (see above). Likewise, wage and pension income of the employees or former employees of foreign governments will usually be subject to withholding under domestic law in the same way as other wages and pensions, but tax treaties may remove the levy of this tax.

F. Company and Shareholder Taxation

The relationship of taxation of company and dividend income in the international setting raises a number of special issues. One major distinction is between direct and portfolio investment. Direct investment refers to the case where the investor in a company has a large enough interest to influence the operations of the company, while portfolio investment is the opposite case of no influence. This distinction often runs throughout the laws and commercial practice of a country (in such areas as takeovers, investment, banking, and accounting, as well as taxation) and may be defined differently for different purposes, although often the taxation definition is affected by treatment in other areas of the law. It is usually defined in terms of owning a certain percentage of the capital or controlling a certain percentage of the votes in a company, with 10 percent and 25 percent or more for direct investment being the most common in taxation laws. The OECD Model uses 25 percent of the capital, while a number of industrial countries use 10 percent of voting power in their tax treaties.99 The discussion that follows will commence with portfolio investment and then move on to direct investment.

1. Integration Systems

The simplest tax system for companies and shareholders is the separate system; that is, the company is taxed on its income, and then dividends paid by the company are taxed as part of the income of the shareholder without reference to any tax paid by the company. Whatever the method of tax collection under this system in a domestic case (where a resident company pays a dividend to a resident investor), frequently a flat-rate withholding tax is levied on dividends paid by resident companies to nonresidents. Tax treaties will often reduce the rate contained in domestic law, the OECD Model, and most tax treaties specifying 15 percent for portfolio dividends.

In recent years, many countries have moved away from the separate system because of its well-known potential for distorting economic decisions by companies and shareholders in the domestic context. Such “integration” systems may consist of some form of imputation, a split corporate tax rate, or a zero or low tax rate on dividends (in all cases with or without some form of equalization tax on dividends to ensure that corporate tax has been paid on distributions of company profits). Domestic tax laws usually confine the full integration benefits to resident shareholders and often continue to tax nonresident shareholders under a separate system with flat-rate withholding taxes.100

Most recently, with the growth of international investment, attention has become focused on the potential for international economic distortions from integration systems of these kinds. This issue has led some countries to extend some of the benefits of integration to nonresident shareholders unilaterally or by tax treaty, for example, by partly removing withholding taxes on nonresidents101 or by giving imputation credits partly to nonresidents.102 Some countries have sought to go further and completely equalize the treatment of residents and nonresidents. A simple approach is to align or approximate the corporate and maximum individual tax rates and to exempt dividends from further taxation whether paid to resident or nonresident shareholders.103 From the point of view of the source country (where the company paying the dividend is resident), neutrality may be achieved with such a system. For nonresident portfolio investors, however, neutrality is unlikely because their residence country will almost invariably tax them on the dividends without any benefit of whatever integration system that country has for its resident companies (if any) and with a foreign tax credit only for any withholding tax levied on the dividend by the source country (as distinct from the corporate tax levied on the company paying the dividend).104

Hence, there is still a bias in the international tax system for resident shareholders to invest in resident companies that other countries cannot prevent under this or any other form of integration. This bias is now providing policy support for the separate system of company and shareholder taxation, as such a system does treat residents and nonresidents more or less alike if the country of residence of the company taxes shareholders resident there on dividends received and if other countries tax shareholders resident there on the dividends, with a credit for any source country withholding tax.105 In fact, the position is more complex, as a large proportion of international portfolio investment is made by institutions that are taxed under special regimes in their residence country.

From the point of view of developing and transition countries, a fairly standard treatment of nonresident portfolio shareholders with a flat-rate withholding tax and a tax treaty rate limit of 15 percent is the simplest solution. Any attempt to extend integration benefits to nonresidents generally is likely to produce a transfer of tax revenue to capital-exporting industrial countries without providing any incentive to invest to the nonresident (or rather without removing the disincentive to invest abroad that arises from the residence country tax system).106 Even if it is decided to extend integration benefits to nonresident portfolio shareholders, it is better to do this unilaterally rather than in tax treaties (even if the domestic law confines the benefit to countries with which there is a tax treaty), because such treaty provisions can lock the country into the form of integration it has adopted. As integration (in the past at least) has been primarily a domestic tax policy issue, integration benefits in tax treaties can become the international tail that wags the domestic dog.

A removal of dividend withholding tax on foreign tax-exempt pension funds as part of a regime of reciprocal recognition of the special tax arrangements that many countries use to encourage private pension schemes may be considered. This is usually done outside tax treaties (though note the comments above in relation to tax treaty provisions dealing with contributions to pensions schemes by expatriates) and across all types of investment income, rather than just for dividends.107

A country employing an equalization tax as part of its integration arrangements108 must take care in drafting it to ensure that it does not conflict with tax treaties. Often, such a tax will be effectively at the corporate tax rate and will be triggered by the payment of dividends. It can therefore be viewed as a withholding-type tax on the dividends, in which event there is potential for the tax rate limits in tax treaties to reduce the amount of the tax and so defeat or at least blunt its purpose. There are well-accepted drafting devices to ensure that such a tax is not regarded as a withholding tax on dividends.109 First, no primary or secondary tax liability can be imposed on the shareholder in relation to the equalization tax, so that it is clearly a tax on the company rather than on the shareholder. Second, it helps to use the dividends simply as a measure for the amount of the tax and not to express the tax as being levied on the dividends as such. Technically, the tax also needs to be at the corporate rate on the amount of the dividend plus the tax, which is most easily done by expressing the tax rate as


where t is the corporate tax rate.

The drafting arrangements for the U.K. advance corporation tax provide a model that can be used to ensure that there is no conflict between the equalization tax and tax treaties (although the basic rate of tax and not the corporate rate is used in the United Kingdom).110

2. Reduction of Dividend Withholding Tax on Direct Investment

In the case of dividends generated by direct investment, the international tax position is very different from portfolio investment from a number of perspectives. A foreign direct investor (assumed in what follows to be a company) generally has a choice as to the legal structure of its investment in a country. It can establish a branch (permanent establishment) or a subsidiary (i.e., a separate company).111 The residence country of the direct investor will grant relief for double taxation by way of a credit or an exemption for corporate tax levied on a branch by the source country (where the branch is situated). It will generally extend this relief to corporate tax levied on a subsidiary when dividends are paid to the direct investor so as not to produce a tax bias in the form of investment.

In its turn, the source country will, by various means, approximate the tax treatment of branch and subsidiary for the same reason. The major likely difference in source country tax treatment in the absence of special provisions in the domestic law or treaties will be that dividends paid by a resident subsidiary to a nonresident parent company are subject to flat-rate dividend withholding tax, while remittances by a branch to its head office (the functional equivalent of dividends) are not subject to any tax. The source country can address this disparity by reducing the tax on direct investment dividends, or by taxing branch remittances, or by a combination of both.

Although it is possible for domestic law to provide a lower tax rate on direct as opposed to portfolio dividends paid to nonresident shareholders, until recently this reduction was most commonly only effected by tax treaties (with 5 percent being the OECD Model norm). Developing and transition countries need not be too concerned with accepting such arrangements for direct investment in treaties, especially where an equalization tax is in place, but it is noticeable that a number of such countries (along with some smaller industrial countries) do not draw the portfolio/direct investment distinction in the dividend article of their tax treaties and apply the same rate of tax to both. Unlike the case of portfolio investment, a lower rate of tax on dividends on direct investment does not usually operate as a transfer of revenues to industrial countries because of the different tax regime in most of them for dividends on direct investment (exemption or underlying foreign tax credit). A small but positive tax treaty rate in the source country also provides some incentive for reinvestment of profits (a major source of investment) by foreign investors without unduly distorting the tax position in the residence country of the investor.

There is now a more general international trend for reducing withholding taxes on dividends paid to nonresident direct investors outside tax treaties. One effect of the tax reform that took place in many countries in the late 1980s was to more closely align the tax base and tax rate applied to companies in industrial countries. This meant, for direct investments through subsidiaries, that the corporate tax in the country of the subsidiary would approximate the corporate tax that the same amount of profit would attract in the country of the investor. As that country would relieve double taxation for the corporate tax paid by the subsidiary, the net effect was to wipe out any corporate tax in the residence country of the investor whether a credit or an exemption system was used, but the dividend withholding tax would remain as an additional tax levy above the residence country tax.

A number of major econometric studies in the early 1990s suggested that such withholding taxes were the main factor accounting for a bias against cross-border investment, and hence some pressure has developed for their removal, even though tax treaties typically contain lower tax rates on dividends from direct investment.112 The fact that the United States typically demands for its resident investors a share of the action in integration systems adopted by foreign countries has also been an influence here. Developing and transition countries that do not have tax treaty networks may therefore wish to consider setting the cross-border dividend withholding tax rate on direct investment at a lower rate (say, 10 percent) than the traditional and typical 20-30 percent tax rate that has been adopted across the board by many countries for dividends, interest, and royalties. There is, however, little reason to adopt a selective zero tax rate on dividends in domestic law as part of regimes of tax incentives for foreign direct investors.113 As the benefit is only likely to operate long after the initial investment occurs, it has little impact on initial investment decisions and does not encourage reinvestment of profits.

A similar pressure to reduce cross-border dividend taxes may arise when countries form a free trade bloc, given that one of their longer-term objectives is usually to remove not just trade barriers but also investment and other barriers to the creation of a common market. This means that taxes applying only at the border (such as a nonresident dividend withholding tax) become targets of the institutions of the common market. Thus, the EU after many years’ debate has adopted a directive that will remove cross-border dividend withholding taxes in the case of direct investment.114

This trade bloc reasoning also applies to other income flows within corporate groups, and the EU has a draft directive extending the same treatment to interest and royalties in direct investment cases.115 However, the reasoning here is very different from the more general argument in relation to dividends and does not make sense outside a trade bloc. The reasoning is that interest and royalties will be taxed in full in the residence country, which is a member of the bloc, and, as long as investment flows are balanced among the countries in the bloc, the revenues of members do not suffer (alternatively, government-to-government reimbursement mechanisms can be devised if flows are not equal), while at the same time the border impediment is removed.

For developing and transition countries, investment flows are not usually in balance with other countries (even in the Commonwealth of Independent States (CIS), the loose trading bloc formed by most of the countries of the former Soviet Union), and interest and royalties are payments that reduce the tax base (as they are usually deductible in the calculation of taxable profit), with significant potential for causing problems for the taxation of direct investment. Hence, the advice given in relation to these payments above was to maintain reasonable levels of tax at relatively uniform rates in both domestic tax law and treaties. The existence of a trade bloc does not change that advice.116 More generally, developing and transition countries need to be very cautious in studying the tax arrangements in trading blocs of industrial countries, especially the EU, even where they have ambitions to become members of the bloc. Where a group of developing or transition countries forms a trading bloc, care should be used in extending special free trade arrangements to taxes, as the countries may not have the capacity to deal with the more sophisticated rules often involved. For example, the international value-added tax and excise rules within the CIS have been an ongoing problem.117

3. Branch Profits Tax

In the case of direct investment in the form of a branch, the branch profits tax represents a strategy to even up treatment of branches and subsidiaries. To produce precisely the same outcome, it is necessary to define branch remittances that equate to dividends and to tax them at the same rate that applies to dividends on direct investment.

While the statement of the principle is easy enough—the amount of remittance can be determined by comparing the branch’s tax balance sheets at the beginning and end of the tax year—in practice, the elaboration of the principle has generally proved very complex, even though to some extent it is based on the same information used to determine the taxable profit of the branch. Some countries therefore use a simpler but rougher measure, namely, the after-tax taxable profit of the branch. To take account of the fact that subsidiaries typically do not repatriate all of their after-tax profit as dividends, the rate is often set lower than the dividend withholding tax rate based on an assessment of the typical payout ratio of subsidiaries of foreign investors in the country (a tax rate of one-half the dividend rate or less being appropriate in most cases).

Certainly, some rules for calculating the amount subject to branch profits tax need to be set out in domestic law. It is neither sensible nor transparent to introduce the tax through the back door by defining all branches for tax purposes to be subsidiaries so that remittances (presumably) become dividends and, thus, subject to dividend withholding tax. The tax administrations of developing and transition countries will not be able to detect remittances as they occur (the possibilities of method of remittance being infinite and the only practicable measurement device being comparison of tax balance sheets at the beginning and end of the tax year). Although, in some cases of more exotic legal entities, there will be difficult cases of characterization as branch or subsidiary, this is not a reason for the tax law to impose an arbitrary rule that is contrary to generally accepted international norms of taxation in clear cases.

A number of developing and transition countries are considering or have enacted branch profits taxes, in some cases without apparent regard to their tax treaties. Treaties based on the OECD and UN Models override the levy of a branch profits tax,118 and the treaties in question do not generally contain the necessary modifications to the dividend and nondiscrimination articles to accommodate such a tax. Although new treaties that are negotiated can contain these modifications, the existing treaties will encourage treaty-shopping to short-circuit the effect of the new treaties, and it will be many years before replacement treaties can be put in place.

Further, it is not possible to either tax all effective remittances or achieve in practice the close approximation of the tax treatment of branches and subsidiaries that the branch profits tax is aimed at, because of the interaction of the tax treatment of dividends and capital gains in the context of the branch or subsidiary. Both the dividend withholding tax and a branch profits tax based on remittances can be avoided by not paying dividends or remitting profits, as the case may be—that is, by reinvesting the profits. The gain in each case can then be realized by selling the shares in the company operating the branch or in the subsidiary (or in a holding company in the corporate group). This gain will usually not be taxable in the source country because of either tax treaties or the inability of domestic law to reach sales of holding companies based in other countries (not to mention the lack of the capital gains tax in many developing and transition countries).

Sale of shares in this way thus achieves an effective remittance of reinvested profits of the branch or subsidiary, but in practice it will be more difficult for a branch to achieve such a sale because the branch will usually be just one part of the operations of the company, with the result that sale of the shares will amount to much more than a realization of the reinvested profits of the branch. Further, as far as capital gains (in excess of those arising from reinvestment of profits) have been made on the investment, the tax treatment of branch and subsidiary will usually differ in practice for the same reason that disposal of the shares in the company operating the branch will often not be a practical possibility. Disposal of the branch will usually be effected by the sale of its assets, which will be subject to the capital gains tax of the country where the branch is situated (if any), while the profits on the sale of the shares in the subsidiary will not be taxed.

Hence, the value of a branch profits tax is doubtful. The tax pales into insignificance when compared with some of the other problems of protecting the tax base of the source country against the base-erosion techniques that are explored below. The main reason why it is sometimes thought to be important for developing and transition countries to have a branch profits tax is to fully tax income from natural resources where many foreign investors typically operate in branch form mainly because of the generous treatment of the early year start-up losses under their home country (especially U.S.) tax law.

4. Branches and Subsidiaries in Transition Countries

The transition countries face a special set of issues in the branch and subsidiary area, which demonstrates once again the problems caused by the lack of clear rules and by departures from international norms in these countries. Under the commercial laws of nontransition countries, there is a generally clear understanding of what is meant by a body corporate (company, corporation) and of when an entity recognized by the law has separate legal personality and when it does not.119 However, the commercial laws of several transition countries are still in the developmental stage, and it is often not clear when a separate legal person exists or, more important, whether in a particular situation there are two legal persons (parent company and subsidiary) or one legal person with a number of operations (head office and branch).120

When a foreign legal person commences operations in a transition country, it is usually required to “register” to do business under the commercial laws of the country. In some of the countries, registration is regarded as the creation of a legal person, because this is how the creation of a legal person is effected in a purely domestic case or, perhaps more accurately, registering to carry on a business in a purely domestic case of itself creates a separate legal person (as the registration is to get approval to do business, and the creation of a separate legal person is a by-product of registration). Representation offices of foreign persons are usually recognized and are not treated as separate legal persons (a separate registration procedure is required in this case), but the functions that such offices can perform under the laws of the transition countries are generally strictly limited, as befits their name.

Before 1989, the question of registering foreign legal persons under domestic procedures did not arise for many transition countries because the only way a foreign legal person could operate a substantial business venture in the country was through the creation of a joint venture with foreign participation, for which special statutes existed. The joint venture in these cases was a separate legal person under the statutes, and the foreign joint venturer a substantial shareholder along with the state-owned enterprise also involved in the venture.

Moreover, in several transition countries (especially members of the CIS), the profits tax is not levied on a legal person as such, but on the separate operational units of the legal person (which may in turn be linked to separate registration of the operational units with the local or regional authority of the area where they are located).121 Thus, if a state-owned enterprise has a glass factory in one city and a television factory in another city, both the factories will often be taxed separately. This may affect rates of tax because, in many of the countries, there are varying tax rates depending on the nature of the business of an operational unit or the region where it is operating, and, more important, it may affect the treatment of losses, as a loss incurred by one operational unit may not be offset against the profit of another operational unit. This fact makes it less necessary under the systems of some transition countries to distinguish in a particular case whether one legal person is involved or two. Again, this system grew up in the closed days of central planning so that international issues did not intrude. Hence, putting aside the case of the representation office, questions did not arise as to whether a branch of a foreign legal person was taxed in this way (assuming that a branch was possible under the system in question) and as to whether operational units (including those of foreign legal persons) were taxed on their worldwide profits.

These rules have a number of important implications for international taxation and tax treaties in cases of direct investment by industrial country resident companies in transition countries. In many of them, what the industrial country resident regards as a branch (permanent establishment) will often be treated as a subsidiary by the transition country because it is registered in that country. Indeed, in one unusual case, this result was regarded as arising from registration for turnover tax purposes. The “subsidiary” will be taxed as a resident legal person by the transition country, and distributions to the industrial country resident will be treated as dividends and subject to any tax treaty accordingly (although some of the transition countries have no taxes on dividends).

If the legal system of the transition country in question characterizes an operation within its borders as a separate legal person, then the private international law rules applied in most industrial countries will lead to the recognition of this characterization by the general law and usually the tax law of the industrial country in question. However, in many cases, the industrial country resident will not be aware of either the legal intricacies involved or the very different legal structures in some transition countries.

In a number of transition countries, the concept of a branch has become fully accepted for both commercial and tax law purposes, although, even then, exchange controls may make operation in branch form impractical. In most countries, the extension beyond the case of the representation office is piecemeal (e.g., banks and building sites) and seems to require special procedures separate from the business registration procedure. In some countries, the representation office is being put under a lot of pressure as nonresident taxpayers try to establish branches for various reasons. Part of the pressure results from the fact that the transition countries generally find it difficult to deal with cases where the taxpayer breaches the law—in this case, when representation offices engage in activities not legally permitted to them.

If the transition country in question taxes each operational unit separately, then further tax issues arise for the industrial country resident direct investor, whether branches are permitted generally or in special sectors or not at all. In the branch case, the industrial country investor may find that losses on one branch operation will not be offset against profits of another branch operation in the same country, which will be contrary to the expected treatment. This has been a problem in some transition countries, particularly in the oil and mining sector where each drilling rig or mine site is taxed separately.

There does not seem to be anything in article 7 of the OECD Model that precludes this outcome (indeed, the Model seems to follow the approach of treating each permanent establishment separately), and, as the same treatment is applied to domestic enterprises, nondiscrimination is unlikely to be an issue. It is certainly the assumption of industrial countries, however, that legal persons are taxed as a whole and not separately on operational units, although in the source country only profits attributable to the permanent establishment are taxable, and not the worldwide profits of the legal person.122 A potentially more difficult question arises for the calculation of expenses. Treating each branch separately in the calculation of tax may naturally lead to the disallowance of head office expenses as deductions of the permanent establishment. The separate treatment of the operational units for tax purposes in transition countries does not seem, however, to produce the consequence that payments between them or to the head office receive dividend treatment.

This range of issues has been the cause of considerable confusion among industrial country investors, however (the precise legal situation varies from country to country), and has had an additional chilling effect on foreign direct investment in a number of transition countries and on the development of tax treaty networks with industrial countries. One alternative has been for foreign investors to enter into special tax contracts with the governments of transition countries that guarantee them a relatively normal tax treatment by market economy standards. While these contracts solve the problems of the particular direct investor, they are already complicating tax reform and tax treaty development in a number of transition countries.

In general, it is recommended that transition and developing countries refrain from entering into special tax contracts or at least limit the effect of the contracts to a relatively short time before reviewing them. Further, transition countries should seek to ensure that their commercial and tax laws accord with the general international distinctions between branches and subsidiaries and that the tax position of an investor with more than one branch in the country is aggregated across the branches. Several transition countries have already taken these steps in recent years.

G. International Tax Avoidance and Evasion

While the source country may be concerned with ensuring that direct investors are taxed in a way that does not bias the form of the investment and with collecting its fair share of tax from both direct and portfolio investors, nonresident taxpayers may seek to escape source taxation altogether or at least to minimize that tax. They may do so through techniques to avoid or minimize tax—that is, arranging their affairs so that under the law of the source country the tax is minimized—or through tax evasion—that is, deliberately not complying with the law of the source country even though income is taxable under that law.123 As with the issues of company and shareholder taxation discussed above, it is helpful to draw a distinction between direct and portfolio investors; indeed, much of the discussion under this heading stems from a number of the points already made. The discussion below initially focuses on nonresident direct investors and then canvasses to what extent the techniques outlined are available to nonresident portfolio investors and to resident investors.

Within an international group of companies investing directly in various countries, what generally matters to the managers and the ultimate shareholders is the after-tax profit of the group; in other words, the corporate group usually has an economic incentive to reduce its total tax payments and is economically indifferent as to the countries to which it pays tax. In some cases, especially where the residence country of the parent company in the group operates an imputation system that ties tax credits available to shareholders to the company tax paid in that country by the parent and local subsidiaries, the economic incentive may rather be to pay as much tax as possible in the residence country. In any event, multinational companies investing in developing or transition countries are likely to have an economic incentive to reduce the tax burden in those countries, either as part of reducing tax burdens worldwide (i.e., reducing tax in both residence and source countries) or as part of moving the tax burden to a country that offers the greatest advantages to the ultimate shareholders of the company group.

This economic incentive may not always lead to tax avoidance or evasion. Cultural, ethical, and nontax commercial factors may act as a counterbalance. With the globalization of trade and investment, deregulation in many areas of international business law, and international financial markets that focus on the “bottom line” and are beyond the reach of any single government, the countervailing factors are likely to weaken in influence over time. Most large multinational companies will nevertheless want to conduct their tax planning within the law; that is, they are more likely to practice tax avoidance or tax minimization than tax evasion. Tax evasion internationally and domestically is more of a problem with small or closely held businesses and individual taxpayers (see the discussion of capital flight above for the problem of evasion in relation to resident taxpayers).

The simplest way to minimize tax is to make payments from the branch or resident subsidiary to a related nonresident company that are deductible in determining the amount of profit subject to corporate tax and that are not subject to withholding tax. Alternatively, as a second-best option, payments can be made that are deductible under the corporate tax and are subject to a low rate of withholding tax.

In the past, two basic strategies (which can be combined) have been mainly used to achieve these ends: increasing the prices of payments and changing the type of payments. To take some simple examples, a local subsidiary operating an assembly plant can pay inflated prices for the components and the technical and management services it purchases from related companies; or a nonresident parent company can invest in the subsidiary by way of loan capital rather than share capital and receive interest payments (deductible to the subsidiary) instead of dividends (usually not deductible to the subsidiary). Similar results can be produced by reducing the amount of payments for goods or services to the local branch or subsidiary for goods or services it provides to other (nonresident) members of the group. Recently, international tax planning has become more sophisticated along with the financial markets. The following discussion will start with the simpler methods of tax avoidance and then move to more recent techniques.

1. Transfer Pricing

“Transfer pricing” is the general term used to refer to the problem of allocating profits among the parts of a corporate group. For the group as a whole, all that matters at the end of the day is the after-tax profit of the group rather than of its individual members. The prices charged within the group for goods or services provided and the financing methods used between the members of the group simply serve as means of moving funds around the group and do not in a commercial sense create profits for the group. Hence, there is often no obstacle to charging any price or structuring a transaction in any way within the group, and the fair or proper distribution of the overall group profits among the companies in the group is often a secondary consideration to tax consequences. In financial accounting, which seeks to determine profits for reporting to shareholders and others with financial interests in the group, the response is to require accounts for the enterprise (group) as a whole and to eliminate transactions within the group, as well as (in most countries) accounts for each company in the group.

In taxation, it is necessary to allocate profits among the companies in the group because under international tax norms a country will tax a nonresident only on the profits sourced in that country. While the country can tax a local (resident) subsidiary on its profits worldwide, affairs within a multinational group will usually be arranged so that the subsidiary only has profits sourced in that country. In theory, this allocation of profits can be effected in one of two main ways. A country can take the worldwide profits of the group and allocate some portion of those profits to a source in that country, thus bypassing the need to consider the pricing and nature of transactions within the group. Alternatively, the country can seek to determine the profits of a local branch or subsidiary separately from the rest of the group on the basis of the pricing and nature of the transactions engaged in by the branch or subsidiary with the rest of the group. In the former case, it is necessary to have allocation rules based on formulary criteria like relative assets, revenues, or salaries (and so this method is often referred to as formulary apportionment), while in the latter case rules are needed to deal with the problems arising from the special nature of transactions within the group.

While arguments range back and forth as to which method is preferable, in practical terms countries pursuing a policy of negotiating tax treaties are automatically tied into the separate accounting method because articles 7 and 9 of the OECD and UN Model treaties operate on the basis of taxing each company within the group separately and dealing with problems of pricing and the nature of transactions on the basis of the arm’s-length principle. Under this principle, adjustments are made to transactions within the group to reflect the terms and nature of transactions that would have been entered into if the transaction had been made with an independent third party rather than with another part of the group.

A drafting issue for the domestic law is that the arm’s-length principle should be provided for both branches and subsidiaries. This is most easily done by using language similar to that found in tax treaties. Such an approach ensures that there is a basis in domestic law for making transfer pricing adjustments. In many countries, it is not clear whether tax treaties on their own would provide a sufficient basis for such adjustments, and, in any event, it is necessary to have the rules in the case of residents of countries with which there is no tax treaty in force. Using statutory language based on treaties has the added advantage of giving a clear signal that the country intends to follow international norms.

Article 7(4) of the OECD and UN Models provides that a country can maintain a customary method of calculating the profits of branches, so long as the result is in accord with the arm’s-length principle (a further provision in each case provides for the application of the same method from year to year unless there is good reason to the contrary). Some countries use simplified profit calculation methods for branch cases (such as a specific percentage of turnover of the branch). These methods can be retained in the legislation insofar as they reasonably reflect actual profits and can be used in cases where tax treaties are involved. The application of the arm’s-length principle to branches is more complex in one way than in the case of subsidiaries, because the branch and the head office are part of the same legal person, and transactions cannot be sensibly reconstructed in some cases. For example, it is often difficult to allocate notional ownership of property between head office and branch.

Simplified methods in domestic legislation are not generally regarded as consistent with article 9 of the OECD and UN Models in the case of related companies, but this does not mean that countries are confined to making tax adjustments between related companies only in international transactions and on arm’s-length principles. Some countries apply their transfer pricing rules in purely domestic cases; where there are different tax rates for different kinds of income or business, taxpayers can use transfer pricing to move profits to categories of income or business with lower tax rates. There are also a number of reasons why countries may wish to have special pricing rules for specific transactions. For example, some countries treat all disposals of property without consideration as having been made for market value—whether between related parties or not—while others treat gifts of property to charities as having been made for the higher of cost or market value. These rules do not directly deal with transfer pricing issues. To the extent that they can apply to international transactions between related parties, they will not generally be contrary to tax treaty arm’s-length pricing rules. How all these rules are coordinated within the tax legislation depends on the specific rules adopted and should be reviewed carefully in each country.

To achieve the application of the international arm’s-length principle in practice, the tax administration starts with the accounts of the local branch or subsidiary, makes the usual adjustments to reflect differences between financial accounting and tax rules, and then makes such further adjustments in accordance with the arm’s-length principle as necessary. Nontax considerations may lead to the group preparing its branch or subsidiary accounts on this basis in any event. For management purposes, the group will wish to know the real profitability of its separate parts, local employees may be remunerated in part on the basis of the local contribution to group profit, and local accounting rules will likely require that the financial accounts give a proper view of the profits of the branch or subsidiary. In practice, the tax administration may use simplified methods and various financial ratios that are similar to formulary apportionment in order to test whether the profits reported by a local branch or subsidiary fall within acceptable boundaries. These methods frequently operate as a means of selecting taxpayers for further checking (audit). The use of such administrative methods will not be contrary to tax treaty rules so long as they are being used as a means to the end of establishing the arm’s length price.

The increasing integration of the activities of corporate groups, the growing importance of unique intragroup intangibles and services, and the sophistication of their financing operations mean, however, that application of the arm’s-length standard is becoming more difficult, both conceptually and practically. The problems have been addressed in part by the OECD, which has updated and expanded its guidance on this issue.124 The OECD standards represent the internationally accepted norms giving content to the arm’s length principle.

Transfer pricing adjustments on the arm’s-length principle have traditionally been viewed as involving price only (as the name suggests) and not the reconstruction of transactions in the sense of disregarding the nominal transaction between the related parties and substituting another arrangement for tax purposes. The transfer pricing guidelines,125 while recognizing that adjusting prices of actual transactions is the norm, do permit tax administrations to recharacterize transactions in two exceptional circumstances: first, “where the economic substance of a transaction differs from the form,” and second, where the “arrangements made in relation to the transaction, viewed in their totality, differ from those which would have been adopted by independent enterprises behaving in a commercially rational manner and the actual structure practically impedes the tax administration from determining an appropriate transfer price.”126 The example of thin capitalization is given for the first category (see next heading) and, for the second, the outright transfer of intangible property before its value is fully known when independent parties could have been expected to enter instead into a continuing research agreement (under which payments would not be irrevocably fixed in advance).

Increasingly, countries are enacting general provisions in their tax laws directed against tax avoidance, which give powers to reconstruct transactions.127 It seems to be increasingly accepted by the OECD that such rules are not in conflict with tax treaty obligations and can be applied to international transactions.128 While such rules in conjunction with transfer pricing rules expressed in the general terms suggested above can deal with many problem situations, they can leave taxpayers uncertain as to their position. Accordingly, countries are increasingly enacting more specific provisions to deal with particular cases and to spell out the rules in more detail, as shown, for example, under the next heading.

2. Thin Capitalization

Thin capitalization is the practice of excessively funding a branch or subsidiary with interest-bearing loans from related parties rather than with share capital.129 The fact that interest is usually deductible for the borrower and taxed to the nonresident lender at a low rate of withholding tax (or not at all in some cases) while in most cases company profits funding dividends are fully taxed makes the practice attractive taxwise to a nonresident investor. Although it is possible to deal with these problems under the arm’s-length principle, taxpayers and tax administrators often want more guidance on the level of permissible loan funding for a subsidiary than to be told that related-party loans can be made up to the point and on the terms that an independent third-party lender would allow, having regard to the other liabilities of the subsidiary. Thin capitalization rules seek to deal with this problem by denying deductions for interest in defined cases (and possibly recharacterizing the payments of interest as dividends).

Tax law provisions in this area can be drafted in a large variety of ways. One important constraint is the nondiscrimination article in tax treaties. In its typical OECD and UN form, this article overrides thin capitalization rules that apply only to payments of interest to related nonresidents by resident enterprises or by branches of nonresidents unless the rules are applied in accordance with arm’s-length principles. Another constraint arises from tax administration concerns. If loans by certain lenders only, such as related nonresidents, are affected by the rules, it is possible to get around this limitation through back-to-back loans.130 Rules can be drafted to deem such loans to have been made by the parent company—and so subject to the thin capitalization limits—but it is very difficult for the tax administration in the country of the subsidiary to detect such transactions, especially if the bank is located in a country with strict bank secrecy laws. One possible solution to problems of this kind is to make the rules generally applicable to all loans for which interest deductions are claimed. Hence, although the specific problem arises in the context of foreign direct investment, the solution for practical reasons may be across the board for all investment.

Further issues relate to the way in which the denial of interest deductions is calculated. One common approach is to provide express ratios of loan capital to share capital beyond which interest deductions are denied (debt-to-equity rules).131 Another is to limit interest deductions by reference to a proportion of the income of the taxpayer (earnings-stripping rules).132 What the appropriate financial ratios are in each case is also an issue (anywhere between 1.5:1 and 3:1 being common for debt-equity rules), as is the application of the rules to financial institutions whose business consists in borrowing and lending and that typically operate at much higher debt levels than other businesses. The following draft suggests a possible approach to these issues.

  • (1) A taxpayer, other than a bank or a financial institution, is denied a deduction for interest in excess of the product of three times the net income-producing assets of the taxpayer and

    • (a) in the case of a loan denominated in the currency of X, 110 percent of the interest rate charged on loans by the Central Bank of X to commercial banks on the last day of the preceding tax year; or

    • (b) in the case of a loan denominated in a foreign currency, 110 percent of the interest rate charged by the U.S. Federal Reserve on U.S. dollar loans to U.S. banks on the last day of the preceding tax year.

  • (2) The net income-producing assets of a taxpayer are assets giving rise to income that is included in the gross income of the taxpayer less liabilities relating to those assets, each averaged between the beginning and end of the tax year.

  • (3) With the prior written permission of the tax administration, a taxpayer may

    • (a) calculate net income-producing assets on an alternative basis; or

    • (b) in the case of a loan denominated in a foreign currency other than U.S. dollars, use a different interest rate based on the interbank rate of the central bank responsible for that currency.

  • (4) Any excess interest that is not allowed as a deduction in a tax year solely as a result of the application of this provision is treated as interest expense of the taxpayer in the following tax year.

Paragraph (1) limits the interest deduction to an amount obtained by multiplying a specified interest rate and three times the net income-producing assets (equity) of the taxpayer. This draft thus provides effectively a 3:1 debt-equity ratio. No limitation in terms of related parties or nonresidents is contained in the provision for reasons already given. Banks and other financial institutions are excluded from the provision altogether, but it would be possible to specify an alternative ratio for this case based on the prudential rules of the central bank for commercial banks. Rather than specifying the amount of loans on which interest is deductible, which is the approach many countries take, the provision directly calculates the amount of interest. This method has two effects. It eliminates the need to calculate the amount of loans, which can be complicated in certain cases, and deals with the problem of excessive interest rates being charged on loans between related parties rather than leaving this as a separate issue for the transfer pricing rules. If an explicit rule is to be provided for thin capitalization, it may as well spell out all the elements.

The interest rate is specified in paragraph (1) using the base rate charged by the central bank at the end of the previous tax year. The provision deals with loans in foreign currency by setting an interest rate based on the international reserve currency, the U.S. dollar, but also permits in paragraph (4) the use of other (major) currencies with the permission of the tax administration. Whether a reference to foreign currency is necessary depends on the rules adopted for dealing with foreign currency in the tax legislation, which are discussed in chapter 16. If, for example, foreign currency conversions are dealt with by recalculating foreign currency assets and obligations at the end of each tax year, giving rise to income or expense accordingly, then no rule for foreign currency is necessary in this provision. The interest rate is set by reference to the end of the previous tax year so that taxpayers know the operative rate at the beginning of the relevant tax year. The central bank rate is marked up by one-tenth on the assumption that most borrowers will not be able to obtain funding at central bank rates.

The net income-producing assets of the taxpayer are defined to include only assets that give rise to income that enters the calculation of taxable income. It is assumed in this draft that interest will be deductible only to the extent that it relates to the production of income included in the calculation of taxable income, which is not the position in all countries. Generally, the calculation is effectively the total assets less liabilities averaged between the beginning and end of the year. For resident companies, the capital and retained profits in the tax balance sheet are effectively equivalent to assets less liabilities (assuming there are no major categories of exempt income for such companies, such as foreign business income). Individuals will often not have a balance sheet as such, and so the calculations of assets less liabilities needs to be made specifically for each case. For nonresidents, only assets giving rise to income sourced in the country and taxed on a net basis after deductions enter the calculation along with their accompanying liabilities. Given that many countries employ final gross withholding taxes on the income of nonresidents, except for income from real property and business (see above), the loans to which the rules apply for nonresidents are likely to be limited.

Where the income-producing assets are shares, the appropriate treatment can be more complicated. Although dividend income is taxed in some countries at a final rate of tax on a gross basis for resident and nonresident shareholders, shares should be included for the purpose of this draft if interest expenses relating to such income will be allowed as a deduction. It is possible, for example, for intercorporate dividends to be exempted so as to eliminate the cascade of company taxation through chains of companies, but for interest expenses relating to the dividend income to be deductible. The rules here need to be coordinated with the interrelationship of the taxation of dividends and the allowance of interest deductions relating to dividend income, but beyond this general caution it is not possible to be specific.

The tax administration may give permission under this draft to vary the calculation of net income-producing assets where the calculations required above are difficult to apply in the particular circumstances of the taxpayer. Interest disallowed as a deduction under this draft is not permanently disallowed, but is carried forward and treated as an interest deduction of the succeeding year. The same calculation is then done for that succeeding year under this draft, and it may turn out that the interest deduction is allowed in that year. The disallowed interest is not treated as a dividend or some other form of payment. Country practices vary widely on this aspect of thin capitalization. Where countries also have general rules relating to characterization of investments as share capital or loans, it will be necessary to consider how those rules should be coordinated with the thin capitalization area. Tax treaties also have standard definitions of interest and dividends that do not provide clear guidance in the thin capitalization area. The OECD Commentaries seem to indicate that it is permitted but not obligatory to recharacterize as dividends interest that is disallowed under thin capitalization rules.133

The carryover of interest deductions in the draft will need to be coordinated with the general carryover of losses in the tax legislation, although co-ordination is likely to be automatic. The general rule for loss carryover is likely to apply only to deductions that are allowed in a particular tax year and that exceed income; because interest in excess of the permitted amount under this draft for a particular tax year is not allowed as a deduction in that year, it cannot enter such a carryover loss.

While a number of detailed issues would require elaboration in the practical application of the draft (e.g., the identification and valuation of assets and liabilities that are used in the calculation of net income-producing assets), what the draft conspicuously fails to do is to define “interest.” The reason is that, with the advent of modern financial instruments, interest is an increasingly difficult concept. We turn now to these instruments.

3. Modern Financial Instruments134

The previous heading left open the definition of interest, that is, the characterization of payments as interest or something else. Other issues that have to be considered in the international domain are the source rules, nonresident withholding taxes, and deductions available to the payer for payments under modern financial instruments. The specific focus of the discussion for the moment is direct investment involving related parties.

Because of the complexity of these issues, OECD countries are still searching for solutions. Hence, it is impossible to provide widely accepted methods that may be of use to developing and transition countries, but it is possible to suggest some partial solutions and note the problems that remain. The solutions that are adopted will need to be closely related to the more general question of how modern financial instruments are dealt with in purely domestic cases to ensure that the domestic and international regimes are consistent.

If a narrow definition of interest is adopted in domestic tax law (which is the typical case where new financial instruments have not been specifically addressed in the tax system), then it will be a simple matter for the taxpayer to use some financial instrument that does not generate interest but that is a functional equivalent, so as to avoid rules that refer to interest. For example, an interest swap arrangement can be structured to be the equivalent of a loan, but swap payments are not regarded as interest in many countries. (An interest rate swap is a financial transaction in which two parties agree to make streams of payments to each other calculated by reference to an underlying or notional principal amount; in its simplest form, it involves an agreement between two parties to make each other’s interest payments on their respective loans.) In this case, any attempt to deal with thin capitalization will be aborted unless some general antiavoidance rule is applied to recharacterize the payment as interest in the particular circumstances.

If a broad definition of interest is adopted to deal with this and similar problems (such as any payment or accrual under a financial arrangement widely defined), then it is necessary to adapt the source, withholding, deduction, and related-party rules to the broader scope. One particular problem arises from the fact that tax consequences—inclusions in income or deductions—under regimes dealing with modern financial instruments often occur on an accrual basis (i.e., on an internal-rate-of-return calculation or a mark-to-market rule) without any actual payment. Lack of a payment poses problems for withholding taxes, for example.

Take the relatively straightforward case of a zero-coupon bond where the issuer receives $100,000 on issue and undertakes to pay $161,051 on redemption of the bond in five years. This transaction is the equivalent of a five-year loan of $100,000 at 10 percent annual compound interest with interest payment only on redemption. In a number of countries, this transaction would be treated as giving rise to interest income and deductions in a purely domestic case for the five tax years of $10,000, $11,000, $12,100, $13,310, and $14,641, respectively. If the holder of the bond is a nonresident, it is difficult to collect tax annually because there is no payment to subject to withholding tax, although it is possible to require the issuer to pay tax annually as if a payment had been made. If nonresident withholding tax is postponed until the end of the five-year period, the nonresident may sell the bond to a resident before redemption and avoid the tax (assuming that there is no final interest withholding tax on payments to residents). If tax is collected from the issuer annually and the nonresident sells the bond to another nonresident, the buyer and seller will have to be aware that the issuer has been paying tax and take account of the tax in the pricing of the bond; if the buyer and seller are resident in different countries, the issuer may have to adjust the amount of tax withheld after the sale because of different tax-rate limits in the tax treaties involved.

Mechanical solutions can be devised to deal with the problems of withholding and timing of deductions—for example, levying the nonresident withholding tax only on payment and postponing deduction for the issuer until that time. These solutions usually bring with them practical enforcement problems and borderline issues where different regimes are being applied in different cases. What, for example, is the effect on accrued deductions and inclusions in income in this case where the nonresident transfers the bond to a resident and an accrual system is in place between residents? Not surprisingly, even in the relatively simple case of a zero-coupon bond, there is little agreement as to the appropriate international tax regime and a large amount of diversity in practice. Some countries do not even assimilate such a payment to interest, let alone deal with timing and withholding issues.

For more sophisticated instruments, such as interest rate swaps and currency hedges, withholding tax can make legitimate transactions uncommercial for reasons similar to those discussed above in relation to interest withholding taxes on ordinary loans. Many countries have therefore not extended their interest withholding tax to these cases. This limitation creates a relatively simple way to avoid withholding tax, especially between related parties. If, in response to this problem, withholding taxes on new financial instruments are directed to transactions between related parties, problems arise with back-to-back transactions.135

As to source of income, it is possible to create special rules for payments under new financial instruments, even if the payments are not characterized as interest. Without special source rules or recharacterization as interest so that the interest rule applies, the source rules for whatever category the payments are placed in will govern (e.g., business income or capital gains), which may allow nonresident parties to avoid source tax on what is equivalent to interest by manipulating the source rules.

Modern financial instruments may also allow parties, especially by combining different instruments, to make an equity position look like debt and vice versa. In the international context, this could lead to substantial erosion of the corporate tax base in relation to subsidiaries in developing and transition countries for what is essentially equity investment. The thin capitalization discussion above dealt with this problem in the case of related parties, where what is pure debt in a formal sense can be viewed in effect as share capital because no independent third party would have made a loan in the situation. Modern financial instruments open this position up more generally even for portfolio investors (see below) and allow related parties in many cases to escape thin capitalization rules.

Tax treaties further complicate the international situation because they were framed before the era of financial innovation and use traditional categories. The OECD Commentary on the interest article was recently changed to clarify the issue as follows:136

The definition of interest in the first sentence of para. 3 does not normally apply to payments made under certain kinds of nontraditional financial instruments where there is no underlying debt (for example, interest rate swaps). However, the definition will apply to the extent that a loan is considered to exist under a “substance over form” rule, an “abuse of rights” principle, or any similar doctrine.

The import of this paragraph seems to be that hedges and swaps will not be regarded as giving rise to interest, except when a transaction has been deliberately manipulated to substitute a future or a swap for what would otherwise have been a normal borrowing operation and when domestic law has recharacterized the transaction to give rise to interest under an antiavoidance measure. Although the Commentary does not say so, payments of discounts under zero-coupon bonds seem to be accepted as interest for the purposes of tax treaties. The result is that many payments under modern financial instruments to nonresidents will be characterized as business profits, capital gains, or other income, and, for treaties in OECD Model form, the result is that the source country will not be able to levy tax unless the payments are connected with a permanent establishment of the nonresident in the country. To partially address concerns in the related-party area, the OECD Commentary on article 21 has also been revised to include a suggested treaty provision that will allow recharacterization of payments in this kind of case.137

Against this complex backgound, what action can a developing or transition country take to protect itself against the sophistication of modern financial markets and multinational enterprises? A number of factors suggest a focus on the deduction area in the form of a general thin capitalization rule, combined with a comprehensive definition of interest for the purposes of the rule to catch all payments under modern financial instruments to the extent they would otherwise be deductible. First, this approach is not contrary to tax treaties, whereas the scope for action in the withholding area is clearly limited by treaties; second, back-to-back problems with related parties can be avoided; and third, the problem of characterization between debt and equity in relation to direct investors is addressed. How the definition of interest is framed for this purpose will depend on whether the country has comprehensive rules dealing with modern financial instruments for general domestic purposes, in which case definitions from that regime can be adopted. Experience to date suggests that a definition framed in general terms is preferable to a list of the kinds of instruments that are covered, given that the number of available instruments increases daily.

As regards withholding tax and source rules, the tax treaty position means that all that can be done directly is to maintain the traditional withholding tax on interest and consider extending it to zero-coupon bonds and similar instruments, although a number of technical problems will arise in doing so, as discussed above. The extract quoted from the OECD Commentary above strengthens the case for including a general antiavoidance provision in domestic law so that it can be applied (especially in the case of related parties) to recharacterize payments under modern financial instruments as interest and can subject them to the interest withholding tax accordingly. Back-to-back transactions may make the involvement of related parties difficult to detect, but the possibility of applying the antiavoidance provision and any resulting penalties may provide an incentive for related parties to fund local branches and subsidiaries by ordinary loans up to the limit permitted by the thin capitalization rules and to pay withholding tax on the resulting interest, thus limiting the problems.

4. Payments to Tax Havens

Where an industrial country resident makes a direct investment in a subsidiary in a developing or transition country, the residence country of the parent will be operating either an exemption system or a foreign tax credit system to relieve double taxation. At first sight it seems, in an exemption system, that there is no residence country concern if income is shifted out of the source country to the residence country, because the income will be exempt. This, however, is not the outcome. If the profit shifting involves transfer pricing, whereby the parent charges inflated prices for the goods or services that it provides to the subsidiary, the increase in the parent’s profits will be taxable in the residence country. This result follows because the profits will not usually be regarded as income sourced in the source country that attracts the exemption but rather as an increase of income sourced in the residence country (e.g., increased manufacturing profit) and so taxed there.

Similarly, if thin capitalization of the subsidiary by the parent is used to shift profits out of the source country, the interest received will probably not be exempt to the parent because the exemption usually extends only to business profits of a branch and dividends on direct investments in subsidiaries, and not to interest taxed by low-rate gross withholding in the source country (especially where the tax rate has been limited by a tax treaty). While there may be reasons why a parent company would find it advantageous to shift profits from the source country to the residence country (such as an imputation system in that country that bases tax credits to shareholders on residence country tax paid), often this form of profit shifting will effect little tax saving. In a residence country operating a foreign tax credit system, shifting profits out of the source country to the residence country as a means of lowering tax in the source country will usually lead to a corresponding increase in residence country taxation.

Accordingly, tax planning by multinational company groups is likely to be directed simultaneously to reducing source country and residence country taxation, which means in many cases that a third country needs to be found to which the profits can be shifted. Tax havens will be used for this end. In the transfer-pricing case, one possibility would be for the parent company to sell the goods to a related company in a tax haven for cost plus an artificially small profit (thus shifting part of the profit out of the residence country). The tax haven company then on-sells the goods to the subsidiary in the source country at an inflated price that leaves little profit to that subsidiary and most of the profit with the tax haven company. Similarly, in the thin capitalization case, the parent company may invest in a tax haven company by way of share capital (equity), and that company then lends to the subsidiary. Interest paid to the tax haven company will not be taxed in the tax haven. In each case where the country of the parent company is an exemption country, the tax haven subsidiary may be able to pay a dividend tax free to the parent so that the profits end up with the parent company having suffered very little tax. If the residence country of the parent company is a foreign tax credit country, the profits can often be retained in the tax haven company and used for group operations in other countries without attracting tax in the parent’s residence. Again, the overall result is payment of little or no tax in the source or residence country.

Given that both the residence and source countries are suffering from this tax haven activity, action can be expected from both. The source country may deny deductions for payments by resident companies or by branches of nonresident companies to tax havens or may permit deductions subject to special conditions.138 A rule of this kind has a number of problems. It is necessary to have a list of countries that are treated as tax havens, and, although such lists are readily available, they need frequent updating. The rule reintroduces the problem of the back-to-back transaction, in that the tax-haven-related company in the thin capitalization case above can, for example, route the loan through a bank in a country that is not a tax haven. Such a rule may also affect quite legitimate payments to tax havens (which in a number of cases are major financial and trading centers in their own right). Finally, if a tax treaty is in force with the tax haven, the rule may fall foul of the nondiscrimination provision in the treaty (obviously, great care is needed in negotiating a tax treaty with a tax haven).

Nevertheless, a rule focusing on payments to tax havens should be considered in some form—for example, a tax clearance system for payments that, to the knowledge of the payer, are made directly or indirectly to tax haven entities. Another possibility is to require all companies and branches in the country to report selected information on transactions with tax havens or more broadly on international transactions.139

5. Double-Dipping

Alternative techniques for reducing source and residence taxation that have been used in recent times seek to double up on favorable tax rules in both source and residence countries (generally referred to as double-dipping). A variety of methods are used.

One method is to exploit differences in the tax law treatments of the same transaction in the source and residence countries. A common example has been the financial lease of equipment. Some countries recharacterize finance leases for tax purposes as purchases and loans, while other countries treat them in the same way as operating leases (i.e., the lessee is treated as paying rent and the lessor as being the owner of the equipment).140 The result is that two countries can end up treating two separate taxpayers (one country the lessor and the other country the lessee) as the owner of equipment and entitled to depreciation and interest deductions. Given that rent in economic terms is equivalent to depreciation and interest, the difference in treatment should not produce a substantial tax variance, but many countries have tax incentives for investment in capital equipment in the form of accelerated depreciation, investment credits, or allowances. Where two different taxpayers are treated as the owner of the equipment in different countries and each is entitled to these incentives in one of the countries, the taxpayers effectively double up on the incentives in a way not intended by either country.

It is not clear, however, which country is being disadvantaged in tax terms and which might therefore be expected to take remedial action. One of the affected countries could enact a rule that investment incentives will not be available under its law when similar incentives are being obtained in respect of the equipment under the law of another country, but the rule will lead to circularity if both countries adopt it. Alternatively, a country may limit investment incentives to equipment used in the country, which will work in most cases, although not for mobile equipment like airplanes. Another solution is for each country to do away with or reduce the investment incentives, as in fact happened in many industrial countries during the 1980s (for more general policy reasons having little to do with the problems of international tax avoidance). Where a developing or transition country adopts this kind of investment incentive, a rule limiting the benefit of the incentive to equipment used in the country is probably the easiest way to ensure that it does not suffer unduly from double-dipping of this form.

Another form of double-dipping that has been much exploited involves dual-residence companies. Some countries permit grouping of the income and losses of commonly owned resident companies (often achieved by permitting the transfer of tax losses to related companies). If the same company is resident in two such countries and has borrowed to finance group operations (whether in those countries or elsewhere), it may be able to deduct the interest in each country. If it has little or no current income, a loss will arise from the interest deductions that may be able to offset the income of two related companies, one in each country where the loss company is resident. Again, it is not clear which country is the loser from this transaction. Nevertheless, a number of countries have enacted rules that prevent the losses of dual-residence companies arising from financing transactions being used to offset the income of any other related company in the country; that is, the losses can be used only to offset future income of the dual-residence company.141 If a developing or transition country does not permit the transfer of losses within a group of companies, it is unlikely to suffer from this particular double-dipping problem. It follows that care should be exercised in permitting transfer or consolidation of losses for tax purposes among commonly owned resident companies.

The deduction of the same expense in two countries is not of itself a cause for concern. Where a resident of a foreign tax credit country has a branch in another country, it will typically get deductions for the same expenses in the source and residence countries. These deductions will generally be offset, however, against the same income that each country is taxing, with the residence country giving double tax relief. The double-dipping problem usually involves the offsetting of the same deductions against different income of different taxpayers. As there are probably as many ways for taxpayers to exploit differences in tax systems of different countries as there are differences, and as the outcome is often ambiguous in terms of whether tax avoidance is involved and which country is suffering an unfair reduction in tax, it is likely that double-dipping will continue to be a difficult international tax problem without a clear solution.

6. Treaty-Shopping

Tax treaties themselves may become the object of tax avoidance activities, even though they often express the purpose of preventing tax avoidance. This possibility was of course never intended by the original framers of model tax treaties and is not in itself sufficient reason for a country to reject the negotiation of tax treaties as their benefits usually outweigh the detriments. The possibility of abuse arises from two features of the tax treaty network—its incomplete coverage of the world and its bilateral structure. The former feature flows from the latter because it is not possible to negotiate with virtually all of the countries of the world at once (as contrasted, say, to the Uruguay Round of multilateral trade negotiations of the General Agreement on Tariffs and Trade); the latter is regarded as flowing from the wide variations in tax systems around the world, so that it is felt necessary for each country to handcraft a tax treaty accommodation with other countries one by one.

A resident of a country that does not have a tax treaty with a particular developing or transition country can simply incorporate a subsidiary in another country that does (usually one with which the investor’s country also has a treaty) and route its investment through that subsidiary, which will be entitled to the reduced tax rates and other protections available under the treaty. Alternatively, a resident of a country with which the developing or transition country does have a treaty may seek what it regards as better tax treatment under another tax treaty by the same route.

For example, the treaty between the investor’s country and country X may have a 10 percent rate limit on royalty payments. If that investor can find another country that has a tax treaty with country X that contains a zero tax rate on royalties, then it will be possible to route a licensing transaction through a subsidiary in that country and eliminate the source country royalty tax; if that third country in turn has a treaty with the investor’s country containing a zero tax rate on royalties, it will be possible in turn to pay the royalty on to the investor without tax in that country. These examples assume that the royalties are deductible in each country by the person who is paying them. The nondiscrimination article of tax treaties will normally ensure that they are deductible on the same basis as royalties paid domestically, so that the assumption will be correct in most cases.

This kind of practice is known as “treaty-shopping.” A country can prevent treaty-shopping by seeking to ensure that its treaties with other countries are uniform in their main elements, especially the tax rate limits on interest and royalties and the definition of permanent establishment. If all treaties to which country X is a party have a 10 percent tax limit on royalties, for example, the planning in the second example in the previous paragraph would not be possible. Many countries have been able to achieve this consistency in their treaty negotiations and have thereby reduced the problems of treaty-shopping. Nonetheless, the possibility remains that residents of nontreaty countries will get treaty benefits through related companies in treaty countries.

One way to deal with the problem is to insert a general provision in tax treaties denying treaty benefits in such cases. The United States is the only country to practice this approach in a comprehensive way142 although some other countries routinely insert more limited treaty abuse provisions in specific articles of treaties.143 The Commentary on article 1 of the OECD Model contains a number of possible provisions for this purpose.144 Developing and transition countries may instead prefer to rely on general antiavoidance provisions in domestic legislation to deal with treaty abuse. While a view is developing that such provisions are not inconsistent with tax treaties, it is probably safer to spell out in the negotiations that the general antiavoidance provision of domestic law will be applied to treaty abuses and to ensure that the general priority rule for tax treaties in domestic law makes this relationship clear.145

7. Combinations of Tax Avoidance Techniques

International tax avoidance in many cases will utilize a combination of the techniques outlined above. Thus, treaty-shopping activities will often go hand in hand with the use of tax havens and the interaction of tax treaties and domestic law. For example, and by way of extension of the case of treaty-shopping in the royalties area discussed above, some countries do not in their domestic law charge withholding tax on payments of royalties by residents to nonresidents for reasons that have been discussed earlier. If such a country has a treaty with a developing or transition country containing a zero royalty rate, a nonresident investor can incorporate a company there to receive royalties from the developing or transition country and then arrange to have the royalties paid to a tax haven company. If the royalties paid to this company equal the royalties received by the company in the tax treaty country, no tax will be collected in that country because the deduction for the royalties paid will wipe out the royalty income received and no withholding tax will be levied on the outgoing royalties. Hence, the result will be achieved of no tax at all being levied on the royalties (unless the residence country of the ultimate owner of the tax haven company has a controlled foreign company regime of the kind discussed below).

The financing of company groups often involves variations on double-dipping, treaty-shopping, and use of a tax haven. For example, a parent company may borrow to finance investment by way of share capital in a tax haven finance subsidiary, which in turn lends to an operating subsidiary in a developing or transition country through a back-to-back transaction with a bank in a country that has a tax treaty with the developing or transition country, lowering the rate of withholding tax on outgoing interest. If the residence country of the parent is an exemption country but nonetheless permits deduction of the interest paid on the loan taken out to finance the investment in the tax haven subsidiary, it is likely that the dividends received from the tax haven subsidiary (representing the interest received by that subsidiary) will be exempt in the parent’s residence country, and yet interest deductions may have been obtained in two countries (that of the parent and the operating subsidiary) for offset against different income of different taxpayers.

Because of the sophistication of international tax planning and its frequent combination of domestic law, tax havens, and tax treaties, the taxation of nonresident direct investors by developing and transition countries is not an easy task. An array of provisions in domestic legislation (such as provisions on transfer pricing, thin capitalization and tax haven payments, and a general antiavoidance rule) and great care in the negotiation of tax treaties will assist in dealing with the differing kinds of tax planning.

A developing or transition country should make clear through explanatory or administrative material that it does not intend to use (the threat of) multiple taxation to penalize taxpayers. If it is felt that the problems of international tax avoidance justify severer penalties than normal, then the tax penalty regime should provide for this directly. Similarly, if higher levels of disclosure of information are required in the international area, the legislation should provide for such disclosure explicitly. Care in drafting such provisions is necessary to ensure that they are not in breach of nondiscrimination provisions in tax treaties. For this reason as well as the considerations raised in relation to residents below, the provisions should apply to both inward and outward investment cases.

Stricter enforcement regimes may be viewed adversely by foreign investors and temper their willingness to invest in the country in question. Each developing and transition country has to judge this issue for itself. The same reluctance may also be triggered by legislation directed at international tax avoidance practices. One possible response (effectively giving in to the difficulties of enforcing international tax rules) is to not include antiavoidance provisions of the kinds outlined above in domestic tax legislation and so to allow nonresident direct investors to determine their tax level for themselves. Alternatively, such provisions can be coupled with special tax regimes for foreign investors conferring tax holidays and other tax privileges on them in specified cases. The alternative approach nominally gives control over the targeting of the tax benefits for foreign investors to the developing or transition country, although in practice, as noted in chapter 23, it generally leads to other forms of tax avoidance.

8. Nonresident Portfolio Investors

The position of nonresident portfolio investors differs substantially from that of direct investors. Because direct investors have control over the transactions undertaken within the company group, they can engage in transfer pricing, thin capitalization, and the like in ways not generally available to the portfolio investor. It is possible for portfolio investors to employ tax havens and treaty-shopping in some of their activities (e.g., using tax treaties to obtain lower withholding tax rates on interest), but generally the scale of the investment in a particular company by a particular investor is unlikely to justify elaborate tax avoidance of the kind that may be practiced by direct investors.

The tax planning of the portfolio investor is likely to consist of portfolio choice. For example, purchasing shares in a company resident in a developing or transition country exposes the investor to the corporate tax and the withholding tax on portfolio dividends, which is probably higher than the tax on dividends paid on direct investment. Some portfolio investors (e.g., tax-exempt pension funds) will be tax privileged in their residence countries and so may not benefit greatly from double tax relief in this case. Failing special provisions in the law of the source or residence country (or both, possibly through a tax treaty) to deal with the international implications of its special tax position, the portfolio investor may adopt a different investment strategy, such as investing in (profit-related) debt of the company resident in the source country and options over the unissued capital of the company. In this way, it benefits from increases in the value of the company’s shares and a share in its income stream without being exposed to much if any tax in the source country (interest withholding tax probably being the only tax applicable) while enjoying its tax privileges in the residence country. With the advent of modern financial instruments, more sophisticated strategies are available to the portfolio investor who wishes to be exposed to a particular market or company without the accompanying source tax liabilities.

Often, the influence of the portfolio investor will be felt not directly but indirectly in the source country through the foreign direct investor responding to the needs of the portfolio investor. Since portfolio investment opportunities are often limited in developing and transition countries (because of the lack of a stock exchange or very thin trading in whatever stock market exists), portfolio investors more often than not will invest in the multinational direct investors operating in the countries (and many other countries) as a way of exposing themselves to investment in many countries and at the same time minimizing risk by having multicountry coverage in the investment. As the totality of investment in many multinational companies will be dominated by institutional portfolio investors, such as pension schemes, banks, and insurers, the companies are likely to adapt their tax profile to suit the institutional investors. If the preference of the institutional investors is to have low-taxed returns because of their privileged tax position in their residence countries, the optimal tax strategy for the multinational company may be to reduce its tax liabilities in all countries, which takes us back to the beginning of the discussion in this section of the chapter.

Increasingly in recent times, international investment has developed many tax niches, with offshore funds offering specialized investment products designed to appeal to particular kinds of investors. The discussion of capital flight above dealt with one kind of such fund in the case of resident portfolio investors. Here, the discussion concerns similar funds designed for nonresident portfolio investors. To the extent that the multinationals do not respond to the tax situations of their different classes of portfolio investor, it is often possible to find a fund that consolidates portfolio investors with a similar tax and investment profile and develops investment products that suit that profile. While it often may not be worthwhile for a single portfolio investor to use treaty-shopping and tax havens for its operations, it is for such offshore funds. Hence, the other major effect of portfolio investors in developing and transition countries is being felt through the operations of such funds (with many specializing in investment in (particular) developing or transition countries). To the extent that this investment is highly tax sensitive, there is little that a developing or transition country can do to prevent offshore fund tax planning in which the nonresident portfolio investor decides to invest in a particular fund or to withdraw from the fund.

The major problem that the nonresident portfolio investor poses is the potentially deleterious incentives that the investor’s tax position creates for direct investors to reduce source country taxation. Hence, additional antiavoidance provisions are not necessary in the law to deal with the tax position of the portfolio investor beyond what has been canvassed in earlier discussion. If the source country specifically wants to cater to the tax problems faced by foreign pension funds and the like, some suggestions have already been made.

9. Resident Investors

Although developing and transition countries will most often encounter the forms of tax avoidance outlined above in the case of nonresident direct investors, many of them are equally available to residents of the developing or transition country (resident in the sense that the ultimate investor is a resident of the country and is not foreign owned). It is necessary in this case to distinguish two investment situations: first, where the ultimate investment is made by the resident overseas and, second, where the ultimate investment is made in the developing or transition country itself.

The first case is the more obvious but less frequent in practice from the point of view of the developing or transition country. A resident engaging in direct investment overseas can engage in transfer pricing and the other kinds of activities outlined above for the dual purpose of reducing the tax in the country of source (where the investment is made) and in the country of residence. Avoidance of resident country tax will often involve the diversion of income from the country of residence to tax havens, combined with manipulation of the system of double taxation relief. Avoidance of source country tax in these kinds of cases is not directly the concern of the residence country, its main concern being to protect its own taxing rights. The mechanism to control tax haven use by residents that is increasingly being used in advanced market economies is to tax residents on their share of low-taxed foreign income derived by nonresident companies controlled by the residents. These “controlled foreign company” regimes are usually very complex in operation.146

As regards the second case, initially it seems unlikely that a resident of a developing or transition country would get involved in international tax avoidance for investment in that country. In fact, however, a resident direct investor can easily appear as a nonresident direct investor by channeling investment into the country through a nonresident company that the resident owns. By this means, the possibilities of reducing tax in the country through transfer pricing, thin capitalization, tax havens, and treaty-shopping become possible in a similar way as for true nonresident direct investors. In addition, the resident can also by this route seek to enjoy any tax concessions given specifically to nonresident investors.

This possibility of residents assuming the guise of nonresidents is closely linked to the capital flight issue canvassed earlier, at least in transition countries. Capital flight was the first reaction of many wealthy people there to the uncertainty and instability created by the transition. As the situation has clarified to the extent that it is possible to conduct profitable business operations in the countries, it seems likely that a number of these people have reintroduced their capital into businesses as disguised foreign direct investment. While the capital is thus reexposed to the risks that it was originally fleeing, the ownership of the capital is usually hidden (as it often originates in tax havens), tax concessions may be available, and further flight remains possible to the extent that the capital remains mobile. There is some evidence that much of this capital is invested in import-export and similar activities that do not anchor the capital in the same way that investing in large-scale plant and equipment would.147

For both kinds of tax avoidance by residents, the provisions discussed above—such as rules on transfer pricing, thin capitalization, tax havens, and general antiavoidance—are available to deal with some of the problems. It was partly because of the need to control circular transactions of residents taking capital offshore and reintroducing it into the country that it has been suggested that many of the rules should not be directed to nonresidents. Eventually, controlled foreign company regimes will be needed, despite the difficulties that they entail. For the time being, the embryonic measure that has been outlined above for the capital flight case can at least be adopted to signal that a developing or transition country is aware of the international tax avoidance techniques that residents may use and proposes to combat them.

VIII. Tax Treaty Issues Not Covered in Domestic Law

A number of provisions found in tax treaties are not usually reflected in domestic law. This section briefly describes these provisions, together with their effect on domestic law, specifically nondiscrimination, exchange of information and assistance in collection, and the mutual agreement procedure.

A. Nondiscrimination

The nondiscrimination article of tax treaties is designed to ensure that foreign investors in a country are not discriminated against by the tax system compared with domestic investors. The OECD Model nondiscrimination provision is narrower, however, than similar provisions found in other areas of international law, such as trade. This difference is necessary because the international tax system operates on the residence and source principles and so necessarily distinguishes the tax position of residents and nonresidents. Hence, it is not usually regarded as discriminatory to collect flat-rate gross withholding taxes from a resident of the other state without a permanent establishment when a resident is taxed on the same income on a net assessment basis.

The first paragraph of the nondiscrimination article in the OECD Model provides against discrimination on the basis of nationality, but makes it clear that distinctions on the basis of residence will not be regarded as giving rise to nationality discrimination (in other areas, such as EU law, residence distinctions can amount to nationality discrimination).148 Hence, to breach this provision it is necessary for a country to treat a resident who is a national of the other state less favorably in the levy of tax or procedural requirements than a resident national, or a nonresident national of the other state less favorably than a nonresident national. Such forms of discrimination are rare in domestic tax laws. The second paragraph of the OECD nondiscrimination article applies a similar rule to stateless persons; this provision rarely appears in actual tax treaties.149

The third paragraph of the nondiscrimination article in the OECD Model requires that a permanent establishment of a resident of the other state shall not be less favorably taxed than enterprises of residents carrying on the same activities. This is the most important provision of the article in practice and, combined with the other articles of the Model—especially the business profits article—means that the profits attributable to a permanent establishment have to be taxed on a net basis150 and that the permanent establishment must otherwise be taxed under the same rules as domestic enterprises. The article deals with the amount of tax liability and not connected requirements so that it is possible, for example, to apply withholding taxes on income derived by a permanent establishment of a nonresident even though such taxes are not applied to a domestic enterprise, so long as the ultimate tax is on a net basis (i.e., any withholding taxes are not final, are credited against the ultimate tax liability, and are refunded if there is an excess). If withholding taxes are applied to income derived by domestic enterprises, there is no question of breach of the nondiscrimination article in applying them to nonresidents, but even if the taxes are final for a resident enterprise, they cannot be for a permanent establishment of a nonresident because of the requirement of the business profits article that taxation be on a net basis.

The exact extent of the nondiscrimination obligation under this paragraph is not clear in all cases, especially as regards application of progressive rate scales to companies, tax relief for intercorporate dividends, and the granting of foreign tax credits to permanent establishments for any foreign tax levied on income attributable to the permanent establishment. The Commentary to the OECD Model contains a lengthy discussion of these issues.151 The second sentence of the third paragraph does make clear that it is not necessary to grant personal allowances to a nonresident individual carrying on business through a permanent establishment (this sentence often appears as a separate paragraph in actual treaties). Thin capitalization rules that are applied to a permanent establishment borrowing from related parties but that are not applied to resident enterprises may be contrary to the paragraph, depending on how the rules are framed. The example of thin capitalization rules given above will not be contrary to nondiscrimination rules because they apply to all enterprises. Branch profits taxes may be contrary to the terms of this paragraph, so that its terms need modification if a country wishes to levy a branch profits tax.

The fourth and fifth paragraphs of the OECD nondiscrimination article ensure that resident enterprises whose capital is wholly or partly owned or controlled by a resident of the other state are not subject to discrimination. The fourth paragraph refers specifically to deductions for interest, royalties, and other disbursements and makes clear that deductions can be denied through the application of the arm’s-length principle by way of exception to the requirement for the same treatment. If a developing or transition country adopts a rule denying deductions for payments to tax havens, the rule will generally be overridden by the fourth paragraph if a tax treaty is in effect with the tax haven. Hence, the caution above about negotiating tax treaties with tax havens. The fourth paragraph is more general, preventing heavier or different taxation or connected requirements than for other similar enterprises. While it can cover the same ground in part as the fourth paragraph, the fifth is more specific and therefore prevails in the event of overlap. Thin capitalization can be an issue under the fourth paragraph, but not if the rules are applied generally to all enterprises. The fifth paragraph would prevent, for example, a local subsidiary of a parent in the other state from being subjected to a higher tax rate than other companies.

The final paragraph in the OECD Model provides that, unlike the other provisions of a tax treaty, the nondiscrimination article applies to all taxes levied by a state. This provision is often omitted from actual tax treaties or altered to make clear that it applies only to taxes covered by the treaty (which it is not strictly necessary to state).

Because tax treaties are enacted in one way or another as part of domestic law and prevail over other taxing provisions, the nondiscrimination provision is self-executing and overrides domestic rules that conflict with it. Because of the general terms of the nondiscrimination article, it is necessary to be aware of its operation when drafting domestic rules. There is generally little point in devising domestic rules that are contrary to the nondiscrimination rules, except in the case of tax haven provisions.

The nationality paragraph aside, at first sight the nondiscrimination article seems to have a residence state bias because its provisions operate effectively only on the source state (where the permanent establishment or subsidiary operates). This view is not accurate if the structure of tax treaties is looked at broadly. It was noted above that the foreign tax credit system in particular may create an incentive for the source country to increase its taxation on nonresidents (or subsidiaries of foreign parent companies) up to the level of tax in the residence country. While tax treaties impose rate limits on source taxation or exclude source taxation altogether in some cases, for income of a permanent establishment or a subsidiary there are no such limits. Hence, the nondiscrimination article ensures that source countries do not target higher taxes to these cases and prey on the relief system of the residence country.152 The equivalent undertaking of the residence country is in its treaty obligation to relieve double taxation for source taxes levied in accordance with the treaty. The residence country could not satisfy its obligations under this paragraph by levying tax rates on foreign investment that are higher than those on domestic investment and then purporting to relieve double taxation through a tax credit.153 The nondiscrimination article does not prevent a country from discriminating in favor of nonresidents (as with tax holidays or other incentives that apply only to foreign investors). Nor does the article prohibit provisions in the domestic law that favor the location of investment in the country; for example, a country can have special tax incentives for research and development conducted in the country or for plant and equipment used in the country, as long as these locational incentives are not confined to residents or locally owned companies.

B. Exchange of Information and Assistance in Collection

Most countries have a domestic law rule that they will not directly or indirectly assist another country in the collection of its taxes.154 This rule means that exchange of tax information and other forms of assistance in collection of taxes are not possible without a tax treaty that overrides this rule in domestic law. The tax secrecy rules of many countries also prevent the exchange of information. Exchange-of-information provisions are found in virtually all tax treaties, but other forms of assistance are less commonly provided for.

The standard OECD and UN Model exchange-of-information article requires a country to obtain information for its treaty partner where the information is necessary for carrying out the provisions of the treaty or of the country’s domestic tax law. Exchanged information is required to be kept secret in accordance with the secrecy rules of domestic law of the recipient country and in accordance with the express treaty rules on this topic. In addition, the standard treaty article provides that information need not be exchanged when it involves commercial or trade secrets. Tax secrecy is often not as strong an institution in developing or transition countries as it is in industrial countries and so can be a very sensitive topic in tax relations between treaty partners. It is implicit in the exchange-of-information article, however, that a country cannot refuse to give information to its treaty partner because of its own tax secrecy laws.

The exchange-of-information article also serves as a test of the lowest common denominator for procedures of collecting information. Information need not be collected if it cannot be obtained under the procedures of either country. For example, the information being sought may be kept at the home of a taxpayer. If the tax procedure law of either treaty country forbids entry of domestic (as opposed to commercial) premises to obtain information, then there is no obligation to obtain the information. If, however, the impediment arises under the law of the country making the request and if the country that has received the request for such information is able to obtain the information under its laws, that country may (but is not obliged to) forward the information to the other country under the exchange-of-information article.

Unlike other articles of tax treaties, the exchange-of-information article is not limited in application to residents of the treaty partners. For example, one country can request the other to obtain information from a permanent establishment in that state of a resident of a third state. Although the Models do not so provide, information is being increasingly extended to taxes other than income taxes for the practical reason that many countries use the same tax officials to enforce a number of different taxes (e.g., income tax and value-added tax), and it is difficult for an official who has received foreign information to use it only in relation to one tax when it is relevant to several taxes.

The OECD provides considerable practical guidance on exchange of information.155 The use of computers in tax administration is spilling over into this area, and the sophistication of the exchange process has increased rapidly. The OECD has developed a standard computer format for exchange of information.156 In recent years, the exchange article has given rise to some novel extensions of its use, such as for simultaneous audits of the same or related tax-payers by each party to a treaty (and even by more than two countries through the use of exchange provisions in a number of treaties). The OECD has developed a model agreement for tax administrations to formalize the process.157 Whether developing or transition countries will be able to participate in these recent developments will depend on their level of computerization and audit capacity.

In addition, provisions for assistance in collection are increasingly being included in tax treaties. Under these provisions, each country undertakes to collect the taxes of the other. As no OECD or UN Model provision currently exists for this purpose, the following text is provided as a sample.

Article 27. Assistance in Collection

  • (1) The competent authorities of the Contracting States undertake to lend assistance to each other in the collection of taxes, together with interest, costs, and civil penalties relating to such taxes, referred to in this article as a “revenue claim.”

  • (2) Requests for assistance by the competent authority of a Contracting State in the collection of a revenue claim shall include a certification by such authority that, under the laws of that State, the revenue claim has been finally determined. For the purposes of this article, a revenue claim is finally determined when a Contracting State has the right under its internal law to collect the revenue claim and the taxpayer has no further rights to restrain collection.

  • (3) A revenue claim of a Contracting State that has been accepted for collection by the competent authority of the other Contracting State shall be collected by the other State as though such claim were the other State’s own revenue claim as finally determined in accordance with the provisions of its laws relating to the collection of its taxes.

  • (4) Amounts collected by the competent authority of a Contracting State pursuant to this article shall be forwarded to the competent authority of the other Contracting State. However, except where the competent authorities of the Contracting States otherwise agree, the ordinary costs incurred in providing collection assistance shall be borne by the first-mentioned State, and any extraordinary costs so incurred shall be borne by the other State.

  • (5) No assistance shall be provided under this article for a revenue claim of a Contracting State in respect of a taxpayer to the extent that the revenue claim relates to a period during which the taxpayer was a resident of the other Contracting State.

  • (6) Nothing in this article shall be construed as imposing on either Contracting State the obligation to carry out administrative measures of a different nature from those used in the collection of its own taxes or that would be contrary to its public policy (ordre public).

  • (7) Notwithstanding the provisions of article 2 (Taxes Covered), the provisions of this article shall apply to all taxes collected by or on behalf of the Government of a Contracting State.

Whether the last paragraph is included will depend in part on a similar extension being made to the exchange article. On the grounds of administrative capacity, developing and transition countries may not consider such an article appropriate to their circumstances (and, equally, industrial countries may not be willing to agree with them on this article). More elaborate stand-alone treaties dealing with tax administration have been developed, and the Multilateral Treaty on Mutual Administrative Assistance, which covers exchange of information, service of documents, and assistance in collection, is open for signature to those countries that join the Council of Europe or the OECD.158 It entered into force on April 1, 1995.

C. Mutual Agreement Procedure

The final provision of tax treaties that requires comment is the article on the mutual agreement procedure. Under the Model versions, this article performs three functions: it provides a dispute resolution mechanism in relation to the application of the provisions of tax treaties to specific cases; it allows the countries to settle common interpretations and applications of their tax treaty; and it allows them to resolve cases of double taxation not otherwise dealt with by their treaty. Some countries find that the third function and often the second are difficult to reconcile with their domestic laws and procedures and therefore omit them from their treaties. In practice, it is dispute resolution for the specific case that predominates, whatever the precise form of the article.

The ground on which the taxpayer can invoke this procedure is that the actions of one of the states result or will result in taxation not in accordance with the treaty. The taxpayer has three years to invoke the procedure from the first notification of the act complained of. The states are obliged under the article to consult on the problem raised by the taxpayer if the state with which the problem is raised is unable or unwilling to resolve it unilaterally, but they are not obliged to resolve the case. If a resolution is agreed to by the states, then under the Models it is to be implemented notwithstanding domestic time limits on amending tax assessments. Some countries are unwilling to agree to such overriding of domestic time limits in their tax treaties.

No specific procedure is provided, but it is made clear that the tax administrations can make contact directly and do not need to go through diplomatic channels. The major issue that arises in practice is the relationship between domestic appeal procedures provided for in tax laws and the treaty dispute resolution mechanism. To avoid competition or conflict between domestic appeals and the mutual agreement procedure, some countries provide in their tax laws or procedures that the taxpayer must waive or suspend appeal rights under domestic law, while other countries will not actively pursue the competent authority procedure until domestic appeal periods have expired and the taxpayer has not utilized them.

The mutual agreement procedure has also been the subject of novel uses in recent times. The main development concerns advance pricing arrangements under which the mutual agreement procedure is used to agree to a transfer price in advance, so that taxpayers and tax administration are spared disputes after the event. This is a sophisticated procedure that for the moment is probably relevant only to industrial countries.159 Taxpayer dissatisfaction with the mutual agreement procedure has led some countries to adopt arbitration procedures in their tax treaties for cases where it is not possible for the competent authorities to resolve disputes. The main purpose of such provisions is to put pressure on the tax administration to resolve international disputes rather than to actually engage in arbitrations.160

IX. International Tax Priorities for Developing and Transition Countries

It will be evident from this chapter that the construction of the international elements of the income tax system in domestic law and tax treaties is a complex topic. Among developing and transition countries (as among industrial countries), there will be wide differences in the capability of the tax administration to deal with international tax issues. While priorities will vary from one country to another, this concluding part of the chapter indicates a line of development that should suit many developing and transition countries.

The priority of any tax system will always be to tax the domestic income of resident taxpayers.161 With the increasing internationalization of economic relations, however, even this goal means that attention must be given to international income tax issues. For better or worse, the globalization of the world economy impinges on developing and transition countries, and it is not possible for a country to isolate itself or its tax system. The interdependence of market economies is a new phenomenon, and transition countries in particular retain a residual belief in the ability of regulation to deal with problems. In some developing countries also the capacity of economic regulation in the current economic environment is overrated. Developing and transition countries face similar problems of international taxation as industrial countries, which means that, whatever may have been the case in the past, it is not possible to adopt the attitude that international issues can wait.

The incentives for capital flight are strong in developing and transition countries even apart from the tax system. If a country operates the source principle only, then it is necessary to have robust rules for the source of income to ensure that the source-based tax is not avoided. Even with such rules, there will be a strong incentive for residents to move income offshore in order to avoid taxation, which will be a relatively simple matter for passive portfolio income (by investment choice). The residence principle should be adopted to prevent this form of tax avoidance. Once the residence principle is adopted, then measures for the relief of double taxation by way of exemption or a simple foreign tax credit are also necessary. At this point of development, the country has satisfied the basic norms for international tax rules on which tax treaties depend.

The ability of residents, again by simple investment choice, to derive foreign-source passive income through nonresident taxpayers (such as offshore mutual funds) indicates that further measures are necessary even for the simple goal of protecting the domestic tax base in the case of residents not engaged in active businesses. A simple provision indicating an intention to levy tax in these cases, together with enforcement efforts directed at tax evasion using foreign bank accounts, is the best that can be achieved to deal with the various kinds of capital flight. Residents involved in purely domestic business activities can also use the international tax system to avoid taxes. In this case, investments will be looped offshore and back into the country, creating the potential for such techniques as transfer pricing, thin capitalization, and profit stripping to move profits out of the country, usually to tax havens. The simplest approach for dealing with such problems is a brief provision levying tax on the resident owners of the offshore entities. Such provisions are necessary today simply to ensure collection of tax on the domestic income of residents.

With provisions in place to secure the domestic tax base, probably the next priority should be tax treaties. These marginally increase the capacity to enforce taxation of the domestic income of residents through exchange of information (although the use of tax havens for much of the offshore activity limits the effectiveness of tax treaties). Most important, they signal to foreign investors the country’s intention to play by the generally accepted rules of international taxation and not to discriminate against foreign investors while leaving room (if negotiated in an appropriate form) to extend domestic taxes to foreign investors. Except in the increasingly unusual case of a country deciding not to pursue the negotiation of tax treaties, the contents of tax treaties overshadow the way in which a country should frame its tax laws for the taxation of foreign investors. It has been suggested throughout this chapter that the rules of tax treaties should generally be followed in domestic law for greater transparency and simplicity in the application of the tax law where a tax treaty is operative.

Taxation of foreign investors in developing and transition countries is a politically divisive issue. On the one hand, there is a natural resentment against the economic resources of a country being owned and exploited by foreigners. In the past, this attitude contributed in many developing countries to restrictions on foreign-owned operations. On the other hand, the need for foreign capital, technology, and management skills is increasingly felt as more and more countries compete for what is available, especially since the transition countries have entered the picture. The result is policy and administrative ambivalence to taxation of foreign investment.

Many countries offer tax incentives for foreign direct investors. While the efficacy of these incentives in attracting increased foreign investment may be doubted, any attempt to tax foreign direct investors effectively involves formidable problems of drafting the law and administering it. The basic provisions for taxing nonresidents consist generally of withholding taxes on passive and employment income and collection by assessment on business income. The investment choices for portfolio foreign investors and the tax avoidance techniques available to the foreign direct investor mean that such provisions are not adequate and that rules in domestic law on transfer pricing, thin capitalization, and tax havens are required. These will by no means cover the tax avoidance strategies available. A general antiavoidance provision or doctrine will assist the tax administration to cope with international tax avoidance, but requires considerable effort to implement. In short, any serious attempt to collect tax from foreign direct investors is fraught with drafting and administrative difficulties, while taxation of portfolio investors may simply induce them to move their investment out of the country. For these reasons, the taxation of foreign investors is probably the last international taxation issue that a developing or transition country should seriously tackle.

The number and significance of the international tax problems that confront the income tax are reasons why developing and transition countries do well to rely on alternative tax bases in addition to the income tax as a major source of tax revenue. The value-added tax, excises, social security, and property taxes generally present fewer international difficulties of drafting and enforcement than the income tax.