It is also obvious that the type of rules which we have been discussing, although they are unquestionably rules of binding law, have in no way the character of religious commandments, laid down absolutely, obeyed rigidly and integrally.… The bundles of fish, the measures of yams, or bunches of taro, can only be roughly assessed, and naturally the quantities exchanged vary according to whether the fishing season or the harvest is more abundant.
—Branislaw Malinowski
I. Introduction
A. In General
Two of the perennial issues in tax policy debates are whether a specific tax should be formally imposed on enterprise profits and collected from enterprise earnings, and, if so, how it should be constructed. Levying a separate tax on the earnings of large corporations is almost universal practice,1 often existing at both national and subnational levels, and in fact predates the imposition of universal income taxes on individuals in some jurisdictions.2 Yet, the tax is not without its detractors, and suggestions for its reform or even repeal are often heard.
Given the prevalence of a tax on enterprise profits, it may seem curious that there is such a debate, and so this introduction will outline some of the main points of contention before their more detailed examination in the body of the chapter. At the policy level, critics have pointed to the perceived deficiencies of the tax, and their list is indeed long and daunting—the so-called economic double taxation of enterprise income (when enterprise profits are taxed, as are distributions of those taxed profits), the conjectural incidence of the tax (because its effects might be shifted to shareholders, workers, or consumers), the indeterminate and discretionary amount of tax payable (because the amount of tax varies with such factors as the capital structure of the enterprise and the timing and proportion of distributions and retentions), the apparent incentive for business enterprises to finance their activities through debt (because interest is usually deductible while dividends are not) and to retain earnings (because retained earnings are not taxed as dividends but are usually taxed at a lower effective rate as capital gains), and the possible tax induced distortions of the way economic activities are organized and conducted (avoiding a particular legal form for less transparent but more lightly taxed alternatives).
Despite the shortcomings of the tax—the significance of which is still the subject of much debate—its supporters have pointed to several important benefits. They argue that it could approximate the economist’s ideal tax on pure “economic rents”—that is, a tax on the excess of revenue over the enterprise’s total input cost, including the cost of capital. Such a tax would have no distortionary effects because it taxes pure profit. Even detractors of the tax acknowledge that there are serious obstacles to removing it, not the least of which is the substantial windfall that would be conferred on holders of equity interests if the tax were removed. Other important obstacles to its removal arise from interdependencies: changing the enterprise tax would also require changing the personal income tax and international tax systems. First, under personal income taxes as they typically currently operate, the enterprise tax is the principal means of preventing deferral of tax and arbitrage of ordinary income into preferentially taxed capital gains. In other words, the enterprise tax is necessary to protect the tax base of the personal income tax. Next, it is an important source of revenue from nonresidents under the existing allocation of taxing rights between countries. The current system allocates rights to tax capital income from equity investments primarily to the source country, and rights to tax capital income from debt investments primarily to the residence country. Unilaterally abandoning the right of the source country by repealing the enterprise tax would confer an unrequited windfall on the residence country.
The other aspect to the debate is more operational—if the tax is to exist, how should it be constructed? At one end of the spectrum, the “separate entity view” would construct and operate the personal and enterprise taxes independently. The other, the “conduit” view discussed below, would adjust the taxes to recognize the existence and operation of both. At present, and on balance, the conduit view has probably emerged in a majority of countries as the most satisfactory theoretical paradigm for imposing tax on income derived from an equity investment in a business enterprise.3 Ultimately, if the conduit view is accepted, its implementation must lead to one of the three theoretical options—imposing tax at the investor level only, imposing tax at the enterprise level only, or imposing tax at both levels, with the corollary of adjusting one tax for the effects of the other.
This chapter considers these policy and operational questions. Section II examines imposing tax at the investor level only and discusses why a tax imposed at the enterprise level, rather than at the shareholder level only, has generally been regarded as necessary. Section III considers in more detail whether important benefits can be achieved by a tax imposed at the enterprise level only. Section IV examines some possible adverse consequences likely to flow from the decision to use an enterprise-level tax that makers of tax policy and tax officials must be aware of and, more important, be prepared to manage. Section V examines systems that have both enterprise-level and investor-level taxes, and some of the options for structuring the interaction between them. Section VI discusses the typology and effect of distributions in their interactions with different enterprise and investor systems. Section VII examines how to define the enterprise taxpayer. Section VIII draws some general conclusions.
B. Relationship Between Enterprise Income and Investor Income
Most countries permit various legal structures for organizing profit making enterprises. These include sole proprietorships, different types of partnerships, companies, and trusts. Modern financial theory views each form of enterprise as a group of investors acting collectively under one or more legal structures. Those structures are all based on contract law and include partnership law, trust or foundation law, and company law.4 The traditional forms of organizing an investment include equity, debt, and leases5 of movable, immovable, and intellectual property. Each different form of investment, whether stock or partnership interest, bank loan or bond, or lease, creates for the investor a different type of claim to the income and property of the joint business enterprise.6
Traditionally, equity holders receive their income in the form of a mixture of periodic payments (partnership distributions or dividends) and increases, or perhaps decreases, in the value of their investment (capital gains or losses). Depending on the local law, equity investments in enterprises can take other traditional forms, such as preferred stock.7 The bondholder or banker is typically entitled both to a fixed rate of return on his or her loan and to repayment of the original amount invested; he or she may sue the enterprise if these amounts are not paid.8 A typical creditor generally has no direct say in managing the enterprise.9 However, a creditor may also experience increases or decreases in the value of his or her investment depending on changes in the creditworthiness of the enterprise, as well as on changes in interest rates or in rates of inflation.10
A lessor is in a legal position similar but superior to that of a creditor whose note is secured by assets. He or she may also see the value of the investment vary on the basis of the value of the security or leased good, as well as the general creditworthiness of the enterprise.11 A lease can shift the risk of loss from the owner to the person who leases the asset, while the owner may retain the opportunity of increase in value.
Different legitimate market-based reasons exist for packaging investments in different economic forms.12 Risk is among the most important. As risk increases, investors will demand compensation for assuming that risk. Unsecured loans are riskier than secured loans or leases; partnership interests and common stock may have the greatest risk of all, but also the greatest opportunity for gain.
Changes in the value of the interest of any particular investor should be equal to the investor’s share of the change in the total value of the enterprise. In other words, the value of an investor’s interest will equal the total change in value of the enterprise minus everyone else’s share. However, to make this calculation, one must first determine the income or loss of the enterprise. Forms of investment that have been traditionally referred to as “debt” or “leases” periodically pay or accrue interest, rent, or royalties to the investor. Therefore, the change in the value of the taxpayer’s investment in the enterprise can be determined largely from the amount of interest, rent, or royalties that has been paid or accrued to him or her. Debt that is accrued over time, but that is not currently payable, can be recalculated so that the “reinvested” portion of the unpaid interest is included.13 However, as noted above, the value of the debt or lease investment may not be completely reflected in the stated interest or rent. Even in simple debt relationships, changes in the creditworthiness of the enterprise, or changes in interest rates, will affect the value of the underlying indebtedness.14
The legal structures of investments have become increasingly varied and complex and have mixed many of the traditional attributes of equity, debt, and lease.15 Examples include debt with call options or contingent interest, shared appreciation mortgages, and notional principal contracts. Instruments that allocate risk in different ways are constantly being created.16 In the more advanced economies, these instruments have a long history. However, even in developing markets, there has been a proliferation of forms of financial instruments representing different types of investments in for-profit enterprise. The internationalization of finance and financial advice has resulted in the prompt spreading of those diverse investment forms throughout the world, or at least to those jurisdictions whose legal structure can accommodate them. As investments become more complicated, the difference between the stated “current yield” and the actual net income of the investor can become quite great.17
The various types of equity investors are entitled to the income of the enterprise minus the amounts paid or accrued to creditors and lessors. Because of the occasionally bewildering forms of equity participation, exactly how this income is to be divided may be completely clear only to the lawyers who draft the forms of participation. Nevertheless, the earnings of the equity participants are the principal object of the taxation of enterprises.
II. Enterprise Income Taxation as a Withholding Tax on Investors
In principle, it would be possible to tax all income of a business enterprise directly to its equity investors by treating all enterprises as “flow-through” entities and allocating income to investors on a yearly basis.18 However, treating the enterprise as a separate taxable entity has a number of advantages over flow-through treatment. One problem with flow-through treatment is that it may be difficult to allocate earnings among a large number of increasingly bewildering types of equity holdings. Another is that, as the number of equity investors increases, allocation becomes more difficult.19 As a result, no country has implemented such an approach.
Some or all of the enterprise earnings may be paid out to the equity investors, often at the option of enterprise managers or the investors themselves.20 Taxing such distributions does not pose any great difficulty; they can simply be added to the income tax base of the shareholder.21 A problem arises with the earnings retained by the enterprise. Most of the value of retained earnings is expressed primarily as increases in the value of the interests of those equity investors who have the legal right to the earnings retained by the enterprise.22 It would, in theory at least, be possible to tax the equity investor on the change in the value of the equity participation.23 The tax base of such a system would include not only the amount of retained taxable earnings of the enterprise, but total economic income as well, including earnings not typically included in the income tax base, such as unrealized gains in the value of assets.24 It would also include changes in the value of the equity interest that are not related to the economic income of the enterprise, for example, a systemic shift in stock market prices if the interest is a traded share.25 Such a change in value would have to be assessed as part of the tax base annually; if not, the taxpayer would benefit from the time value of money on the deferred taxes.26
A number of suggestions have been made in favor of such accrual taxation of gains (and losses) on ownership interests in business enterprises. Yet, while such systems might be practicable for equity interests that are regularly traded with enough liquidity to determine a price, they would be difficult indeed for other interests.27 For other equity interests, it might be possible to make periodic valuations and to adjust them for errors once the interest is actually traded, or whenever a fair market value can be ascertained with certainty.28 However, a system of periodic valuation would ignore the possibility that the value of the interest can shift, perhaps even wildly, during the holding period.29 This can create difficult problems for tax administration.30
Even if one could accurately determine changes in the value of equity interests on an annual basis, including these changes in the tax base would still be problematic. First, and perhaps most important, is the problem of collection. Taxes on capital gains are among the more difficult to enforce and collect. Except where there is direct reporting to the tax administration from exchanges or from broker-dealers, each individual taxpayer must voluntarily disclose the amount of gain or loss. The taxpayer must then remit the correct amount of tax. Even in the United States, which has a relatively effective tax administration, taxpayer compliance in reporting capital gains on traded securities and remitting the required tax is low relative to other areas. The most likely reasons are the lack of withholding for such tax and the absence of accurate, easily usable information reporting.31 And, because there may be a plethora of equity investors, the problem of administration is magnified. Obviously, the less sophisticated the tax administration, or the less likely taxpayers are to report income and pay tax voluntarily, the worse this problem becomes.
A second problem would result from any difference in tax treatment between ownership interests in business enterprise and business income earned directly or through flow-through entities. As noted earlier, accrual taxation of ownership interests would include not only what is commonly accounted for as taxable income, but also accrual taxation of unrealized gains in the value of assets held by the enterprise.32 While proposals have also been made for accrual taxation of all assets used in the course of business,33 no tax jurisdiction has adopted the rule. If the effective tax rate on the income of business enterprise were substantially different depending on the form of the enterprise (flow-through or non-flow-through), there would be a tax-induced preference to operate in the form that produced the lower effective tax rate. For equity, efficiency, and administrative reasons, such tax incentives should usually be avoided.
The problems inherent in taxing equity interests on an accrual basis can be avoided by levying a tax on the income of the business enterprise as a surrogate for the tax that equity participants would pay if all enterprise income were distributed. In that way, the enterprise would not be able to defer tax simply by not paying dividends. If the enterprise tax were levied at the same effective rate as the tax paid on dividend income by the owners of the enterprise, there would be no deferral benefits. Because most direct equity investors are likely to be taxed either at the top marginal personal rate or at a final schedular rate, business enterprises should also be taxed at this top or schedular rate.34 A single-rate tax on business income would obviate the need to distinguish among different types of equity holders and, at least initially, to allocate income among such holders.
Income already taxed at the enterprise level may eventually be distributed to physical persons. If the tax system elects to levy a schedular, final tax on income from capital, the tax paid at the enterprise level can serve as that final tax on any distributions. If, instead, such income is taxed at progressive marginal rates, the tax paid at the enterprise level can serve as a withholding tax, for which a credit can be given the investor and for which a refund can be paid if necessary.35
Using a separate enterprise tax as either a final schedular tax or a withholding tax has proved to be an effective way of collecting tax on business income. Such a tax is typically levied as a separate legal liability on the enterprise. Because there are fewer enterprises than there are investors, and because they are more easily identifiable, having an enterprise as the principal taxpayer makes administration much easier than having only the investors as legal taxpayers. It also makes it much easier for the tax administration to distribute or collect adjustments resulting from audit.36
A number of technical reasons also favor a separate business enterprise tax over flow-through treatment. The first is that losses suffered at the enterprise level may be prevented from flowing through to equity investors and will then not be set off against other, taxable income.37 Second, transactions that might not typically be viewed as giving rise to taxable income in the case of flow-through entities can more readily be deemed to do so under a separate enterprise tax. The most important of these is that the entity may make distributions to its investors of economic income that was not taxed at the enterprise level. For example, income that benefits from tax incentives, or from the inexact science of tax bookkeeping (such as unintended acceleration of depreciation or income from unrealized capital gains), would not normally be taxed currently. However, under a separate system of business enterprise taxation, if such income were distributed as a dividend it might then be subject to tax.38
However, treating a business enterprise as a separate taxable entity, even if the tax raised is then treated as a prepayment or withholding of tax on income eventually received by the equity investor, can create a number of serious administrative problems. These problems vary depending on (1) whether investors who are physical persons are to be taxed at graduated rates on a global income basis or at a final schedular rate, and (2) how close a connection is made between the tax on the business enterprise and the tax due at the level of both the equity investor and the nonequity investor. The elaboration of these problems, and how they might be dealt with in an income tax law, constitutes the principal subject of this chapter.
Most jurisdictions employ a system of flow-through taxation for certain types of business enterprise and separate taxation for others.39 The choice of which business enterprises to subject to business entity tax and which to tax on a flow-though basis depends on a number of considerations. The greater the difference in outcome between the separate entity tax and flow-through treatment, the greater the incentive for taxpayers to engage in tax planning by selecting the more favorable form.40 Such tax planning may make tax administration more difficult and may affect the economy adversely if more efficient legal forms of business enterprise are eschewed in favor of those that are less efficient, but tax preferred. While inefficiencies may result from requiring small partnerships to be treated as separate taxpayers, as a general matter, making the net of inclusion for separate business enterprise tax as wide as possible will, in most instances, ease tax administration. For example, including all legal persons and entities engaged in business or profit-making activities (depending on how such organizations are defined under the applicable law), unless they have a very small number of owners, may be preferable to providing flow-through treatment for all partnerships.
III. Separate Taxation of Business Enterprises and of Distributions to Investors
While theory suggests that, in general, an enterprise income tax should be levied only as a withholding tax or a final schedular tax for business income of enterprises, a number of major income tax systems continue to apply the “classical” system of separate taxes on the income of certain business enterprises, resulting in double taxation of that income.41 A number of arguments have been advanced in favor of the classical system.
A. Tax on Economic Rents
Some economists have supported the imposition of a separate business enterprise tax in order to capture “economic rents,” or pure profits, which an investor earns in excess of the “cost of money”—that is, the given risk-free rate of return to capital.42 In theory, economic rents can be taxed without adversely affecting investment because they represent a return in excess of that otherwise required to make the investment.43
However, taxing both enterprise income and enterprise distributions would not necessarily tax economic rents. Instead, a system of double taxation would more likely result in inaccurate taxation of the amount paid to equity investors. With equity investments, the risk of low dividends or capital losses will largely be offset by the possibility of dividends or capital gains that are higher than the risk-free cost of money. Therefore, over time, earnings will include compensation for that risk. In other words, part of the excess of the return to equity investment over the risk-free average cost of capital is likely to be not economic rents, but risk premium. Taxing this risk premium as economic rent could cause substantial distortions.44 Methods have been developed to tax such rents; however, these methods appear seriously flawed.45
B. Subsidy Recapture
Some commentators have tried to justify a separate enterprise tax as a surrogate levy for the cost of government goods and services provided to those enterprises.46 The argument is that all investors operate in an environment deeply affected by free government benefits and that the level of government spending may increase the profitability of economic activity. This additional charge would, it is reasoned, ensure that the government does not distort the market allocation of resources.
It is difficult to see how the cost of government services provided to enterprises can be realistically related to additional income that is attributable to the equity holders of those enterprises. The argument seems to be that the total benefits provided free to the enterprise would equal the total enterprise income tax collected and that a particular enterprise’s share of the benefits would equal its share of the tax. This relationship seems highly implausible. Specially designed excise taxes (or, alternatively, charges for the services) would be a more efficient way of compensating for such benefits.
C. Increased Vertical Equity
Some have argued that, because enterprises tend to be owned by the wealthy, enterprise taxes should constitute a separate tax so that the vertical equity or fairness of the overall tax system can thereby be increased. A higher tax on all forms of income from capital would increase overall progressivity. Presumably, the most effective way of increasing the vertical equity of an income tax system is to increase its progressivity on all forms of income. However, under double taxation, only equity investments are subject to a separate tax. Removing existing tax benefits that favor the wealthy or imposing a more progressive income tax rate structure or, perhaps, a wealth tax would be more likely to raise the overall progressivity of the tax system than would taxing the income from equity capital twice.
D. Retention of Existing Double Taxation
Some have argued that, while a separate enterprise tax regime is not preferable in theory, if one is established it should be retained, at least for existing equity.47 There are two principle arguments: first, eliminating double taxation would reduce revenues,48 and second, elimination could result in a windfall to current equity holders because the effect of double taxation would already have been capitalized by a reduction in the price of equity.49 Equity investors will have demanded that other investors compensate them for the double tax burden they bear, so that the after-tax rate of return on equity would equal the after-tax rate of return on other investment forms. If the separate enterprise tax were removed, equity investors would receive a windfall as the value of equity increased.
While these arguments have merit, the effort to preserve a separate tax on old equity is unlikely to be worth avoiding the inequity of potential windfall benefits. First, potential losses in revenue can be made up with higher rates, or by eliminating investment incentives and other tax expenditures; jurisdictions that have recently eliminated double taxation have relied primarily on the latter technique.50 Second, investors take many forms of risk: one is that the tax system will change in a manner that affects them. If total revenues from capital income do not decrease, then eliminating double taxation will shift wealth from debt investors to equity investors, a risk that both forms of investors would probably have anticipated and for which some discounting may already have occurred. In addition, some have suggested that the burden of the separate enterprise tax is often substantially reduced through tax planning and that the amount of windfall shifting would therefore be small.
IV. Problems with Retaining Double Taxation of Enterprise Income
A. Aggravation of Tax Planning
A system of double taxation of equity income creates incentives to avoid double taxation through tax planning, and may involve opportunities to avoid tax altogether. Among the techniques are the following:
(1) Choosing business forms that are not subject to double taxation, where feasible. For example, commercial laws may allow taxpayers to set up entities that will be treated as transparent for tax purposes—partnerships or trusts might be available options—and that have enough of the properties investors want, particularly limited liability and free transferability of interests.
(2) Raising capital through legal forms that allow deduction of payments to investors, such as rental payments and interest. As was mentioned above, modern financial instruments give taxpayers opportunities to structure their investment in a form that is classified as debt, so that the return on the investment reduces the corporate tax base, but that has enough of the desired attributes of equity, particularly the opportunity to share in the potential gains from corporate success.
(3) Distributing earnings to equity investors through techniques that do not give rise to the second tax.51 Corporate law may allow corporations to return amounts to investors through a variety of devices, including the redemption or purchase of stock, and partial reductions of capital. These amounts will be a desirable substitute for a distribution of a similar amount that is labeled a dividend.
(4) Making deductible payments to investors (and their associates) in their capacity as directors or employees. Small businesses in particular will be able to reduce the corporate tax base by paying salaries to owner-managers and to family members, thus achieving two benefits—a reduction in the enterprise tax base and the splitting of income within the family.52
(5) Retaining, rather than distributing, profits. The event that triggers imposition of the personal income tax will usually be the payment of a dividend or the sale of shares. Insofar as it is possible for the corporation’s managers to delay triggering this event by retaining profits, the shareholder tax can be delayed and thus reduced.
Such incentives can lead to inefficiencies in the operation of enterprises. How significant these inefficiencies prove to be depends upon the circumstances—a more benign view of these devices regards them as self-help remedies employed by investors to alleviate the problem of double taxation informally.
These issues are symptomatic of the problems that arise whenever one legally defined form of investment is taxed at a higher rate—taxpayers will usually try to recharacterize that investment as a form taxed at a lower rate. This legal recharacterization is time consuming, expensive for the tax administration to prevent, and frequently a losing battle. These problems do not necessarily go away once taxation of equity income is limited to a single level of tax. As long as the final, effective rates of tax on different types of income from business enterprise are different, these problems will exist. While eliminating the second level of tax generally reduces the incentive to recharacterize interests to a preferred type, as long as final effective tax rates differ, the incentives will remain, albeit in a less virulent form.
That is to say, even if a system is formally designed to tax business enterprise income once, the way in which the system is designed may result in income being taxed differently depending on whether capital is in the legal form of equity or debt or lease, or on whether earnings are distributed to equity investors other than through normal distributions.
B. Profit Retention
More generally, two concerns have been expressed about the overall economic impact of the enterprise tax on economic activity. One is that it tempts enterprise managers to retain rather than distribute profits. The other is that it encourages the financing of investment through debt rather than equity.
The effect of the enterprise tax on required rates of return and, by inference, on the cost of capital can be illustrated in the following example. Assume an enterprise tax rate of 30 percent and a personal income tax rate of 40 percent. Assume also that investors have enough other investment opportunities with the same risk profile and that the corporation needs to provide investors with an after-tax return of 6 percent to induce them to part with their savings. In these circumstances, the investment would need to offer a pretax return of 8.5 percent [6/(1 − 0.30)] if it could be financed out of retained earnings and the enterprise did not need to distribute profits, 10 percent [6/(1 − 0.40)] if the investment were to be financed through debt, and more than 14.3 percent [6/(1 − 0.3)(1 − 0.4)] if the enterprise needed to finance the investment with new equity and investors expected to receive dividends.
Not surprisingly, therefore, commentators have tried to assess the consequences of the incentive to retain profits. Insofar as retained earnings are used to finance enterprise expansion, they serve as a substitute for raising that capital through formal borrowing, leasing, or further equity issues. These substitutes are likely to be conducted under the scrutiny of the market—bankers will examine the enterprise’s solvency and cash flow before making further loans, underwriters and investment houses will examine the prospectus for a further share or bond issue, and so on. This scrutiny is not applied when the managers of the enterprise can choose how much profit to retain. The fear is that the tax system will encourage managers of existing mature companies to retain funds unnecessarily and invest in projects that are less than optimal in order to use the excess funds. These retained earnings should instead be liberated for the use of fast-growing innovative enterprises.53 Some argue, however, that retained earnings are a source of additional private savings within an economy.
There is no unequivocal evidence that these outcomes have occurred systematically and that, where they do occur, it is the tax system rather than independent corporate financial, policy that is the motivating cause. There are competing visions of the economic effects of these incentives. According to the so-called traditional view, the increased tax cost associated with dividend payouts is likely to be significant, and corporations will therefore tend to rely on retained earnings. When retained earnings prove to be inadequate, and enterprises have to issue further equity, they will have to raise their payout ratios to meet the added tax costs, increasing their cost of capital. But an important qualification to this prognosis is the recognition that systemic factors may prevent excess profit retention from becoming a problem—nonfiscal considerations may outweigh the fiscal advantages. The market may not allow shareholder distributions to be deferred indefinitely, and shareholders may insist on receiving some return as an indication of the ongoing soundness of the enterprise.54
An alternate vision, the so-called new view of enterprise taxation, argues that because the tax disadvantages of dividend payouts are well known, enterprises will indeed finance their activities largely through retained earnings. Paradoxically, however, the higher taxation of dividends will be of little consequence.55 According to this theory, shareholders might save the personal income tax on the dividends they would otherwise have received, but, adopting a longer-term view, they have simply converted the immediate tax on distributions into a deferred capital gains tax liability that will be triggered on the disposal of their investment. If this is so, buyers of the security will discount it to reflect the deferred liability, and so the additional tax is capitalized into the price of the share. The additional tax is a real cost that the original holder of the share will bear regardless of whether the distribution is paid. The buyer of the share recoups the additional tax because he or she has been able to buy the share at a reduced price, reflecting the implicit tax liability. If this new view is correct, the additional tax on distributions becomes almost irrelevant for mature enterprises because the existing shareholders are affected in both cases and the buyer is not affected in either case.
C. Debt and Equity
The final question is whether the enterprise tax system encourages firms to finance their investments excessively through debt. If so, it is feared that firms would be vulnerable to bankruptcy in times of economic downturn and that increased numbers of bankruptcies would exacerbate the destabilization of the national economy during such a period.
Whether there is an incentive to finance new investments through debt or retained earnings, and how significant it is, will depend on the differences between the tax treatment of the investor under the personal income tax and the enterprise tax—that is, the enterprise tax rate on retained earnings must be compared with the personal income tax rate on interest income.
In an international context, the substitution of debt for equity has additional consequences. For an individual country, it implies the diminution of the domestic tax base because the return on enterprise equity is taxed in the source country through the imposition of the enterprise tax on the resident enterprise. The return on enterprise debt, by contrast, is often taxed only in the residence country because the interest reduces the domestic enterprise tax base and there will often be no withholding (or only limited withholding) on interest payments paid to another country. The result is that the investor’s country of residence instead of the source country will tax the interest payment.56
V. Relationship Between Enterprise Income and Investor Income
A. Single Schedular Tax on Income from Capital
1. Equity Interests
As noted earlier, there are two basic systems for taxing income derived from an equity investment in a business enterprise. The first is to tax the income at a single rate that is applied to all investors and applied on a schedular basis to the net income of the enterprise. The second is to tax the income at different rates (typically graduated) depending on the circumstances of each investor; that is, the particular rate applied is determined by reference to the investor’s total net income. Typically, then, the income from the business enterprise is added to the total net income of the investor, and the appropriate rate is applied on the basis of that total net income. Which system is chosen and how the system is implemented are exceptionally important to the operation of the enterprise income tax. A related issue is which system is used to tax income from deductible debt (or lease) investments.
Chapter 14 discusses a number of issues surrounding the choice between taxing income at schedular rates or at multiple (typically graduated) rates applied to a global income base. In addition to those issues, there is often considerable support for exempting not-for-profit organizations and pension funds from income tax. However, it should be noted that the economic and social arguments in favor of such exemptions are often rather less than fully convincing.57 In addition, existing bilateral double taxation agreements might provide for varying rates of tax depending on the residency of the investor. It is, however, quite possible, and perhaps even advisable, to exempt nonresidents from any withholding tax (in addition, that is, to enterprise-level tax).58 There have also been a number of concerns that a schedular tax is less equitable than a graduated tax. However, a schedular tax on income from property or capital, with a progressive tax on income from labor, has been advanced as a technique that would combine the added fairness of progressive taxation with the simplicity of schedular taxation.59 In spite of these arguments, it might still be difficult to fully implement a policy of a single, schedular rate of tax on income from business enterprise.60
The main administrative benefit of a single schedular rate is that the tax can largely be levied at the enterprise level, without reference to the investor. As will be discussed later in this chapter, in particular, when a single schedular tax rate is combined with a highly effective enterprise tax and full imputation, problems involving levying taxes on distributions at the shareholder level more or less disappear. As can be imagined, this makes for perhaps the easiest type of enterprise-shareholder tax system to implement. There are two possible exceptions to this rule. The first involves the taxation of capital gains and losses realized by the investor when he or she sells the equity investment. However, as will be discussed, the more effective the enterprise tax, the less important the investor-level capital gains tax. In such cases, one can probably exempt most investors most of the time from tax on such gains without much loss of revenue or equity.
The second exception is that, if income from each separate business enterprise is taxed on a schedular basis, the losses associated with an investor’s share in an unprofitable enterprise will not be used to offset the income of a profitable enterprise. This will raise concerns of tax administration and taxpayer equity, as investors try to accomplish this offset by other means. This problem is, by and large, not completely solvable unless losses of all business enterprises are flowed through along with income. Because levying a tax on business income is designed in part to avoid having to allocate earnings and losses among equity investors, the problem of schedular taxation is shared by all separate enterprise income tax systems.
The most obvious problem with multiple rates is that discussed briefly earlier: to avoid the deferral problems that the enterprise tax is designed to combat, the rate of enterprise income tax must be the highest rate at which investors are taxed. Thus, some investors will be taxed at rates higher than the marginal rate that applies to their other income. The problem does not come up if there is a single rate of tax.61
A single rate of tax on all earnings from equity investments is clearly preferable from a tax administration viewpoint. The problems that need to be addressed under a multiple-rate system are discussed later in the chapter.
2. Debt and Lease Interests
Taxing all income, including that from debt and leases, at the same schedular rate eases administration markedly. First, one of the more difficult and complex areas in tax administration involves distinguishing equity interests from debt. While limiting taxation of income from equity investments to a single level of tax is an essential step toward equal taxation of equity and debt, it is not the only issue. If income from both equity and debt investments is taxed once, but the income from one or the other is taxed at a different rate, an incentive will still exist to design the legal form of the investment to fit the category of income that is taxed at that reduced rate. It is in part for this reason that the U.S. Treasury Report recommended that the income from both equity and debt investments be taxed identically, at a single schedular rate.62 The usual administrative response to different tax treatment for income from equity and debt interests is to establish rules of “thin capitalization” or “earnings stripping.”63
Even if the income from both debt and equity investments is taxed formally at the same rate, the method by which an enterprise income tax is typically levied causes the tax treatment of income from equity and debt investments to differ in some cases. This is because the general treatment of income accruing to a debt investor is to allow the enterprise to deduct the interest accrued or paid.64 The deduction at the enterprise level ensures that the interest is not subject to tax at that level. Instead, the interest can be taxed as income to the lender.65 By contrast, income on equity investments is generally taxed at the enterprise level and perhaps also at the shareholder level.
The tax administration problem arises under three circumstances. One of these arises from the mismatching of preference income and deductions: the income of the taxpayer entity is largely tax exempt, while deductible interest (or lease) payments that can be used against income that is not tax exempt are allowed. This issue will be addressed below.66 The others arise because of the different treatment of income from equity investment and debt investment, shareholder and creditor; even if these categories are taxed formally at the same schedular rate, they are not taxed at the same effective rate.
Deductions are clearly worth more to taxpayers who are in higher tax brackets.67 If a deduction by one taxpayer is followed by an equivalent inclusion for another, as is generally the case with interest payments, overall taxes will be reduced if the taxpayer paying the deductible amount is in a higher rate bracket than the recipient of the payment. There will then be an incentive for those paying at the higher rate to accrue as many deductions as possible. They can then share the benefits with those paying the lower rate. Of course, if the borrowing can be structured so that a deduction by the borrower is not followed immediately by an inclusion by the lender, a tax benefit will accrue even if the lender is taxed at the same rate.
The borrowing taxpayer can structure investments in a number of ways to increase or accelerate deductions. First, the taxpayer can overstate total amounts of interest. As discussed earlier, payment of interest is directly related to risk.68 Even if inflation risk is eliminated from taxation through adjustment,69 differences in default risk will result in different interest rates being paid by different borrowers.70 Therefore, it can be difficult for tax authorities to determine how much of a payment would constitute actual economic interest and how much a return of invested capital.
Debt instruments can also be designed to accelerate interest payments in early years. One way is to structure the instrument so that it pays interest through discount. The payer can then seek to accrue interest on an annual basis without including the compounding effects of the discount. Although both of these avoidance techniques can be countered with proper accounting rules for interest imputation,71 financial product innovation has made such accounting increasingly difficult.
If, however, both the borrower and the lender are taxed at the same schedular rate on this periodic income, the incentive to shift income is eliminated. Any benefit that the debtor might derive by mischaracterizing interest in order to take a deduction when one is not legitimately due is canceled by the taxation of such income to the creditor.72 Also, and of great importance, the schedular tax on the creditor can be levied at source.
Identical rates of schedular tax on equity and interest earnings (as well as among different types of interest)73 would not end all problems of the allocation of payments between interest and principal or of interest over time. This is because it is impossible to effect a single, schedular tax at the enterprise level. Normally under an enterprise-level tax, losses at the enterprise level do not flow through to the investor.74 Therefore, if the enterprise has no taxable income, and if any carrybacks for current losses do not result in a refund, an interest deduction at the entity level may not be worth any current tax benefit.75 In such cases, there will be an incentive for the creditor to reduce or eliminate interest payments by mischaracterizing interest as principal or by delaying interest deductions.76
There are two major reasons that an enterprise may have no taxable income. First, it may be a for-profit business enterprise, but have no taxable income77 either because it has no economic income or because it benefits from tax preferences—deferral or complete exemption. If the enterprise has no economic income, mischaracterization of interest and principal or delay in accruing interest should not be a significant tax policy concern. The reason that the entity would mischaracterize interest and principal is that the genuine, economic loss resulting from the payment of interest could not be effectively reflected in a reduction in the tax base. Under a Haig-Simons analysis, a decline in wealth should be reflected in a decrease in the taxpayer’s tax base.78 It is only the practical operation of the enterprise-level tax that prevents this loss from being accrued. Therefore, with certain exceptions,79 the equity owner is unfairly penalized for being unable to realize the value of the deduction for any interest actually accrued or paid. If there is an offsetting reduction in the tax owed by the creditor, there will be no net loss to the exchequer; any shifting of tax benefits between enterprise and creditor can be adjusted by the two actors. If the creditor’s tax is collected through withholding at source, the adjustment could be implemented quite easily.80
Second, the entity may be a governmental, charitable, or other entity that is statutorily exempt from tax.81 Pension funds are also typically exempt from income taxation. The problem posed by exempt entities may be reduced by taxing them on their investment income, which may well be advisable from a purely economic perspective as well. Such taxation would create a tax base from which the entity could deduct interest expenses.82
It is also possible that there is economic income at the enterprise level, but that this income is “tax preferred,” such that no tax is currently due. The preference can be intentional; for example, provisions in the law may exempt some income from tax, tax some income at a lower rate, or delay the inclusion of some income until a later time. Some income may be unavoidably subject to a timing preference because of the deferral of tax on unrealized capital gains. In these instances, there will also be an incentive to understate deductions at the enterprise level, so as to have a mirror understatement of income at the creditor level. Unlike the earlier case, the tax administration should be concerned about understatements of income tax at the creditor level.
These problems can be minimized by reducing or eliminating special tax benefits. However, if a realization system of taxation is retained for most capital gains, the problems will never be eliminated, although this chapter will argue that the distortions caused by the realization event system can be greatly minimized by marking certain financial assets to market, and by taking into income currently total borrowings in excess of the total adjusted cost of assets. However, if these ideas are not implemented or are only partially implemented, the tax administration will have to ensure that interest accrues to the creditor.
As noted earlier, it is difficult to impute interest on debt whenever a risk premium is due. It is also increasingly difficult to impute interest on many financial instruments. One possible solution is to require a minimum imputation of interest on all debt instruments, based on the amount of capital invested. This minimum imputation could be based on a provision in the income tax code that would give the tax administration the authority to impute an interest component on any debt obligation of an enterprise.83
One possible technique for more completely equating the tax treatment of income from equity and debt might be to extend deductibility treatment to returns on equity investments. Some proposals have arisen in the past to do so, particularly partially to integrate enterprise and investor taxes.84 Extending deductibility treatment would not solve the problems discussed above. If the entity had no taxable income, there would still be a benefit to the debt investor and therefore, by extension, to the equity investor who might share the benefit. However, with regard to distributed income, equivalent treatment would certainly prevail. No major income tax system currently affords such treatment, for a number of reasons. Two of the most important are the passing along of foreign tax credits to investors and the treatment afforded nonresident equity investors under most double taxation agreements.85 Dividend deduction models are largely missing from the world tax scene.
Some have suggested that equation of treatment could be reversed; instead of allowing a deduction for payments of earnings on equity investments at the enterprise level, deductions for interest expense could be disallowed. The 1991 U.S. Treasury Report on integrating enterprise and investor taxes recommended taxing income from both equity and debt investments entirely at the enterprise level by denying a deduction for interest payments. This would turn the enterprise tax into a comprehensive business income tax.86 The problems of understating interest or delaying its payment or accrual would disappear in that the interest deduction would no longer formally be part of the income tax system. However, the net economic result would be the same; the recipient would be able to defer tax on interest income when the entity had economic, but not taxable, income.87 Once again, the obvious answer is, where possible, to apply a system whereby taxable and economic income most closely approximate each other.
Eliminating deductions for interest could be a technique for ensuring that, at least with regard to interest accrued, such payments would be taxed at the same rate, that is, the entity rate. In this sense, disallowing deductions is analogous to integration schemes that tax enterprise income only and exempt the distribution from tax at the shareholder level. Such treatment would certainly reduce debt-equity and earnings-stripping problems. However, such a system has not yet been attempted in any major tax jurisdiction (except partially, as part of a regime to prevent earnings stripping.)88 Once again, much of the reason for this may stem from the existence of double taxation agreements and the problems that would arise if all interest income were effectively taxed to nonresidents at the enterprise rate of tax, rather than at the rates specified by those agreements.89 However, treating interest in this way is logically consistent with taxing equity income only at the entity level and—if a decision is made to tax income from capital at a single rate and if the international dimensions can be negotiated—may constitute the easiest system of entity taxation to administer.
3. Treatment of Preferred Income
A. Deductions for Interest Expense
As noted earlier, problems arise with regard to the like tax treatment of debt and equity, or of different types of debt and different types of equity, when the entity has tax-preferred income. The tax effect on the equity investor may also be problematic. If the tax on the “preferred” income is deferred or if this income is tax exempt, there will be an incorrect tax result, either a reduction in tax (because of the time value of money) or a complete exemption. However, the taxpayer may be able to finance the investment with borrowed money. Normally, the payment or accrual of interest leads to a real decline in wealth. As noted earlier, under a Haig-Simons analysis, a decline in wealth should be reflected in a decrease in the taxpayer’s tax base.90 However, a taxpayer with no taxable income cannot benefit from the deduction. The benefit of the exclusion afforded the taxpayer on the preferred income would be reduced by the denial of interest deductions. While it is unlikely that denying the interest deduction would have the same effect as the benefit afforded through the preferred income, it would have the effect of a partially compensating disallowance.91 However, if the taxpayer has other taxable income, the taxpayer may use the interest deductions against this income, thereby avoiding tax on this income as well.92 The effect would be to eliminate the (only partially) compensating distortion caused by the inability to benefit from the deduction of interest, which may compound the problem of having preferred income in the first place.
Tax policy analysts normally recommend, for reasons of economic efficiency and administrative ease, eliminating tax preferences whenever politically possible. If all preferences were eliminated, in theory at least, this mismatching problem would cease to exist. And, although many tax systems have attempted to travel far in the direction of eliminating preferences, the problem of tax deferral under the mixed accrual/realization event system of accounting is unlikely to go away entirely. This means that a taxpayer can borrow against assets that have appreciated in value, but on which the gain has not been taxed, while still being able to deduct interest. This situation has been described as allowing the taxpayer to “realize” the gain (by borrowing against it) without having to pay tax on it.93
Chapter 16 discusses techniques of quarantining or otherwise disallowing interest deductions when they relate to the financing of tax-preferred income.94 An alternative approach is to recharacterize not the deduction of interest, but the borrowing itself.
As noted earlier, tax-preferred income can take a number of different forms, the main ones being statutory incentives (which the legislature can avoid enacting) and preferences related to the realization event (which are difficult to avoid). The benefit of the latter is not permanent, but is related to timing. In effect, by delaying the taxation of accrued gains, the taxpayer benefits from the time value of the deferred tax. When the taxpayer realizes the gain by selling or transferring the gain asset, tax is incurred; obviously, if possible, the taxpayer would prefer to avoid such a taxable event.
However, if the taxpayer needs cash, he or she can instead borrow the money. In effect, borrowing the money is analogous to selling or transferring the asset. Instead of quarantining interest by disallowing a deduction in a “compensating distortion,” it would be possible to treat the borrowing as a realization event, at least to the extent that the borrowing exceeds the adjusted cost (book value or written-down value) of the asset. If the borrowing is secured by a single asset, the amount of gain can be determined on the basis of that asset alone. To the extent that the borrowing is not secured by a single asset, the amount of the gain can be determined on the basis of all assets held by the taxpayer.
For example, if the taxpayer holds a single asset with an adjusted cost of $10 and borrows $20, he or she would include $10 in taxable income. The asset’s cost basis would be increased to $20, and the full amount of interest due on the $20 debt would be deductible. If the taxpayer has a large number of assets, with a total adjusted cost of $100,000, and borrows $200,000, he or she would include $100,000 in income, and the adjusted cost of all assets would be increased by $100,000. However, because each individual asset would have to be adjusted, the $100,000 increase would have to be apportioned among all the assets, for example, on a proportional basis.95
Although no jurisdiction currently treats borrowing as a realization event, the logic of such an approach seems compelling. It would have additional benefits with regard to taxing economic earnings at the enterprise level that are distributed to the equity investor; this issue will be dealt with at greater length below.
In addition to this technique, other accounting methods can be used to reduce the amount of accrued but unrealized capital gains (and losses as well). The most important of these is to require enterprises to mark assets to market whenever reasonable. In particular, such marking to market could be done for foreign exchange, precious metals, and securities and financial derivatives for which a listed price could be easily obtained.
B. Taxation of Equity Distributions from Preference Income
Tax-preferred income of flow-through entities can normally be distributed to the investor without any immediate additional tax consequence, although typically there are consequences for the taxation of capital gains and losses when the entity interest is transferred.96 For investors in entities without flow-through treatment, similar tax rules could apply; all income tax would be levied at the entity level on the preferred income.97 If the preferred income were of the permanent, or exclusion variety, no tax would ever be paid. If the preferred income were of the deferral type, tax would be paid when the preference expired at the entity level.
Such tax treatment of the distribution of preference income to equity investors is extremely rare.98 In the vast majority of jurisdictions, tax is levied on distributions out of income that was not already fully subject to tax at the entity level. Tax can be levied in various ways, and those variations have considerable effect on the administration of the tax (see discussion below). However, levying tax on such untaxed income has two basic and important effects on tax administration, one somewhat positive and one quite negative. Therefore, before examining the specific effects of different forms of implementation, we discuss why distributions of preference income should be taxed in the first place.
Various arguments have been raised as to why distributions from preference income should be taxed.99 One reason is that specifically enacted tax preferences may have been designed to encourage investment. Whenever a business entity distributes such income, the implication may be that it is to be used for consumption and not for investment. Therefore, a tax should be levied. Another argument, which applies only to systems that tax all distributions of economic income not previously taxed, is that any income arising from timing preferences that can be realized without otherwise incurring entity income tax (primarily through borrowing) should be taxed as if the distribution were a realization event. Finally, if equity holders are taxed at different rates, distributions should always be subject to tax to ensure that a higher-rate investor will pay tax at the higher rate on such distributions.100
Again, of course, these problems are reduced, or may disappear entirely, as the amount of preference income is curtailed or eliminated.
B. Multiple Taxes on Income from Capital
A single schedular tax on capital income derived through legal persons may be considered unacceptable for a variety of reasons, not the least of which is the difficulty of finding an acceptable single rate. The conduit view is not fully implemented if the single tax rate imposes higher tax burdens on low-income investors or reduces the tax burden of high-income investors.101
Consequently, many countries find it necessary to operate in tandem both the enterprise-level profit tax and the investor-level personal income tax. This section reviews some of the more commonly used options for reducing or eliminating the double taxation of capital income that these systems can induce. It describes in detail some of the principal interaction mechanisms between the enterprise and personal income tax used in the taxation systems of various countries. Except for one case, the interaction mechanisms attempt to deal only with the double tax on distributions and leave untouched the double tax on retained corporate profits.102 Given that the issue they try to resolve is the double taxation of equity, they deal only with returns on equity investments, not interest or rent.
Even with these restrictions, it is common to see a wide variety of idiosyncratic mechanisms for integrating the enterprise and individual tax, and so this section will describe the most important features of a few representative types seen in practice.103 The design of these interaction systems involves many issues, but the most important are the following:
The level at which the relief is to be provided. The mechanism for reducing double tax on dividends can operate at either the enterprise or the shareholder level.
The form that the relief is to take. Generally, the options are to use a tax deduction, a tax credit, or an exemption system. A subsidiary issue is whether the application of relief is to be made conditional upon some tracking or verification of other tax payments.
Whether the relief is to be afforded to nonresident shareholders. The extent to which nonresidents can be further burdened, especially by withholding taxes, will often be controlled by any applicable tax treaties, but the extent to which they may be benefited is largely a matter for domestic law.104
Whether relief is to be afforded to income from equity investments derived by tax-exempt investors.
Whether different types of shareholders are to be treated differently. These mechanisms often distinguish corporate from individual shareholders, and resident from nonresident shareholders, but other possible distinctions might differentiate holders of controlling interests from holders of portfolio interests.
The treatment of enterprise tax preferences. Tax preferences can be preserved in full for the benefit of shareholders, be preserved but at a reduced value, or be recaptured entirely at the shareholder level.
The treatment of foreign-source income. This income can be viewed as raising issues similar to those surrounding enterprise tax preferences. In both cases, domestic enterprise-level tax is not paid on income that is to be distributed or retained (although for foreign-source income, some foreign tax may well have been paid), but the arguments about imposing tax on the distributions are slightly different.
The analysis concentrates on the major aspect of the problem and the topic of this chapter—the treatment of income earned by resident individual shareholders—but is expanded, where relevant, to examine the position of nonresident individual shareholders and income earned by intermediaries, such as other enterprises.105
The idiosyncrasies of the mechanisms that countries have adopted (not to mention the peculiarities of nomenclature)106 make it difficult to generalize. Nonetheless, once the classical system is abandoned, the mechanisms for recognizing the impact of both enterprise and shareholder-level income tax can be combined into a few illustrative groups: split-rate systems and dividend-paid deduction systems, dividend-received exemption or dividend-received deduction systems, dividend-imputation systems, and full integration systems (see Table 1).107 The first three are often referred to as systems for dividend relief—adjusting the combined tax rate on distributions—while the last, integration, is more ambitious—reducing the combined tax rate on all enterprise profits.
Interaction Systems
Interaction Systems
| Dividend-Relief System for Distributed Profits | Integration System for All Profits | ||||
|---|---|---|---|---|---|
| Classical System | Corporate level | Shareholder level | |||
| Dividend-paid deduction system | Split-rate system | Dividend-received exemption/deduction system | Imputation system | ||
Interaction Systems
| Dividend-Relief System for Distributed Profits | Integration System for All Profits | ||||
|---|---|---|---|---|---|
| Classical System | Corporate level | Shareholder level | |||
| Dividend-paid deduction system | Split-rate system | Dividend-received exemption/deduction system | Imputation system | ||
There are virtues and vices to each interaction system, which explains why a standard regime has not emerged.108 To illustrate, systems that reduce the enterprise’s primary tax liability (such as dividend-paid deduction systems) will benefit both resident and nonresident shareholders equally, a result the source country may dislike.109 Systems that reduce the enterprise’s primary tax liability would also reduce the implied tax paid at the enterprise level by tax-exempt investors.110 Dividend-paid deductions would need to be targeted to deny the tax benefit when distributions are paid to tax-exempt entities or nonresidents if the double taxation of dividends is to be sustained for these groups. Dividend-received deduction systems and some imputation systems will not ensure that the enterprise has actually paid any tax on the dividend received by the shareholder although they do preserve the full nominal value of enterprise tax incentives for shareholders.111 Some systems can result in overtaxation of the enterprise when the tax collected exceeds the enterprise’s own tax liability, while others require elaborate record keeping.112 Integration systems that tax shareholders on the value of retentions can cause solvency problems for individual shareholders when distributions are small but profits are large and are generally considered impractical for large enterprises in part because of the difficulties of administering them113 and because the substantial international treaty network assumes that nonresident shareholders are not currently taxed on retentions.114
Not surprisingly, therefore, in a 1993 study, the O E C D found representative types of almost all possible systems among the corporate tax systems of its (then) 24 member countries (see Table 2 for a classification of the systems then existing).
Degree of Reduction of Economic Double Taxation (Central Government)
Degree of Reduction of Economic Double Taxation (Central Government)
| Reduction of Economic Double Taxation | Elimination of Economic Double Taxation | |||||
|---|---|---|---|---|---|---|
| None or Very Little | Corporate level | Shareholder level | Corporate level | Shareholder level | ||
| Classical system | Split-rate system | Partial dividend- deduction system | Partial imputation system | Partial shareholder- relief schemes | Zero-rate system | Full imputation system |
| Belgium | Germany | Iceland | France | Austria | Greece | Australia |
| Luxembourg | Spain | Ireland | Canada | Norway | Finland | |
| Netherlands | Sweden | Denmark | Germany | |||
| Switzerland | Iceland | Italy | ||||
| United States | Japan | New Zealand | ||||
| Portugal | Turkey | |||||
Degree of Reduction of Economic Double Taxation (Central Government)
| Reduction of Economic Double Taxation | Elimination of Economic Double Taxation | |||||
|---|---|---|---|---|---|---|
| None or Very Little | Corporate level | Shareholder level | Corporate level | Shareholder level | ||
| Classical system | Split-rate system | Partial dividend- deduction system | Partial imputation system | Partial shareholder- relief schemes | Zero-rate system | Full imputation system |
| Belgium | Germany | Iceland | France | Austria | Greece | Australia |
| Luxembourg | Spain | Ireland | Canada | Norway | Finland | |
| Netherlands | Sweden | Denmark | Germany | |||
| Switzerland | Iceland | Italy | ||||
| United States | Japan | New Zealand | ||||
| Portugal | Turkey | |||||
The variety of interaction systems that existed in 1993 suggests that the effects of each system on the after-tax returns to shareholders and the cost of enterprise capital will differ. This section analyzes the classical system and four systems of enterprise and shareholder interaction. The models described are stylized to capture the fundamental relationships of the systems discussed, rather than being precise descriptions of the exact rules employed in any particular jurisdiction.
1. Separate (or Classical) System of Enterprise Tax
The pure classical system is declining in industrial countries’ tax systems.115 Among the countries of the EU, Belgium, the Netherlands, Luxembourg, and Sweden116 retain the classical system for distributions to individual shareholders.117
A. Resident Individuals
In a pure classical system, there is no formal interaction between the enterprise and individual income taxes, and each is levied without explicit regard for the operation of the other. But even in a classical system, there may be implicit recognition of the dual operation of both taxes in the rate imposed under either tax or in the definition of its base. For example, a lower marginal rate imposed upon an individual’s capital income or substantial investment concessions offered to industry may each be a method for recognizing the existence of the two layers of tax. The first reduces total tax by encouraging retention of profits by the enterprise and extraction of gain by the individual selling the shares, while the second reduces the total tax collected from the enterprise.118
The first option can be seen in the following example.
Example
Assume that the enterprise tax rate is 33⅓ percent, the top marginal rate under the personal income tax is 50 percent, and dividends are taxed at a flat rate of 25 percent imposed on individual shareholders (or perhaps collected by with-holding at source). This rate alignment offers a good approximation of the after tax return [$100*(1 − 0.33)(1 − 0.25) = $50.25] that would be earned if the income from the investment had been earned by a high-income shareholder directly ($50).
Such a system would probably, however, have several serious consequences. First, it might discourage distributions of enterprise profits—indeed, some rule would probably be needed to oblige distributions. If distributions were not obliged, serious strain would be placed on the administration of the capital gains tax as the means of collecting the deferred tax on retained earnings. It would also deliver a sizable benefit to tax-exempt institutions because the enterprise tax is the principal tax that they pay on capital income.
The position of a shareholder in such a system can be expressed algebraically as follows. Under the classical system, the enterprise pays tax (Tc) on its taxable profits (P), and the individual resident shareholder pays income tax (Tt) at progressive marginal rates on the proportion (d) of after-tax profits distributed by the enterprise as dividends. Retained profits (1 − d), reflected as accretions to the value of the shares, are taxed as capital gains (Tg) on a deferred basis when the shares are sold by the shareholder, and are sometimes taxed at a lower nominal rate.119 Given an enterprise tax system bearing these features, the return (R) to an individual shareholder after payment of enterprise tax on all profits and personal tax on distributions and retentions is120
R = dP(1 − Tc)(1 − Ti) + (1 − d)P(1 − Tc)(1 − Tg).
B. Preference Income
Because the enterprise and the shareholder are taxed separately, this system applies also to distributions of untaxed income, such as income that enjoys tax preferences or foreign-source income that is not subject to tax in the residence country.121 Earnings that are untaxed or not fully taxed at the enterprise level would, nevertheless, be subject to full shareholder tax:
R = dP(1 − Ti) + (1 − d)P(1 − Tg).
c. Nonresident Shareholders
As the example illustrates, one can achieve the same result as under conduit treatment by reducing the enterprise rate below the top individual rate and imposing a withholding tax on distributions. Another reason for adjusting the personal tax rather than the enterprise tax is the position of nonresident shareholders. For nonresidents, the enterprise tax rate is an important determinant of the total tax that they will pay to the source country, and reducing it will confer a substantial benefit on them. Treaties generally allocate the enterprise tax to the source country and limit its ability to impose substantial withholding taxes on payments to nonresident shareholders.122 Consequently, a country’s treaties and the international norms for taxing enterprise income will most likely make it impossible for the country to impose substantial additional taxes on a nonresident shareholder to compensate for a low enterprise tax rate.
2. Dividend-Paid Deduction System
A dividend-paid deduction system operates at the enterprise level to impose different rates on an enterprise’s distributed and undistributed profits. The system achieves this result by giving to the enterprise a tax deduction for distributions made and then imposing tax on the distribution at the shareholder level.123 A tax deduction for distributed profits means that the profits incur no tax at the enterprise level and are effectively taxed as if they were payments of interest by the enterprise.124
A. Resident Individuals
Under a dividend-paid deduction system, the enterprise is able to reduce its taxable profits by the amount of any distribution. The enterprise therefore pays no tax on distributed profits, but pays tax at the enterprise rate on retentions. This system has some of the same effects as a split-rate system (discussed below) under which the rate on distributed profits is set at zero. The individual shareholder who is a resident pays income tax at marginal rates on the proportion of profits distributed by the enterprise as dividends, and any retained profits already taxed to the enterprise are taxed as a capital gain on a deferred basis to the shareholder. No deduction or tax credit is given to the shareholder for taxes paid by the enterprise. The after-tax return of an individual shareholder after payment of enterprise tax on all profits and personal tax on distributions and retentions is
R = dP(1 − Ti) + (1 − d)P(1 − Tc)(1 − Tg).
Again, tax-exempt investors will benefit from the elimination of all tax on distributed earnings.
B. Preference Income
Other forms of untaxed income, such as income enjoying tax preferences or foreign-source income that is not taxable in the residence country, would be subject to enterprise tax only if retained.
Example
Assume a corporation has operating profits in the current year of $30,000, but taxable profits of $24,000 (prior to making any dividend payment) because of a $6,000 enterprise tax preference. Its potential tax liability at a 25 percent rate is thus $6,000 if it makes no distributions. If, in the current year, the company retains $12,000, its tax liability is $3,000 [25 percent of $12,000]. The shareholder will pay tax at personal rates on the $12,000 of taxable profit that is distributed. Tax will thus be collected currently on $24,000, part from the corporation at enterprise tax rates and part from the shareholders at their marginal rates. The other $6,000 of nontaxable operating profit will be taxed as a capital gain to the shareholder when the gain is realized, reducing the value of the tax preference from a permanent to a temporary reduction of tax.
Such a system raises a few stacking and ordering issues, which the example below addresses:
Example
Assume a corporation has operating profits in the current year of $30,000, but taxable profits of $20,000 (prior to making any dividend payment) because of a $10,000 enterprise tax preference. Its potential tax liability at a 25 percent rate is thus $5,000 if it makes no distributions. If the corporation distributes $25,000, two related questions arise: what is the treatment of the $5,000 in the hands of the shareholder, and does the enterprise generate a carryover loss from this transaction? It has an additional deduction of $25,000, which exceeds its taxable profits of $20,000.
The answer should depend on whether the government wishes the enterprise tax preference to be lost. If it is to be recaptured, the shareholder should be taxable on the $25,000, and the enterprise should not recognize a loss from this transaction. If the preference is to be preserved and enjoyed immediately, the shareholder should be exempt from tax, but the enterprise should not have a further deduction. There is an intermediate point, however, that would preserve the preference, but at a reduced value. That position would tax the entire $25,000 distribution to the shareholder under the personal income tax, but allow the enterprise to carry forward the $5,000 as a loss against future income.
A further complication would arise if the enterprise had retained profits from prior years.
Example
Assume a corporation has operating profits in the current year of $30,000, but taxable profits of $20,000 (prior to making any dividend payment) because of a $10,000 enterprise tax preference. Its potential tax liability at a 25 percent rate is thus $5,000 if it makes no distributions. The corporation has $2,000 in retained profits from a prior year. The corporation distributes $32,000.
The same questions would arise: What is the treatment of the $12,000 in the hands of the shareholder, and does the corporation generate a carryover loss from this transaction, given that it has a further deduction of $12,000 that exceeds its taxable profit? The added complication is that the $2,000 of retained earnings was presumably already taxed to the corporation the previous year and perhaps ought not be taxed again if distributed now. An ordering rule would be necessary to resolve this question, one that would identify (and perhaps immunize) the amount paid from taxed retained profits and then identify and deal with the $10,000 paid from the preference income.
c. Nonresident Shareholders
One important qualification to the desirability of a dividend-paid deduction system is the position of nonresident shareholders. Current international tax practice is to allocate the enterprise tax to the source country, while a dividend-paid deduction system will effectively abandon any entity-level tax on distributed earnings. Consequently, substantial withholding taxes on a nonresident shareholder would be needed to compensate for the reduction in the enterprise tax. Such an option might be limited by treaties.125
3. Split-Rate Systems
Split-rate systems reduce the enterprise tax payable on distributed profits or formally impose tax only on retained earnings.126 The same effect can also be achieved with a tax surcharge on undistributed enterprise profits.127
A. Resident Individuals
Under a split-rate system, the enterprise pays tax on its retained profits (Tcr), but generally faces a lower rate of tax (Tcd) on the proportion of pretax profits distributed as dividends.128 A resident individual pays income tax at ordinary marginal rates on distributions (Ti), while retained profits are taxed as capital gain (Tg) on a deferred basis to the shareholder, offsetting to some extent the higher rate paid by the enterprise when the profits are earned.129 The after-tax return to a resident individual shareholder is130
R = dP(1 − Tcd)(1 − Ti) + (1 − d)P(1 − Tcr)(1 − Tg).
A lower tax on distributions, combined with an increased personal income tax on distributions, is likely to suffer from some of the same problems alluded to in the discussion of the classical system—it might discourage distributions of enterprise profits. Again, tax-exempt institutions would benefit from the lower tax on distributed capital income.
B. Preference Income
As in a dividend-paid deduction system, distributions of other forms of untaxed income, such as income that enjoys enterprise tax preferences or for eign-source income that is not taxable in the residence country, would be subject to enterprise tax only if retained. The same stacking and ordering issues would also arise. The after-tax return on preference income would be
R = dP(l − Tt) + (1 − d)P(1 − Tcr)(1 − Tg).
c. Nonresident Shareholders
As with a dividend-paid deduction system, split-rate systems can confer benefits on nonresident shareholders that withholding taxes may not be able to offset.131
4. Dividend-Received Deduction (or Dividend-Exemption) System
The systems discussed above all reduce the enterprise-level tax paid on distributions. Dividend-exemption or dividend-received deduction systems operate at the shareholder level,132 by giving the shareholder a deduction from income for some or all of the distributions received or by exempting some or all dividends received from tax.133 This type of system was in place in the United States until 1986 for a limited amount of dividends received by individuals134 and is still retained in many countries as the means for adjusting the total tax paid on dividends flowing through chains of enterprise135 or as a general integration mechanism.136 Because these systems leave the enterprise’s tax liability untouched, they solve some of the problems surrounding tax-exempt and nonresident shareholders mentioned in the prior discussion, but they also raise new issues.
In a dividend-received deduction system, as under the classical system, the enterprise still pays tax on the profits it derives during the year. The shareholder includes in income dividends received. A deduction from income is, however, given to resident shareholders for enterprise distributions received, which may be as much as the amount of dividends received but is sometimes limited.137 In a dividend-exemption system, a percentage of (or all) dividends received are exempt in the hands of the shareholder. In both systems, retained enterprise earnings are taxed to the enterprise and to the shareholder under the capital gains tax, with no adjustment for the enterprise tax already paid.
A. Resident Individuals
If the deduction available to the shareholder is for the entire amount of the dividend received, the position of the shareholder after payment of enterprise and personal income tax on the profits is
R = dP(1 − Tc) + (1 − d)P(1 − Tc)(1 − Tg).
For tax-exempt shareholders, the position now changes from that reached under the systems described above. Because the enterprise tax remains intact and the adjustment occurs under the shareholder’s tax, no benefit is conferred on tax-exempt shareholders through the interaction mechanism—the benefit of their exemption is not increased because no tax deduction is available to a tax-exempt entity.
B. Preference Income and Foreign-Source Income
When a resident enterprise distributes preference income, the value of the preference is retained and passed through to the shareholders—the enterprise pays no tax on this income because of the incentive, and the investor pays no tax because of the dividend-received deduction.
When a resident enterprise is distributing foreign-source income, the effect of the foreign tax credit system (or system of exemption for foreign income) will in most cases replicate the outcome for preference income; that is, the resident shareholder will receive the dividend income free of further (residence country) enterprise tax and is entitled to a deduction for the amount of the dividend received. The position of the shareholder becomes
R = dP + (1 − d)P(1 − Tc)(1 − Tg).
But when the shareholder invests directly in a foreign enterprise, there is an added complication. In many countries, dividends received do not qualify for the dividend-received deduction if the paying enterprise is not also a resident. Where this is the rule, and any withholding tax on the dividend is fully creditable in the residence country, the position of the shareholder approximates the position of a shareholder under the classical system, with the important exception that the enterprise tax paid is the foreign enterprise tax (Tfc), rather than the domestic enterprise tax. The after-tax position thus becomes
R = dP(1 − Tfc)(1 − Ti) + (1 − d)P(1 − Tfc)(l − Tg).
This result need not be the case, of course. The tax system of the residence country might simply include dividends received and then allow a deduction as its means of eliminating double taxation. The effect would be similar to an exemption system for foreign dividends.138
c. Nonresident Shareholders
When a nonresident shareholder invests in a domestic enterprise, the relationship between the (source country) enterprise tax and the (residence country) investor-level tax is principally a matter for the residence country to resolve because the international norm allocates the enterprise tax exclusively to the source country and limits the ability of the source country to impose withholding taxes on dividends. Indeed, some countries will take that position to its logical conclusion and choose not to impose any withholding taxes on dividends paid to nonresidents.
5. Imputation Systems
The four systems just described adjust the double taxation of distributed earnings by effecting changes at either the enterprise or the shareholder level. Many countries now operate tax credit or tax imputation systems139 that retain both the separate enterprise tax and the personal tax but treat the payments of one tax as also satisfying a tax liability arising under the other. They achieve this by giving a tax credit of some amount, either to the shareholder or to the enterprise, reflecting more or less accurately the amount of tax that the profits have already borne.140 This section examines three versions of the wide variety of imputation systems.
Imputation systems should not be confused with simple withholding systems in which the enterprise is obliged to withhold tax on distributions and the tax withheld is credited to the shareholder. The difference between imputation and withholding systems is that a pure withholding system is simply a collection mechanism on behalf of the shareholder, and not an attempt to change the consequences of the separate or classical system. For example, most European jurisdictions see the need both to impose a withholding tax at a constant rate on enterprise distributions and to have some other interaction mechanism, such as an imputation system that attributes payments of the enterprise’s own tax liability to the shareholders or a dividend deduction system. Even the Netherlands, which retains the classical system, has a withholding system in which tax is collected from the enterprise on distributions. The tax is creditable to the shareholder, but does not further reduce the total tax payable by either the enterprise or the shareholder.
Although all imputation systems have elements in common, within this broad framework, there are also many differences. Common to all systems are the survival of the separate enterprise tax, the attribution to shareholders of at least some enterprise tax paid on distributed profits, and the denial of a credit for enterprise tax paid on retained profits. Differences are manifested, for example, in the accuracy with which imputation systems take account of enterprise tax payments.141 In some systems, the amount of tax credited to the individual shareholder may not reflect the total tax paid by the enterprise. At one extreme, the Canadian system simply increases the amount of any distribution by a constant amount to represent enterprise tax paid and then gives the shareholder a credit for a portion of the grossed-up amount. The gross-up and credit occur regardless of whether tax has actually been paid at the enterprise level.142 The United Kingdom’s ACT system is slightly more careful to ensure that the tax has been paid, but occasionally at the expense of collecting payments of ACT that exceed the enterprise’s own “mainstream” (i.e., enterprise) tax liability.143 Of the three systems modeled, the most accurate is that used in Australia and New Zealand. It attempts to track the amount of tax an enterprise actually pays on its profits and attributes only those payments to the profits distributed.
Which system a country chooses to put in place will depend upon many factors, but probably the most important are the desired treatment of enterprise level tax preferences,144 the treatment of exempt shareholders, the preferred treatment of nonresident shareholders, the importance of fairness concerns, the treatment of foreign-source income, and administrative convenience.
A. Automatic Imputation Model
The first imputation system to be described, which is based on the system used in Canada, appears to be the least accurate. It will be seen, however, that except in some unusual circumstances, the alleged accuracy of some systems may be more apparent than real.145 This system increases the shareholder’s distribution by an amount assumed to represent some of the enterprise’s tax payment on the distribution and then gives to the shareholder a credit for a proportion of that assumed enterprise tax. For the purposes of this chapter, the automatic operation of the system is the interesting element of the interaction mechanism.146
In the Canadian system, the enterprise pays tax on its taxable profits, whether distributed or retained, at the enterprise tax rate.147 The shareholder must include in income the amount of distributed profits increased by a multiple representing the enterprise tax that is assumed to be paid on the distribution.148 The factor by which the dividend is increased is set at a constant rate, which is currently 25 percent. The shareholder then pays personal tax on the amount of increased distribution and is given a tax credit against this liability for an amount that is a proportion (currently 66 percent) of the grossed-up amount.149 Retained earnings are taxed to the enterprise, and the balance after enterprise tax (to the extent reflected in the sales price) is taxable to the shareholder on realization as a capital gain.150 All the steps involved in making the interaction between enterprise- and investor-level taxes are effected by the operation of the imputation system at the shareholder level.
1. Resident Individuals
Rather than be distracted by the complexities of the system as it actually operates,151 this section will abstract a little from reality and concentrate on the effect of the automatic interaction mechanism. We therefore treat the imputation process as if the shareholder is given credit under the imputation system at full rates: the gross-up occurs at the full enterprise rate, and the tax credit against the personal income tax occurs in the same amount. If the Canadian system is Owners modeled in this way, and the level of the enterprise tax is lower than the investor-level tax, the after-tax return to the shareholder becomes
R = dP(1 − Ti) + (1 − d)P(1 − Tc)(1 − Tg).
This treatment comes about from the following steps. The shareholder includes in income the amount of the dividend received increased by an amount set by reference to the enterprise tax rate:
dP(1 − Tc) + [dP(1 − Tc) × Tc/(1 − Tc)],
which can be simplified to dP(1 − Tc) + dPTc = dP.
This total is then subject to investor’s income tax (dPTi), and the shareholder is entitled to a credit against the investor income tax liability of the same amount that was included by the grossing-up procedure. The net tax at the shareholder level is thus (dPTi − dPTc). The procedure operates on the assumption that enterprise tax (dPTc) was collected from the enterprise and gives effect to the goal of taxing distributed profits (dP) ultimately at the investor’s income tax rate (Ti) although the tax is collected at two points. No gross-up and credit system operates for retained profits and they are taxed as under the classical system.
Several interesting design questions arise under such a system. The first is the question of surplus credits—what happens when the assumption that the entity-level tax is lower than the investor-level tax is relaxed? That would be the case, for example, with investors who are tax exempt or have carryover losses, and with individuals who are taxed at a low marginal tax rate, most commonly individuals in retirement who are living off the dividend income from their savings. The tax credit is usually conceived as partly satisfying the shareholder’s liability for tax on the dividend income, but the shareholder might not have a tax liability on that income. The following example demonstrates the point:
Example
A corporation pays a dividend of $7,500 to a shareholder who is a resident individual. The shareholder has no other income. The enterprise tax rate is 25 percent. The personal income tax has a tax-free zone of $10,000 a year. Income over $10,000 is subject to a 20 percent rate. The shareholder will report income of $10,000 and will be entitled to a tax credit of $2,500, but will still have no tax liability.
Tax credits that exceed the taxpayer’s current need for credit can be dealt with in many ways. It would be possible to refund the excess to the investor in cash; deny cash refunds, but allow the taxpayer to carry forward any excess credits to future years; allow the taxpayer to transfer (or perhaps even sell) the credit to another taxpayer, such as a related corporation in a corporate group; or deny any further benefit.
A second, though slightly different, issue is spillover—what would happen if the taxpayer had derived other income? Here, the taxpayer could benefit from the tax credit to reduce or eliminate the tax on the other income.
Example
A corporation pays a dividend of $7,500 to a shareholder who is a resident individual. The shareholder has other interest income of $4,000. The enterprise tax rate is 25 percent. The personal income tax has a tax-free zone of $10,000 a year. Income over $10,000 is subject to a 20 percent rate.
The shareholder will report income of $14,000 and will be entitled to a tax credit of $2,500. The shareholders tax liability is $800 (20 percent of $4,000), which could be fully satisfied by the tax credit with a surplus of $1,700. The interest income would be effectively shielded from tax by the credit for corporate tax paid.
While this transaction may seem innocuous, a variation on this example will show how these enterprise tax credits can be used in tax-sheltering activities.
Example
A corporation pays a dividend of $7,500 to a shareholder who is a resident individual. The taxpayer has a deductible interest expense of $10,000 incurred for the purchase of the shares (the taxpayer obviously assumes capital growth in the value of the shares, which is presently untaxed). The shareholder has employment income of $22,500. The enterprise tax rate is 25 percent. The personal income tax has a tax-free zone of $10,000 a year. Income over $10,000 is subject to a 20 percent rate.
The shareholder will report taxable income of $22,500. The shareholder’s tax liability is $2,500 (20 percent of $12,500), and the shareholder will be entitled to a tax credit of $2,500. All the salary income is effectively shielded from tax by the credit for corporate tax paid.
There are several solutions to this problem, assuming it is seen as a problem. One solution, a rule that the tax credits for enterprise tax are quarantined and can be used only to satisfy the tax liability on dividend income, would address both this example and the prior one.152
2. Enterprise-Level Tax Preferences
A second series of issues arises from the automatic nature of the process. An automatic gross-up and credit mechanism automatically passes through to shareholders the benefit of preference items offered to enterprises. This is because distributions of untaxed preference income come with tax credits attached; the enterprise pays no tax on this income because of the preference and the investor pays no tax because of the automatic tax credit. The following example demonstrates the outcome:
Example
The corporation has pretax financial profits of $30,000. It is entitled to a special tax deduction of $6,000 for making an investment and therefore has taxable profits of $24,000. The enterprise tax rate is 25 percent and so the corporation pays $6,000 in tax. The corporation pays a dividend of $24,000 to a shareholder who is a resident individual. The personal income tax rate is 25 percent on income up to $40,000 and 40 percent thereafter.
The shareholder reports as income $32,000—the sum of the dividend of $24,000 and the gross-up for assumed corporate tax on a dividend of that size of $8,000 ($24,000 × 0.25/0.75). The taxpayer has a tax liability of $8,000 (25 percent of $32,000) and a tax credit of $8,000. The total tax paid is $6,000—one corporate tax payment of $6,000 and no further shareholder tax payment.
In other words, the outcome is just as if the investor, like the company, had faced the following tax rates:
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | 25 | 6,000 | |
| Preference income | 6,000 | 0 | 0 | |
| Total | 30,000 | 6,000 | ||
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | 25 | 6,000 | |
| Preference income | 6,000 | 0 | 0 | |
| Total | 30,000 | 6,000 | ||
The same outcome would have occurred if, instead of being a tax incentive, the difference between commercial profit and taxable income had been brought about because the enterprise earned $6,000 of foreign-source income that was treated as exempt in the residence country. It would also have happened if the $6,000 of foreign-source income had been taxable but the enterprise was entitled to a tax credit of $1,500.
If, as is more common in industrial economies at the moment, the investor’s rate is higher than the enterprise rate, the outcome changes in this way:
Example
The corporation has pretax financial profits of $30,000. It is entitled to a special tax deduction of $6,000 for making an investment and therefore has taxable profits of $24,000. The enterprise tax rate is 25 percent and so the corporation pays $6,000 in tax. The corporation pays a dividend of $24,000 to a shareholder who is a resident individual. The personal income tax rate is 30 percent on income up to $40,000 and 40 percent thereafter.
Again, the shareholder reports as income $32,000—the sum of the dividend of $24,000 and the $8,000 gross-up for assumed enterprise tax. The taxpayer now has an initial tax liability of $9,600 (30 percent of $32,000) but still has a tax credit of $8,000. The total tax eventually paid is $7,600—one corporate tax payment of $6,000 and a further shareholder tax payment of $ 1,600.
In this case, the outcome is just as if the investor had faced the following tax rates:
| Income (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | 30 | 7,200 | |
| Preference income | 6,000 | 6.66 | 400 | |
| Total | 30,000 | 7,600 | ||
| Income (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | 30 | 7,200 | |
| Preference income | 6,000 | 6.66 | 400 | |
| Total | 30,000 | 7,600 | ||
The outcome in each case occurs because the computation made at the shareholder level is based not on the amount of enterprise tax paid, but on the enterprise tax rate, and occurs automatically. That is, the shareholder must gross up at the rate of
Tc/(1 − Tc).
There are, of course, other options. One is a nonautomatic tracking system that traces only the amount of tax payments made, a system discussed in more detail below.153 Automatic systems do have the advantages of simplifying somewhat compliance and administration. However, tying the imputation process to a stipulated rate can lead to a problem, sometimes referred to as overintegration. If the gross-up and tax credit are simply a constant proportion of the dividend received, and are not tied to the current enterprise tax rate, the system can become misaligned, for example, when there are multiple enterprise and investor rates:
Example
The corporation has pretax financial profits of $100,000 but has taxable profits of only $94,000. The enterprise tax rate is 10 percent up to $50,000 and 25 percent thereafter. The corporation pays $19,000 in tax ($50,000 × 10 percent plus $44,000 × 25 percent). The corporation pays a dividend of $81,000 to a shareholder who is a resident individual. The investor’s tax rate is 30 percent on all income up to $40,000 and 40 percent thereafter.
The shareholder reports as income $108,000—the sum of the dividend of $81,000 and the gross-up computed as one-third of the amount of the dividend $27,000 ($81,000 × ⅓). The taxpayer has an initial tax liability of $37,200 ($40,000 × 25 percent plus $68,000 × 40 percent) and a tax credit of $27,000, leaving a net liability of $10,200. The total tax paid is $29,200—one enterprise tax payment of $19,000 and a further shareholder tax payment of $10,200. This is an average tax rate of 29.2 percent.
The alignment of rates in the example may appear bizarre, but common circumstances and plausible arguments can lead to these kinds of situations. The low rate in the enterprise tax might be intended as an incentive or concession for small business. The marginal rates in the investor tax might be intended to reflect government goals about progressivity and wealth redistribution. The choice of a constant ⅓ ratio to represent the gross-up could have been made because it is the right gross-up rate for the higher enterprise rate. The ⅓ ratio is too high a gross-up for the lower enterprise rate—for a 10 percent enterprise rate, the correct gross-up should be 1/9—but it might be a deliberate decision intended to ensure that the benefit of the small business rate is permanent, like the tax incentives discussed above, and is not recovered when the small business distributes its profits to its shareholders. But this benefit is itself subject to the proviso that the progressive rate scales in the investor-level tax will be allowed to operate thereafter to recapture some, though not all, of the benefit delivered to high-income earners. This juxtaposition of policies, each of which may have some merit in isolation, explains how profit of $100,000 can become subject to an average rate of 29.2 percent.
3. The Equalization Tax Variant
The problem highlighted above in relation to untaxed enterprise income and the potential for overintegration can be solved, even within the broad parameters of an automatic tax credit system, with a common European variant of the automatic credit process just described.
Many countries in Western Europe—in particular, France, Germany, and Italy—apply an additional withholding amount or “equalization tax” on distributions out of accounting income that have not borne enterprise tax.154 Its principal point is to collect tax on distributed income not taxed at the enterprise level. However, while these systems do try to recapture some enterprise preferences, they need not attempt to levy a compensatory tax on all distributions of economic income. In particular, foreign-source income distributed by a resident enterprise to resident shareholders will typically not trigger the equalization tax.
The system in France is typical of this variant.155 In France, the précompte operates within the framework of the basic automatic imputation system. The enterprise pays tax at 33⅓ percent. Every dividend paid by an enterprise carries a tax credit, the avoir fiscal, of 50 percent of the amount of the dividend. The shareholder grosses up the dividend by the amount of the tax credit and is taxed on the total with an automatic credit in the manner described above.156
Unlike the Canadian variant, however, the automatic process of gross-up and credit is not intended to have the effect of passing enterprise tax preferences through to investors. So, for distributions of untaxed income from domestic sources, and for distributions of profits retained for more than five years, the précompte can apply.157 The rate of précompte varies and operates as a supplement to the actual rate of enterprise tax paid, so that the total of enterprise tax and précompte equals 33⅓ percent—in other words, the amount needed to fund the avoir fiscal.158 The automatic process is unaffected and functions in the usual way at the shareholder level, but the imposition of the précompte at the enterprise level has been interposed to correct for some of the problems noted above.
Example
The corporation has pretax financial profits of $30,000. It is entitled to a special tax deduction of $6,000 for making an investment and therefore has taxable profits of $24,000. The enterprise tax rate is 33⅓ percent and so the corporation pays $8,000 in enterprise tax. The corporation pays a dividend of $20,000. The corporation will be liable to précompte of $2,000 (33⅓ percent of $6,000).
The automatic process can now resume at the shareholder level. The personal income tax rate is 40 percent. The shareholder reports as income $30,000—the sum of the dividend of $20,000 and a further one-half of the dividend—the amount of the avoir fiscal The taxpayer has a tax liability of $12,000 (40 percent of $30,000) and has a tax credit of $10,000. The total tax eventually paid is $12,000—$8,000 enterprise tax, précompte of $2,000, and the shareholder pays a further $2,000.
In order to operate the précompte, the French system requires enterprises to keep accounting records to determine whether the income being distributed has borne enterprise tax at the full rate. However, that process of tracing taxed and untaxed profits raises important administrative questions that recur throughout the remainder of this discussion:
Because the précompte is triggered by payment of a dividend out of untaxed profits, can the enterprise avoid the tax by retaining all profits?
If it does not want to retain all profits, can the corporation choose which profits are being distributed and to whom, allowing a process generally referred to as “streaming”?
If not, what are the “stacking” rules; that is, what rules determine the order in which various types of profits are distributed?
These questions also arise in the examination of the remaining systems, and are discussed below.
B. Advance Corporation Tax Model
An advance corporation tax (ACT) system, modeled on the system used in the United Kingdom, uses a distribution-related tax as both a collection mechanism and the interaction mechanism between the enterprise tax system and the personal income tax.159 The essence of the ACT mechanism is that a flat-rate tax is imposed on the enterprise making a distribution, and this tax is then credited against both the enterprise’s liability for enterprise tax payable on its taxable income and the shareholder’s liability for tax on the distribution. While the system used in the United Kingdom is not actually a withholding tax, at least not for the purposes of international tax treaties, an ACT can best be understood as a withholding tax that is credited twice—once for the benefit of the enterprise making the payment and again for the benefit of the investor receiving the payment.
The ACT system described below also abstracts from reality in order to identify more clearly the major policy choices involved. The mechanism of the system operates in these steps. Each dividend distribution made by an enterprise is subject to ACT at a flat rate, and the enterprise subtracts the ACT payment made during the year from its own liability for enterprise tax on its profits.160 The balance of enterprise tax remaining to be paid after the credit for ACT payment is usually referred to as the mainstream corporation tax (MCT) liability, and it can come about either because the ACT rate on dividends is less than the adjusted enterprise tax rate on distributed profits or because the enterprise has elected to retain some profits. Where the enterprise retains profits, there is no ACT payment and hence no change to the classical system’s consequences for the enterprise and the shareholder. If we assume that the ACT rate (Ta) is less than the enterprise tax rate (Tc), the position of the enterprise after payment of tax is therefore
dP[1 − Ta − (Tc − Ta)] + (1 − d)P(1 − Tc) = dP(1 − Tc) + (1 − d)P(1 − Tc).
1. Resident Individuals
The shareholder is treated in respect of distributed profits in the same way as under other imputation systems. The shareholder is taxed on the net distribution increased by the amount of ACT and then receives a credit for the ACT.161 The shareholder includes in income
dP[(1 − Ta − (Tc − Ta)] × [(1 + Ta)/(1 − Ta)].
Where (Ta) is set at a lower rate than (Tc), the position for distributions is equal to
dP[(1 − Tc)] × [(1 + Ta)/(1 − Ta)].
The shareholder receives a credit equal to the amount of ACT (dPTa). If the ACT rate (Ta) is set at the same rate as the investor’s rate (Ti),162 and the ACT is lower than the enterprise rate, the after-tax return of the shareholder becomes
R = dP(1 − Tc) + (1 − d)P(1 − Tc)(1 − Tg).
There is obviously a lot of importance to be attached to the rate alignments under such a system; that is, what are to be the relative sizes of the ACT rate, the enterprise rate, and the personal tax rate? At the enterprise level, dividends will effectively be taxed at the higher of the two enterprise rates as the following examples of different rate alignments show:
Example
The corporation has pretax financial profits of $24,000. The enterprise tax rate is 25 percent, and so the corporation is in principle liable to pay $6,000 in mainstream corporate tax. The corporation pays a dividend of $18,000 to a shareholder who is a resident individual. The ACT rate is 15 percent of the amount of dividends paid, and so the corporation is liable to pay $2,700 in ACT. The personal income tax rate is 15 percent on income up to $40,000 and 40 percent thereafter.
The corporation pays ACT of $2,700 and MCT of $3,300 ($6,000 − $2,700). The shareholder reports as income $21,176—the sum of the dividend of $18,000 and the $3,176 gross-up for ACT ($18,000 × 0.15/0.85). The taxpayer has a tax liability of $3,176 (15 percent of $21,176) and has a tax credit of $3,176. The total tax eventually paid is $6,000—ACT of $2,700 and MCT of $3,300, with no further shareholder tax payment.
Where, as here, the enterprise rate is higher than the ACT rate, it is the higher enterprise rate that is collected on distributed profits, but the gross-up occurs at the shareholder level only at the lower ACT rate. Indeed, in the United Kingdom, the ACT rate is approximately the same as the rate charged on taxable enterprise profits under the enterprise tax to avoid some of these problems.163
Where the ACT rate is higher than the enterprise rate, the ACT is the amount that is collected before the dividend is received by the shareholder:
Example
The corporation has pretax financial profits of $24,000. The enterprise tax rate is 20 percent, and so the corporation is in principle liable to pay $4,800 in mainstream corporate tax. The corporation pays a dividend of $18,000 to a shareholder who is a resident individual. The ACT rate is 33⅓ percent of the amount of dividends paid, and so the corporation is liable to pay $6,000 in ACT. The personal income tax rate is 33⅓ percent on income up to $40,000 and 40 percent thereafter.
The corporation pays ACT of $6,000 and MCT of 0 ($4,800 − $6,000 = −$1,200). The shareholder reports as income $27,000—the sum of the dividend of $18,000 and the $9,000 gross-up for ACT ($18,000 × 0.33/0.66). The taxpayer has a tax liability of $9,000 (33⅓ percent of $27,000) and has a tax credit of $9,000. The total tax eventually paid is $6,000—an ACT payment of $6,000, no MCT payment, and no further shareholder tax payment.
2. Enterprise-Level Tax Preferences
Because the tax is imposed upon distributions, it is collected whether or not the source of enterprise profits from which the distribution has been paid has borne tax or is even taxable. Nor, usually, does the reduction in the enterprise’s tax liability for payments of ACT generate a refund of enterprise tax if the enterprise distributes more than its taxable profits, or if it is taxable at less than the ACT rate on its profits.164 Effectively, tax is collected from the enterprise at the higher of the ACT rate or the enterprise tax rate on distributed profits and at the enterprise tax rate on retentions. An ACT system can thus generate the consequence that all distributions are reduced by an amount of ACT, while some will be reduced by the enterprise tax rate if that is higher. This outcome is especially important for the treatment of tax preferences, which will reduce the enterprise’s MCT tax liability by reducing either its taxable income or its tax.165 Unfortunately, however, under an ACT system fashioned in this way, these items have no effect on the enterprise’s ACT liability. This problem is referred to as “surplus ACT”; that is, the enterprise can distribute more profit than its own tax payments would indicate. In such a case, surplus ACT is generated on the difference—the amount by which the ACT on distributions exceeds the enterprise’s own MCT liability.
One question that arises is, what should be done with these surplus ACT credits? As was discussed above, with all tax credits it is possible to refund them, allow the taxpayer to carry forward any excess credits to future years, allow the taxpayer to transfer (or perhaps even sell) the credit to another taxpayer such as a related corporation in a corporate group, or simply deny any further benefit. Each option will obviously have different consequences under an ACT system. If the excess credits are lost, it means that tax preferences are effectively recaptured at the enterprise level, but are taxed at the ACT rate, not at the personal or the enterprise rate.
If the enterprise reduces its primary enterprise tax liability, for example, by using domestic tax preferences, the after-tax position of the shareholder remains the same for distributions of declared earnings. But lower enterprise tax means that the enterprise’s managers can attribute the ACT payment on distributions of undeclared earnings toward the mainstream enterprise tax liability on declared earnings. The gross-up and credit procedure occurs automatically as in the Canadian system, but on the basis that ACT has actually been collected on distributions. No enterprise tax will be collected on retained earnings where tax has been successfully reduced, and no ACT will be collected because profits have been retained. Consequently, only capital gains tax will be collected on the sale of the shares.
Example
The corporation has pretax financial profits of $30,000. It is entitled to a special deduction of $6,000 for making an investment and therefore has taxable profits of $24,000. The enterprise tax rate is 25 percent, and so the corporation is in principle liable to pay $6,000 in mainstream tax. The corporation, which could pay a dividend of up to $24,000, decides to pay a dividend of $18,000 to a shareholder who is a resident individual. The ACT rate is 33⅓ percent of the amount of dividends paid, and so the corporation is liable to pay $6,000 in ACT. The personal income tax rate is 33⅓ percent.
The corporation pays ACT of $6,000 and MCT of 0 ($6,000 − $6,000 = 0). The shareholder reports as income $27,000—the sum of the dividend of $18,000 and the $9,000 gross-up for ACT ($18,000 × 0.33/0.66). The taxpayer has a tax liability of $9,000 (33⅓ percent of $27,000) and has a tax credit of $9,000. The total tax eventually paid is $6,000—an ACT payment of $6,000, no MCT payment, and no further shareholder tax payment. The shareholder’s shares will have grown in value by an amount related to the $6,000 retained profits.
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | 25 | 6,000 | |
| Preference income | 6,000 | 0 | 0 | |
| Total | 30,000 | 6,000 | ||
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | 25 | 6,000 | |
| Preference income | 6,000 | 0 | 0 | |
| Total | 30,000 | 6,000 | ||
Under an ACT system, there is no problem with allocating tax credits to particular shareholders or groups of shareholders. There is, however, a different allocation question—about the “spillover” of ACT credits to preference income—which this example demonstrates. ACT is collected on distributions, but can ACT payments be used to offset all MCT liabilities, even the MCT on retained earnings? Or are ACT payments to be quarantined, so that they can be used to reduce the MCT only on distributed earnings?
The example above shows how, by retaining the $6,000 in preference income, the enterprise paid ACT only up to the point where the MCT liability was completely eliminated. Thus, the ACT in this case does not ensure that the correct enterprise tax is actually paid when the enterprise enjoys tax preferences but makes distributions. Rather, the ACT simply permits the enterprise’s managers to reduce the amount of any final MCT to be paid on declared profits. More enterprise tax will be paid under the ACT mechanism when the enterprise proposes to report less taxable profit than the amount of profit (both taxed and untaxed) that it proposes to distribute. But in the reverse situation—when the enterprise’s managers propose to retain the untaxed profits—the ACT system does not recapture preferences.
At the shareholder level, the rate alignment question involves the relationship between the ACT rate and the shareholder’s personal rate. Where the amount of ACT is less than the individual shareholder’s tax liability, the shareholder can be made to report the deficiency and make a top-up payment, as was shown in prior examples. Where the amount of ACT exceeds an individual shareholder’s tax liability, there is a further question about the treatment of the excess ACT paid, as far as the shareholder is concerned. In the United Kingdom, the ACT rate is set at a level equal to the basic personal income tax rate so that the ACT is the effective collection mechanism for the personal tax liability.166 However, for low-income shareholders who receive dividend income, typically retired individuals, the ACT rate may exceed their own personal income tax rate. When this occurs, there is again the issue of the proper treatment of the surplus credits. It is possible to refund them, allow the taxpayer to carry forward any excess credits to future years, allow the taxpayer to transfer (or perhaps even sell) the credit to another taxpayer, or simply deny any further benefit. In the United Kingdom, the ACT credit, if it exceeds the shareholder’s tax liability, is refundable to the shareholder.167 When the ACT rate is set equal to the highest personal rate rather than the lowest, and there is no intention of refunding “excess” credits to low-rate shareholders, the gross-up and credit procedure achieves nothing, and it is possible to simply exempt dividends received from further tax.168
c. Tax-Tracing Model
The final imputation system to be modeled is similar to that used in Australia and New Zealand.169 Of the three imputation systems discussed, it appears to be the most accurate measure of the interaction of the enterprise and personal income tax, at least on distributed income. The Australian system tries to trace tax payments actually made by the enterprise and to attribute tax credits for those payments to individual shareholders only to the extent that verified tax payments have been made by the enterprise. The stylized Canadian system modeled above assumes that enterprise tax has been paid on all distributions; in other words, it disregards the possibility that distributed profits may not have borne tax. The U.K. system forces the payment of a tax on all distributions through the ACT mechanism, even when no MCT is owed. The Australian system traces the tax actually paid by the enterprise on its profits and attributes only tax actually paid to the profits distributed. It does permit untaxed profits to be distributed, but identifies them as such in the hands of the shareholder.
As under the previous imputation systems, the enterprise still pays tax on its taxable income, whether distributed or retained (PTc) and will have a balance available for distribution [P(1 − Tc)]. The net amount distributed to the shareholder [dP(1 − Tc)] is increased by the gross-up that represents corporate tax, effected by multiplying a fraction of the net dividend170 by the factor (Tc/1 − Tc) and adding this amount to the net dividend. A resident shareholder pays income tax at marginal rates on the proportion of after-tax profits distributed by the corporation as dividends,171 and a tax credit is given to the shareholder for the amount of the gross-up.172 Retained profits are still taxed as a capital gain when the shares are sold by the shareholder; no explicit credit against capital gains tax is given for enterprise tax already paid on retained profits—the tax paid on reported profits that are retained is effectively lost.173 In this respect, the Australian imputation system, like the other imputation systems discussed, operates in a way similar to the classical system for retained profits.
1. Resident Individuals
The gross-up and credit procedure operates in several steps, which are recorded through entries in a notional account maintained by the enterprise to trace enterprise tax payments made and consumed. First, the enterprise’s payment of tax on its taxable profits creates for the enterprise a credit in the enterprise’s account (PTc).174 The enterprise can then attach a tax credit of that amount to a dividend. When the enterprise declares a dividend of some portion of the after-tax profits (dPTc), it also reduces the account by this amount, leaving a balance of [(1 − d)PTc].175 In contrast to the ACT system, none of these steps affects the enterprise’s own tax liability—they merely serve as a record, permitting subsequent calculations to be effected.176 The shareholder reports the portion distributed increased by the gross-up for enterprise tax (dPTc).177 However, this gross-up does not occur as a simple increase of the net dividend by a constant rate. Rather, it is calculated on the amount that has been debited to the enterprise’s franking account, which may or may not correspond to the net amount of dividend distributed, as will be shown later. Where the full taxable profits have been declared, this step becomes
dP(1 − Tc) + [(1 + Tc)/(1 − Tc)] = dP.
This amount is then subject to personal income tax (dPTi) and the shareholder is entitled to a credit against the personal income tax liability of the same amount that was included by the gross-up procedure (dPTc).178 The net tax at the shareholder level on distributed dividends is thus (dPTt − dPTc). The aftertax return to the shareholder is
R = dP(1 − Ti) + (1 − d)P(1 − Tc)(1 − Tg).
The total tax is paid in two parts: enterprise tax (dPTc) is collected from the enterprise, and when the enterprise rate is less than the personal income tax rate, the deficit (dPTt − dPTc) is collected from the shareholder.
Example
The corporation has pretax financial profits of $24,000. The corporate tax rate is 25 percent and so the corporation pays $6,000 in enterprise tax. The corporation pays a dividend of $18,000 to a shareholder who is a resident individual. The payment of $6,000 in enterprise tax will be recorded as a credit in the corporation’s tax-paid account, and the payment of the dividend of $18,000 will be a debit of $6,000 to its tax-paid account.
The personal income tax rate is 33⅓ percent. The shareholder reports as income $24,000—the sum of the dividend of $18,000 and the $6,000 debited to the corporation’s tax-paid account. The taxpayer has a tax liability of $8,000 (33⅓ percent of $24,000) and a tax credit of $6,000. The total tax eventually paid is $8,000—corporate tax payment of $6,000 and a further shareholder tax payment of $2,000.
When the amount of enterprise tax paid is more than the shareholder’s own personal liability—for low-income shareholders who receive dividend income, there is again the issue of the proper treatment of the surplus credits—it is possible to refund the surplus results to the shareholder, allow the taxpayer to carry forward any excess credits to future years or to transfer (or perhaps even sell) the credit to another taxpayer, or simply deny any further benefit.
2. Enterprise-Level Tax Preference Income
Even if the two rates are identical, a further shareholder payment also comes about when for some other reason, such as the existence of foreign tax credits or enterprise-level tax preferences, all of the profits distributed by the enterprise’s managers have not borne domestic tax at the full enterprise rate.179 One purpose of the system is to provide tax credits only for enterprise tax actually paid in the country where the enterprise is resident.180 This has the consequence that enterprise-level tax preferences or foreign tax credits are recaptured at the shareholder level.
Example
The corporation has pretax financial profits of $30,000. It is entitled to a special deduction of $6,000 for making an investment and therefore has taxable profits of $24,000. The enterprise tax rate is 25 percent and so the corporation pays $6,000 in tax. The corporation pays a dividend of $24,000. The payment of $6,000 in enterprise-level tax will be recorded as a credit in the corporation’s tax-paid account, and the payment of the dividend of $24,000 will create a debit of $6,000 to its tax-paid account.
The personal income tax rate is 33 percent. The shareholder reports as income $30,000—the sum of the dividend of $24,000 and the $6,000 debited to the corporation’s tax-paid account. The taxpayer has a tax liability of $9,900 (33 percent of $30,000) and a tax credit of $6,000. The total tax eventually paid is $9,900—a tax payment of $6,000 and a further shareholder tax payment of $3,300.
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | |||
| Corporate tax component | 25 | 6,000 | ||
| Shareholder tax component | (33–25) | 1,920 | ||
| Distributed preference income | 6,000 | |||
| Shareholder tax | 33 | 1,980 | ||
| Total | 30,000 | 9,900 | ||
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | |||
| Corporate tax component | 25 | 6,000 | ||
| Shareholder tax component | (33–25) | 1,920 | ||
| Distributed preference income | 6,000 | |||
| Shareholder tax | 33 | 1,980 | ||
| Total | 30,000 | 9,900 | ||
This example involves the same problem of the “leakage” of tax benefits that was discussed in relation to the ACT system: Is it possible to use tax credits from taxed but retained profits to immunize untaxed distributed profits from tax? If we relax the assumption that the enterprise distributes all enterprise-level profits, the treatment of untaxed profits distributed as dividends will depend on how the account is debited and how the tax credits are attached to dividends—an issue similar to that raised both in the ACT system and under the equalization tax variant. It is clear that the balance in the account (PTC) would be insufficient to permit the enterprise’s managers to distribute a dividend with full tax credits greater than [P(1 − Tc)]. But if the enterprise’s managers retain a proportion of the profits [(1 − d)P(1 − Tc)], the “unused” credits in the account (representing tax on taxed but retained profits) can be applied against the undeclared but distributed profits. If so, distributed undeclared profits can also be distributed tax free to shareholders under this system, as they are under an ACT system.
Example
The corporation has pretax financial profits of $30,000. It is entitled to a special deduction of $6,000 for making an investment and therefore has taxable profits of $24,000. The enterprise tax rate is 25 percent, and so the corporation pays $6,000 in tax. The corporation, which could pay a dividend of $24,000, decides to pay a dividend of $18,000 to a shareholder who is a resident individual. The payment of $6,000 enterprise-level tax will be recorded as a credit in the corporation’s tax-paid account, and the payment of the dividend of $18,000 will create a debit of $6,000 to its tax-paid account.
The personal income tax rate is 33⅓ percent. The shareholder reports as income $24,000—the sum of the dividend of $18,000 and the $6,000 gross-up for the amount debited to the corporation’s tax-paid account. The shareholder has a tax liability of $8,000 (33⅓ percent of $24,000) and a tax credit of $6,000. The total tax eventually paid is $8,000—an enterprise-level tax of $6,000 and a further shareholder tax payment of $2,000. The shareholder’s shares will have grown in value by an amount related to the $6,000 retained profits.
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | |||
| Corporate tax component | 25 | 6,000 | ||
| Shareholder tax component | (33⅓–25) | 2,000 | ||
| Undistributed preference income | 6,000 | 33⅓ | 0 | |
| Total | 30,000 | 8,000 | ||
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Nonpreference income | 24,000 | |||
| Corporate tax component | 25 | 6,000 | ||
| Shareholder tax component | (33⅓–25) | 2,000 | ||
| Undistributed preference income | 6,000 | 33⅓ | 0 | |
| Total | 30,000 | 8,000 | ||
3. Streaming of Taxed and Untaxed Dividends
The systems in operation actually allow more flexibility than this simple matter of leakage would suggest—hence, the comment above that the accuracy of the crediting mechanism may be more apparent than real. That flexibility raises an issue that is commonly referred to as “streaming”—that is, directing the tax credits to shareholders who can use them most advantageously. The examples above assume that the enterprise’s managers will (and must) debit the tax-paid account with its full credit balance if the size of the dividend being paid permits them to do so. But if such a rule does not exist—in other words, if the enterprise’s managers have discretion about how much of the credit balance in the account to use and when—the possibility of streaming arises. For example, if the enterprise has shareholders in both high and low tax brackets, it might try to direct the credits predominantly to the former group (and not to the latter, where excess credits at the shareholder level might be unusable). Various devices and techniques would be needed, but having shares with differential rights or declaring successive dividends might be feasible tools, especially in closely held companies.
To reduce this problem, the systems in place try to ensure that all dividends carry the same proportion of credits, where there are insufficient credits to cover all dividends to be declared in a year.181 But this rule does not apply if the enterprise’s managers plan to make distributions from undeclared profits up to the amount of retained declared profits, that is, if the total distribution is less than the balance in the account. This would mean that all dividends, whether out of declared or undeclared profits, up to that amount could effectively be distributed tax free to shareholders. If, however, the enterprise’s managers plan to distribute all of the declared profits and some portion of the undeclared profits, the rule does apply and all dividends will carry only fractional credits.182 The enterprise’s managers can effectively attach tax credits to a distribution up to an amount of taxed profits regardless of whether some portion of the amount distributed has actually borne tax. If there are insufficient credits, the gross-up and credit procedure described above will still operate for taxed profits but not for untaxed profits. If the profits are not taxed, they carry no tax credit, and the shareholder simply includes the distributed portion of untaxed profits in income with no gross-up or credit and is taxed in the same way as under a classical system.
4. Nonresident Shareholders
The position of nonresident shareholders raises a few novel questions in the context of an account-based imputation system.
One reason for the imputation systems described above is that benefits can be, although they need not be, confined to resident shareholders. But if benefits are so confined, the issue will arise as to whether the tax-paid account must be debited in the case of dividends paid to nonresident shareholders because the shareholders will derive no benefit from the tax credit. Indeed, this is an area where streaming could be expected to occur—allocating all the credits for enterprise tax paid to resident shareholders where the system does not afford any benefit to nonresident shareholders.
International experience in this area is not uniform, although most, but not all, countries choose to confine the benefits of their imputation system to resident shareholders. One interesting exception is Singapore, which levies no additional withholding tax on distributions at all, whether to residents or nonresidents. In addition, some of the treaties negotiated by France, Ireland, and the United Kingdom levy withholding taxes, but allow partial or full credits to flow to nonresidents by refunds in cash.
For example, Australia’s tax treaties with France, Ireland, and the United Kingdom allow Australian resident shareholders who are individuals some access to foreign imputation credits. Article 9(6) of the Australia-France treaty provides that an Australian resident who is an individual and receives a dividend from a company resident in France is entitled to a payment from the government of France equal to 85 percent of the tax credit (avoir fiscal) that would be attached to the dividend if received by an individual resident of France. Fifteen percent remains in France by way of withholding tax.
6. Full Integration System
The last option to be explored is a full integration system for enterprises.183 An integration system operates at the shareholder level and attributes the enterprise’s income, whether distributed or not, to the shareholders who are taxable on all the enterprise’s profits. There are two varieties of full integration systems. One, usually referred to as the partnership version of the integration system, implies that no tax is imposed on the enterprise’s profits, unlike the other systems already discussed. Under the other version, where the enterprise remains taxable, the enterprise’s tax is then credited to the shareholders as a credit against their liability on the attributed profits.
Schedular integration systems that are not based on a single rate are intended to offset the effect of the enterprise income tax entirely so that all enterprise profits are ultimately taxed at individual marginal rates in the current year, regardless of whether the profits are distributed. This system promises the model treatment to which the other systems aspire, because all profits are taxed at exactly the shareholder’s personal income tax rate (although portions of the total tax might be collected from both the enterprise and the shareholder), and there is no gain to the taxpayer from deferring the recognition of income by retaining profits within the enterprise.184
The so-called partnership-style integration achieves this result by eliminating the enterprise tax altogether and taxing the shareholders as if they were in partnership—all enterprise profits are included in the individual’s taxable income. The United States permits shareholders to elect this treatment under Subchapter S of Chapter I of the Internal Revenue Code for domestically controlled corporations with few shareholders and little foreign-source or passive income.185 One consequence of the election is that the benefits of corporate losses and tax preference items are passed through to the shareholders.186 For the reasons discussed in section II above, this style of integration is generally considered unfeasible as a model for all enterprises and will not be considered further in this chapter.
With this exception and for the reasons referred to earlier, no country has adopted a full integration system for the taxation of domestic enterprises and their resident shareholders, despite the support of many commentators and several government reports.187 But, somewhat surprisingly, a second style of integration system is more common for taxing nonresident enterprises controlled by resident shareholders, where the system is usually referred to as a controlled foreign corporation (CFC) tax system. In this context, the system is used not because it approximates the economist’s ideal of eliminating the double taxation of enterprise profits, but rather as an antiavoidance mechanism to prevent the accumulation of untaxed passive or tax-sheltered income offshore.188 This section considers a theoretical CFC-type system, but for domestic enterprises.189 Such a system is not in use in any domestic tax system, but the one described here would achieve the central element of an integration system, taxing shareholders currently on all declared enterprise profits, but with the innovation of retaining the enterprise tax as a pure withholding mechanism.
A. Resident Individuals
For this integration system, it is assumed that the enterprise still pays tax on its profits and that the individual shareholders pay income tax at progressive marginal rates on all the taxable profits of the enterprise, whether or not they are distributed. A tax credit is then given to the shareholders for the entire enterprise tax paid. Again, a decision would have to be made as to whether the tax credit mechanism traces actual payments of enterprise tax or operates automatically. Any retained profits are taxed to the shareholders at the appropriate personal rate when earned (and appropriate credits are also attributed). The shareholder’s cost in the shares is increased by the amount of profit taxed to the shareholder, to avoid double taxation when the shares are disposed of.190 Any further capital gain beyond the value of retained taxed earnings is taxed in the usual way as capital gain. When the enterprise’s managers report and pay tax on the enterprise’s full profit, the after-tax position of the shareholder is
R = P(1 − Ti).
This result would be achieved in several steps. First, enterprise tax (PTc) is collected from the enterprise. The amount of any distribution is included in the shareholder’s income together with the usual gross-up for enterprise tax:
dP(1 − Tc)[(1 + Tc)/(1 − Tc)] = dP.
This approach generates a tax liability at the personal income rate (dPTi), and the shareholder receives a credit (dPTc) against this tax liability for the enterprise tax paid. The element that makes this system different from those described earlier is that retained earnings and a further gross-up for enterprise tax on the earnings would also have to be included in the shareholder’s current assessable income:
(1 − d)P(1 − Tc)[(1 + Tc)/(1 − Tc)] = (1 − d)P.
This creates a tax liability of [(1 − d)PTi and a credit of [(1 − d)PTc] is set off against the tax liability.
The capital gains tax is retained to capture items not taxed on a current basis, such as unrealized enterprise profits, tax preferences, or stock market gains.191 A further adjustment is necessary to reflect the fact that some of the retained profits reflected in the price of the shares will already have been taxed. The adjustment involves annually increasing the shareholder’s cost in the shares by the amount of retained earnings taxed in that year.192 If the taxpayer realizes only the accumulated value of retained taxed profits, the shareholder’s basis equals this amount and no capital gain arises.
The annual increase in the shareholder’s cost in the shares comes about through a series of steps.193 Given that the enterprise’s managers have distributed some after-tax profits, the enterprise still retains an amount [(1 − d)P(1 − Tc)] on which enterprise tax has already been paid. The shareholder’s cost in the total retained earnings is calculated in the following way. First, the shareholder’s cost is increased by the amount of taxable profits remaining after enterprise tax [P(1 − Tc)]. Then the shareholder’s cost is reduced by the amounts already “liberated” from the enterprise for the benefit of the shareholder—the distributed taxed profits and the tax attaching to all profits. Thus the taxpayer’s basis in the earnings is increased by [(1 − d)P(1 − Tc)]. The system operates in this fashion:
Example
The corporation has pretax financial profits of $24,000. The enterprise tax rate is 25 percent and the corporation is liable to pay $6,000 in enterprise tax. The corporation pays a dividend of $15,000 to a shareholder who is a resident individual and retains $3,000. The personal income tax rate is 30 percent on income up to $40,000 and 40 percent thereafter.
The corporation pays tax of $6,000. The shareholder reports as income $24,000—the sum of the dividend of $15,000 and a tax credit attached to it of $5,000 ($15,000 × 0.25/0.75) and the retained earnings of $3,000 and the tax credit attached to it of $1,000 ($3,000 × 0.25/0.75). The investor is liable to gross tax of $7,200 and has a total tax credit of $6,000. The shareholder’s cost is increased by $3,000.
B. Preference Income
The treatment under such a system of the untaxed enterprise profits, such as enterprise-level tax preferences or foreign income, raises several policy issues. The decision about the crediting mechanism will be the basis for the answer. If the decision is made that preference income is washed out at the shareholder level, then only actual tax payments, rather than an automatic credit, should be used for computing the tax credit. The shareholder’s cost is increased by the amount of taxable profits (P) remaining after enterprise tax [P(l − Tc)] and reduced by the amount of profits distributed without further personal income tax [dD(1 - Tc) + d(P − D)], up to the amount of the taxable profits. Thus, the taxpayer’s basis in the earnings is increased by only [(1 − d)D(1 − Tc) − d(P − D)]; that is, untaxable but distributed profits effectively reduce the increase in basis by the amount distributed. This means that the capital gains tax calculation becomes
{(1 − d)D(1 − Tc) + (1 − d)(P − D)−[(1 − d)D(1 − Tc)−d(P − D)]}(1 − Tg),
which becomes
[(P − D)(1 − Tg)].
The eventual after-tax position of the shareholder becomes
R = D(1 - Ti) + (P − D)(1 − Tg).
Example
The corporation has pretax financial profits of $30,000. Because of a special incentive, it is entitled to a special tax deduction of $6,000. It therefore has taxable profits of $24,000. The enterprise tax rate is 25 percent, and so the corporation is liable to pay $6,000 in tax. The corporation pays a dividend of $15,000 to a shareholder who is a resident individual and retains $3,000 of its taxable profits. The personal income tax rate is 30 percent on income up to $40,000 and 40 percent thereafter.
The corporation pays tax of $6,000. The shareholder reports as income $24,000—the sum of the dividend of $15,000 and a tax credit attached to it of $5,000 ($15,000 × 0.25/0.75) and the retained earnings of $3,000 and the tax credit attaching to it of $1,000 ($3,000 × 0.25/0.75). The investor is liable to gross tax of $7,200 and has a total tax credit of $6,000. The shareholder’s cost is increased by only $3,000.
If the enterprise distributes some of the untaxed profits, the deficiency will be recaptured in the taxation of dividends. But because the tax preference income is not taxable income, it is difficult to see how it could be attributed to shareholders and taxed on a current basis. Therefore, it remains to the capital gains tax to collect tax, albeit deferred, on this income.
Example
The corporation has pretax financial profits of $30,000. Because of a special incentive, it is entitled to a special tax deduction of $10,000. It therefore has taxable profits of $20,000. The enterprise tax rate is 20 percent, and the corporation is liable to pay $4,000 in tax. The corporation pays a dividend of $24,000 to a shareholder who is a resident individual and retains $2,000. The personal income tax rate is 30 percent on income up to $40,000 and 40 percent thereafter.
The corporation pays tax of $4,000. The shareholder reports as income $28,000, the amount of the dividend and the tax credit attached to it of $4,000 (because of the tracing process). The investor is liable to gross tax of $8,400 and has a tax credit of $4,000. The shareholder’s cost is not increased so that the retained untaxable earnings of $2,000 will be taxed as capital gain only. Total tax paid is $8,400, $4,000 by the corporation and $4,400 by the investor.
In this case, the outcome is just as if the investor had faced the following tax rates:
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Distributed nonpreference income | 20,000 | 30 | 6,000 | |
| Tax on nonpreference income | 4,000 | 30 | 1,200 | |
| Distributed preference income | 4,000 | 30 | 1,200 | |
| Retained preference income | 2,000 | 0 | 0 | |
| Total | 30,000 | 8,400 | ||
| Amount (In units of domestic currency) | Rate (In percent) | Tax (In units of domestic currency) | ||
|---|---|---|---|---|
| Distributed nonpreference income | 20,000 | 30 | 6,000 | |
| Tax on nonpreference income | 4,000 | 30 | 1,200 | |
| Distributed preference income | 4,000 | 30 | 1,200 | |
| Retained preference income | 2,000 | 0 | 0 | |
| Total | 30,000 | 8,400 | ||
c. Nonresident Shareholders
Although on the basis of the discussion in previous sections it might be thought that a full integration system is eminently desirable, its apparent virtues are subject to one major and probably insuperable impediment—the difficulties presented in taxing nonresident shareholders.
As has been mentioned already, the general consensus that has developed on the appropriate international allocation for taxing enterprise profits is that the source country is able to tax the enterprise in full and has limited rights to tax dividends paid out of those profits. A system of full integration challenges this consensus by attempting to tax the nonresident shareholder on undistributed profits, an option that tax treaties do not explicitly countenance.194 It is unclear whether the source country has the right to tax this amount prior to distribution, what rate would be applied, and, correspondingly, whether the shareholder could insist that the tax system in its country of residence give relief for the tax so collected.
Taxing undistributed profits is a challenge not only to the existing tax base orthodoxy, but also to a tax administration. The ability to tax resident shareholders on undistributed income is facilitated by having both the enterprise and the shareholder as residents—any top-up tax on undistributed profits can be collected when the resident files a return. For nonresident shareholders, the tax administration would have to collect both the profit tax and the tax on undistributed profits from the enterprise directly because there is no dividend to tax, and the shareholder is not necessarily within the jurisdiction of the source country’s tax administration. While this result is technically feasible, it is not clear what rate should be applied because there is no real information about the marginal tax rate applicable to the nonresident.
Moreover, as a practical matter, in a world where taxes are an important factor in decisions about locating real investments, no country can afford to be the sole country to tax resident enterprises on such a basis.
VI. Distributions
A. Typology of Distributions
For the purpose of this section, distributions are defined as any payment made by an enterprise to its shareholders with respect to the shareholders’ capital investment. Distributions can take various forms, the most common of which are amounts paid by companies as dividends and amounts paid either to repurchase company shares or to purchase the shares of a subsidiary of the company. However, inventive finance and tax experts are constantly developing new techniques for making company distributions to shareholders. In addition to assuming various different forms, distributions can have different economic origins. They can be paid out of profits that have been taxed at the company level, out of profits that have not been taxed at the company level, or out of no profits at all (meaning, they constitute a return of capital).
B. Tax Consequences of Distributions from Different Origins
The tax consequences of a distribution arising from one of these three different origins will vary significantly depending on the type of tax system in place. One constant among income tax systems, however, is that shareholders do not include as income distributions that constitute a return of capital. In addition, shareholders whose tax base includes capital gains and losses on the sale or transfer of their shares must make a downward adjustment to their share cost in an amount equal to such a distribution.195 Therefore, all tax systems are concerned with whether a distribution constitutes a return of capital.
In addition, fully integrated tax systems are concerned with techniques whereby enterprises may declare taxable bonus shares or with other techniques that allow shareholders whose tax base includes capital gains and losses on the sale or transfer of their shares to, in effect, increase the cost of their shares by the amount retained. In addition, in integrated systems where the enterprise tax is at a higher rate than that of at least some shareholders, these techniques may also allow the shareholder to receive a credit for the difference between the two rates.196
The different treatment accorded distributions made from taxed income and from untaxed income will vary depending on a number of factors. The most important is whether there is an integrated enterprise-shareholder tax, where the enterprise rate is equal to or higher than the shareholder rate. In such a system, distributions of income fully taxed at the enterprise level need not be taxed at the shareholder level, although depending on the integration system the shareholder may be entitled to a refund of all or part of the accompanying credit. However, as discussed in the previous section, any distribution from income that was either not taxed at all or not fully taxed at the enterprise level raises the question of whether tax should then be levied.
Obviously, this question can arise only when enterprises are not taxed on a base that closely approximates their economic income or at a rate that is less than the top shareholder rate. As was discussed at length in the previous section, the tax system can either not tax distributions from untaxed income or can tax them in some way. Some of the more common techniques employed to tax distributions from untaxed or partially taxed income include levying a compensatory-type tax on payment at the enterprise level, levying tax on receipt at the shareholder level, or combining the two techniques into a hybrid system.
In jurisdictions without full integration, the issue is different still. These jurisdictions impose tax on distributions from both taxed and untaxed income, although typically at the same rate. While the reasoning for imposing double taxation on enterprise income is not particularly compelling, it requires levying a tax on all distributions other than those that constitute returns of capital. Withholding or shareholder-level taxes, or a hybrid of both, can be used to levy this additional tax on the distribution.
C. Implications of Different Tax Consequences for Distributions
As can readily be seen, it would benefit an enterprise’s shareholders if it could make distributions of untaxed or partially taxed income without drawing additional tax. For this reason, and depending on the rules the particular jurisdiction has in place, enterprises may attempt to disguise distributions that draw additional tax as distributions that do not draw additional tax. For example, in integrated systems, an enterprise may try to make a dividend look as if it were paid from previously taxed income. In both integrated and unintegrated systems, an enterprise may try to make a dividend or a redemption appear as if it constitutes a return of capital.197 And, as noted earlier, inventive finance and tax experts are constantly developing new techniques for making enterprise distributions to shareholders, techniques that may not be adequately addressed by existing rules or that may not be sufficiently understood by hard-pressed tax administrators.
D. Simplified Systems
Given these incentives for enterprises to avoid tax, and the inventiveness with which they may try to do so, it would greatly simplify the design and implementation of enterprise-shareholder tax systems if it were unnecessary to tax distributions. If nearly all income were taxed at the enterprise level in a fully integrated system, the question would not arise as to whether any distributions should be taxed because all distributions either would be paid out of taxed income or would represent returns of capital to shareholders. And, if a single-rate schedular tax at the shareholder level equal to and integrated with the tax of the enterprise level were applied, then no distribution would need to be treated as taxable as income by the shareholder.198 In such a simplified system, the only tax effect a distribution would have would be on those shareholders subject to capital gains tax, who would have to determine whether the distribution were a return of capital; if so, they would be required to adjust downward the cost of the share by the amount of the distribution.
In a nonschedular integration system, it would be necessary to distinguish only between distributions carrying imputation credits (which would be included as income, along with a credit) and those that did not (which would be treated as a return of capital).199
E. Rules for Distinguishing Between Distributions of Income and Returns of Capital
In the absence of a schedular system with a single tax rate, it may sometimes be necessary to determine whether a distribution constitutes a return of capital. In addition, in a system where a fair amount of income can escape enterprise tax or where integration is absent or incomplete, there is an incentive to describe distributions as nontaxable returns of capital. However, in a system where substantially all enterprise income is already fully taxed, and where there is complete integration, there is no such incentive. In fact, if anything, there may in some cases be an incentive to disguise distributions of capital so as to avoid reducing the adjusted cost of the share for capital gains purposes. While the latter incentive would presumably be considerably less of a problem than the former, it would still be helpful to have simple techniques for determining what constitutes a return of capital and what does not.
Unfortunately, most jurisdictions have systems in which a fair amount of income can escape enterprise tax, or where integration is nonexistent or incomplete, and are therefore more concerned with proving distributions to be taxable than not to be taxable. One technique for policing distributions is to rely, in effect, on the operations of corporate law.200 Corporate law governs the circumstances and manner by which a company may make distributions to its shareholders. Under the corporate law of many jurisdictions, distributions to shareholders are subject to a number of restrictions designed to protect the rights of creditors. In the most restrictive company law regimes, distributions are restricted to dividends paid out of company profits (as determined by special corporate accounting rules), and to the redemption of certain limited types of stock (usually preferred);201 other types of stock redemptions, including the purchase of shares in a subsidiary, are prohibited. In these cases, only the price paid for the redemption of preferred stock would be treated as a return of capital.
However, corporate law rules concerning shareholder distributions, at least in many jurisdictions, are being liberalized. The repurchase of nonredeemable shares is now often permitted, as is, in some jurisdictions, the payment of dividends out of capital. In these cases, it may still be possible to rely on the corporate law rules to define a return of capital for tax purposes. In particular, the rules would have to determine when a dividend is not made out of company income and how much of the purchase price of a share buyback would have to be considered a return of capital.
A helpful modification of this approach may be to combine corporate law rules with special tax rules, particularly with regard to determining a corporation’s income. If comprehensive income tax rules are applied, taxable income can, for example, be substituted directly for traditional definitions of corporate “profits.”202 However, the rules concerning what constitutes a repayment of capital in the case of stock redemptions would continue to apply.
In a fully integrated system that effectively taxes nearly all of a company’s income, such rules should be relatively easy to apply.
F. Complex Systems
However, it is a different story in systems that do not capture most income through the enterprise tax, or where the shareholder-level tax is equal to and integrated with the tax at the enterprise level. Considerable additional care on the part of tax administrations will be required if they wish effectively to capture distributions from untaxed income (in integrated systems) or from both untaxed and taxed income (classical systems). This is because the incentives to make otherwise taxable distributions look like nontaxable distributions will be greater. In these cases, corporate law rules may be too easily manipulated and may require additional tax rules to prevent tax avoidance. For example, if there is untaxed income at the enterprise level, it will always be preferable to make distributions to shareholders through redemptions if those payments are treated as returns of capital and therefore not taxable.
G. Examples
As described in the previous section, France has a partial imputation system.203 It levies a compensatory tax on distributions out of income that is not fully taxed at the enterprise level. The tax system does so in what is essentially a two-step process. It determines first if the distribution is from profits (whether taxed or untaxed) and next if the distribution is from income already subject to full tax. If so, the précompte mobilier is applied.
Essentially, any distribution to shareholders (other than bonus shares that represent capitalization of reserves or earnings) is deemed to be out of profits unless it qualifies for treatment as a redemption or a liquidation.204 The enterprise keeps track of what profits it has retained and which have borne full tax, and a stacking rule provides that distributions come first from after-tax profits, and then from untaxed profits.205
The French rules on the treatment of redemptions are rather complicated. Where redemptions are permitted, a portion of the distribution may be deemed to be taxable. A number of steps must first be followed to determine what portion. For those shareholders not subject to capital gains tax, the portion of the distribution that exceeds the greater of the shareholder’s actual gain or the amount of the share’s paid-in capital is taxable as a dividend, up to the extent of the enterprise’s accounting profits. All amounts paid out of untaxed profits are subject to the précompte mobilier. A different, and more complicated, rule applies to shareholders subject to capital gains tax.
These tax rules appear to allow enterprises to borrow against appreciated assets and then to use the proceeds of the borrowing to pay exempt dividends or to make exempt redemptions. However, French corporate law mitigates these options substantially. It prohibits companies from making distributions of dividends except out of accounting profits, and severely restricts the ability of companies to make redemptions. However, if corporate law were to change, so too might these conclusions.
As described in the previous section, the United Kingdom has only a partial imputation system, and uses the ACT as its primary technique for capturing distributions from untaxed income. Under this system, it is necessary first to determine which distributions are from income (whether taxed or untaxed), for those distributions will attract ACT. Distributions from capital do not attract ACT. Next, it is necessary to determine whether the distribution is from taxed income. This is done by keeping track of total taxes paid and by assuming that distributions are made first from taxed income and then from untaxed income. A partial credit for the ACT is then given against the corporate tax paid. If the ACT credit were given in full, the net effect would be that only those distributions from untaxed income would be subject to tax.
The U.K. law includes in the definition of taxable distribution any dividend allowed under corporate law, as well as any other distribution unless defined as repayment of capital.206 U.K corporate law allows dividends to be paid from unrealized capital gains, which are taxable. However, a dividend from capital is not permitted. As a general matter, amounts returned to the shareholders in a redemption of capital in excess of the paid-up capital allocable to the shares in question are deemed to be noncapital distributions.207 There is also a rule that where a company repays share capital and, at any time thereafter, issues any share capital as paid up other than through the receipt of new consideration, then the amount so paid up will be treated as a distribution.208 A distribution of bonus shares is not treated as a distribution, although the bonus shares may be subject to what is in effect a special tax at the shareholder level.209
The U.K. rules do capture borrowings against appreciated assets that are paid out to shareholders as dividends. However, the rules also make it possible for enterprises to turn otherwise taxable distributions into nontaxable returns of capital through share redemptions. While corporate law limits the ability of companies to make redemptions, the opportunities are still greater than under French law.
Canada maintains a partial imputation system without levying additional tax on any distributions. This is due in part to the presumption that distributions to shareholders in the form of dividends have already born enterprise level tax. Dividends are therefore deemed to have been paid out of taxed income, except those that are deemed to be returns of capital.210 Because the imputation system is only partial, additional tax may be due at the shareholder level on dividends. However, if the distribution is deemed to be from capital, there will be no additional tax.
The term “dividend” is not defined by statute, but has been interpreted by the courts and the tax administration as meaning any distribution except as an authorized reduction of capital. In addition, the statute defines dividends to include stock dividends. The statute treats all or a portion of distributions made during share redemption or reduction in capital as nontaxable returns of capital. The amount treated as a dividend is the amount distributed in excess of the paid-up capital allocable to the shares in question.
H. Taxable Bonus Shares and Constructive Dividends
Enterprises may wish to retain earnings rather than to distribute them to their shareholders. These retained earnings will be reflected in an increase in the value of the enterprise’s shares. If these earnings have been subject to tax, the increase in shareholder value will represent already taxed gains. In a fully integrated tax system, there will be a tax disincentive for retaining these earnings unless the shareholder who is subjected to capital gains taxation is not taxed on these gains. Therefore, such systems typically allow enterprises to take measures to ensure that such shareholders are not so taxed. In addition, in integrated systems where the enterprise tax is at a higher rate than for some shareholders, these techniques may also allow the shareholder to receive a credit for the difference between the two rates.
Two typical methods include allowing enterprises to declare either taxable bonus shares or what has sometimes been termed “constructive dividends.” Taxable bonus shares are typically shares paid as dividends that represent capitalized earnings through the issuing of additional shares of stock,211 although there is no particular tax reason why capitalization under company law should be required. The value of a dividend distributed as a bonus share equals the proportionate amount of capitalized earnings. The result is a decrease in the value of the existing shares equal to the cost of the new shares, which is itself equal to the amount of retained earnings. Systems must ensure only that the bonus share represents after-tax income.212
Another technique is to allow enterprises to declare “constructive” dividends.213 These are notional dividends that are declared but not actually paid and are designed to allow shareholders to increase their share cost by the amount of the retained earnings. This can be effected by an enterprise simply reporting to a shareholder the per share amount of after-tax income the enterprise has retained; a shareholder subject to capital gains taxation can then increase cost by this amount.214
VII. Defining Which Business Enterprises Should Be Subject to Separate Corporate Tax
An enterprise tax law must spell out, usually at the beginning of the statute, which entities are subject to tax. As with the individual income tax, a distinction must be drawn between residents and nonresidents, nonresidents typically being taxed only on income sourced in the jurisdiction. The definition of residence is discussed in chapter 18.
One definitional technique that is often used in civil law countries is to rely on an entity’s legal status. Under this approach, if an entity is considered a legal person under the civil code, then it will be subject to enterprise tax. This rule may then be supplemented by listing specific forms of legal persons that are subject to tax, listing as taxpayers certain entities that are not legal persons, and excluding certain legal persons from tax. For example, the German corporate income tax law lists the most common types of commercial companies, adds “any other legal persons under private law,” as well as certain entities (such as Stiftungen) that may not be legal persons, and also includes enterprises administered by entities that are legal persons under public law (even when the enterprise may not itself be a legal person).215
The French approach is broadly similar. The law lists certain forms of company and then refers to “any other legal person carrying out an exploitation or operations of a profit making nature.” It then lists certain forms of companies that are subject to corporate tax on an elective basis.216
Although, as illustrated in the above examples, corporate tax laws in civil law countries typically start from the status of entities as legal persons, these countries do not uniformly subject entities to corporate tax if and only if they are legal persons.217
Common law countries take different approaches. Canada relies on legal personality, imposing the income tax on any “person.”218 The United States imposes the corporate income tax on “every corporation,” but corporation is defined as including “associations.”219 In turn, the regulations have adopted a test of corporate resemblance, holding that entities with sufficient corporate characteristics are taxed as associations. Hybrid entities can now elect whether to be treated as a corporation or as a partnership.220
In the United Kingdom, corporation tax is imposed on “profits of companies.”221 Company “means … any body corporate or unincorporated association but does not include a partnership, a local authority or a local authority association.”222 This differs from the U.S. approach in that partnerships cannot be recharacterized as associations and therefore treated as corporations. In addition, some noncorporate entities such as unit trusts are taxed in essentially the same manner as companies, only at different rates and with more complete integration.
Some transition countries treat as taxpayers under the corporate tax not just legal persons, but separate divisions of legal persons.223 This practice arises from the treatment of these divisions as separate enterprises under the former command economy. The fact that these enterprises were not separate legal persons may have been of little importance in the past. However, their treatment as separate taxpayers under the profit tax can be problematic. In particular, how can systems designed to tax dividends operate when the dividends are paid not by each separate division but by the legal person? How are transfers of property among divisions to be accounted for? Although taxing divisions separately may not fit very well with a market economy-type corporate tax, there has in some countries been resistance to changing the system of taxing divisions separately. The divisions may be accustomed to keeping separate accounts, and tax officials may also be accustomed to auditing and dealing with divisions separately (corruption may be involved here). Local governments may be used to receiving their share of the revenues from the divisions located in their jurisdictions (they are often entitled to a share on this basis under laws governing the division of revenues from taxes). Eventually, however, as revenue-sharing laws are adjusted, it can be expected that these special rules treating divisions as separate taxpayers will be abandoned.
The opposite issue is consolidation of taxpayers. Consolidation is allowed, for example, in the United States under extremely complicated rules. A few transition countries also allow consolidation.224 Generally, countries whose tax system is not highly developed should steer clear of allowing consolidation.
Some transition countries impose a tax not on legal persons or corporations but on enterprises, which in some cases can include sole proprietorships. For example, in Latvia, taxpayers of the enterprise income tax are defined as enterprises, with a cross-reference to the Law on Taxes and Fees.225 That law in turn defines as resident an entity that is “registered” in accordance with the legislation of Latvia.226 The enterprise income tax excludes from the definition of taxpayer individual enterprises that are not required to submit annual reports in accordance with the Law on Annual Reports of Enterprises. Therefore, sole proprietorships that are required to submit such reports are taxed under the enterprise income tax. And in Vietnam, the new Business Income Tax applies generally both to individuals engaged in production and trade and to business entities. The law itself does not provide for flow-through treatment for partnerships, thereby leading to some confusion when business is carried out in partnership form (particularly when the partners themselves are companies). Who is the taxpayer in that case? The partnership, the partners, or both? The enterprise income tax law of China likewise taxes enterprises, which are defined as state-owned enterprises, collective enterprises, private enterprises, joint-venture enterprises, and “any other organizations deriving income from production and business operations and other income.”227
Imposing tax on enterprises as described in the preceding paragraph can be faulted for lack of clarity. The basic problem is that “enterprise” is generally not a clear legal concept.228 It is much better technique for the law to refer to legal persons because it will be clear whether an entity is a legal person.229 However, one can see the counterargument. If everyone carrying on a business is required to register as an enterprise, it seems an attractive proposition to tax separately each registered enterprise, regardless of its legal status.230 Again, the same arguments can come up as with corporate divisions. Enterprises may be registered locally. An administrative mechanism may have grown up around the concept of enterprise registration. The basic problems with this approach are that (1) a single legal or physical person may have more than one registered enterprise or branch, and the boundaries around these enterprises may be difficult to draw; and (2) a person may carry on a business without registering it. Using instead the concept of legal person provides for greater certainty because it derives from the legal personality of the taxpayer as defined in the civil code.
The definition of taxpayer also needs to specify exemptions. Government agencies, but not government-owned enterprises, are typically exempt. Also typically exempt are various forms of nonprofit organization, whose definition will differ from country to country. When a system of incorporation and registration of such organizations exists outside the tax law, it may be possible to simply make a cross-reference, rather than to put all the necessary qualifications into the tax law. It is necessary to determine which agency (e.g., the tax agency or some other licensing agency) will be responsible for ensuring that the entities in question qualify as nonprofit. While some countries completely exempt certain organizations from tax, others tax nonprofit organizations on their business income if they carry on a business that is not related to their nonprofit purpose. The United States has developed quite detailed rules and practices on what is known as “unrelated business taxable income.”231 A more aggressive approach would be to tax nonprofits not only on their business income but on all their business and investment income. One advantage of such an approach is that it is not necessary to distinguish between business and investment (e.g., how would rental activity be classified?). Whether such an approach is taken is very much a political decision because of possible reluctance to impose tax on entities that are considered to be carrying on good works.
VIII. Concluding Remarks
This chapter began with a discussion of the merits of an income tax system including a separate enterprise tax and continued with recommendations as to how such a tax should be structured. It elaborated a number of arguments in support of a system that taxes enterprise income once, at the highest shareholder marginal rate, and that collects such tax to the greatest extent possible at the enterprise level. In addition, it advocated an enterprise-level tax that sought to capture, as accurately as possible given practical constraints, all income as it accrued and at the same tax rate.
The arguments favoring such a system were primarily rooted in economics, that such a system was likely to result in the fewest distortions and would allow the market to function with greater efficiency. However, an important by-product would be that the system would be far simpler to administer and also considerably less prone to tax avoidance. The primary reasons for this are that, with nearly all economic profits taxed at the enterprise level, there is no need to levy dividend taxes or for rules to determine what consitutes a distribution. Incentives to make nonequity payments would be greatly reduced.
In addition, where there is little untaxed income at the enterprise level, there is a corresponding reduction in the need to tax capital gains at the shareholder level: more of any share’s increase in value due to the enterprise’s economic income will already have been subject to tax at the enterprise level. This reduces the need to administer a capital gains tax at the level of the individual shareholder (including having to provide for the adjustment of the cost of shares for amounts of retained earnings or distributions of capital), a difficult undertaking in industrial countries and correspondingly more difficult in developing or transition economies.232
Next, rules providing for the stacking of income or against the streaming of distributions become unnecessary because all income bears the same rate of tax. Finally, the chapter argues that if the enterprise tax is final, meaning that the tax on enterprise income is schedular, there is additional improvement in administrative ease and a reduction in tax avoidance possibilities because the tax system does not need to tax distributions at the shareholder level.
It may be argued that it is impossible to implement a completely effective enterprise-level tax on economic income. Even if this turns out to be the case, to the extent that preference income can be reduced by ending as many intentional enterprise-level tax preferences as possible (such as investment credits, accelerated depreciation, and the like), and deferred capital gains can be reduced by marking to market as many assets as possible for which objective values can easily be ascertained (such as precious metals, foreign exchange, quoted securities, derivatives, and the like), and by including in enterprise income total borrowings net of written-down asset value, the need to capture distributions from untaxed economic income will be reduced. This would mean, in effect, that some or all of the elaborate mechanisms described in section V to capture untaxed income or to allow shareholders to be taxed at marginal rates might raise so little additional revenue as to be necessary only in occasional cases. It would also mean that any of the elaborate rules described in section VI to distinguish among different types of distributions would become similarly less necessary or less important. With significant progress having been made toward such a simplified system, the methods described in those sections can be selectively enacted, or applied, or both, as required.
It may also be argued that it is politically difficult to tax enterprises at the highest shareholder rate. But to the extent that enterprise income and shareholder income were taxed at as close to the same rate as possible, any incentive by enterprises to retain earnings or to make nontaxable distributions to take advantage of the lower enterprise rate would be reduced.
Finally, tax designers may also argue that it is unfair to implement a schedular tax on enterprise income. However, to the extent that a nonschedular tax can be limited to the fewest taxpayers possible, the need to file returns, or for enterprises to shift ownership to those taxed at lower rates, would be reduced.
Therefore, even if complete adherence to a simplified system is impossible, there is still considerable merit in designing a tax system with as many features of the simplified system described above as possible.
While these arguments apply in varying degrees to all economies, they have particular relevance to developing countries and to economies in transition. Even in these jurisdictions, many enterprises, particularly larger companies or companies with foreign management, may have developed considerable tax-planning expertise. The globalization of sophisticated tax planning ability, and therefore tax avoidance, has been a remarkable—and another perhaps unexpected—consequence of the general globalization of markets and financial information. The authors have experienced, in a number of cases, developing and transition countries with complex systems, in which a surprisingly large amount of tax administration resources were dedicated to attempting to prevent sophisticated schemes designed to avoid income tax. However, because of an inadequacy of resources, these tax administrations were less likely to be able to design and implement the rules necessary to operate their complex systems without diverting administrative resources from other tasks. These other tasks, while perhaps more mundane, were also more likely to be productive in the collection of needed revenue.
Therefore, in these circumstances it is perhaps best to design the most effective simple system possible, and to direct limited bureaucratic resources not to trying to capture the relatively meager income that will escape through the tax avoidance net, but to more productive, if less intellectually challenging, activities.