21 Fiscal Transparency

Victor Thuronyi
Published Date:
June 1998
  • ShareShare
Show Summary Details
Alexander Easson and Victor Thuronyi 

Men may put on the habiliments of a partnership whenever it advantages them to be treated as partners underneath, although in fact it may be a case of “The King has no clothes on” to the sharp eyes of the law.

Felix Frankfurter

I. Introduction

As discussed in chapters 14 and 19, income tax systems invariably draw a distinction between physical persons and legal persons. In some systems, income tax is imposed by separate laws—an individual income tax law and a corporate (or enterprise) income tax law; in others, physical and legal persons are taxed under the same law, but are governed by separate rules and rate schedules. Some business or investment income, however, is not earned directly by such taxpayers, but is earned through entities or arrangements that—depending on the legal system—may or may not be separate persons. In that case, it is necessary to decide whether to tax the entity as a separate physical or legal person, or to provide for fiscal transparency, whereby the entity’s income flows through to its owners. Pure transparency would mean disregarding the entity altogether, which is sometimes done.1 However, more commonly, the entity is recognized as existing for tax purposes, but rules are devised so that the entity’s income is taxed not to it but to its owners. This chapter explores the circumstances under which fiscal transparency (also called flow-through treatment) applies for income tax purposes and the rules by which transparency is given effect.

The topic is a confusing one to investigate on a comparative basis, for two reasons. First, the nature of legal arrangements to which transparency can be applied differs considerably from one legal system to another. In general terms, there is a big difference between common law and civil law countries, although there are also differences within the groups of common and civil law countries. Because tax law must apply to the economic rights that are specified in a country’s civil and commercial law, these legal differences have strongly influenced the tax rules in various countries. Given these fundamental differences, it is difficult in this area to generalize and point to an optimal set of rules for the income tax. Second, even laying aside the differences in underlying legal systems, most industrial countries have not formulated rules for transparency in a thorough and consistent fashion. Developing and transition countries that are formulating rules to deal with transparent entities must therefore rethink approaches to the issue that have been employed elsewhere.

A further problem arises from the fact that, while in some countries the tax status of an entity is determined by its status (as a legal person or otherwise) under civil law, in many systems the tax status of an entity is established by the tax law and does not always coincide with its status under private law.2 In some cases, the entity is a legal person but is not treated as a separate taxpayer for purposes of the income tax on legal persons. In other cases, the converse is true—the entity is not a legal person but is regarded as such for tax purposes.3 Such a difference in status should not necessarily be considered a defect in the overall legislative scheme; there are perfectly valid reasons why an entity might be regarded as a legal person for purposes of registration or of civil liability, but not for purposes of taxation.

This chapter is primarily concerned with the income tax treatment of those entities that are neither legal nor physical persons, and with entities that are legal persons under the general law but are not treated as such for purposes of the income tax.

There are a great variety of ways in which ownership interests in investment property or in a business may be split up among different participants. In the case of investment property, there can be a pure co-ownership arrangement (such as a joint tenancy), in which two or more persons each own a fraction of the property. In such a case, minimal rules are needed to specify the taxation of the income from the property: each owner is taxed on his or her fractional share of the income.4 Such joint-ownership arrangements are not considered further in this chapter.

Arrangements for the joint operation of a business can be referred to generally as partnerships, although the term does not mean precisely the same thing in different legal systems. The tax treatment of partnerships is discussed in section II. Another important joint-ownership arrangement in common law countries is the trust, which has some analogues in civil law countries. Taxation of trusts is considered in section III. Finally, section IV deals with a number of other business entities that are accorded special tax treatment in various countries.

One general approach to taxing such entities is to provide some form of transparent treatment, whereby the income is taxed at the level of the owners rather than at the level of the entity. Precisely how this may be done is considered below. An alternative approach is to accord only partial flow-through treatment to the entity; income that is distributed or allocated to the beneficiaries or owners is taxed to them, with the remainder being taxed at the entity level. This method is commonly adopted for trusts and is designed to ensure that all the income of the entity is taxed once.

II. Partnerships

A. Introduction

1. Legal Nature of Partnerships

The legal concept of partnership exists both in common law legal systems and in civil law countries, although the two concepts are not entirely equivalent.5 The traditional common law definition holds that “[p]artnership is the relation which subsists between persons carrying on business in common with a view to profit.”6 That is, a partnership is a relationship among persons, essentially contractual in nature rather than a “person” in its own right.

Civil law systems generally do not use the term “partnership” but have the concept of what could be literally translated as an association of persons or company of persons.7 This concept is distinct from that of a capital company.8 The precise legal nature and form that companies of persons may take differ depending on the civil and commercial laws of each country.9 In many civil law countries, a distinction is also made between civil law partnerships (governed by the civil code) and commercial partnerships (regulated by the commercial code). Typically, civil law partnerships are those that are engaged in farming or investing in land, or that are carried on by members of the liberal professions—activities not considered to be “commercial.”10 They can also include agreements to split the profits of a business.11

Most countries recognize at least two forms of partnership: the general partnership, in which the partners are jointly liable for the debts of the firm, and the limited partnership, in which the liability of some of the partners is limited.12 In a number of countries, there are more than two forms of partnership, and the tax treatment may vary according to the particular form.13

In some legal systems, partnerships have legal personality, while in others they do not.14 In this chapter, the term “partnership” is used not as a term that corresponds precisely to a concept in the legal system of all countries, but as a general one that encompasses a variety of legal forms. This variety and the differences in treatment under civil law, in particular whether the partnership is considered a legal person, make it difficult to generalize about partnerships. To some extent, differences in tax treatment from one country to another may also have been influenced by differences in civil law.

The term “joint venture” may be even more confusing than partnership. In some countries, joint ventures are transparent arrangements that may be less formal than partnerships.15 The term is also sometimes used in ways that include a number of legal entities, including capital companies.16

2. Partnerships as Taxable Entities

The absence of legal personality of partnerships in many countries may have facilitated transparent treatment for tax purposes, although the fact that a partnership is or is not categorized as a legal person is not necessarily determinative of its tax status. Different approaches are possible. In several countries, partnerships are considered legal persons but are not treated as taxable persons.17 Belgium, Spain, and many Latin American countries treat as taxable persons those forms of partnership that are legal persons, except for specified cases where a fiscal transparency regime applies.18 In common law countries, partnerships generally are not considered legal persons and are not taxed as corporations, although some partnerships that are considered to resemble corporations are taxed as corporations rather than as partnerships, even though they are not legal persons.19 In Indonesia, partnerships are taxed as separate entities even though they have no legal personality.20

In recent years, the civil and commercial laws of many transition countries have undergone changes under which the legal status of various kinds of business entities has been defined or redefined; the tax treatment of such entities has also been in a state of flux. The different patterns that have emerged can be illustrated with some examples. In Kazakhstan and Romania, all legal persons (including partnerships) are subject to income tax as separate entities.21 In Latvia, the enterprise income tax applies to all enterprises, which are defined according to registration requirements,22 except that partnerships are taxed on a flow-through basis23 and physical persons and “individual enterprises” that are not required to submit annual reports under commercial law are taxed on a flow-through basis to the owner.24 Individual enterprises that are required to submit annual reports are therefore taxed as entities even if they are not legal persons. Similarly, in China, all enterprises are subjected to enterprise income tax as separate entities regardless of whether a given entity is a legal person.25 In Estonia, the entity-level tax applies to legal persons.26 General and limited partnerships are taxed under the entity-level tax, except that general partnerships consisting of no more than 10 partners who are all resident physical persons are taxed on a flow-through basis.27

Taxing partnerships as entities has the advantage of administrative simplicity, as it is generally easier to collect tax from a single entity than from the individual participants. Income tax returns of partners who are physical persons are kept simple, as they do not include income received through the entity,28 and complicated rules for the taxation of flow-through entities can be largely avoided. A further advantage of taxing partnerships as entities is that it avoids discrimination between different forms of business organization and eliminates “entity shopping.”29 However, the disadvantage is that the income will then normally be taxed at a flat rate rather than at the marginal rates applicable to the individual partners.

If partnerships are taxed as entities, it must also be decided whether they should be treated the same as corporations in all respects. For example, should partnership distributions be treated as dividends, and should all the rules governing transactions between corporations and shareholders apply to partnerships as if partners were shareholders? The answer may depend on what system is used for taxing corporations (classical, imputation, or other).30

3. Defining Which Entities Are Subject to Which Regime

A threshold question in designing the income tax on business and other entities is the determination of which entities should be subject to the tax on legal persons31 and which should be subject to flow-through treatment. As previously noted, this determination does not necessarily depend on whether for other purposes the entity is a legal person. Entities that are not legal persons may still be taxed as if they were, and entities that are legal persons may receive flow-through treatment.

Even in systems that impose a single enterprise tax on business entities or on all legal persons, there are usually some situations where an exception to the general rule is made, and a business arrangement between two or more participants gives rise to income that is allocated and taxed to the participants; that is, it is given flow-through treatment. In civil law countries, such arrangements are usually provided for under the civil or commercial code.32 They do not normally give rise to a separate registration requirement and are not separate legal persons. A typical example is the arrangement commonly referred to as a joint venture,33 in which each participant (itself often a legal person) is taxed separately on its share of the venture profits.34

Depending on what tax regimes are provided, definitions must be framed to allow distinctions among different entities. For example, the law might provide for three different regimes: (1) entities taxed as corporations, (2) entities taxed on a flow-through basis with income determined at the entity level, and (3) entities or arrangements with full transparency (the distinction between (2) and (3) is explained below). How the definitions are framed may depend on the civil and commercial law. For example, it might be provided that all entities that have legal personality under the civil law are taxed as corporations, that all entities (other than legal persons) required to keep books of account under the commercial law are taxed under regime (2), and that all other entities or arrangements are taxed under regime (3).35 Whether it makes sense to frame the definition in this way depends on the civil law. Sometimes it is difficult to frame the definition in general terms and resort is had to listing types of entities.36 Whether a list is resorted to or not, the definition is most often framed in terms of the status of the entity under the civil law. Some countries, such as the United States, have adopted an independent definition for tax purposes.37 Recently, the United States has made the rule elective, so that most foreign entities that are not stock companies can elect whether to be treated as a partnership or as a corporation for U.S. income tax purposes.38 This raises the possibility of an entity being treated as a taxable person in its country of residence, but obtaining flow-through treatment for U.S. tax purposes, with consequent tax-planning opportunities that exploit the inconsistent treatment by the two countries.39

4. Partnerships as Flow-Through Entities

With some exceptions noted previously, most countries provide flow-through treatment for partnerships; that is, they do not treat partnerships as taxable entities, but rather tax partnership income only in the hands of the partners themselves according to their respective shares in that income. The remainder of this part of the chapter will assume that partnerships are treated as flow-through entities.

5. Tax Obligations Imposed on Partnerships

The fact that partnership income is flowed through to the partners does not necessarily mean that the tax system entirely ignores the existence of a partnership. In many countries, a partnership is required to file a return of partnership income, even though the tax is imposed on the partners themselves.40 It may also be appropriate, especially where most personal income is taxed at a flat or standard rate, to have the partnership pay tax at that rate on the total partnership profits;41 this operates as a form of nonfinal withholding, and the tax is paid on account of the individual partners. This type of system may be useful for taxing the share of a nonresident partner.42

B. Allocating Partnership Income to Partners

There are basically two ways of thinking about a partnership. Both imply flow through of partnership income, but the meaning of the flow through is different in each. The first view is that the partnership is an entity separate from the partners. The income of the partnership is therefore to be determined separately, and this income can then be allocated to the partners. This “entity theory” may be particularly strong in jurisdictions where the partnership has independent legal personality.

The second view, which is more consistent with the private law view of partnerships in common law and other jurisdictions where the partnership does not have legal personality, is that the partnership is simply an aggregation of the partners whereby each partner is treated as an owner of a fraction of all the assets of the partnership.43 This may be called the “aggregate” or “fractional” theory of the partnership. Under this view, the partnership does not exist independently of the partners. There is no need to determine income at the entity level. Rather, each partner is simply allocated the partner’s fractional share of partnership receipts and outgoings, and the tax consequences are determined in the hands of each individual partner. Different systems implicitly or explicitly adopt for tax purposes either the entity or the aggregate approach or, more often, a hybrid of the two.44

Systems (such as the United States) adopting a hybrid approach can end up with a particularly convoluted set of rules governing partnerships.45 The reason for this is that either of the polar approaches—entity or aggregate—is internally coherent and allows one to solve new problems through logical application of the approach to the new situation. For example, the aggregate theory holds that when a partner leaves the partnership, the partner disposes of his or her interest in the partnership assets to the other partners. It may be complicated to perform the necessary accounting but there is no conceptual difficulty involved. By contrast, under the entity theory, the partner is treated as disposing not of his or her fractional share of the partnership assets, but of the partner’s partnership interest. This leaves the cost base of the partnership assets unaffected.

While appealing from the point of view of logical coherence, strict application of either the entity or the aggregate theory may lead to undesirable consequences. A hybrid approach may be chosen to avoid these, but this loses the benefits of logical coherence and leads to a situation where instead of being able to apply a coherent theory to new situations, each new situation will require an ad hoc response, resulting in an inconsistent and complicated set of rules and little reference point when gaps must be filled in.

Whether a country adopts the entity or the aggregate approach, or some hybrid of the two, a number of general issues can be identified as to the mechanism for allocating partnership income to partners. First, there is the question of elections (including election of accounting methods) in the determination of taxable income (e.g., there may be an election as to whether to claim expensing for certain assets or what method of depreciation to use). These elections could be made at the partnership level or by individual partners. It is almost always simpler to require that elections be made at the partnership level. Second, there is the question of whether taxable income is to be determined at the partnership level. The extreme possibilities are (1) to make the determination at the partnership level and then allocate the net amount to individual partners, or (2) to make no determination at the partnership level and to allocate the component elements of the calculation (items of receipt, expense, and credit) to the partners. Third, there is the issue of how to make the allocation to partners (i.e., which partner gets which share? Can different partners get different shares of different items?). Fourth, when income or deductions are allocated to individual partners, how is their character determined? Fifth, if there is a partnership loss, can it also be allocated to individual partners or can it be used only to offset future profits of the partnership? These issues are obviously interrelated, but the number of combinations in the actual practice of countries46 and the detailed rules sometimes involved are such that a full review is beyond the scope of this chapter. The main possibilities are sketched out below.

1. Allocation According to Partnership Agreement

The first inclination is to follow the allocation of partnership income that is adopted for accounting purposes. Accounting standards will normally provide for the allocation of the income to the partners in accordance with the partnership agreement. This allocation may be directly proportionate to capital contributions or may take into account other factors, such as the amount of expertise or effort that particular partners are expected to bring to the business or the fact that they have contributed different property.47

Once the partnership income has been allocated to the partners, each partner includes his, her, or its share in total taxable income and is taxed accordingly. Thus, two partners may pay tax on their shares of partnership income at markedly different rates, such as when one partner has a substantial amount of other income and the other partner does not, or when one partner is a legal person and pays tax at the corporate income tax rate and the second partner is a physical person who pays tax at the individual income tax rate.

It should almost go without saying that, in a flow-through system, partners should be taxed on their share of partnership income regardless of whether the income has been distributed; otherwise, the tax on this income would be deferred. Care should therefore be taken in drafting any rule for the taxation of partners to refer to income “allocated” to the partner rather than to income “distributed” to the partner.

2. Deductions

It would be possible to calculate the share of partnership income attributable to each partner by allocating to the partners an appropriate share of gross receipts and expenditures (the fractional approach). In many flow-through systems, however, the net profits of the business are calculated at the partnership level and are then allocated to the individual partners (the entity approach).48 Thus, expenses incurred by the partnership for the purposes of earning income will normally have been taken into account in determining a partner’s share. For example, interest on money borrowed by the partnership for the purpose of earning income is deducted in computing the partnership profits. Where the money has been borrowed from a partner, the interest paid by the firm is the income of that partner.49

In the case of deductions that must be specifically claimed (such as depreciation or capital cost allowances), the entity approach would require that the deductions be taken at the partnership level. That is to say, the partners decide among themselves whether or not to claim the deduction in a particular year. By contrast, under the aggregate approach, each partner would separately choose whether to claim his or her share of the total allowable deduction.50 Even where the aggregate approach is preferred in general, the entity approach seems much simpler to apply in this type of situation.51 The same goes for other elections.

Sometimes partnership agreements make provision for a “salary” to be paid to a partner. One view is that the salary should not be deductible in computing the profits of the partnership, given that its true nature is that of an advance share of profits paid to the partner; that is to say, it is received by the partner as a share of partnership profits and is usually characterized as business income. This view is supported by the aggregate theory, on the basis that a partner cannot be his own employee.52 Alternatively, under the entity theory, the salary could be deducted in determining partnership profits, in the same way as a salary paid to an employee, and be included as a separate component of the partner’s total income. An analogous issue arises in the case of other transactions between the partner and the partnership, such as loans or leases of assets.53

Expenses incurred by individual partners on their own accounts do not enter into the computation of partnership profits and should be claimed by the partners themselves. For example, where a partner borrows money in order to buy a share of the partnership, the rules applicable to the deduction of interest expense by individuals will govern the deductibility of the interest.

3. Losses

An important issue is the treatment of partnership losses, in particular whether a partner may deduct a share of a partnership loss against other income for that year. A simple, though harsh, solution would be to treat the partnership in the same way as a legal entity for this purpose and to deny any deduction by the partners themselves; that is to say, a partnership loss could be carried forward (or back) only against partnership profits of other years.54 Logically, however, under a flow-through system a partnership loss should be allocated proportionately among the partners, and each partner should be entitled to claim a deduction in the same way as for any other business loss, carrying the loss forward or backward against income of other years if necessary.55 It may nevertheless be appropriate to restrict the amount of loss that may be claimed to the amount of the tax cost of the partner’s partnership interest.56

4. Taxable Year

It is customary to specify that partnership income be included in the income of the partner for the partner’s taxable year in which the partnership taxable year ends. This makes sense from a practical point of view because it is only when the partnership closes its books for its taxable year that it knows exactly how much income and expenses it had. There is no problem if everyone, including partnerships, must use the same taxable year. But if partnerships are allowed to choose their own taxable year, then they can be used as tools for deferring tax. For example, if the partnership chooses a taxable year ending on January 31, there will be an 11-month deferral of tax. For this reason, some countries have restricted the freedom to select a taxable year that differs from the taxable year of the principal partner or partners.57 However, given the complexity of such rules, the preferable approach is to require all partnerships and taxpayers to use the same taxable year.

5. Antiavoidance Rules

Partnerships between persons who do not deal at arm’s length provide obvious opportunities for tax avoidance. In particular, partnerships between spouses or between parent and child provide opportunities for income splitting. An initial question is whether such an arrangement constitutes a genuine partnership at all; a partnership may exist on paper but not in fact.58 Even where a true partnership does exist, the tax legislation may specify that the agreed-upon allocation of profits may be disregarded when the parties are related, and a reasonable allocation substituted.59

When partners deal at arm’s length, it will normally be appropriate to accept for tax purposes the allocation of profits and losses provided for in the partnership agreement.60 However, special allocations that are not based on capital or work contributed may be used as a tax avoidance device. For example, suppose that, under the income tax, charitable organizations are taxed on business income but not on investment income. A charity that owns a factory used in a manufacturing business, with respect to which it pays tax on the income, could contribute the factory to a partnership that it enters into with an investor who owns an office building. Under the partnership agreement, the rental income is allocated to the charity and the business income to the investor. The result is to convert the charity’s taxable income into nontaxable investment income.61 There are different mechanisms by which this result may be precluded. One is to stipulate that partnership allocations will be accepted for tax purposes only if they have substantial economic effect.62 In the above example, if the amount of income allocated to the charity is limited to the rental income so that the charity has no economic stake in the performance of the factory, this allocation would have economic effect and would be regarded as legitimate. However, if the agreement requires the investor to reimburse the charity, in one way or another, for deficits in expected rental income, or if the arrangement allows the charity to benefit indirectly from higher manufacturing income, then the allocation of investment income to the charity would be a formal matter only and should not be respected for income tax purposes. An alternative, more strict approach would allow the tax authorities to disregard the parties’ allocation of profits—and to substitute what they consider to be a reasonable allocation—even when the arrangements have substantial economic effect, if the principal reason for the arrangements is the reduction of tax.63

In addition to antiavoidance rules focusing on the allocation of partnership income and deductions, more general antiavoidance rules may apply to partnerships. For example, the U.S. Treasury Department has promulgated regulations that give the Internal Revenue Service a broad power to attack transactions involving partnerships. One of the rules provides that “the provisions of subchapter K [the subchapter dealing with partnerships] … must be applied in a manner that is consistent with the intent of subchapter K. . . . Accordingly, if a partnership is formed or availed of in connection with a transaction a principal purpose of which is to reduce substantially the present value of the partners’ aggregate federal tax liability in a manner that is inconsistent with the intent of subchapter K, the Commissioner can recast the transaction for federal tax purposes, as appropriate to achieve tax results that are consistent with the intent of subchapter K. . . .”64 A second rule allows the Commissioner to treat a partnership under the aggregate theory if entity treatment is being abused: “The Commissioner can treat a partnership as an aggregate of its partners in whole or in part as appropriate to carry out the purpose of any provision of the … Code …” unless a “provision of the … Code … prescribes the treatment of a partnership as an entity, in whole or in part, and … that treatment and the ultimate tax results, taking into account all the relevant facts and circumstances, are clearly contemplated by that provision.”65 These rules are of uncertain scope and have been criticized as overly broad.66 They were no doubt motivated, however, by the difficulty of designing more specific antiavoidance rules in the context of the intricacies of the provisions relating to partnerships and the ingenuity of tax lawyers and accountants engaged in manipulating those provisions. The fact that such rules were perceived to be needed may also serve as a warning against imitating the rather detailed statutory scheme for partnership taxation in the United States.

C. Flow Through of the Character of Partnership Income

1. General

The issue of special allocations of partnership income is related to the question of the character of partnership income in the hands of the partners. Almost all income tax laws classify various types of income in different ways and may have special rules and limitations depending on the character of the income. When partnership income is allocated to the partners, there are four main possibilities, corresponding to the aggregate and the entity views of partnership and points in-between. Under the pure aggregate approach, each item of income or deduction is treated as if it had been received or incurred by the partner directly. This means that in certain cases a receipt or expenditure of the partnership will be treated differently in the hands of different partners, depending on the activity of the partners (e.g., where the partner is a trader in the type of property disposed of by the partnership).67 Under the second possibility, which is a hybrid entity-aggregate approach, the character of items of income and deduction is determined at the partnership level, and each item is allocated to the partners and retains the same character in their hands as it had in the hands of the partnership. Thus, partners may receive their shares of the total partnership income as business income, dividends, interest, or rental income, as the case may be. Third, under the pure entity approach, taxable income is determined at the level of the partnership, with the net amount being allocated among the partners as a single category of income (most likely, as business income), whatever its original character. Finally, the modified entity approach allows the flow through of specific items (such as dividends or interest).

The second approach (i.e., partnership-level determination of character and flowing the character of the income and deductions through to the partners) is the rule in the United States.68 It goes hand in hand with a highly complex system under which different types of income and expense are subject to special rules and limitations. It is natural under this system to provide for flow through of the character of items of partnership income and expense, because partnerships could otherwise be used as vehicles for avoiding the various limitations.

The alternative, and simpler, entity approach is to treat all partnership income in the hands of the partners as business income, even though it may have been received by the partnership as investment income. This may be justified on the grounds that partnerships are (usually), by definition, business entities;69 consequently, even income such as dividends and rents received by a partnership may be considered as derived from the carrying on of business by the partners. However, where the tax system treats business income more favorably than investment income, the ability to convert investment income into business income by forming a partnership could open up tax avoidance possibilities.70 Alternatively, treatment as business income could be disadvantageous such as, for example, when investment income received by individuals is subject to a flat-rate withholding tax. In addition, even a system that generally treats partnership income as business income may still need to make special provision for the flow through of items (e.g., interest, dividends, capital gains, and foreign-source income) that are subject to special regimes (modified entity approach).71

2. Capital Gains

Even when a partnership is not a legal entity, it may acquire and dispose of assets in the course of its business, giving rise to the realization of a gain or loss. Because most tax systems treat capital gains differently from other types of income, the question arises as to whether and how a capital gain or loss realized by the partnership flows through to the partners and retains that character in their hands. Flow through may be done in either of two ways. One method measures the gain or loss from disposals at the partnership level, with the resulting net gain or loss being shared among the partners and included in their income while retaining its character as a capital gain or loss.72 The other method (the fractional approach), corresponding to a pure aggregate theory, treats each partner as owning a fractional interest in each of the assets of the partnership, so that gains or losses are realized directly by the partners without passing through the hands of the partnership.73 A problem with the latter approach is that, when there is a change of membership of the partnership, there will often also be a change in the fractional interests of the partners, resulting in a disposal and tax liability or in the need for complex rollover rules.

The problem is avoided with respect to business assets if gains and losses on the disposal of business assets are simply taken into account in determining the profits of the business and receive no preferential treatment. This is the situation in a number of countries, notably Germany and the Netherlands.74

3. Foreign-Source Income

A somewhat similar problem arises where a partnership receives foreign-source income. According to the entity theory, that income would simply form a part of the partnership’s total income and, in the hands of the partners, would have the character of business income with a source in the country in which the partnership was resident; that is, in most cases, the income would be converted from foreign-source to domestic-source income. One consequence would be that the partners might lose any relief in respect of taxes paid in the original source country. It is true that, when relief from double taxation is provided through the exemption method, the exemption could be taken at the partnership level. But in countries that employ a mixture of the exemption and the credit methods,75 it would be excessively complex to give relief for some foreign taxes at the partnership level and for others at the level of the individual partners. Consequently, even when there may be a general preference for the entity approach, it seems more appropriate that foreign-source income should retain its character as foreign-source income in the hands of the partners. This in turn raises two problems.

The first is the question of relief for foreign taxes, referred to above. When relief from double taxation is given through a foreign tax credit, the partner should be entitled to claim a proportionate share of the credit. That is, it is not only a share of the foreign-source income that flows through to the partner, but also a share of the foreign tax paid on that income. This procedure involves a certain amount of complexity, in that it requires calculation of the allowable amount of the credit on the tax return of each individual partner. When relief from double taxation on a particular item of foreign-source income is given through the exemption method, the income should retain its exempt character in the hands of the partner, although the amount of that income may still have to be taken into account in determining the partner’s ultimate tax liability if the exemption-with-progression method is used.

The other problem occurs when a member of the partnership is a nonresident. If foreign-source income received by the partnership retains that character in the hands of the partners, the nonresident partner should presumably be exempt from tax on the partner’s share of that income.76 If, however, the income loses its character and becomes converted into business income derived from the partnership, the nonresident partner would be taxable.

D. Disposals of Partnership Interests

A partnership interest is an asset capable of being bought, sold, or otherwise disposed of. Under the aggregate theory of partnership, when a partner disposes of his or her interest in the partnership, the partner is considered to sell a fractional share in all the partnership assets. Gain or loss on the sale of each asset would have to be computed and its character determined separately. Because of its complexity, this approach is followed in only a few countries.77

An alternative is to treat the partnership interest as a separate asset.78 Depending on the rules for taxing capital gains, a gain on the disposal of a partnership interest may or may not be taxable or a loss allowable.79 If it is, then it will be necessary to provide rules for determining the tax cost of the partnership interest.80 This amount will not necessarily be the amount originally contributed by the partner, because in the meantime the partnership may have earned income that has not been distributed. Since the partner will already have been taxed on that income, it should be added to the tax cost in order to prevent double taxation. More specifically, the tax cost should be

(i) the original cost of the partnership interest (including the partner’s share of partnership debt),
plus(ii) any additional contributions made by the partner to the partnership,
plus(iii) the partner’s total share of partnership income for the period during which he or she was a partner,
less(iv) all partnership income distributed to the partner during that period,
and(v) the partner’s share of partnership losses (if a deduction is allowed for such losses).

“Income” in the above formula, should include exempt income of the partnership, because otherwise this income would be taxed in the form of capital gain. The partner’s share of debt81 will depend on whether the partner is a general or a limited partner. Recourse debt is typically allocated to the former and nonrecourse debt to the latter.

It will also be necessary to establish rules for determining the proceeds of disposal of the interest. Although a partnership interest may be sold for a lump sum to some other person who will take the vendor’s place in the partnership (usually subject to the agreement of the other partners), it is common for partnership interests to be disposed of in return for a sum payable by installments or for a share of future profits payable over a number of years. Sometimes, it may be specified in the partnership agreement that on the death of a partner the partner’s surviving spouse will receive a share of future profits. One possibility is to treat the proceeds of disposal as an amount equal to the present value of the future payments; the payments would then be treated in the same manner as installment payments on the disposition of any other property. The disadvantage, for the continuing partners, is that the payments will presumably be regarded as capital payments for the purchase of the deceased partner’s interest and will not be deductible in computing their income from the partnership. Alternatively, the future payments may be taxed as income in the hands of the recipient, in which case the proceeds of disposal must be adjusted accordingly.82

E. Formation or Liquidation of a Partnership

Again, depending on the general rules for taxing income and capital gains, there may be a question as to whether contributions of property to a partnership or distributions in liquidation of a partnership give rise to taxable gains or allowable losses or to the recapture (or terminal loss) of depreciation allowances. Property contributed by a partner to a partnership may be property previously used by the partner in the partner’s own business, in which case any gains might be treated as business gains of the partner. Whether to defer taxation of such gains should probably be resolved in the same way as for formations of legal persons generally. It should be noted that, if a rollover is permitted, one effect may be to transfer potential tax liability for a proportion of any accrued gain to the other partners.83

Similarly, when a partnership is liquidated, its property will be disposed of, giving rise to possible capital gains or losses.84

In legal systems in which a partnership is not a legal entity, hut is merely a relationship between persons, there may be a further problem in that, whenever a partner dies or retires, or a new partner is admitted, the partnership is technically dissolved and replaced by a new one. It would be most inconvenient if every change in membership were to result in a disposal of partnership property and of the interests of all the partners; consequently, it seems advisable to specify that the new partnership should be treated as a continuation of the old one wherever there is a sufficient commonality of membership.

F. Partnership Distributions

Assuming that all partnership income is, in one way or another, taxed to the partners currently, then distributions of cash by the partnership to the partners should not be taxed. They represent either a withdrawal of capital or previously taxed income. As to distributions of property, systems differ substantially on the extent to which gain recognition is required on appreciation of the property. Nonrecognition (rollover) is provided for to varying degrees in Canada, the United Kingdom, and United States.85 On the other hand, in Germany and other countries that follow a similar conceptual approach, the distribution of partnership property to a partner is treated as a withdrawal of property from the business, which will generally be taxable unless the property is integrated into a business of the partner.86

G. Adjustment to Cost Base of Partnership Assets

Under a pure aggregate theory, a partner does not have a separate cost base in his or her partnership interest. However, most systems adopt either an entity or a hybrid view under which partners do have such a cost base, which can be referred to as “outside” cost base, the “inside” cost base being the partnership’s cost base in its assets. The inside cost base (i.e., the partnership’s total cost base in its assets) is initially equal to the total of the “outside” cost bases of all the partners, and remains so if the partnership interests do not change hands.87 Suppose, however, that the value of the partnership increases and that a partner sells his or her partnership interest to a new partner at a gain. The new partner’s cost base will now be greater than that of the old partner, thus upsetting the equality of inside and outside cost base. This can be a problem because it could cause the partners to be taxed on gains realized by the partnership for which the exiting partner has already paid tax. The remedy is conceptually simple but practically difficult. When the new partner is admitted, the cost base of the partnership assets can be increased with respect to the transferee partner to reflect the gain of the retiring partner.88 Whether to provide such rules depends on the general approach taken to taxing partnerships. If an entity approach is taken, transactions in partnership interests could be considered as unrelated to the inside cost base. Given the complexity of adjustment, an alternative would be to provide for adjustment only upon termination of a partnership. Termination could be provided for in cases where a substantial shift in partnership interests takes place over a specific period.

H. Territorial Application of Partnership Rules

In most jurisdictions, partnerships are not taxable entities and the question of the residence of a partnership does not arise directly.89 Because partnership income is flowed through to the partners, it is the determination of their residence that is important.90 As a general rule, a country will assert the right to tax a resident partner on worldwide income, which includes both domestic- and foreign-source income from both domestic and foreign partnerships;91 a nonresident is taxable only on income derived from a source in that country. A variety of situations may exist:

All members of the partnership are resident in country A. In this case, all partnership income allocable to each partner is taxable in country A.

No member of the partnership is resident in country A. In this case, the partners are taxable only in respect of partnership income sourced in country A in the same manner as nonresidents generally.

Some members of the partnership are resident in country A; others are not. In this case, the resident partners are taxable on their entire allocable shares of the partnership’s income; the nonresident partners are taxable only on the portion of their shares that is derived from a source in country A.92

These rules are simple to state, but may be difficult to apply.93 Their application depends largely on (1) whether foreign-source income retains that character when flowed through to the partners, and (2) how foreign tax credits are treated. Those questions have been considered in section C above.

In this context, it should also be noted that the taxation of different partners may differ depending on how the partner’s country of residence considers the partnership. For example, a partnership doing business in country X may be taxed by this country as a resident business entity. On the other hand, country Y, the country of residence of one of the partners, may treat the partnership on a flow-through basis. In this case, the partner should be able to take a credit in country Y for the tax paid in country X.

I. Conclusion

Several options are available for taxing partnerships. We have already discussed the option of taxing partnerships as separate entities. When this method is not used, some form of flow-through treatment must be prescribed. Given the complexity of this area, one approach for a developing or transition country would be to model its rules on those of another country with a similar legal system. A drawback of doing this is that, as discussed in this chapter, those rules may not be completely coherent, simple, or elaborated. The chief reason for this incoherence is that few countries have adopted a pure aggregate or entity approach to taxing partnerships. The aggregate approach, while coherent, is complex. It is complex from an administrative point of view because it depends on compliance by individual partners; individual compliance complicates return filing and can lead to enforcement problems that cannot be dealt with by tax administrations that are otherwise weak. The aggregate approach also requires complex calculations for distributions and transfers of partnership interests, because these are considered as involving fractional shares of all the partnership assets. However, somewhat paradoxically, the statutory rules required to implement a pure aggregate rule are not complex. All that would have to be provided is that each partner is considered to be the owner of a fractional share of the partnership assets and income according to the partnership agreement. The partnership itself would not be considered a person for purposes of the income tax. Despite the statutory simplicity, the practical difficulties preclude the adoption of the pure aggregate approach as a general rule in developing and transition countries, although it can be reserved to deal with those forms of co-ownership that are not subject to the general partnership rules.

An alternative to be considered by developing and transition countries therefore would be to adopt as pure an entity approach as possible.94 This means that income would be determined at the entity level and flowed through to the partners as business income. Limited exceptions might be made for income that receives special income tax treatment. For example, foreign-source income might be broken out separately in order to allow partners to claim the foreign tax credit with respect to such income. Interest and dividends might be flowed through separately if these are subject to special rules (such as being taxed in the hands of individuals through a low-rate final withholding tax). While partners could still manipulate such a system to some extent to minimize tax, the opportunity to do so is limited if the types of income that flow through to the partners are limited. Consideration should also be given to providing for carryover of partnership losses to be used against future income of the partnership, instead of allowing losses to be flowed through to the partners. Such a provision would minimize tax shelter opportunities and is consistent with the taxation of corporations, which also are not allowed to flow losses through to their shareholders. Adoption of an entity approach would solve a number of issues discussed in this chapter. For example, disposition of an interest in a partnership would be treated as disposition of a separate asset, not a disposition of a fraction of the partnership assets. Wages paid to a partner would be deductible by the partnership and taxable as wages to the partner. Under an entity approach, it would be clear that a partnership is a “person” for income tax purposes.95

As discussed in section A(3) above, if this modified entity approach is adopted, it will be necessary to specify which entities are subject to this rule. The form of the definition will depend on the legal forms of partnership in the country concerned. There will probably be co-ownership or joint-venture arrangements that are not legal persons, do not require commercial registration, and would not be subject to this type of entity treatment. For these, a pure aggregate approach may be most appropriate; that is, the joint-ownership arrangement is not treated as a separate person for tax purposes, and the joint owners are treated as directly earning their share of the income.

III. Trusts

A. Introduction

A trust is an arrangement, peculiar to common law systems,96 whereby legal title to property is vested in a trustee or trustees, but the income from the property (and ultimately the remaining property of the trust, known as the corpus) is or may be distributed to specific beneficiaries. A trust is created by a settlor or a grantor transferring property to the trustee to hold in trust for stipulated purposes and may be created inter vivos or on death, by will (testamentary trust).97

Trust arrangements can be very flexible.98 In the simplest case, sometimes referred to as a “bare” trust, the trust property is held for the sole use and benefit of a single individual, who may terminate the trust at any time and take possession of the property; this is in effect the same as having property held by a nominee. Almost as simple is the case wherein there is a single beneficiary, who is not immediately entitled to end the trust, being a minor or under a legal disability. Under the traditional family trust, the property might be held on trust to pay the income from the property to the settlor’s spouse, for life, and then to be divided among the surviving children. In such a case, the spouse would have a present income interest, and the children would have a future capital interest. In other, more elaborate cases, the settlor may direct that the trust income be accumulated (e.g., until a child reaches majority), or the trustee may have discretion as to which of a number of specified beneficiaries should receive the income or capital. Although trusts are most commonly used to hold income-producing property, it is possible for a trust to carry on business, and, in some countries, trusts have been used as a vehicle for family businesses.99

Common law jurisdictions will need to include provisions for the taxation of trusts in their income tax laws. Civil law jurisdictions may also provide such rules, given that trust arrangements are also being incorporated into the legal systems of some civil law countries. Developing and transition countries that are civil law jurisdictions probably do not need a detailed set of rules for the taxation of trusts except to cover some of the situations described below. However, even civil law countries whose legal systems do not provide for the existence of trusts should consider providing rules for taxing of income from foreign trusts, because a wealthy individual can easily establish such a trust in a foreign tax haven jurisdiction. Situations may also arise where a person resident in a civil law country is a beneficiary under a trust established in a common law jurisdiction, for example when a person formerly resident in country A (common law) marries and becomes resident in country B (civil law).100

B. Flow Through of Trust Income to Beneficiaries

1. General

Trusts raise a similar problem to partnerships in that it is necessary to decide whether to allocate the income of the trust to the beneficiaries for tax purposes and, if so, how. In theory, a trust could be treated as a separate taxable entity and be taxed on the entire amount of the income from the trust property without regard to amounts distributed to beneficiaries, who would presumably receive such amounts free of tax. The objection to that approach is that the rate of tax borne by the trust (whether progressive or flat) would bear no relationship to the income of the beneficiaries. The rate of tax would have to be high (probably equal to the top marginal rate for individuals); otherwise, tax avoidance would be too simple. However, a high rate would be grossly unfair if the income were distributed to a low-income beneficiary. This unfairness can be mitigated by giving the beneficiary a refundable credit for tax paid by the trust, in which case the end result would be much the same as under a flow-through system.101

A flow-through system, such as that applicable to partnerships, is an obvious alternative. However, the problem is more difficult than for partnerships in the sense that there is not necessarily an allocation of the trust’s current income to the beneficiaries. Some of the income may be accumulated by the trustee for future distribution to beneficiaries at the trustee’s discretion, so that the ultimate recipients are not currently known. Consequently, a hybrid system is usually adopted, under which a beneficiary who receives trust income is taxed on that income, while income accumulated by the trustee, to which no beneficiary is currently entitled, is taxed in the hands of the trustee. There may thus be only a partial flow through of trust income.102

This system is somewhat artificial and does not necessarily correspond to economic reality. For example, if one beneficiary holds an income interest in a trust and another holds a remainder interest, then in economic terms the holder of the remainder interest has economic income each year because the present value of the remainder interest increases, but is not taxed on that income under generally accepted rules. However, it would be difficult to design rules that more closely correspond to economic reality, and such a goal should in any event not be a matter of priority for developing or transition countries. Accordingly, the generally applied approaches to taxing trusts will be reviewed, because these serve as the most likely models.

2. Method of Taxing Beneficiaries

Beneficiaries may be taxed on their shares of trust income either directly or indirectly. According to one method, a beneficiary includes in his or her income for the year income received (or income to which he or she is entitled) from the trust and pays tax on that income in the normal manner. The trustee is taxed only on the residual undistributed income of the trust.103 If and when that income is subsequently distributed to a beneficiary, it is received tax free. Under the other system, the trustee is initially taxed on the entire income of the trust. A beneficiary who receives (or is entitled to receive) income from the trust includes that income (grossed up at the rate paid by the trust) in his or her annual return, but is given a credit for the tax already paid on that income in the hands of the trustee. In other words, the system operates as a form of withholding.104

3. Allocating Trust Income to Beneficiaries

There are again two alternatives for allocating trust income to beneficiaries: beneficiaries might be taxed only on income actually distributed to them, or they might be taxed on any income that they were entitled to receive, whether distributed to them or not, in much the same way as partners are taxed.

The first approach has the apparent advantage of simplicity, in that only actual distributions are taxed. However, it opens up the possibility of tax avoidance unless the trust rate is equal to the highest individual tax rate. A beneficiary could simply leave his or her income to accumulate in the trust, withdrawing only what is needed for immediate consumption. Consequently, most countries tax trust beneficiaries on the amounts that they are entitled to receive. For example, in the United States, allocation is on the basis of the amount of the trust’s “distributable net income” that is required to be, or is in fact, distributed to beneficiaries during the taxable year (or within 65 days thereafter, at the election of the trustee).105 Beneficiaries are taxed on the trust’s “distributable net income” to the extent of distributions they receive or are legally entitled to receive. This approach calls for taxing beneficiaries on amounts accumulated for their benefit (if they are legally entitled to receive those amounts)106 in addition to amounts actually distributed to them. The United Kingdom adopts an essentially similar approach.107

The difficulty with this approach is that it requires a determination of the entitlement of the beneficiaries under the trust instrument, an exercise that involves interpreting the trust instrument, as opposed to simply observing how much has actually been distributed. It is, however, consistent with the principle that a person should be taxed on income accruing to him or her, whether or not it is actually received.

A trust might direct the trustees to maintain the former family home for the benefit of a surviving spouse and to pay for the upkeep of the home, or to pay for the maintenance or education of a beneficiary. Normally, the value of benefits of this nature will be included in the beneficiary’s income.108

4. Flow-Through Character of Trust Income

When a trust receives different types of income that are taxed under different rules, the question arises as to whether income flowed through to a beneficiary retains its original character, for example, as a dividend, a capital gain, or foreign-source income. The problem is essentially the same as that encountered with partnerships,109 and one would expect the legislation to deal with both situations in the same way. However, this is not always the case.

In the United States, the character of distributions is determined on a pro rata basis with reference to the composition of the “distributable net income.”110 Thus, for example, a nonresident beneficiary would pay no tax on foreign- source income deemed distributed to him or her. Although an income beneficiary is normally not entitled to receive a capital gain, the proceeds of a disposal of part of the trust capital may on occasion be paid to a beneficiary (e.g., when there is a power to encroach on capital for the benefit of a beneficiary), and in such a case a capital gain may flow through to the beneficiary.111 The position is essentially similar in Australia; for example, franked dividends flowed through to a beneficiary retain that character and are consequently free of tax.112 In Canada, income received by a beneficiary, or to which a beneficiary is entitled, is generally regarded as income from property; thus, income derived by the trust from carrying on business would not be considered earned income in the hands of a beneficiary.113 However, dividend income, capital gains, and foreign-source income are expressly stated to retain their original character when distributed.114

The position is somewhat less clear in the United Kingdom. It appears that foreign-source income retains its character when paid to a nonresident beneficiary.115 However, in a case in which a trust provided for the payment to a beneficiary of an annuity of a fixed annual amount, and the trust income was insufficient to support the payment with the result that the difference was paid out of capital, the entire amount was held to be income in the hands of the annuitant; that is, the capital nature of the payment did not flow through to the beneficiary.116

When a trust is treated as a conduit, to the extent that a beneficiary is entitled to income, all types of income (or capital payments) should in principle retain their original character when flowed through.117 This is especially important in the cases of tax-exempt income, dividends (if an imputation credit applies), and income that has been subjected to a final withholding tax. It is also necessary to consider whether income from each source should be divided proportionately among the beneficiaries entitled, or whether the trustees, or trust instrument, may allocate income from different sources to different beneficiaries.118

C. Taxation of the Trust

1. Liability of the Trustee

Whether the entire income of a trust or only the undistributed part is to be taxable in the hands of the trustee, it is necessary to determine in what capacity the trustee is taxable; in particular, it is necessary to indicate whether the trust is to determine its income according to the rules that generally apply to physical persons or to those that generally apply to legal persons. Often hybrid rules may be appropriate, given that all the rules for physical or legal persons, as the case may be, may not be appropriate for trusts.

The usual practice is to tax the trustee (or trustees, jointly) as a separate physical person.119 This will be the case even if the trustee is a legal person such as a bank or trust company. Thus, the trustee is taxed entirely separately on (1) income accruing to the trustee in the trustee’s personal capacity and (2) trust income in respect of which the trustee is taxable. Normally, the trustee is required to file a return of trust income even though no tax may be payable.120

2. Income on Which Tax Is Payable

As previously noted, there are basically two systems for taxing trust income. In one (Ireland and the United Kingdom), the trustee is taxed on the entire income of the trust and the beneficiary is entitled to a credit for the tax so paid. In the other, the trustee is liable for tax only on income retained in the trust. This is achieved by allowing the trustee to claim a deduction in respect of income distributed, or required to be distributed, to a beneficiary.121

Although the trustee is not generally permitted to claim personal deductions,122 the usual deductions are normally allowed for expenses incurred in earning trust income—for example, repairs to rental properties or interest on borrowed funds.123

3. Rate of Tax

A basic problem with the income taxation of trusts is the rate of tax to be charged. Although trusts are normally treated as separate taxpayers and as physical persons, the application of a graduated rate schedule is inappropriate, because a trust may have a number of beneficiaries (with widely different incomes), and the amount of undistributed income may bear no relationship to the incomes of those beneficiaries.

Trusts provide a variety of opportunities for minimizing taxation, depending very much on the rate or rates at which undistributed income is taxed. If the trust rate is lower than that at which a beneficiary would be taxed, it will be advantageous to accumulate income in the trust, thereby splitting income between trust and beneficiary.124 If the trust rate is lower than that at which the same income would be taxed to the settlor, there will again be advantages in transferring property to a trust over which the settlor retains some control. In addition, because there are no limits on the number of trusts that a person may create, it becomes advantageous to create multiple trusts if trusts are taxed at progressive rates.

Problems with the use of trusts for tax avoidance can be minimized by specifying that all trust income that is not flowed through to beneficiaries should be taxed at a flat rate equal to the top marginal rate applicable to physical persons.125 That approach is probably satisfactory if the rate is a moderate one; if it is very high, then it may not be acceptable because it will tax at a high rate income that may be destined for a beneficiary in a much lower rate bracket. Even if rates are moderate, the proposal can be criticized on the basis that it will be unfair in some cases. Inevitably, there will be some trusts accumulating income for the benefit of beneficiaries in low brackets. Some such unfairness is inevitable and is the price of simplicity. The simplicity resulting from such a rule is considerable: there will be no need for multiple trust rules or special rules governing delayed distributions from trusts. The unfairness will be minimal in a country where low-bracket beneficiaries of trusts are likely to be rare.126

A suggested general rule, therefore, would be that all accumulated income of a trust be taxed at the top marginal rate for physical persons. Distributed income would be taxed to the individual beneficiaries. However, if certain kinds of investment income are subject to a final flat rate of tax, then it would be unfair to tax that income at the top marginal rate in the hands of a trust where it will ultimately be distributed to beneficiaries who are physical persons. Therefore, the trustee should be allowed to exclude such income as if the trust were a physical person. To prevent abuse, it may be necessary to restrict this rule to cases where the only beneficiaries of the undistributed income are physical persons, as is the case with most trusts. Trusts with corporate beneficiaries do exist, and they should not benefit from a flat withholding tax on investment income if corporations are taxed on such income at the same rate that applies to other corporate income; nor should they be taxed on dividends received through a trust if intercorporate dividends paid directly would be exempt from tax.

An exception to the above rule might also be justified where the trust has only one beneficiary, or where the trustee (or some other person) has the power to vest the corpus or income of the trust in herself or himself.127 The reason for this rule is that if the trust income is being accumulated for the benefit of a single beneficiary, it makes more sense to tax that income at the possibly lower marginal rate of the beneficiary than at the top marginal rate that would apply to the trust.

In practice, few of the countries that have well-elaborated rules for taxing trusts do impose tax at the top individual rate.128 In the United States, for example, residual trust income is taxed according to a graduated-rate scale, although the rate scale was compressed by the Tax Reform Act of 1986.129 In the United Kingdom, where the trustee is taxed on the total income of the trust at the “standard rate,” an additional tax is imposed on accumulated income, which reduces—but does not entirely eliminate—the opportunities for tax avoidance.130 As a consequence, virtually all of the countries in which trusts are common have found it necessary to enact antiavoidance rules of varying complexity.

D. Antiavoidance Legislation

1. Grantor Trusts

In the case of some trusts, it will be appropriate to ignore the existence of the trust for income tax purposes, that is, to treat it as ineffective and to tax its income to the original settlor or grantor. A trust is generally treated as ineffective when the grantor has retained control over the trust or has retained benefits from the trust.

The United States has a rather elaborate and hypertechnical set of rules governing the circumstances under which a trust will be treated as a “grantor trust.” These rules were formulated at a time when a substantial tax benefit could be obtained by creating a trust (by taking advantage of the separate taxation of each trust under a progressive rate schedule). They therefore contain a number of safeguard provisions; ironically, they also contain a number of loopholes through which careful estate planners are able to structure arrangements so as to avoid grantor trust treatment. If trust income were taxed at the top marginal rate, as previously suggested, then the definition of grantor trust could be simplified because it would be less critical to catch all possible situations in which grantor trust treatment might be justified, given that the tax benefits from setting up a trust would be minimized.

Under the U.S. rules, the grantor is treated as the owner of a trust in which the grantor has a reversionary interest if, as of the inception of the trust, the value of the interest exceeds 5 percent of the value of the trust.131 The grantor is also treated as the owner of a trust whose beneficial enjoyment is subject to a power of disposition exercisable by the grantor or a nonadverse party without the approval or consent of an adverse party.132 “An adverse party” is a person with a beneficial interest in the trust who would be adversely affected by the exercise of the power that the other adverse party possesses.133 However, a number of exceptions are provided for certain powers that the grantor may hold without running afoul of this rule. These include

  • the power to apply income to the support of a dependent, as long as the income is not actually so applied;

  • a power the exercise of which can only affect the beneficial enjoyment of the income after the occurrence of an event that is sufficiently remote;

  • a power exercisable only by will, with limited exceptions;

  • a power to allocate among charitable beneficiaries;

  • a power to distribute corpus that is limited by a reasonably definite standard and certain other powers to distribute corpus;

  • certain powers to withhold income temporarily;

  • a power to withhold income during legal disability or minority of a beneficiary;

  • a power to allocate receipts and disbursements between corpus and income;

  • certain powers exercisable by independent trustees; and

  • a power to distribute, apportion, or accumulate income to or for a beneficiary or beneficiaries, exercisable by trustees who are not the grantor or grantor’s spouse, if the power is limited by a reasonably definite external standard.134

The grantor is also treated as owner of a trust over which the grantor has certain administrative powers, including

  • a power to deal with the trust for less than adequate consideration; and

  • a power to borrow from the trust without adequate interest or security.135

The grantor is also treated as owner of a trust when

  • the grantor has borrowed from the trust and has not completely repaid the loan, unless the loan provides for adequate interest and security and is made by a trustee other than the grantor or a related party;

  • a power of administration is exercisable in a nonfiduciary capacity by any person;136

  • the grantor has the power to revoke the trust, or a nonadverse party has the power to revest title to the property of the trust in the grantor;137 or

  • the income of the trust is—or, at the discretion of the grantor or a nonadverse party, may be—distributed or accumulated for the grantor or the grantor’s spouse without the consent of any adverse party.138

While the above set of rules can appear daunting (and note that the description is only a simplified summary), it is necessary to have some guidance for when a trust will be treated as a grantor trust. As noted, under a regime that taxes accumulated trust income at the top marginal rate, a simpler set of grantor trust rules can be envisaged. For example, the following set of grantor trust rules for domestic trusts was proposed by the U.S. Treasury Department in 1985:

The grantor would be treated as the owner of a trust to the extent that (1) payments of property or income are required to be made currently to the grantor or the grantor’s spouse; (2) payments of property or income may be made currently to the grantor or the grantor’s spouse under a discretionary power held in whole or in part by either one of them; (3) the grantor or the grantor’s spouse has any power to amend or revoke the trust and cause distributions of property to be made to either one of them; (4) the grantor or the grantor’s spouse has any power to cause the trustee to lend trust income or corpus to either of them; or (5) the grantor or the grantor’s spouse has borrowed trust income or corpus and has not completely repaid the loan or any interest thereon before the beginning of the taxable year. For purposes of these rules, the fact that a power held by the grantor or the grantor’s spouse could be exercised only with the consent of another person or persons would be irrelevant, regardless of whether such person or persons would be characterized as “adverse parties” under existing law.139

Although the U.S. rules on grantor trusts are considerably more complex than those found in most jurisdictions, more limited rules to similar effect are found in the laws of other countries. For example, where a trust may be revoked, it is commonly provided that the income from the trust is attributed back to the settlor or grantor.140 Other provisions are found that attribute the trust income back to the settlor if the income is paid or payable to the settlor’s spouse or minor children.141 Provisions of this kind may be found in that part of the legislation that deals with trusts or may be contained in attribution rules of general application. For example, in Canada income and capital gains may be attributed to an individual who “has transferred or lent property … either directly or indirectly, by means of a trust or by any other means whatever …” to or for the benefit of a spouse or a minor who is a relative.142

2. Multiple Trusts

Because there is no limit on the number of trusts that a person may create, there developed in some countries the phenomenon of multiple trusts, whereby property was split among a number of identical or substantially similar trusts so as to take advantage of progressive rate schedules applied to each trust separately. (No advantage will be obtained, of course, if all trusts are taxed at the top marginal tax rate applicable to individuals.)

In Canada,143 New Zealand,144 and the United States,145 the legislative response was to provide rules for aggregating multiple trusts in certain circumstances and to narrow the rate brackets, limiting the amount of income taxed at lower rates.

E. Disposals of Trust Property and Trust Interests

1. Trust Property

According to usual tax principles, a capital gain or loss may occur (1) when property is transferred to a trust, (2) when the trust itself disposes of property, and (3) when the trust is liquidated.

In the first situation, the principal issue is whether the transferor (e.g., the grantor or settlor) incurs tax liability or can claim an allowable loss.146 In case (2), assuming that a taxable gain or allowable loss is realized, the question is whether the gain (or loss) accrues to the trust or to the beneficiary. In most circumstances, the benefit of a gain accrues to the ultimate capital beneficiaries and the gain is consequently taxed in the hands of the trust. However, when the trustee encroaches on capital for the benefit of an income beneficiary or makes an advance of capital to a capital beneficiary, it is usually permitted to flow the gain through to that beneficiary, and the gain preserves its character when taxed in the hands of the beneficiary.147 A capital loss, by contrast, should not flow through because it cannot be distributed. In case (3), if the trust property is sold on liquidation of the trust, the position should be as in (2), except that both gains and losses should flow through to the beneficiaries who receive the proceeds of sale. If, instead, trust property is distributed in specie to a beneficiary, it may be appropriate to provide for a rollover.148

2. Trust Interests

An interest in a trust is property that may be alienated. In some cases (e.g., the prospective share of a potential beneficiary under a discretionary trust), it may be difficult to determine the market value, and thus the cost base, of the interest. However, a vested life interest or residuary capital interest can be valued with a reasonable degree of accuracy. For example, if property worth $1 million is settled in trust for person X for life, remainder to person Y, the value (and cost base) of person X’s life interest will depend on his or her life expectancy and on the anticipated future earnings from the property. Given that, at the time of the settlement, the combined values of X’s and Y’s interests must add up to $1 million, the value of Y’s interest is also revealed. If X or Y subsequently disposes of an interest, a gain or loss may accrue. However, the calculation of this gain or loss is complicated by the fact that the change in value of the individual’s interests will be affected by two factors: (1) any change in the value of the underlying trust property, and (2) the fact that X’s life interest decreases in value over time (as does his or her life expectancy), and the value of Y’s interest increases correspondingly. While it may be legitimate to tax gains attributable to (1),149 gains or losses attributable to (2) should probably be ignored, because the reduction in the value of X’s life interest is offset by the increase in Y’s capital interest, and because the calculation would become impossible in more complicated cases involving the trustee’s discretion.

F. International Aspects of the Taxation of Trusts

1. General

According to general principles, a country would normally claim the right to tax resident trusts and resident150 beneficiaries of both resident and foreign trusts on their worldwide income. Nonresident individuals are taxed only on income sourced in the country, and, because trusts are normally taxed as individuals, the same rule should apply to nonresident trusts.151 In practice, a nonresident trust is likely to be taxed only through the withholding of tax on its investment income.152

2. Residence of Trusts

Determining the residence of a trust is obviously important, because residence renders the trust liable to tax on foreign-source income. However, that determination may be a difficult matter because a trust is not a legal person and is not required to register in order to be recognized. Various factors may be taken into account, including

  • the residence of the trustee;

  • the place of management or administration of the trust;

  • the location of the trust assets;

  • the residence of the beneficiaries; and

  • the residence of the grantor or settlor.

Generally, the first two factors will be the most important,153 but it is possible that none of them will be determinative. A trust might have three trustees, each resident in a different country; meetings of the trustees might be held in various locations, as might the trust assets; there might be a large number of beneficiaries, resident in various countries; and the settlor might well be dead. For these reasons, a number of countries have considered it necessary to adopt special rules to prevent tax avoidance through the use of nonresident trusts.154

3. Foreign-Source Income

When a resident trust receives foreign-source income, the question arises as to whether the income retains that character when distributed to a beneficiary. For example, investment income from a source in country A, received by a trust resident in country B, and paid to a beneficiary resident in country C might be regarded as sourced in country A (investment income) or in country B (trust income). In the latter case, it will be taxable in country B; in the former, it will not.155 In the former case, a further question arises as to whether the trustees, or the trust instrument itself, may allocate foreign-source income to nonresident beneficiaries in order to avoid or reduce tax liability. If the foreign-source character is flowed through to a resident beneficiary, then that beneficiary should also be entitled to claim a credit for foreign tax paid.156

4. Nonresident Beneficiaries

Apart from the flow-through question discussed in the preceding paragraph, the main concern will be to ensure that tax is paid on trust income distributed to a nonresident beneficiary. When the trustee is taxable on the entire income of the trust, as in Ireland and the United Kingdom, this presents no problem; in those countries in which the trustee is taxed only on the undistributed income of the trust, a nonresident beneficiary’s share can be taxed by requiring the trustee to withhold tax.157

5. Nonresident Trusts

Foreign trusts (i.e., trusts wherein the trustee is a nonresident) pose a problem because the trustee is beyond the country’s taxing jurisdiction. This means that a foreign trust (like a foreign company) can be used to defer a country’s tax on foreign-source income even though residents of the country are beneficiaries of that income.158 Provided that foreign-source income is accumulated in the trust, tax is deferred until the accumulated income is either distributed to a resident beneficiary or realized as a capital gain on disposal of the interest in the trust.

Two types of foreign trust may be used to defer tax on foreign-source income. The first is a trust structured as a “roll-up fund,” in which beneficiaries purchase an interest (such as units in a unit trust) of a type that carries an entitlement only to capital. A beneficiary can realize his or her interest in the trust either by selling it or by having it redeemed by the trustee. In either case, the beneficiary effectively realizes the income of the trust as a capital gain and, therefore, obtains the benefit of both deferral and the conversion of income into capital gains (which may be concessionally taxed). Because these trusts are structured in essentially the same way as companies, some countries subject such trusts to the same antideferral rules that apply to companies.159 An alternative approach, adopted in the United Kingdom, is to discourage investment in such trusts by taxing the gain on disposal of the interest in the trust as income rather than as a capital gain.160

The second type of trust that may be used to defer tax on foreign-source income is a nonresident discretionary trust. The elimination of deferral for this type of trust poses particular difficulties for tax designers because, in the tax year in which the trust derives the income, it may not be known with any certainty which beneficiaries will ultimately benefit from the income. In other words, there are difficulties in identifying a taxpayer who may be subject to current taxation in respect of foreign income accumulated in such a trust. An initial line of attack against taxpayers transferring property out of the jurisdiction is to tax the transferor on the gain on any appreciated property transferred to a foreign trust. For example, in the United States, a 35 percent tax is imposed on the unrealized appreciation of property that is transferred by a citizen or resident of the United States to a foreign corporation, partnership, estate, or trust.161 In Canada, any transfer of property to a trust, other than to a resident “spousal trust,”162 constitutes a disposal for capital gains purposes. These rules, however, will not act as a deterrent to transferring property to a foreign trust when the property has not appreciated in value, nor will they discourage transfers on death in countries that do not treat death as a taxable event.

Another technique that in effect keeps the trust within the taxing jurisdiction is to treat it as a grantor trust. This means that the grantor will be taxed on the trust’s income. This requires applying more expansive grantor trust rules to foreign trusts than to domestic trusts in cases when the grantor is a resident taxpayer. For example, a U.S. citizen or resident who transfers property to a foreign trust is treated as the owner of the portion of the trust attributable to the property, unless no part of the income or corpus of the trust may be paid or accumulated to a U.S. person.163 For purposes of this rule, a foreign corporation, partnership, trust, or estate is considered a U.S. person if, in the case of a corporation, more than 50 percent of the stock is owned or is considered as owned by a U.S. person; if, in the case of a partnership, a U.S. person is a partner; and if, in the case of an estate or trust, a U.S. person is a beneficiary.

The U.S. grantor trust rule for foreign trusts is a broad one, but it does not cover trusts when the grantor has died or is not a U.S. citizen or resident. In these cases, special rules for foreign trusts may be needed to deal with the problem of tax deferral in cases where the trust is located in a tax haven jurisdiction. While it is possible to tax beneficiaries on their share of distributed income of the foreign trust, they cannot be taxed on trust income accumulated for the benefit of presently unknown beneficiaries. One solution is to impose at the time of a distribution to a resident beneficiary an extra tax, determined by applying an interest rate to the difference between the foreign income tax paid by the trust and the marginal rate that would have applied domestically.164 Simply taxing the beneficiaries on distributions from the trust as received would not do. The distributions may represent corpus, which should not be taxed at all, or they may represent income that was taxed at a very low rate abroad, so that even full taxation at distribution would confer a substantial tax benefit.

An alternative approach, adopted in Canada, is to simply deem the trust to be a resident and make the trustees and resident beneficiaries (including discretionary beneficiaries) jointly liable for tax on its income.165 The rules are complex, can have harsh consequences, and appear to be designed less to ensure that a fair tax burden is imposed on nonresident trusts than to deter the creation of such trusts altogether, it being assumed that the most likely motive for their creation is tax avoidance.

Australia, New Zealand, and the United Kingdom have also introduced grantor trust regimes applicable to nonresident trusts. The Australian and U.K. regimes are broadly similar to the U.S. regime described above.166 The design of the New Zealand regime is different, although the practical effect is the same. When a New Zealand resident has transferred value to a nonresident trust, the trustee of the trust is liable to New Zealand tax on the foreign-source income of the trust.167 If the trustee is not a resident, then the trustee is liable for tax as if he or she were a resident. In recognition of the difficulty of enforcing this liability against a nonresident trustee, it is provided that a resident person who has transferred value to the trust is liable to tax as agent of the trustee.168

IV. Other Flow-Through Entities

Partnerships and trusts are by far the most common of the entities that are given flow-through treatment, but various other types of business and investment entities that may be taxed in that manner merit a brief mention.169 A distinction may also be drawn between those entities that are automatically taxed on a flow-through basis and those cases where the flow through is optional and is permitted on an elective basis.

A. Automatic Flow-Through Treatment

1. Joint Ventures

The term “joint venture” can be a confusing one, because it can cover a variety of legal forms. A distinction is commonly made between “equity joint ventures,” in which the parties incorporate a separate joint subsidiary corporation, and “contractual joint ventures,” which are closer in nature to partnerships and are generally taxed on a flow-through basis. For example, when two companies establish a joint venture for a particular purpose, the normal practice is to tax each of the companies upon its share of the profits from the venture rather than to tax the venture as a separate entity.170

A special form of joint venture—the European Economic Interest Grouping (EEIG)—was introduced in the member states in 1985.171 The EEIG is formed by contract and established by registration, which confers on it legal personality. It is intended as a means of cooperation between individuals or entities that otherwise wish to maintain their independence. The profits from the grouping’s activities are treated as the profits of the members themselves and are taxed only in their hands.172

2. Other Entities Given Flow-Through Treatment

It is frequently considered appropriate to tax various other types of business or investment entity on a flow-through basis. For example, in Spain, the impuesto sobre sociedades (corporate income tax) generally applies to all legal entities, but an exception is made for unquoted portfolio investment entities and for certain family-owned investment-holding companies.173 In the United States, flow-through treatment is given to investment vehicles such as real estate investment trusts and real estate mortgage investment conduits.174

In civil law countries, trust agreements, whereby assets are entrusted to a trustee for the carrying on of business, are typically taxed on a flow-through basis.175 These are similar to joint ventures and do not involve the complexities of taxing common law trusts because the shares of the beneficiaries are specified.

Certain types of companies are also taxed on a flow-through basis. Sometimes this is done in recognition of their essentially personal nature, as in the case of the one-person company (entreprise unipersonelle á responsabilité limitée) in France, or their resemblance to a partnership, as with the limited-liability company in the United States.176 Flow-through taxation of companies may also be adopted as a method of counteracting tax avoidance, for example, when the income of a controlled foreign corporation is allocated among resident shareholders and taxed in their hands whether or not dividends are paid.177 This treatment may also be appropriate to prevent undue tax deferral through the use of personal holding companies, where the standard corporate tax rate is substantially lower than the top rate of individual income tax.178

B. Elective Flow-Through Treatment

Depending on tax rates and the system of taxing corporations, incorporation either may confer a tax advantage (as noted in the preceding paragraph) or may result in a heavier tax burden. The latter is particularly likely to occur when the “classical” system is adopted. Relief from economic double taxation may be given by permitting certain corporations to elect to be taxed on a flow-through basis. For fairly obvious reasons, this solution is appropriate only in the case of relatively small corporations. A well-known example is the U.S. “S Corporation” rules, under which a corporation that has 35 or fewer shareholders, all of whom are individuals resident in or citizens of the United States, may elect to be taxed as a flow-through entity.179 The flow-through taxation of S corporations is simpler than that of partnerships, in part because S corporations are allowed to have only one class of stock. Thus, allocation of corporate income among the shareholders is straightforward because it can be allocated in proportion to share ownership. Elective flow-through treatment may also be granted to other types of entity that are otherwise normally taxed as legal entities.180

In designing appropriate election rules, policymakers should consider the nature of the entity and the rights of its participants. It may be unfair, for example, to allow the directors of a closely held corporation to elect to have its income taxed in the hands of its shareholders when, as a consequence, a minority shareholder might find himself paying tax on income that he might never receive. In such circumstances, it might be more appropriate to require the election to be made unanimously or by a special majority of the shareholders. Consequently, elective flow-through treatment is normally only appropriate for small businesses or for associations with relatively few participants.

V. Conclusion

Whether and how particular legal entities or arrangements are given flow-through treatment will depend on the precise nature of such entities under the civil and commercial law of a country, as well as on basic income tax policy considerations, such as the desire to simplify individual income taxation. The legal forms differ substantially, particularly between common law and civil law countries. Therefore, one can expect substantial differences in the tax rules from country to country, and a single uniform solution cannot be prescribed. Nevertheless, as this chapter shows, it is possible to identify common approaches that provide guidelines for developing and transition countries, even though the details of the solutions adopted will not be uniform.

    Other Resources Citing This Publication