Income tax discrimination: still stuck in the labyrinth of impossibility.
ESSAY CONTENTS INTRODUCTION I. EFFICIENCY AS THE NORM II. THREE EFFICIENCY STANDARDS FOR INTERNATIONAL TAX NEUTRALITY Ill. THE ASSUMPTION THAT TAXPAYERS CANNOT CHANGE RESIDENCE IV. THE NORMATIVE CASE FOR COMPETITIVE NEUTRALITY V. THE TAX LAW REQUIREMENTS OF COMPETITIVE NEUTRALITY VI. THE PROPOSED CONSTITUTIONAL VERSION OF COMPETITIVE NEUTRALITY VII. THE POSITIVE CLAIM FOR COMPETITIVE NEUTRALITY VIII. COMPETITIVE NEUTRALITY AS A WAY OUT OF THE LABYRINTH OF IMPOSSIBILITY CONCLUSION
The foundational treaties of the European Union establish a unique system of government. (1) In general, they leave decisions about how to levy income taxes and at what rates to the member states. If the member states agree unanimously--a rare occurrence indeed--the European Commission, Council, and Parliament can together issue income tax directives, but so far these few directives have been limited to rather technical matters. (2) Within Europe, as elsewhere, cross-border transactions involving income taxation are also governed by an extensive network of bilateral income tax treaties that, while reflecting many common principles, often vary in their details. (3)
In this context, the European Court of Justice (ECJ) is charged with ensuring, to the extent appropriate and practicable, that the member states' income tax laws do not interfere unduly with the "four freedoms" guaranteed by the Treaties: free movement of goods, (4) services, (5) labor, (6) and capital. (7) These freedoms of movement were intended to create an economic market relatively free of internal barriers, as well as greater social and political union within Europe.
There is considerable tension inherent in this structure, in which each member state retains a veto over European income tax legislation, including proposals that would promote the cohesion of the internal market, while the ECJ reviews the tax laws of the member states to ensure that they do not violate the Treaties' guarantees of free movement. The national income tax laws at issue vary across the Union, generally providing an important source of revenue and implementing national distributive and economic policy goals in light of each member state's economic and social conditions, as well as internal political dynamics and conflicts. Variations in income tax laws and rates across Europe affect taxpayers' decisions about where to work, live, and invest, as well as tax planning efforts about where to locate income and deductions to minimize income tax burdens. (8)
In the 1980s, the European Court of Justice began deciding income tax cases with an aim to strengthening the Union and to limiting the member states' ability to favor their own residents or to favor domestic over foreign investments? Although there is considerable doctrinal confusion in the decided cases, the essential construct used by the ECJ to achieve its goals is the concept of discrimination against cross-border transactions as compared to purely domestic transactions. (10) While a number of commentators, including us, have criticized these decisions as being incoherent in terms of tax policy, doctrinally confusing, sometimes conflicting, and constitutionally questionable in terms of democratic decisionmaking, those decisions have no doubt contributed to the economic, social, and political union in Europe?' In recent years, the European Union has expanded to twenty-seven members, enlarging the membership of the court and creating even greater diversity among the member states' economies and income tax laws. After that expansion and the rejection of a proposed European constitution, the court has become less aggressive in striking down aspects of member states' income tax laws, accepting justifications offered by the member states that in earlier times the court would have rejected. (12)
Competition in Europe has had a notable effect on income tax structures and rates in the member states. Although efforts to harmonize income taxes in Europe to alleviate downward pressures have been advanced since the 1960s, (13) such harmonization efforts have proved unavailing; income tax rates and bases differ markedly throughout the EU. Indeed, one recent quantitative study concludes that "tax competition is stronger in the EU than the rest of the world." (14)
In two previous articles, we criticized the court's income tax jurisprudence. (15) We argued that its reliance on nondiscrimination as the basis for its decisions did not (and could not) satisfy commonly accepted tax policy norms, such as fairness, administrability, economic efficiency, production of desired levels of revenues, avoidance of double taxation, fiscal policy responsiveness to economic circumstances, inter-nation equity, and so on. In addition, we argued that the court could not achieve coherent results by requiring nondiscrimination simultaneously in both origin and destination countries when goods, services, individuals, or capital move from the first country to the second. With regard to the latter point, we contended that-in the absence of harmonized income tax bases and rates--the court had entered a "labyrinth of impossibility." (16)
Let us restate what we mean by this labyrinth of impossibility. (17) The following three principles cannot hold simultaneously in a consistent and coherent way in the absence of harmonized tax bases and rates:
Principle 1--Sovereignty in the Origin Country: The origin country (18) can choose how and at what rates to impose income taxes on its citizens or residents. The origin country, for example, may decide to tax all of its residents or citizens (including individuals, resident corporations, and other business entities, such as partnerships) at progressive rates based on their ability to pay, as measured by their total worldwide income, with whatever personal or family allowances the origin country deems appropriate.
Principle 2--Sovereignty in the Destination Country: The destination country (19) can choose how and at what rates to impose taxes on income earned within its borders. The destination country, for example, may decide to tax individuals (or corporations) on income earned there regardless of whether the earner is local or foreign.
Principle 3--Nondiscrimination: To implement the freedoms of movement, equal treatment of domestic and cross-border income-producing labor, capital, and business activities is required in all member states in their capacity both as countries of origin and destination. (20)
A simple example may help to illustrate the conundrum that the ECJ faces. Assume that one country, let us call it the United Kingdom, taxes income at a 40% rate, while another country, which we shall call Hungary, taxes income at a 15% rate. As a destination country, the United Kingdom may tax Hungarians (as well as Britons) working there at its 40% rate. But this, of course, means that there is an additional tax burden on Hungarians earning income in the United Kingdom, compared with Hungarians earning income at home. Hungarians are taxed differently when they move from Hungary to the United Kingdom. Taking a job in the United Kingdom is not "free" movement for Hungarians (at least in the sense of being costless), even if Hungary does not tax its citizens residing abroad. There is clearly no obligation for Hungary to reimburse its citizens or residents working in the United Kingdom for the 25 additional percentage points of income tax they face in the United Kingdom, nor does any country do so. Making either Hungary or the United Kingdom compensate Hungarians for the additional 25 percentage points of tax they pay when they work in the United Kingdom would violate sovereignty in the origin or destination country, respectively. Now consider a U.K. national who works in Hungary. Forcing Hungary not to tax Britons would violate Hungary's destination-based sovereignty, so Hungary will typically impose its 15% tax on income earned by foreigners working there. (21) Forbidding the United Kingdom from taxing the income earned by a British person in Hungary would violate the United Kingdom's origin-based sovereignty (and the ECJ has said that member states have no obligation under European law to prevent international double taxation (22)), but if the United Kingdom imposes its 40% tax on the income earned in Hungary-even if it allows a deduction or credit for the Hungarian tax paid-Britons working in Hungary will bear a higher tax burden than Hungarians working there.
While rate differences are important, it is not only tax rates that produce such disparities between cross-border and purely domestic activities. Consider, for example, a charitable organization, organized in one country (the country of origin) that has some activities in a second country (the country of destination) when the two countries have different criteria for favorable income tax treatment, such as eligibility for a charitable deduction or exemption from income taxation. (23) Exemption in the destination country might, for example, be based on the premise that the activities of the organization relieve that country from having to provide certain public services to its residents. (24) From the perspective of the origin country, requiring the organization to qualify with additional requirements in the destination country would burden cross-border activity more than domestic activity. From the perspective of the destination country, there is, however, no extra burden because all charities operating there must satisfy the same criteria. Attempting to eliminate barriers to cross-border activity by requiring nondiscrimination from the perspective of both the origin and destination countries does not therefore provide an answer to the question of whether a guarantee of free movement across borders means that such an organization must or must not meet the requirements of the destination country in order to operate there. Rather, a court presented with that question must choose between the two perspectives.
As a final tax example, consider countries that reduce the burden of double taxation of corporate income by providing a credit to shareholders for corporate taxes paid on corporate earnings that are then distributed to the shareholders as dividends taxed again at the shareholder level)s When the company operates in one country (the destination of the capital) and the shareholder resides in another (the origin of the capital), which country should extend the credit? If the first country refuses credits to foreign shareholders, it is arguably discriminating against foreign investors. If the second country refuses credits for shares in a foreign corporation, it is arguably discriminating against investment in foreign countries. On the other hand, if both countries grant credits, cross-border commerce will be favored over domestic commerce. Our point is simply that requiring nondiscrimination in both destination and origin countries is not a satisfactory tool for resolving the conflicts between nonharmonized income taxes and the four freedoms (or, indeed, for resolving other basic issues of international taxation). (26)
These kinds of cases are not, of course, limited to taxation. More than thirty years ago, the ECJ famously decided that Germany could not refuse importation of a French liqueur because it did not meet a requirement of minimum alcohol content applicable to both domestic and foreign products in Germany. (27) To do otherwise, the court decided, would restrict the free flow of goods produced in another member state, infringing what is sometimes called the principle of mutual recognition. (28) More recently, the European Commission, Council, and Parliament have struggled to find an acceptable answer to the question of when certain service providers (such as plumbers) licensed in one member state should be permitted to offer their services in other member states with different licensing standards. (29) The list could go on and on. (30)
This is not to say that there is no way out of the labyrinth we have described. In fact, there are three ways. One might, for example, choose to impose only the requirements of the origin country. Europe does this, for example, with drivers' licenses and certain professional qualifications. (31) Alternatively, one might impose only the requirements of the destination country. European value-added taxation does this, ceding the power to tax goods and services to the destination country. (32) Finally, one could harmonize the origin and destination countries' tax laws and rates. But, absent harmonization, one simply cannot have both origin and destination income taxation along with consistent neutrality or equality between cross-border and domestic activity from the perspective of both origin and destination countries.
Nevertheless, Ruth Mason and Michael Knoll claim that they have discovered a single normative nondiscrimination criterion-which they label "competitive neutrality" - that the ECJ should be using to decide income tax cases under the Treaties. (33) In addition, they aver that this norm in fact best describes the criterion that the court is using to decide such cases. They insist that their competitive neutrality criterion offers the ECJ a way out of the labyrinth of impossibility, and, for good measure, contend that it would also enable the Supreme Court of the United States to craft sensible and coherent doctrine implementing the Privileges and Immunities and Commerce Clauses of the U.S. Constitution. For Mason and Knoll, competitive neutrality is the holy grail of tax discrimination.
The argument of their article can be summarized in eight propositions, which we will consider in turn. (1) The paramount role of the tax nondiscrimination principle in common markets such as the European Union or the United States is to promote economic efficiency. (2) Three familiar efficiency standards for the taxation of capital income can appropriately be translated into efficiency standards for taxing labor income. (3) It is reasonable to analyze tax discrimination on the unrealistic assumption that taxpayers cannot change their state of residence. (4) Given that assumption, one of the three efficiency standards, competitive neutrality, is superior to the other two. (5) Properly understood, competitive neutrality requires a precise set of tax rules that courts have no power or ability to promulgate. (6) Nonetheless, courts should interpret tax nondiscrimination to require a particular partial version of competitive neutrality in common markets such as the EU or the United States. (7) The foundational treaties of the European Union and the Constitution of the United States, as well as the decisions of the European Court of Justice and the U.S. Supreme Court, are best understood as actually imposing that partial requirement of competitive neutrality. (8) Because competitive neutrality would impose obligations on both origin and destination countries, it provides a way out of the labyrinth of impossibility.
Mason and Knoll advance their conclusions about tax discrimination by analyzing decisions in several ECJ labor income cases, decisions that, as they say, are relatively easy to explain. (34) The exact scope of their claims is, however, difficult to pin down. The grand title of their article and its opening pages seem to promise a general examination of tax discrimination in the European Union and United States. However, at some points of the article, capital income seems to be excluded from consideration, and the subject of the article appears to be restricted to labor income. (35) But other parts of their article are definitely not restricted to labor income. When, for example, Mason and Knoll say that 10% of the cases decided by the ECJ are tax cases and imply that their analysis and recommendations apply to all such cases, (36) they certainly are not limiting their claims to labor income cases. (37) Likewise, when they argue that the U.S. Supreme Court should adopt their preferred efficiency norm, (38) they discuss business cases and do not limit their conclusions to labor income cases. (39) In order to understand the implications of the analysis, we will therefore treat their article as making both a weak claim (tax discrimination in labor income cases should be interpreted to further a particular efficiency norm in the EU) and a strong claim (tax discrimination in all income tax cases should be interpreted in that way in common markets such as the European Union and the United States).
I. EFFICIENCY AS THE NORM
Mason and Knoll view the ECJ as engaged in promoting the internal market within Europe and explicitly adopt economic efficiency as the most important norm for deciding tax discrimination cases. (40) (They also urge the same criterion for the U.S. Supreme Court in such cases. (41)) To us, this is much too restrictive a focus for constitutional courts. We obviously agree that efficiency is an important perspective from which to examine judicial decisions involving economic issues. On the other hand, the ECJ's institutional role cannot be so narrowly cabined, particularly in tax cases, where its decisions can constrain a fundamental sovereign power and will often result in the assignment of a part of a tax base (and the resulting revenue) to one member state, rather than another.
Nor do we agree with Mason and Knoll that the ECJ has declared economic efficiency to be the most important underlying value in resolving these tax cases. (42) While we have criticized the court for its institutional role in democratic decisionmaking, as well as for not sufficiently taking into account legitimate national fiscal and tax policy considerations beyond discrimination, (43) the ECJ has clearly been concerned with more than just efficiency, striking down, for example, barriers to free movement to further greater political union across Europe. Moreover, although the court continues to insist that it will not consider its decisions' impact on member states' revenues, it has (as we predicted it might (44)) recently given more weight to member states' defenses grounded in fiscal and administrative concerns, such as preventing taxpayer abuses of domestic tax regimes. (45)
Curiously, given Mason and Knoll's controversial decision to make one particular dimension of economic efficiency the paramount consideration for ECJ (and U.S. Supreme Court) decisionmaking, they fail to advance any strong normative case for doing so. Contrariwise, notwithstanding their assertions that their preferred efficiency standard "would promote the welfare of Europeans," (46) they insist that they are not advancing any claim that the economic efficiency criterion they urge the courts to adopt "would do a better job of promoting economic welfare or any specific notion of the good, justice, or fairness than other possible interpretations that might be imposed on the member states." (47) We are puzzled, given this, why one should endorse their proposals. Nevertheless, in the interest of understanding their argument, let us for now accept their focus on economic efficiency to see where it leads them.
II. THREE EFFICIENCY STANDARDS FOR INTERNATIONAL TAX NEUTRALITY
As Mason and Knoll note, understanding the tax nondiscrimination principle as essentially promoting economic efficiency is insufficient to give content to that principle because there are competing notions of efficiency. (48) Consider a U.K. individual or company operating a manufacturing plant in Germany. Should the plant's income be subject to taxation by Germany, by the United Kingdom, or by both? From the perspective of welfare economics, the argument for taxation by only the United Kingdom, the origin country (or residence country in tax parlance), is that taxation would not distort the owner's decisions about where to locate the plant, achieving what is known as "capital export neutrality." If the destination (or source) country, Germany, also imposed a tax, capital export neutrality could also be achieved if the United Kingdom granted a tax credit for German taxes (including taxes in excess of the U.K. rate, which no country does).
The parallel argument from the perspective of welfare economics for taxation by only the destination (or source) country, Germany in this case, is that the tax distortion between consumption and saving would not depend on the taxpayer's country of residence, achieving what is known as "capital import neutrality." Source-only taxation would also put domestic- and foreign-owned plants on an equal footing in terms of income taxation.
As Mason and Knoll acknowledge, unless countries harmonize their tax systems and rates, capital export neutrality and capital import neutrality cannot be accomplished simultaneously. (49) We obviously agree, as we have used the impossibility of simultaneously implementing capital export and import neutrality as one example of the impossibility of requiring nondiscrimination simultaneously in origin and destination countries. (50)
In recent years, the economists Mihir Desai and James Hines have argued that ownership is another important economic choice that may be distorted by taxation. (51) Their analysis emphasizes that the large portion of foreign direct investment occurring through mergers and acquisitions had largely been ignored in the literature and focuses on productivity gains potentially available through multinational firms. Consider, for example, a U.K. and a German company competing to buy a French enterprise. Suppose that, in the absence of taxation, the U.K. company had such superior management attributes and other valuable intellectual property in the relevant sector that it would outbid the German company, because the French enterprise would be most productive under U.K. management. If the U.K. tax system distorted the outcome so that the German bidder prevailed, worldwide economic welfare would be diminished. Such a tax system would fail to achieve "capital ownership neutrality," which would be accomplished by exclusive source-country taxation. (52)
Mason and Knoll import these three familiar ideas from discussions of capital income taxation into the domain of labor income taxation, relabeling them "locational neutrality" (no efficiency gains from "shifting workers across jurisdictions"), "leisure neutrality" (which obtains when "leisure is allocated efficiently across jurisdictions" because the relevant choice for a worker is between work and leisure, rather than between saving and consumption), and "competitive neutrality" ("when it is not possible to increase productivity by shifting jobs among people"). (53)
They treat the analogue between capital and labor taxation as completely natural, requiring little or no justification for transferring efficiency criteria that have been developed in the former context into the latter. But this is not nearly as straightforward as Mason and Knoll imply. The efficiency norms applicable to capital have been developed largely in the context of multinational corporations deciding where to invest. When labor is at issue, individuals are deciding where to live, whether and where to work, and how much (including how hard) to work. It is, for example, quite reasonable to assume that corporations maximize profits, regardless of their level of income, while individuals are generally thought to have declining marginal utility of income. To be sure, individuals' work efforts-and, as a result, individuals' productivity-are affected by tax burdens, but through both income and substitution effects, making it inappropriate to take workers' productivity as fixed in analyzing different tax regimes. (54) This alone makes Mason and Knoll's sharp distinction between "leisure neutrality" and "competitive neutrality" problematic. Taxi drivers, for example, decide how much to work each day and are well-known to work less when it rains or on holidays. (55) Should we think of these decisions as work-leisure tradeoffs (implicating leisure neutrality) or as affecting productivity (implicating competitive neutrality) ?
The difficulties of insisting on a sharp distinction between individuals' work-leisure tradeoffs and those involving productivity are well illustrated by an economic analysis of the behavior of bike messengers in Zurich, Switzerland in response to a temporary wage increase. (56) In this study, the wage increase produced a large increase in the overall labor supply of the bike messengers, who could freely choose how many shifts to work and how much effort to generate. It turned out, however, that the messengers worked longer hours but decreased their effort per shift, as measured by the number of deliveries per shift, (57) thus confounding leisure neutrality and competitive neutrality in Mason and Knoll's framework. The authors concluded that insights from behavioral economics (which Mason and Knoll ignore completely) offer better explanations of the messengers' behavior than those of neoclassical economics. (58) A variety of behavioral factors, such as the performance of fellow workers and workplace relations between the employer and her employees, affect individuals' decision about both how much and how hard to work. (59) These kinds of considerations are not similarly applicable to corporate decisionmaking about how to allocate capital.
Moreover, individuals cannot simultaneously work full-time in, say, both France and Germany, whereas corporations have great flexibility about their allocation of capital between the two locations. (60) Acquiring a new corporate affiliate in a foreign country does not usually require shedding one elsewhere.
Mason and Knoll's decision to apply capital ownership neutrality to workers is especially problematic. (61) Desai and Hines ground their case for this norm on productive synergies that may be achieved by multinational companies, especially through economies of scope and scale resulting from their intellectual property. While workers may vary in their productivity-a particular pastry chef, for example, may be more productive than another if hired by a particular pastry shop-it is far from obvious that the scope and magnitude of the differences among workers are comparable to those that Desai and Hines postulate for large multinational corporations. Moreover, the productivity of workers in a particular country depends on the amount and type of capital available to workers there, as well as the efficiency of the firm's owners. This makes the location of capital important to the productivity of workers. It is unclear how variations in productivity due to the location or movement of capital affect Mason and Knoll's analysis.
In addition, the capital ownership analysis of Desai and Hines is based on very precise (and controversial) conditions, such as the assumption that a dollar of investment that goes abroad will be offset by a dollar of investment that comes in from abroad. Although Mason and Knoll are aware of these assumptions of the concept they import, (62) they do not indicate whether they are making comparable assumptions about, for example, an exact balance between workers entering and leaving a country.
Mason and Knoll also ignore tax differences between multinational corporations and individuals that should be taken into account in applying analyses of capital income taxation to labor income. For instance, individuals' tax burdens and marginal tax rates on their taxable income often turn on family characteristics, including, for example, the number and ages of children and spousal income. In addition, individual tax systems frequently have progressive tax rates, whereas the corporate rates applicable to large multinational firms are usually flat. Mason and Knoll generally ignore these differences and simply assume that the Desai and Hines framework can be applied to individuals without modification. All of their tables and figures, for instance, assume flat-rate taxation, even though graduated rates are a key feature of labor income taxation.
Finally, the three norms discussed do not, of course, exhaust the dimensions along which either capital or labor income taxation can distort decisions. For example, one could imagine a capital or labor "residence neutrality" norm, designed to reduce tax distortions of a corporation's decision of where to establish residency or an individual's decision of where to reside. Indeed, in previous work, Ruth Mason has herself recognized the importance of tax distortions to where individuals may choose to live. (63) Here, Mason and Knoll avoid consideration of this particular distortion simply by assuming that taxpayers do not change their country of residence, an assumption that we consider below.
Despite our substantial reservations about simply importing capital neutrality norms into the taxation of labor income, we now turn to their analysis based on those norms.
III. THE ASSUMPTION THAT TAXPAYERS CANNOT CHANGE RESIDENCE
As the economists Peter Diamond and Emmanuel Saez have recently emphasized, moving from either a theoretical mathematical model or from calculated examples to policy recommendations--as Mason and Knoll do-requires that the "result should be reasonably robust to changes in the modeling assumptions." (64) Mason and Knoll do not provide any precise model from which they derive their claims. Instead, the mathematical heart of their analysis is a series of examples, two-by-two matrices and tables that explicitly assume taxpayers cannot change their country of residence. (65) Even as to the weak claim of the paper, limited to labor income in the EU, this key assumption is unrealistic. Although the precise rules vary somewhat among the member states, an individual generally changes residence for tax purposes in the EU if she spends more than 183 days in another member state in any given year. As the European Commission puts it on its website, if you "spend more than 6 months in a year in another EU country, you will, in most cases, become a tax resident of that country." (66) Mason and Knoll therefore base their entire analysis on an assumption that EU citizens will not take a job in another country if they have to live in the other country for more than six months. While labor mobility may historically have been more limited in the EU than the United States, (67) the addition of Eastern European member states has increased wage differentials, and, as a result, lower-income workers have increasingly crossed borders in recent years. (68) In addition, migration of highly educated (including high-tech) workers has long been of concern to many governments, including those in Europe. (69) As for mobile, high-income workers, including star soccer players among others, there is ample evidence that Europeans are quite willing to change their residence, specifically in response to lower tax rates. (70)
Whatever one thinks of the assumption that taxpayers cannot change their residence for purposes of the paper's weak claim, that assumption is patently implausible for purposes of any stronger claim. Mason and Knoll do not suggest any reason to suppose that owners of capital (whether individuals or companies) are unable to change their residences (or places of incorporation). Nor do they offer any reason to believe that American workers do not move their residence from one state to another. (71)
Given its implausibility, why do Mason and Knoll assume that residence is fixed? The justification offered is that cross-border discrimination for workers can come about only when a worker resides in one member state and works in another, because moving across the border would subject a worker to residence taxation in the state of his new home. (72) While the situation of "cross-border workers" (the term used by the authors) (73) who do not move is undoubtedly a subject worthy of analysis, that subject is remarkably confined when measured against the grand project promised by the article. Of the ECJ tax discrimination cases, fewer than a quarter (or only about 2% of the ECJ's total cases) involve labor income. (74)
Having made the assumption of fixed residence, Mason and Knoll analyze a series of cases using two-by-two matrices in which workers face different tax regimes. The results in each case depend on the assumption of fixed residence. (75) These matrices do indeed represent the situation of cross-border workers, but we do not see how any general implications about tax discrimination can be drawn from them. The assumption of fixed residence obscures the risks of tax-induced distortions to choices of residence, which should be included in any realistic analysis of the economic efficiency consequences of tax discrimination. As Mason and Knoll put it in their discussion of pure residence taxation, their assumption that residence is fixed relieves them from even considering the very real possibility that individuals or corporations shifting their residence across jurisdictions may increase output. (76) It also means that the different residence-based tax rates in their examples are of no relevance for cross-border workers' decisions about where to work, since no one can change their residence to take advantage of lower rates. (77)
The assumption also obscures the possibility that the location of jobs may change, making it impossible in the absence of a formal model to sort out the effects of such a change from a change in which worker is performing the job. (78) The entire analysis of the article is concerned with enhancing the ability of more productive workers to cross borders to work, but we know that many jobs may move across borders. If one is focused on income tax burdens on labor productivity in Europe, as Mason and Knoll are, one should surely consider cases where production may be relocated. Indeed, in today's global economy, capital moves faster than labor, and there may well be only a relatively small minority of tasks that can be performed only in a specific location. (79) Moreover, the productivity of workers, which is the factor that Mason and Knoll regard as of overriding importance, depends crucially on how much real capital they have to work with. (80) Clearly, Mason and Knoll do not intend their analysis and conclusions to be limited to migrant workers crossing borders temporarily to immobile jobs, such as the olive pickers who come to the Mediterranean states each fall, but such workers are representative of the cases that they analyze and where their crucial assumptions actually hold. (81)
Although the assumption that residence never changes effectively confines the actual analysis of the article to cross-border workers, (82) we want to continue considering the article's argument on its own terms, so we now turn to the case for competitive neutrality, keeping in mind that assumption.
IV. THE NORMATIVE CASE FOR COMPETITIVE NEUTRALITY
Since Mason and Knoll's norms of locational neutrality, leisure neutrality, and competitive neutrality lead to different income tax design decisions, how should a policymaker or judge choose among them? Because these norms depend on a welfare economics framework, the choice presumably should depend, at least in part, on how much distortion occurs along each dimension, which, in turn, depends on the underlying supply and demand conditions, including the relevant elasticities. For example, one important long-standing result in economists' comparisons of capital export and capital import neutrality is that the former is more apt when the supply of capital in source and residence countries is fixed, while the latter is more apt if the demand for capital is fixed. (83) When neither of those extreme conditions is met, countries are often described as compromising between these two efficiency norms. (84)
In any event, a policy decision based on an economic efficiency standard should be grounded in evidence as to the magnitude of the various distortions. Accordingly, economists who prefer capital export neutrality over capital import neutrality generally believe that tax-induced locational distortions are greater than tax-induced savings distortions. (85) So, for example, in addressing the question whether the United States should forgo residence-based taxation in order to achieve capital import neutrality in the interests of reducing economic distortions, the staff of the Joint Committee on Taxation has concluded that a "tax rate on outbound investment lower than the tax rate on domestic investment can only increase economic welfare if the improvement in efficiency from the increase in saving is greater than the reduction in efficiency from the misallocation of savings." (86) Not surprisingly, when Mihir Desai and James Hines proposed capital ownership neutrality (on which Mason and Knoll's concept of competitive neutrality for labor income rests), they emphasized empirical studies of the behavior of multinational corporations, suggesting, in that context, that tax-induced ownership distortions are large and important. (87) Surprisingly, Mason and Knoll concede that, even within the context of the three neutrality benchmarks they discuss, many economists would view violations of locational neutrality as having the largest negative welfare consequences. (88)
Peter Diamond and Emmanuel Saez have emphasized that economic analyses are relevant for policy only if the economic mechanism "is empirically relevant and first order to the problem at hand." (89) Importantly, a number of sophisticated economic and legal analysts read the evidence here as inconclusive, not pointing in any clear direction as to which version of capital income neutrality is a more important efficiency norm for tax design, whether by a legislature or a court. (90) In addition, economists have found that the appropriate efficiency norm may vary depending on the extent to which foreign and domestic activities complement or substitute for one another and how capital expenditures are treated. (91)
Mason and Knoll concede that their versions of competitive neutrality would violate locational neutrality or leisure neutrality, giving rise to distortions along those margins. (92) They nevertheless choose competitive neutrality over locational neutrality and leisure neutrality as the superior efficiency standard for judging tax discrimination. Remarkably, they fail to offer any evidence whatsoever that the distortions that would be prevented by competitive neutrality are greater than the distortions that would be prevented by the other two efficiency benchmarks they analyze and reject, even given their unrealistic assumption of fixed residence. (93) Nor do they offer any theoretical reason to conclude that competitive neutrality is the superior standard for efficiency.
We obviously agree that differences in member states' income taxes may distort decisions as to which worker gets which job (or which company owns which enterprise), but, just as obviously, these taxes may distort locational decisions about both where to work or reside and where to locate capital by both individuals and corporations. Such differences may also distort the tradeoffs between work and leisure (which, as we have described, may become conflated with productivity) and between saving and consumption. It is not possible on economic efficiency grounds to privilege one potential set of distortions over others without any comparison of their relative magnitudes. For example, if jobs are in fact more mobile than workers, distortions from the absence of locational neutrality may be more important to economic efficiency than those from the absence of competitive neutrality. Without such evidence, we simply cannot know whether a norm based on enhancing workers' ability to compete for jobs across borders is more welfare-enhancing than the other efficiency norms.
In discussing normative arguments, Mason and Knoll say at several points that they are not arguing that competitive neutrality would best promote EU welfare because more intrusive measures, such as harmonization, might be more welfare-enhancing, (94) in part because such measures might also promote locational and leisure neutrality. Fair enough; the authors are under no obligation to consider every possible tax policy. On the other hand, having explicitly adopted the economist's concept of welfare (95) and having defined and extensively analyzed three competing welfare norms, they surely owe the reader an efficiency-related explanation for choosing one of the three as the superior efficiency norm. The authors frequently characterize their argument as merely "interpretive," (96) but that characterization does not relieve them of the obligation to provide a cogent reason for urging a particular welfare norm. As indicated above, their view is that efficiency is and should be the key focus of the ECJ in deciding tax discrimination cases. Given that view, we cannot understand how they can advocate one of three diverging pathways to greater efficiency without explaining why that pathway is likely to produce superior efficiency results.
Mason and Knoll do suggest several non-efficiency advantages of interpreting nondiscrimination to require competitive neutrality: increased predictability, promotion of representation reinforcement and political unity, avoidance of legislative decisions, resolution of open questions, and, they claim, a way out of the labyrinth. Some of these results--such as increased predictability, greater certainty through resolution of open questions, and avoidance of legislative decisions--would obtain under any of the three neutrality (or other) standards. (97) And some, such as promotion of representation reinforcement, avoidance of legislative decisions, and greater political unity (again offered without any evidence) depart from their insistence that enhancing economic efficiency in an internal market is and should be the court's prime focus. Moreover, none of these asserted advantages addresses the question of why competitive neutrality is the superior benchmark, even if one limits the analysis to the economic efficiency norm they urge.
Despite the foregoing shortcomings of the article's normative claims for the competitive neutrality standard, we shall now examine Mason and Knoll's view of the tax laws that would be required to implement that standard.
V. THE TAX LAW REQUIREMENTS OF COMPETITIVE NEUTRALITY
In one section of their article, (98) Mason and Knoll use concrete examples to argue that certain taxes violate competitive neutrality and therefore reduce economic welfare. To clarify the authors' argument, let us begin by providing a straightforward statement of what we understand to be the underlying logic (which includes the usual assumption that markets will achieve efficient results in the absence of taxation). (99)
Definition: "A tax system is competitively neutral when it is not possible to increase productivity by shifting jobs among people." (100)
Proposition 1: The economy will allocate workers among jobs in different countries on the basis of relative differences in certain "retention' rates and ratios. (101) More precisely, in the simple two-state, two-job, two-worker examples in the article's tables, (102) the relevant ratio is that between the amount of labor income that a worker would retain after all taxes if he worked in one country (call it country A) and the amount that worker would retain after all taxes if he worked in the other country (call it country B). The workers will sort themselves so that the worker with the higher A/B ratio will work in country A, while the worker with the higher B/A ratio will work in country B. (103)
Proposition 2: In the absence of taxation, the foregoing sorting of workers will satisfy the definition of competitive neutrality. (104)
Proposition 3: Certain tax structures will modify the ratio of the retention ratios that would obtain in the absence of taxation, thereby modifying the allocation of jobs among workers. (105)
Conclusion: Such tax structures do not therefore satisfy the definition of competitive neutrality. Q.ED.
To reach the foregoing conclusion, the three propositions obviously have to be demonstrated, rather than merely asserted. Mason and Knoll provide two versions of the argument: an algebraic version in the notes and a verbal version with examples in the text (as supplemented in the notes). We begin with the algebraic version because the authors state that it is more general. (106) As far as we can tell, there is no demonstration in that version of propositions 1 and 2, which the notes simply assume to be true. (107)
We now turn to the verbal argument and the related tables. In the most concrete statement of the argument, we have two workers, Francoise and Gunther, who are said to sort themselves between a job in Germany and a job in France based on the retention ratios described above. There are indeed some verbal examples in the text that are used to argue for propositions 2 and 3 under some circumstances. (108) More importantly, there is a final example (Table 4 on page 1069) with actual retention ratios that is said to demonstrate how a tax that changes those ratios will result in a reshuffling of jobs (and therefore a reduction in economic welfare). (109) Readers who verify the arithmetic will be surprised to learn that the authors' conclusion apparently holds only if the employer does not maximize profits. (110) The authors say nothing about this peculiar result and assert that "employers will hire the worker with the largest relative difference between productivity and wage." (111) But why would an employer do that if it would reduce profits ? (In their example, profits would be maximized if an employer focused on the absolute difference between a worker's output and wage. (112)) For Mason and Knoll, the answer apparently lies in this mysterious sentence found a few pages earlier:
Assuming that there are many Francoises and Gunthers and that employers are not restricted to hiring a fixed number of employees, but rather trying to produce a given output at least cost, then employers will select employees with the greatest relative difference between their wage and their output. (113)
We admit that we are uncertain what the components of the clause beginning with "Assuming" are meant to communicate. Do the many Francoises have the same or different characteristics ? Do we have to assume fractional Gunthers? These questions are not addressed or answered by the authors. Without convincing answers to such questions, the "then" clause of this sentence once again simply assumes propositions 1 and 2 in the argumentative structure we laid out above. (114) It may, of course, be possible to describe an economic model with continuous functions that satisfies the sentence quoted above, but the authors do not do so. It might also be possible to construct a discrete example that illustrates that sentence without assuming that firms do not maximize profits, but the authors have not done that either.
We are not therefore asserting that someone might not be able to write down a demonstration of the propositions identified above. All we have shown is that Mason and Knoll have not done so. Mathematically trained tax economists tend to prefer models (often with continuous functions) that carefully specify assumptions and then reason to conclusions based on mathematical analysis. Verbally trained tax lawyers tend to prefer written arguments that carefully specify assumptions and then reason to conclusions based on verbal analysis (often with discrete examples). Mason and Knoll's argument does neither. The algebraic expressions in the footnotes might lead a mathematically challenged reader to think that the claimed conclusion had been demonstrated, but that algebra simply assumes key parts of the argument. The same is true of the verbal argument, at least as far as we are able to parse it.
Although the tables and algebra in their article do not actually demonstrate Mason and Knoll's proposition that certain taxes violate competitive neutrality, let us nevertheless examine their view of what tax laws would implement that concept. Competitive neutrality, they say, requires either of two income tax systems. One alternative is that all residence countries would adopt worldwide taxation with unlimited credits for income taxes paid to source countries. (115) Whenever the source-country tax rate is higher than the residence-country rate, this system would require the residence country to refund to its taxpayers the additional amounts they pay abroad. As Mason and Knoll note, under such a system, source-country taxation becomes irrelevant to competition. (116) However, no country has such a tax system, nor will any country enact such a system, because no country is willing to sacrifice its revenues from taxing domestic activities to refund higher-rate income taxes paid to another country.
Mason and Knoll's second alternative for achieving competitive neutrality is for all countries to tax both residents and nonresidents on a source basis and then to tax residents again on their worldwide income, with a deduction from that income for the source tax they paid at home or abroad. Mason and Knoll call this "the ideal deduction method," which includes the possibility of simply exempting all foreign income. (117) The deduction and exemption versions of the ideal deduction method are, however, as unrealistic as the unlimited credit proposal. Their proposed mandatory two-tier system does not exist in Europe, the United States, or anywhere else as far as we know. Although dividing taxation of residents into two tiers (source and residence) has been discussed in the past, it has generally been rejected by taxing authorities as too complicated for practical administration. (118) The second possibility, exemption of foreign income, would be simpler, but is equally unrealistic. Many developed countries exempt foreign-source dividends paid by subsidiary corporations to their parents, but foreign-source earnings are not generally wholly exempt from tax in the residence country.
To achieve Mason and Knoll's preferred norm of competitive neutrality, all the member states of Europe (and presumably all the states of the United States) would have to adopt an income tax system that none now has. And all would have to adopt exactly the same system (but each could set its own rates). Of course, as the authors concede, "the choice of how to tax cross-border income is a legislative question." (119) A unanimous vote of the member states would be required for any such change to occur at the European Union level.
Neither the European Court of Justice nor the U.S. Supreme Court has the legal authority to require that states adopt any of these very specific taxing systems. Implementing either system would require the court to harmonize their methods of taxing cross-border income by imposing the same system on all states. Nothing could be more antithetical to the European Treaties' assignment of primary authority over income taxation to the member states, including the unusual requirement of unanimity.
Finally, let us say a word about the limited role of tax-rate differentials in Mason and Knoll's analysis of competitive neutrality. Earlier we indicated that, in the absence of a formal model, the assumption of fixed residence made it impossible to sort out differences between changes in the location of jobs and changes in the worker performing the job. (120) When applied to Mason and Knoll's concept of competitive neutrality, that assumption also obscures the effect of differences in tax rates in any realistic tax system. Under the unlimited credit they postulate, differences in source taxes become irrelevant to foreign investors because of credits and refunds in the residence country. Under their exemption alternative, residence taxes are assumed to be levied at the same (zero) rate. Under the ideal deduction, the assumption of fixed residence assumes away the effect of differences in tax rates on where taxpayers might choose to live or organize their businesses. The diminished role of rate differences under their standard of competitive neutrality is thus a function of unrealistic presuppositions (fixed residence, zero rates, unlimited credits and so on). In any more realistic setting, tax-rate differentials play a much more significant role, as many European commentators have observed with respect to both labor and capital income. (121) The limited role of rate differences is thus surprising in an article that begins by postulating that taxes are the only thing that matters in workers' decisions about "where to work, how much to work, and which job to work." (122)
Of course, the courts--whose decisions concern Mason and Knoll--have no ability whatsoever to affect tax rates or to redress any economic dislocations or effects on EU welfare that differing tax rates may induce. As they acknowledge, "the ECJ has expressly held that cross-border tax disadvantages arising from rate differentials do not constitute discrimination." (123) But since tax rates (or tax rates net of social insurance and other welfare benefits) may be the most important tax factor in the individual and corporate decisions being analyzed, this creates a conundrum both for the ECJ and for Mason and Knoll. For the ECJ (as we and others have shown), it results in confusing, contradictory, and changing patterns of decisions. And it necessarily plunges Mason and Knoll into a second- or third-best world in analyzing and evaluating the efficiency of the ECJ's nondiscrimination jurisprudence.
Up to this point, despite its unreality, we have accepted the operation of Mason and Knoll's two-tier ideal deduction as presented in their article, but that presentation is confined to flat-rate taxes. Earlier, we pointed out some of the ways in which individual income taxes differ from the corporate income taxes analyzed in the Desai and Hines papers on which Mason and Knoll are relying. (124) This is one instance in which those differences preclude simple importation of the Desai and Hines framework, because unlike corporate income taxes, individual income taxes are often progressive and depend on family circumstances, such as the number of children or a spouse's income. The authors never show how their ideal deduction would work with graduated tax rates, so all we can do here is tentatively raise a couple of potential problems.
One problem is that the taxpayer has to identify the graduated marginal rate applicable to each slice of income in order to compute the relevant retention ratios. Although the authors do not say so, the ideal deduction method in their concrete examples seems to boil down to taxing each item of income at a total tax rate ([t.sub.t]), which is the arithmetic sum of the source tax rate ([t.sub.s]) plus the residence tax rate ([t.sub.r]) minus the product of those two rates. (125) The after-tax amounts are then the basis for the retention ratios calculated in the article for Francoise and Gunther. (126) The tax rates used to compute (1-[t.sub.t]) are straightforward in the flat-rate examples used by the authors.
On the other hand, if residence taxes are graduated, (127) identifying the rate applicable to a slice of income can be problematic. As Mason and Knoll are concerned with marginal incentives, they are interested in the marginal rate of residence taxation. Notice, however, that if there is more than one source country, the residence tax under the ideal deduction is a function of the source tax rates in all those source countries, so we are unclear how the marginal rate for each country is to be sorted out. Indeed, a change in one country's source tax rate will affect the total residence tax and therefore possibly affect the amount of residence tax attributable to income produced in another country.
Consider, for example, a taxpayer who earns half her income (10,000 [euro]) at home where the source tax rate is 10% and the other half (10,000 [euro]) abroad where the source tax rate is 5%. Now assume that her residence tax is graduated, so the rate is 15% on the first 10,000 [euro] and 30% on any income above that. Under the authors' ideal deduction, total residence taxable income will presumably be 18,500 [euro] (9500 [euro] from abroad and 9000 [euro] from home), so the residence tax due will be 4050 [euro] (15% of 10,000 [euro] plus 30% of 8500 [euro]). But how much of that residence tax is attributable to the foreign income? Is it half (because the foreign and domestic pretax amounts are equal) or 95/185 (because that is the portion of total after-source-tax that is from abroad)? Alternatively, are we to assume that the foreign income is stacked first (subject to only the 15% rate) or second (subject mostly to the 30% rate)? Without an answer to these questions, the taxpayer cannot make the calculations that Mason and Knoll indicate are a predicate to achieving competitive neutrality.
Now assume that the residence country reduces its source tax to zero. Residence taxable income is now 319,500 [euro] (9500 [euro] from abroad and 10,000 [euro] from home), so the residence tax due will now be 4350 [euro] (15% of 10,000 [euro] plus 30% of 9500 [euro]), for an increase of 3300 [euro]. How much of that increase should the taxpayer attribute to the foreign income when she calculates her retention ratios? Half (because the foreign and domestic pretax amounts are still equal) or 95/155 (because that is now the portion of total after-source-tax from abroad)? Alternatively, if we assume, as above, that the foreign income is stacked first, there is no change in the amount of residence tax attributable to the foreign income because that income is still subject to the 15% rate. If, on the other hand, we assumed that the domestic income was stacked first, some of the increase in the residence tax is attributable to the foreign income, because the entire 15% bracket is now taken up by the domestic income, so that all of the foreign income is now in the 30% bracket. The authors never address how the total tax from earnings in each country is to be calculated when residence taxes are graduated; (128) so we do not know how the retention ratios are to be computed even in this simple kind of case. Readers interested in more complicated cases might think about an opera singer (subject to graduated-rate residence taxation) who has offers to perform in twenty countries (with different rates of source taxation) during a period in which she has time for only ten performances.
It is, of course, appropriate to limit discussion of an idea (such as the ideal deduction) to particular assumptions (such as flat-rate taxation) for purposes of analysis. Here, however, Mason and Knoll go much further and claim that their analysis produces policy recommendations for tax legislation and jurisprudence. (129 One such recommendation is that the ideal deduction provides the easiest pathway to the harmonization necessary for competitive neutrality in the EU. (130) Given the absence of any explanation of how that deduction would operate in tax systems with graduated rates, we do not see how this recommendation is germane to actual EU taxes. Indeed, the disconnect serves as a cautionary tale about the dangers of importing an analytical framework from taxation of multinational corporations (usually at flat rates) into taxation of individuals (often at graduated rates) without working through how that framework would have to be modified to take into account differences in the two environments. (131)
The authors do not discuss the kind of case we put forth here, but in very brief comments on progression late in their article, they do indicate that coupling graduated-rate residence taxation of worldwide income with flat-rate source taxation of nonresidents would not be incompatible with competitive neutrality. (132) Under that standard, graduated-rate residence taxation could be joined with flat-rate source taxation of nonresidents only in countries that had adopted the ideal deduction, which would require a residence country to apply its flat-rate source tax to its own residents before then applying its graduated-rate residence tax to those residents. (133) Given the failure of the article to address more generally the possibility of graduated rates, all we can say with confidence at this point is that the restriction of the examples, tables, and figures to proportional taxes leaves the implications of the analysis for individual income taxation with graduated rates largely unknown.
To summarize, Mason and Knoll present two tax systems that would satisfy their standard of competitive neutrality: worldwide residence taxation with unlimited foreign tax credits and the ideal deduction method (which includes the possibility of exempting foreign income). Neither alternative is realistic: none exists anywhere. Mandatory adoption of either alternative in all twenty-seven EU member states would require unanimous agreement of the member states and implies an unrealistic level of harmonization. As for the ideal deduction method, how it would operate in the context of graduated rates has yet to be explained. The required departure from the political understanding enshrined in the Treaties and the unrealistic assumptions of competitive neutrality confirm our view that constitutional courts should not be making tax policy based on abstract and contradictory principles of nondiscrimination. Mason and Knoll's view is different: after conceding that courts cannot implement the neutrality concept that they prefer, they then go on to suggest that judges may nevertheless promote that concept by interpreting prohibitions against discrimination in a certain way, to which we now turn.
VI. THE PROPOSED CONSTITUTIONAL VERSION OF COMPETITIVE NEUTRALITY
Despite admitting that courts are incapable of fully implementing the requirements of competitive neutrality, Mason and Knoll still argue that both the ECJ and the Supreme Court should strike down legislation by interpreting nondiscrimination requirements as /f competitive neutrality were a constitutional requirement. Specifically, they urge that income tax laws should be struck down whenever they involve "nonuniform" source or residence taxes. (134) There is no indication that this advice is limited to labor income cases involving cross-border workers. To avoid confusion between the actual requirements of competitive neutrality advanced by Mason and Knoll and the related constitutional standard that they urge courts to apply, we will call the latter "partial competitive neutrality." For readers who are unwilling to embrace competitive neutrality, Mason and Knoll also identify partial versions of locational neutrality ("uniform residence taxes") and leisure neutrality ("uniform source taxes" with no residence taxes) that they say constitutional courts could apply.
At this point, we are mystified by what theory of constitutional interpretation Mason and Knoll have in mind. Having chosen efficiency as the paramount norm for constitutional interpretation in this area, they have urged one efficiency concept--competitive neutrality--over competing formulations, albeit without providing any empirical (or theoretical) basis for the choice. Then, after conceding that courts do not have the legal or institutional competence to implement the requirements of the concept they favor, they nonetheless urge courts to elevate competitive neutrality to constitutional status. Mason and Knoll ask courts, to the extent of their ability, to invalidate legislation that fails to conform in a particular, partial way to the authors' competitive neutrality concept. For us, the institutional and conceptual gaps between the assumed constitutional norm (efficiency) and the proposed interpretative standard (uniformity) alone are simply too great to make the connection compelling as a matter of constitutional law. (135)
Finally, is it clear that using partial competitive neutrality as a constitutional standard will necessarily reduce distortions and advance competitive neutrality, even on the assumption that taxes are proportional, rather than progressive? Formulation of this partial concept endorses certain attributes of competitive neutrality (uniformity) and suppresses others (for example, their two-level tax with an ideal deduction). Why then should we conclude that the former necessarily advances the overall goal of improving efficiency, even though the latter cannot be required? Without more analysis, we are not convinced that partial competitive neutrality will always advance competitive neutrality. Indeed, the theory of the second best shows that in the presence of other distortions, it cannot be assumed that reduction of any one economic distortion will always increase overall welfare. (136) Consider, for example, a residence country, let us say Poland, that adopts a limited foreign tax credit. Due to the foreign tax credit limitation (which is not consistent with competitive neutrality, but which is both universal in credit countries and beyond the scope of judicial invalidation in Europe and in Mason and Knoll's framework), Polish individuals and companies may decide not to compete for projects in a higher-tax source country for which they have a comparative advantage, distorting who works at those projects and who owns them. In order to induce Polish and other foreign individuals or companies to bid on projects within its borders, the source country might respond by lowering its income tax only for foreigners (or in a bilateral tax treaty only for Polish individuals and companies). As we understand Mason and Knoll, the ECJ (and the U.S. Supreme Court) should strike down the nonuniform favorable source tax as a violation of partial competitive neutrality, even if the nonuniformity simply offsets another distortion, thereby promoting competitive neutrality overall. (137)
VII. THE POSITIVE CLAIM FOR COMPETITIVE NEUTRALITY
Mason and Knoll argue not only that competitive neutrality should be adopted as the judicial standard of tax nondiscrimination in common markets, but also that the European Treaties' four freedoms and the decisions of the ECJ are consistent with competitive neutrality, but not with locational or leisure neutrality. (138) They claim that this convergence with their preferred norm can be discovered in both the text of the freedoms and the ECJ's jurisprudence, including the court's overall approach to tax cases.
Needless to say, imputing a particular economic concept, such as competitive neutrality, to constitutional documents, such as the EU Treaties, and to an extensive set of constitutional decisions striking down national laws would be a daunting task. To be convincing, the exercise would presumably include examination of preparatory documents, analysis of a multitude of tax cases, comparison with nontax cases decided on comparable grounds, and so on. Mason and Knoll do not undertake that massive task, nor will we. Instead, we restrict our comments here to why we do not find convincing their arguments that the text of the freedoms and the ECJ's approach to tax cases accord with competitive neutrality, but not with locational or leisure neutrality. (139)
To begin with, there are simply too many counterexamples that come immediately to mind. To take but one example, neither the ECJ nor the U.S. Supreme Court invalidates legislation that advantages foreigners over residents or cross-border transactions over domestic ones. Such "reverse discrimination" is very controversial in the literature, (140) but the ECJ has "consistently refused" to strike down such provisions, (141) even though they are clearly nonuniform under Mason and Knoll's definition. This refusal is not surprising. As Mason and Knoll themselves suggest, the nondiscrimination doctrine is a means to remove "direct tax obstacles to the EU common market" and to preclude member states from enacting "with impunity all sorts of nonuniform tax laws that burdened nonresidents and interstate commerce more heavily than residents and domestic commerce." (142) Despite legitimate policy concerns, it is difficult to argue that the nondiscrimination requirement of the EU Treaties is concerned with limiting member state actions that favor nonresidents. Surely, benefits to nonresidents fail to raise concerns about the lack of political representation similar to those that occur when nonresidents are being disadvantaged. Moreover, as Miguel Maduro has observed, accepting the home-country principle (mutual recognition of residence-only regulation or taxation) reflects the "acceptance of reverse discrimination." (143) In income taxation, the so-called "Beckham law" in Spain, which provides favorable income tax rates for foreign professional soccer players, may be the most famous example of reverse discrimination. (144) Nonuniform "patent boxes," which reduce the rate of corporate tax on income derived from patents (and, in some cases, from other intellectual property), are also common in Europe. (145) And since 1991, Denmark has applied a temporary (three-year) top marginal rate of 25%, rather than its standard 59% top rate, to highly paid foreigners who migrate to Denmark. (146) Despite their obvious failure to meet Mason and Knoll's requirement of uniformity, none of these laws has been held to violate the EU Treaties.
The U.S. Supreme Court has likewise sustained discrimination in favor of nonresident corporations over resident corporations, while invalidating discrimination in favor of resident over nonresident corporations, (147) prompting one leading treatise to observe that the constitutional guarantee of equality "was not designed to protect the wolves from the sheep." (148) Although the Supreme Court has occasionally struck down provisions providing incentives for local investments and jobs, (149) the Court has also made it clear that the Constitution "does not prevent the States from structuring their tax systems to encourage the growth and development of intrastate commerce and industry." (150)
In addition, Mason and Knoll assume that the free movement of portfolio investment (by shareholders) is fully sufficient to achieve locational neutrality for direct investments (of corporate capital). On that assumption, they conclude that application of the freedom of movement to both portfolio and direct investment in the Treaties must mean that capital ownership neutrality is being protected, because freedom of movement for direct investment is unnecessary for locational neutrality. (151) Mason and Knoll provide, however, no support from the economics literature for their extraordinary assumption. Nor do they identify the market conditions necessary to reach the conclusion that freedom of movement for portfolio investment (by shareholders) would efficiently eliminate all capital locational distortions (by companies), so that freedom of movement for direct investors would not be necessary to achieve locational neutrality. Those conditions are likely to be unrealistic. As Mason and Knoll indicate, many economists are likely to view violations of locational neutrality as having greater negative welfare consequences than violations of leisure or competitive neutrality. (152) On efficiency grounds alone, a more plausible reading of the freedom of movement is therefore that the goal of locational neutrality for corporate investment cannot be simply ruled out as superfluous.
Mason and Knoll also say that interpreting nondiscrimination as locational neutrality would preclude the exemption method of taxing international income, because locational neutrality requires residence-based worldwide taxation with unlimited tax credits. (153) Similarly, they contend that interpreting nondiscrimination as leisure neutrality would preclude residence taxation, because leisure neutrality allows only source taxation. (154) Since the ECJ decisions clearly permit both the exemption of foreign income and residence taxation without unlimited credits, they conclude that the decisions cannot be regarded as implementing either principle. In the traditional language of international tax policy, the court's decisions are not fully compatible with limiting nondiscrimination to either capital export or capital import neutrality. As indicated above, we obviously agree, as one of the points in our original article was that some cases point in one of these directions and others in the opposite direction, leading to incoherent and inconsistent results. (155)
However, it seems equally obvious that the ECJ cases cannot be read as implementing competitive neutrality, because the court has never indicated that an unlimited foreign tax credit or "ideal deduction" is required by the nondiscrimination standard. In considering locational and leisure neutrality, Mason and Knoll test the ECJ jurisprudence against the actual analytical requirements of those concepts, not against the partial standards they have proposed for judicial interpretation. When considering competitive neutrality, however, they use a different, much more lenient test, focusing on certain language in the ECJ decisions to conclude that they are compatible with the logic behind their partial version of competitive neutrality. (156) If, however, a similar test were applied to all three concepts, the ECJ decisions would be regarded as just as incompatible with competitive neutrality as with locational and leisure neutrality. (157)
Having insisted that their preferred norm of competitive neutrality can be found in the EU Treaties and cases, Mason and Knoll also assert that the dormant Commerce Clause and the Privileges and Immunities Clause of the U.S. Constitution, as interpreted by the U.S. Supreme Court, promote that standard as well. (158) As with the European court, this conclusion is both normative and positive, and the discussion goes far beyond cases involving cross-border workers. We have already indicated why we think the normative case for competitive neutrality is inadequate, even given the authors' efficiency norm. As for the positive claim that competitive neutrality explains the hitherto confounding U.S. decisions, Mason and Knoll's analysis is remarkably limited, spanning only a few pages and discussing only a couple of cases. Indeed, their discussion of tax nondiscrimination in the United States focuses on the Article IV Privileges and Immunities Clause, which they recognize does not even apply to corporations, (159) the principal litigants in interstate tax discrimination cases. Moreover, although there may be Commerce Clause cases involving discrimination that could, with the benefit of hindsight, be cast in a competitive neutrality mode, the overwhelming proportion of interstate tax discrimination decisions cannot fairly be read as saying anything about residence or source. Rather, they reflect the view that Commerce Clause discrimination "simply means differential treatment of in-state and out-of-state economic interests that benefits the former and burdens the latter," (160) for whatever that is worth. Finally, even Mason and Knoll appear to acknowledge that, unless and until the Supreme Court adopts their competitive neutrality analysis, its jurisprudence may fairly be characterized as "a series of confused and incoherent tax discrimination decisions." (161)
A convincing showing that competitive neutrality is the key to understanding the Supreme Court's extensive nondiscrimination jurisprudence would thus require a much fuller analysis of the very large number of relevant decisions. In sum, Mason and Knoll's claims that their norm of competitive neutrality follows from the text of the EU Treaties and the U.S. Constitution and can be found in the ECJ and Supreme Court cases interpreting these founding documents are simply not convincing.
VIII. COMPETITIVE NEUTRALITY AS A WAY OUT OF THE LABYRINTH OF IMPOSSIBILITY
Finally, Mason and Knoll insist that competitive neutrality, unlike locational or leisure neutrality, offers a way out of the labyrinth of impossibility. They sometimes present our point as limited to the conflict between capital import and export neutrality (locational and leisure neutrality in their terms). (162) As we have clearly stated in everything we have written on this subject, (163) our criticism of the ECJ's approach to tax cases is more general: the impossibility of requiring nondiscrimination from the perspective of both origin and destination countries. The well-known impossibility of achieving both capital import and capital export neutrality is just one example, albeit an important one.
As our article was about the ECJ's jurisprudence, we presume that it is the partial version of competitive neutrality that they formulate for courts that is supposed to provide a way out of the labyrinth of impossibility. It is, however, worth our stating why their full version of competitive neutrality does not resolve the impossibility of applying a nondiscrimination principle in both origin and destination countries in the absence of harmonization of tax systems and rates, which is the proposition we reiterated at the beginning of this Response. (164) The full version of competitive neutrality contemplates harmonization of all tax systems, but not rates, into one of two unrealistic systems: (a) worldwide taxation with unlimited foreign tax credits or (b) two-part taxation of source and residence income, accompanied by either: (i) a deduction for source taxes (paid at home or abroad) or (ii) an exemption for foreign income. (165) Such harmonization obviously undermines the sovereignty of the origin and destination countries that we postulated above. Moreover, as Mason and Knoll recognize, (166) their harmonized version (a) eliminates any role for taxation of cross-border income in the destination country, as the country of origin must credit or reimburse all foreign taxes. Version (b)(ii), on the other hand, eliminates any role for taxation of cross-border income in the origin country, which must exempt foreign income. Only version (b)(i) maintains a role for taxes in both countries, but it creates a locational incentive to invest or work at home or abroad, depending on relative tax rates, thereby discriminating for or against cross-border movement of capital or labor. (167) This distortion could, of course, be eliminated by harmonizing tax rates as well as tax systems, but, as we have always said, harmonization of both income tax bases and rates would avoid the impossibility result, but is unrealistic. Without such harmonization, however, different tax rates distort people's choices about where to live and work and companies' decisions about where to locate their subsidiaries and their plant and equipment. And, as Mason and Knoll acknowledge, neither the ECJ nor the U.S. Supreme Court has the power to compel such harmonization.
Turning now to the partial version of competitive neutrality that Mason and Knoll propose for judicial decisionmaking, their claim seems to be that it provides a way out of the labyrinth because there is nothing fundamentally incoherent about imposing requirements of uniform taxation on both source and residence countries from the perspective of partial competitive neutrality. (168) Note, however, that they also assert that uniform residence taxes maintain locational neutrality and that uniform source taxes (with no residence taxes) maintain leisure neutrality, (169) but they admit the impossibility of having locational and leisure neutrality simultaneously. (170) Therefore, if both the origin country and the destination country in a cross-border transaction are included in the analysis, there is no escape from the labyrinth. As a result, Mason and Knoll concede that their approach requires a court to consider only the laws of a single country when evaluating cross-border transactions involving two or more countries. (171) It is hardly surprising that analysts who restrict their consideration to the laws of either the origin or destination country in any given case will find the impossibility of requiring discrimination from the perspective of both countries uncompelling.
As we indicated in our original article, our analysis, like the ECJ's, is not so restricted because we view the nondiscrimination interpretation of the four freedoms as protecting taxpayers against a higher tax burden for cross-border income than for domestic income. (172) We also discussed an alternative, more limited view that was not subject to our impossibility result. Under this alternative view, which has a long intellectual history in the EU, (173) nondiscrimination would preclude a member state only from more heavily taxing income that crosses its borders than income that does not. As with Mason and Knoll's recommendation, such limited alternatives would look only at one member state's laws and would not take into account the tax situation in the other member state, recognizing that each member state imposes income taxes on both a source and a residence basis. As one example of this approach, we cited the conclusion of Wolfgang Schion that nondiscrimination only requires (a) source countries to adopt capital import neutrality and (b) residence countries not to unreasonably hinder the export of capital, whether human or monetary. (174) This prescription is very similar to (if not identical with) Mason and Knoll's partial version of competitive neutrality, which would require (a) source countries to adopt uniform source taxation and (b) residence countries to adopt uniform residence taxation. (175)
As in our original article, we agree that these approaches avoid our impossibility result. The rub is that they require the ECJ to restrict its analysis of nondiscrimination to a single country, ignoring the tax system in any other country implicated in cross-border commerce, a restriction that Mason and Knoll regard as salutary. (176) Our own view is that any serious attempt to identify the tax advantages or disadvantages for cross-border income should take account of the tax consequences in both countries. In any event, as Mason and Knoll concede, (177) it is clear that the court does not consider itself subject to any such restriction, as many of its nondiscrimination decisions involve evaluation of the taxing systems in both source and residence countries. (178) As Joachim Englisch has put it, "the ECJ has repeatedly held that a discriminatory taxation of cross-border transactions in one Member State can be offset by tax privileges granted in the other State with links to the transaction at issue." (179) The limited approach that Mason and Knoll recommend is thus not a way out of the ECJ's labyrinth any more now than was the similar approach of Wolfgang Schon when we wrote our original article.
Moreover, even if the court, contrary to its longstanding practice, were to decide to restrict its requirement of income tax nondiscrimination to a single country, we are skeptical that Mason and Knoll's partial competitive neutrality would, as they claim, always provide "clear direction" (180) in difficult cases. Reconsider two of the tax examples with which we began. Would an EU member state (the destination country) be permitted to deny a tax exemption to a charity organized (and exempt) in another member state (the origin country) if the destination country treated all charities the same by requiring them to meet its own requirements? That would seem to be the result under the requirement of uniform destination-country taxation, but it would also be inconsistent with many ECJ decisions and EU practices that privilege origin-country laws by requiring mutual recognition. Would a country from which an investment in corporate stock originated be required to grant tax credits to the shareholders for corporate taxes paid to another country if it granted such credits for purely domestic dividends? That would seem to be the result under origin-country uniformity, but it would also result in favoring cross-border transactions over domestic transactions if the country from which the dividends were paid also granted a credit.
Consider, finally, a same-sex couple who marry in a member state permitting such marriages and then move (for more than six months) to take new jobs in a member state that does not permit such marriages. Would it be permissible for the destination state to deny the newly arrived residents income tax benefits accorded married couples on the grounds that all same-sex couples are treated the same in that country? If not, the new residents would be better off than same-sex couples that had always lived in the destination country. On the other hand, the denial of benefits could create a disincentive for a same-sex couple to move to another member state to take jobs for which they have a comparative advantage. Mason and Knoll avoid such cases by their assumption that taxpayers cannot change their country of residence. We, however, do not see how a standard of competitive neutrality provides "clear direction" and an obvious answer to this important, realistic question. (181)
Let us summarize our eight principal differences with Mason and Knoll: (1) Their article adopts economic efficiency as the paramount criterion for interpreting tax discrimination, but this is too restrictive a focus for constitutional courts, because it would preclude member states from adequately considering other important tax policy norms in writing tax legislation. (2) The Desai and Hines capital neutrality framework for corporate taxation cannot be imported into the context of individual taxation without much greater analysis of the differences between corporate and individual behaviors and tax regimes. (3) The unrealistic assumption of fixed residence confines the article's actual analysis to discrimination against cross-border workers, so any broader claims about tax discrimination are not grounded in the article's analytics. (4) The article never makes an adequate normative case for competitive neutrality as the preferred economic efficiency criterion, because the distortions that would be eliminated by competitive neutrality are never compared with those that would be eliminated by other efficiency standards, including leisure neutrality or locational neutrality. On its own logic, the article therefore provides no reason to favor competitive neutrality over the competing economic efficiency concepts it discusses. (5) The tax systems required to implement competitive neutrality (universal adoption of either unlimited foreign tax credits or the "ideal deduction" method) are not realistic or practical. (182) The unrealistic assumption of flat rates for individual taxation in the presentation of the ideal deduction method makes any policy recommendations based on that deduction problematic. (6) Given the actual requirements of competitive neutrality developed in the article, the partial version that they urge courts to use is simply too attenuated to serve as a compelling constitutional principle. (7) The positive case that the ECJ and the U.S. Supreme Court are, and have been, actually engaged in promoting competitive neutrality is not convincing in light of the article's limited analysis of judicial decisions and the many counterexamples available. (8) Competitive neutrality fails to provide a way out of the labyrinth because it entails either harmonization, locational distortions, or effective elimination of either source or residence country taxation. Partial competitive neutrality fails to provide a way out because it assumes, contrary to the ECJ's decisions (and, indeed, to our original description of the labyrinth itself), that the law of only a single country is relevant to questions of tax discrimination in cross-border transactions. Even on the assumption that the court would and should restrict its consideration to the law of only one country, neither version of competitive neutrality provides clear and acceptable answers in many difficult cases.
For us, then, competitive neutrality turns out not to be the holy grail of tax discrimination. With respect to tax jurisprudence, the ECJ remains stuck in a labyrinth of impossibility because the logical implications of requiring nondiscrimination in both origin and destination member states remain contradictory and incoherent.
(1.) Throughout this Response, we use the plural to refer to the EU foundational treaties, which continue to evolve. The current version is the Consolidated Version of the Treaty on the Functioning of the European Union, Mar. 30, 2010, 2010 O.J. (C 83) 47 [hereinafter TFEU]. For further discussion of the current institutional arrangements, see generally RUTH MASON, PRIMER ON DIRECT TAXATION IN THE EUROPEAN UNION (2005); and Michael J. Graetz & Alvin C. Warren, Jr., Income Tax Discrimination and the Political and Economic Integration of Europe, 115 YALE L.J. 1186, 1188-94 (2006).
(2.) For a comprehensive list of EU legislation currently in force regarding direct taxation (including the income tax directives), see Directory of European Union Legislation in Force, EUR. UNION, http://eur-lex.europa.eu/en/legis/latest/chapo920.htm (last updated Jan. 1, 2012) (listing European Union legislation on direct taxation). For a comprehensive discussion of directives in force from 1998 to 2007, see Philipp Genschel & Markus Jachtenfuchs, How the European Union Constrains the State: Multilevel Governance of Taxation, 50 EUR. J. POE. RES. 293, 297-300 (2011). See also MASON, supra note 1, at 22-35.
(3.) For the template for these bilateral treaties, see OECD, MODEL TAX CONVENTION ON INCOME AND ON CAPITAL 17 (2010). In the "D" case, the European Court of Justice (ECJ) rejected the judgment of the Advocate General and held that income tax treaties in Europe need not apply a "most favored nation" approach, thus affirming the ongoing status of varying bilateral treaties in Europe. Case C-376/03, D. v. Inspecteur van de Belastingdienst/ Particulieren/Ondernemingen Buitenland te Heerlen, 2005 E.C.R. I-5821, I-5872. For the opinion of the Advocate General, see id. at para. 113 (opinion of Advocate General Ruiz-Jarabo).
(4.) TFEU, supra note 1, arts. 26, 28.
(5.) Id. arts. 26, 56.
(6.) Id. arts. 26, 45, 49.
(7.) Id. arts. 26, 63. Discrimination on grounds of nationality is also prohibited. Id. art. 18.
(8.) See, e.g., Philipp Genschel, Globalization, Tax Competition and the Welfare State, 30 POL. & SOC'Y 245, 259-61 (2002); Philipp Genschel, Achim Kemmerling & Eric Sells, Accelerating Downhill: How the EU Shapes Corporate Tax Competition in the Single Market, 49 J. COMMON MKT. STUD. 585, 587-89 (2011).
(9.) See Genschel & Jachtenfuchs, supra note 2, at 302-03. The vast majority of income tax cases decided by the ECJ have been brought by private litigants-most often corporations--challenging national measures. See Genschel et al., supra note 8, at 598-99. This, of course, tends to reduce national revenues, since private parties litigate only when victory will reduce their tax liability. Id. Between 1986 and 2003, these private litigants succeeded in more than 80% of the cases. Id. at 599.
(10.) See, e.g., Malcolm Gammie, Non-Discrimination and the Taxation of Cross-Border Dividends, 2 WORLD TAX J. 162, 170 (2010) (discussing the "single non-discrimination principle").
(11.) See, e.g., MIGUEL POIARES MADURO, WE THE COURT: THE EUROPEAN COURT OF JUSTICE AND THE EUROPEAN ECONOMIC CONSTITUTION: A CRITICAL READING OF ARTICLE 30 OF THE EC TREATY 70-76, 110-26 (1998) (discussing the political consequences of ECJ decisionmaking); Ruth Mason, Tax Discrimination and Capital Neutrality, 2 WORLD TAX J. 126, 137 (2010) ("The EC non-discrimination principle has been expounded primarily by the ECJ, an institution with a clear political agenda to promote the integration of Europe.").
(12.) See Suzanne Kingston, The Boundaries of Sovereignty: The ECJ's Controversial Role Applying Internal Market Law to Direct Tax Measures, 9 CAMBRIDGE Y.B. EUR. LEGAL STUD. 287, 303-10 (2007). Kingston argues that recent ECJ decisions reflect a "new, more rational balance between direct tax and free movement." Id. at 311.
(13.) For a brief description of these efforts since 1990, see Graetz & Warren, supra note 1, at 1228 n.143. For a description of the European Commission's current project for a "Common Consolidated Corporate Tax Base," see Common Tax Base, EUR. COMM'N TAXATION & CUSTOMS UNION, http://ec.europa.eu/taxation_customs/taxation/company_tax/common tax base/index_en.htm (last updated Nov. 16, 2011).
(14.) Genschel et al., supra note 8, at 595; see also Wolfgang Kerber, Interjurisdictional Competition Within the European Union, 23 FORDHAM INT'L L.J. S217, S234 (1999) (describing competition as "an integral part of the constitutional structure of the European Union").
(15.) Michael J. Graetz & Alvin C. Warren, Jr., Dividend Taxation in Europe: When the ECJ Makes Tax Policy, 44 COMMON MKT. L. REV. 1577 (2007) ; Graetz & Warren, supra note 1.
(16.) Graetz & Warren, supra note 1, at 1243.
(17.) Id. at 1216-23.
(18.) In tax parlance, the origin country is generally referred to as the "residence country"; sometimes it is called the "home country."
(19.) In tax parlance, the destination country is generally referred to as the "source country"; sometimes it is called the "host country."
(20.) For an important article that extends our "labyrinth of impossibility" beyond taxation to a wide range of regulatory contexts, see Alexandre Sayde, One Law, Two Competitions: An Enquiry into the Contradictions of Free Movement Law, 13 CAMBRIDGE Y.B. EUR. LEGAL STUD. 365 (2011). Sayde calls our impossibility result the "negative harmonization conundrum." He describes the dilemma we identified as the impossibility of simultaneously achieving "intra-jurisdictional equality" (from the perspective of the destination country) and "inter-jurisdictional equality" (from the perspective of the origin country) through "negative harmonisation" (nondiscrimination jurisprudence) in the absence of "positive harmonisation" (legislative action). Id. at 388.
(21.) The actual rate may vary if the foreign worker does not stay long enough to become a resident for tax purposes, as the destination country will not have full access to the worker's financial transactions. In that case, countries typically impose a flat-rate final "withholding" tax on domestic income of foreign workers that is designed to approximate the tax burden on domestic workers. See, e.g., I.R.C. [section][section] 881, 1442-1446 (2006) (imposing flat-rate taxes on foreign individuals and corporations).
(22.) See, e.g., Case C-128/08, Damseaux v. Belgium, 2009 E.C.R. I-6823, para. 27; Case C-67/08, Block v. Finanzamt Kaufbeuren, 2009 E.C.R. I-883, paras. 28-31; Case C-513/04, Kerckhaert v. Belgium, 2006 E.C.R. I-10967, paras. 20-24.
(23.) cf. Case c-25/10, Missionswerk Werner Heukelbach eV v. Belgium (Feb. 10, 2011), http://curia.europa.eu/juris/document/document.jsf?text=&docid= 84328&pageIndex=o&doclang=EN&mode=doc&dir=&occ=first&part=1&cid=364959 (ruling that the denial of reduced succession duties by one member state for contribution to a charity organized in another member state, when the donor did not live or work in the country in which the charity was organized, infringed the free of movement of capital); Case C-318/07, Persche v. Finanzamt Ludenscheid, 2009 E.C.R. I-359, http://curia.europa.eu/juris/celex.jsf?celex =62007CJ0318&lang1=en&type=NOT&ancre= (ruling that the denial of a tax deduction by one member state for in-kind contribution to a charity organized in another member state infringed the free of movement of capital); Case C-386/04, Centro di Musicologia Walter Stauffer v. Finanzamt Munchen fur Korperschaften, 2006 E.C.R. I-8203 (ruling that the denial of a tax exemption by one member state for rental income earned in that member state by a charity organized in another member state infringed the free of movement of capital).
(24.) Missionswerk, paras. 30-31 (holding that the close link between recognized charities and the state's activities is not an adequate ground for treating a charity organized in another member state differently).
(25.) These issues are discussed in detail in Graetz & Warren, supra note 15.
(26.) See Graetz & Warren, supra note 1, at 1219.
(27.) Case 120/78, Rewe-Zentral AG v. Bundesmonopolverwaltung fur Branntwein, 1979 E.C.R.. 649 (commonly referred to as Cassis de Dijon). For other nontax examples, see Sayde, supra note 20.
(28.) Kalypso Nicolaidis, Kir Forever? The Journey of a Political Scientist in the Landscape of Mutual Recognition, in THE PAST AND FUTURE OF EU LAW: THE CLASSICS OF EU LAW REVISITED ON THE 50th ANNIVERSARY OF THE ROME TREATY 447, 451 (Miguel Poiares Maduro & Loic Azoulai eds., 2010); Joseph H. H. Weiler, Mutual Recognition, Functional Equivalence and Harmonization in the Evolution of the European Common Market and the WTO, in THE PRINCIPLE OF MUTUAL RECOGNITION IN THE EUROPEAN INTEGRATION PROCESS 25, 25 (Fiorella Kostoris Padoa Schioppa ed., 2005).
(29.) The current resolution is found in Council Directive 2006/123/EC of the European Parliament and of the Council of 12 December 2006 on Services in the Internal Market, 2006 O.J. (L 376) 36. A previous version played a role in the campaign against the EU Constitution, particularly in France, where it was said that foreign service providers, symbolized by a Polish plumber, would work in France without having to comply with French regulations. See Kalypso Nicolaidis, Trusting the Poles? Constructing Europe Through Mutual Recognition, 14 J. EUR. PUB. POL'Y 682, 685 (2007).
(30.) For a more comprehensive discussion of such cases, see MADURO, supra note 11; and Sayde, supra note 20.
(31.) See, e.g., Council Directive 91/439/EEC, 1991 O.J. (L 237) 1 (pertaining to drivers' licenses).
(32.) This is done by taxing imports and exempting exports. Such border adjustments put the sales of goods and services within any particular country on an equal footing regardless of the location of production even when there are variations in tax rates among different member states. Philipp Genschel, Why No Mutual Recognition of VAT? Regulation, Taxation and the Integration of the EU's Internal Market for Goods, 14 J. EUR. PUB. POL'Y 743, 745 (2007).
(33.) Ruth Mason & Michael S. Knoll, What Is Tax Discrimination?, 121 YALE L.J. 1014 (2012).
(34.) Id. at 1037.
(35.) Id. at 1037-38.
(36.) Id. at 1018.
(37.) See also id. at 1022 (stating that their arguments have broad applicability due to the pervasiveness of "legal prohibitions of tax discrimination"); id. at 1030 (stating that the cases they discuss illustrate the controversies surrounding "EU tax discrimination doctrine"); id. at 1086 (stating that their preferred concept of neutrality "would advance the EU goal to integrate the economies of Europe because it would constrain state practices (including tax laws) that decrease competition, hamper specialization, and prevent the exploitation of economies of scale"); id. at 1099 (stating that their interpretation of tax nondiscrimination may be what the founders of the EU hoped to secure "by implementing the prohibition on discrimination").
(38.) Id. at 1106-14; see also id. at 1021 (stating that labor cases "illustrate the kinds of state tax practices that give rise to discrimination challenges"); id, at 1036 (stating that the goal of the article is to get a clearer understanding of what "the tax nondiscrimination principle" requires); id. at 1085 (stating that in the EU single market, nondiscrimination should be interpreted to require competitive neutrality).
(39.) Id. at 1109, 1112 n.274 (discussing capital income tax cases).
(40.) Id. at 1034-36.
(41.) Id. at 1106-14.
(42.) Id. at 1036.
(43.) Graetz & Warren, supra note 15, at 1602-03; Graetz & Warren, supra note 1, at 1212.
(44.) Graetz &Warren, supra note 1, at 1253.
(45.) See Genschel & Jachtenfuchs, supra note 2; Genschel et al., supra note 8, at 600; Kingston, supra note 12; Joachim Englisch, Tax Coordination Between Member States in the EU: Role of the ECJ 10 (2011) (unpublished manuscript) (on file with authors) ("In the last five years ... the ECJ has been particularly inclined to uphold discriminatory tax provisions based on the rule of reason .... This implies that the principles of free market access ... inherent to the internal market concept of the Union will have to be balanced against certain national tax policy preferences.").
(46.) E.g., Mason & Knoll, supra note 33, at 1022. The authors also speak of the goal of the treaties as "promot[ing] welfare" of member states and citizens. Id. at 1024. In a recent article, one of the authors, Ruth Mason, in contradistinction to this article co-authored with Michael Knoll, considers other common tax policy criteria such as equity, administrability, and internation equity to be important to how the U.S. Supreme Court and the ECJ should approach tax discrimination cases. Ruth Mason, Tax Expenditures and Global Labor Mobility, 84 N.Y.U.L. REV. 1540, 1585-93 (2009) (equity); id. at 1593-99 (inter-nation equity); id. at 1599-1604 (administrability).
(47.) Mason & Knoll, supra note 33, at 1086 n.195. They also say that they "do not argue in favor of a competitive neutrality interpretation for tax discrimination from economic or philosophical first principles." Id.
(48.) Id. at 1040-42, 1085.
(49.) Id. at 1020; see also Michael J. Graetz & Michael M. O'Hear, The "Original Intent" of U.S. International Taxation, 46 DUKE L.J. 1021, 1108 (1997) (illustrating the impossibility of simultaneously achieving capital export and capital import neutrality without tax harmonization).
(50.) Graetz &Warren, supra note 1, at 1216-19.
(51.) See, e.g., Mihir A. Desai & James R. Hines, Jr., Evaluating International Tax Reform, 56 NAT'L TAX J. 487 (2003).
(52.) Id. at 494. Capital ownership neutrality could also be accomplished by source- and residence-country taxation as long as the latter granted unlimited foreign tax credits for taxes imposed by the former, but no country grants unlimited foreign tax credits. Id.
(53.) Mason & Knoll, supra note 33, at 1043, 1047, 1053.
(54.) It is not uncommon in the economics literature to assume productivity is fixed. For a discussion of the literature and a model which examines wages differing from marginal product for certain workers, see Thomas Piketty, Emmanuel Saez & Stefanie Stantcheva, Optimal Taxation of Top Labor Incomes: A Tale of Three Elasticities 15-23 (Nat'l Bureau of Econ. Research Working Paper No. 17616, 2011), http://www.nber.org/papers/w17616.pdf.
(55.) For studies of taxi drivers' work decisions, see, for example, Colin Camerer et al., Labor Supply of New York City Cabdrivers: One Day at a Time, 112 Q.J. ECON. 407 (1997) ; Yuan K. Chou, Testing Alternative Models of Labour Supply: Evidence from Taxi Drivers in Singapore, 47 SING. ECON. REV. 17 (2002); and Henry S. Farber, Is Tomorrow Another Day? The Labor Supply of New York City Cabdrivers, 113 J. POL. ECON. 46 (2005).
(56.) Ernst Fehr & Lorenz Goette, Do Workers Work More If Wages Are High? Evidence from a Randomized Field Experiment, 97 AM. ECON. REV. 298 (2007).
(57.) Id. at 310-12.
(58.) Id. at 314-15.
(59.) For a good summary of the literature, see Stefano DellaVigna, Psychology and Economics: Evidence from the Field, 47 J. ECON. LITERATURE 315 (2009).
(60.) Corporations will, of course, differ in their flexibility in this regard depending on their ability to raise capital.
(61.) European analysts sometimes expressly include labor in their analyses of capital export neutrality (CEN) and capital import neutrality (CIN) by relabeling them CLEN and CLIN, respectively. See, e.g., Frans Vanistendael, In Defence of the European Court of Justice, 62 BULL. INT'L TAX'N 90 (2008). Compare Dennis Weber, Is the Limitation of Tax Jurisdiction a Restriction of the Freedom of Movement?, in ACCOUNTING & TAXATION WITH SPECIAL REGARD TO TRADING IN EMISSION RIGHTS & ASSESSMENT OF ECJ CASE LAW IN MATTERS OF DIRECT TAXES AND STATE AID 113, 118-21 (Michael Lang & Frans Vanistendael eds., 2007) [hereinafter ACCOUNTING & TAXATION] (arguing that the ECJ should adopt CLIN as its lodestar for decisionmaking), with John F. Avery Jones, Comments on the Conference Papers, in ACCOUNTING & TAXATION, supra, at 135, 140-41 (arguing that CLEN would be a better approach).
(62.) Mason & Knoll, supra note 33, at 1054 n.142.
(63.) Mason, supra note 46, at 1545 (suggesting "labor residence neutrality" as a criterion); id. at 1566 ("[A]bsent a specific policy objective to encourage or discourage international labor mobility, tax laws should not distort decisions regarding cross-border migration or crossborder work."); id. at 1577 ("Rather than simply working in a lower tax jurisdiction, taxpayers might change their state of residence in order to escape high taxes.").
(64.) Peter Diamond & Emmanuel Saez, The Case for a Progressive Tax: From Basic Research to Policy Recommendations, 25 J. ECON. PERSP. 165, 166 (2011).
(65.) Mason & Knoll, supra note 33, at 1038-39.
(66.) Looking for Work Abroad: Taxes, YOUR EUR., http://europa.eu/youreurope/citizens/work/ job-search/taxes/index en.htm (last updated Nov. 2011).
(67.) A European Commission portal aimed at facilitating job mobility indicates that about 2% of EU citizens currently live and work in a member state other than their country of origin (without specifying how long a period is involved). Jobseekers, EURES: EUR. JOB MOBILITY PORTAL, http://ec.europa.eu/eures/main.jsp?acro=job&lang=en&catId=52& (last visited Sept. 20, 2011). When new member states are admitted to the EU, there is sometimes a transition period that postpones full freedom of movement for a number of years for citizens of the new member state. See id. An important initiative that is expected to improve mobility of university graduates is the standardization of diplomas in the European Higher Education Area, begun by the Bologna Process in 1999. See Mobility, EUR. HIGHER EDUC. AREA, http://www.ehea.info/article-details.aspx?ArticleId=13 (last visited Sept. 20, 2011).
(68.) Jennifer Gordon, Free Movement and Equal Rights for Low-Wage Workers? What the United States Can Learn from the New EU Migration to Britain, U.C. BERKELEY L. SCH., CHIEF JUST. EARL WARREN INST. ON L. & SOC. POL'Y (May 2011), http://www.law.berkeley.edu/ img/Gordon_Issue_Brief_May_2011_FINAL.pdf.
(69.) International Mobility of the Highly Skilled, OECD OBSERVER, July 2002, at 1 (describing important intra-regional migration of the highly skilled in Europe).
(70.) Camille Landais, LeBronomics: Do Taxes Really Affect Location Decisions of High-Paid Workers?, STAN. INST. FOR ECON. POL'Y RES. Clan. 2011), available at http://siepr. stanford.edu/system/files/shared/pubs/Policybrief_jan2011.pdf (finding that the level of top tax rates has a large impact on the migration decisions of EU soccer players); Henrik Kleven, Camille Landais & Emmanuel Saez, Taxation and International Migration of Superstars: Evidence from the European Football Market (Nat'l Bureau of Econ. Research, Working Paper No. 16545, 2010), http://www.nber.org/papers/W16545.pdf [hereinafter Kleven et al., Taxation and International Migration of Superstars] (same); Henrik Kleven et al., Taxation and International Migration of Top Earners: Evidence from the Foreigner Tax Scheme in Denmark (2011) (unpublished manuscript), http://elsa.berkeley.edu/~saez/ kleven-landais-saez-schultz1_1_11denmark.pdf [hereinafter Kleven et al., Taxation and International Migration of Top Earners] (finding large response of foreigners changing residence due to a special lower top rate income tax scheme in Denmark). For a summary of trends of growth in labor mobility, see Mason, supra note 46, at 1547-50.
(71.) Census data suggest that about 13% of the U.S. population moves every year and that about 40% has moved over a five-year period. See Geographical Mobility: 2000 to 2005, U.S. CENSUS BUREAU tbl.1 (Jan. 18, 2007), http://www.census.gov/hhes/migration/ data/cps/files/ cps2005-5yr/tabo1-1.xls; Geographical Mobility: 2009 to 2010, U.S. CENSUS BUREAU tbl.1 (May 2011), http://www.census.gov/hhes/migration/data/cps/files/cps2010/tabo1-01.xls. Over a five-year period, about 7.7% have changed their state of residence.
(72.) Mason & Knoll, supra note 33, at 1038-39. Mason and Knoll claim that the tax treatment of a French resident who moves to Germany is no longer a concern of EU law. Id.
(73.) See, e.g., id. at 1031.
(74.) For a comprehensive list of the ECJ direct tax cases, see European Court of Justice Case Law, EUR. COMM'N TAXATION & CUSTOMS UNION, http://ec.europa.eu/taxation_customs/ common/infringements/case_law/index_en.htm (last updated Nov. 16, 2011). See also Genschel et al., supra note 8, at 598-600.
(75.) Consider, for example, just two of the cases Mason and Knoll examine in detail. In one instance, they analyze German and French residents who work in both Germany and France, when the only tax is an origin-based income tax on a country's residents wherever they work. Mason & Knoll, supra note 33, at 1049 fig.3. This produces a different tradeoff between labor and leisure for German and French citizens when the applicable tax rates are 29% and zero, respectively. This conclusion, however, depends on assuming away the possibility that a German worker would move to France or would remain there for more than six months to benefit from the French tax rate of zero. In another instance, they analyze a German income tax applied only to French residents working in Germany, concluding that the tax will distort only French residents' decisions about where to work. Id. at 1056 fig.4. This conclusion also depends on assuming away any possibility that a French worker would move to Germany or would remain there for more than six months to benefit from the zero German tax rate. Their entire analysis is grounded in such immobility.
(76.) Id. at 1047.
(77.) As Mason and Knoll indicate, see id. at 1047-51, differences in residence tax rates may, of course, affect choices between saving and consumption or between work and leisure. For further discussion of the limited role that tax-rate differentials play in their analysis, see infra notes 121-137 and accompanying text.
(78.) Mason and Knoll do not explicitly consider the possibility of jobs moving until they reach the possibility of nonuniform taxes. Mason & Knoll, supra note 33, at 1055-60. On the other hand, the subsequent numerical example of nonuniform taxes (Table 4) seems to depend on workers changing where they work, rather than jobs moving. See also id. at 1071 n.170.
(79.) Michael Spence & Sandile Hlatshwayo, The Evolving Structure of the American Economy and the Employment Challenge, COUNCIL ON FOREIGN RELATIONS (Mar. 2011), available at http://i.cfr.org/content/publications/ attachments/CGS_WorkingPaper13_USEconomy.pdf (documenting the increasing transfer of jobs out of the United States, other than in the nontradable sector dominated by government and health care).
(80.) See, e.g., Hans-Werner Sinn, EU Enlargement, Migration, and Lessons from German Unification, 1 GER. ECON. REV. 299 (2000), http://www.cepr.org/pubs/new-dps/showdp.asp ?dpno=2174 (providing lessons from German unification--emphasizing capital movements--for European policy regarding labor migration); Assaf Razin & Efraim Sadka, Fiscal and Migration Competition (Nat'l Bureau of Econ. Research, Working Paper No. 16224, July 2010), http://www.nber.org/papers/w16224.pdf (modeling productivity differences among workers based on differences in capital-labor ratios).
(81.) The two ECJ decisions chosen "to illustrate the controversies surrounding EU tax discrimination doctrine" both involve cross-border workers. Mason & Knoll, supra note 33, at 1030-33.
(82.) Late in the article, footnote 170 states that even if the assumption of fixed residence were relaxed, uniform taxes would still implement competitive neutrality under certain circumstances. Id. at 1071 n.170. The reasoning in the footnote is so abbreviated that we cannot evaluate its generality, but if the assumption of fixed residence is not necessary to the authors' argument after all, we wonder why the reader has been led through a multitude of matrices and examples in which that assumption is determinative of the outcome.
(83.) Thomas Horst, A Note on the Optimal Taxation of International Investment Income, 94 Q:J-ECON. 793, 795 (1980).
(84.) Michael J. Graetz, Taxing International Income: Inadequate Principles, Outdated Concepts, and Unsatisfactory Policies, 54 TAX L. REV. 261, 269-73 (2001).
(85.) E.g., U.S. TREASURY DEP'T, THE DEFERRAL OF INCOME EARNED THROUGH U.S. CONTROLLED FOREIGN CORPORATIONS: A POLICY STUDY 23-54 (2000); see also STAFF OF JOINT COMM. ON TAXATION, 102D CONG., FACTORS AFFECTING THE INTERNATIONAL COMPETITIVENESS OF THE UNITED STATES 248 (Comm. Print 1991) ("A policy that reduces all tax rates (applied to domestic and foreign source income equally) is superior [on efficiency grounds] to a policy of equal revenue cost that reduces tax rates only on foreign source income.").
(86.) STAFF OF JOINT COMM. ON TAXATION, supra note 85, at 247.
(87.) See, e.g., Desai & Hines, supra note 51, at 491-92. But see Mitchell A. Kane, Ownership Neutrality, Ownership Distortions, and International Tax Welfare Benchmarks, 26 VA. TAX REV. 53, 60-66 (2006) (contesting the import of the empirical evidence).
(88.) Mason & Knoll, supra note 33, at 1098.
(89.) Diamond & Saez, supra note 64, at 166.
(90.) See, e.g., Am. Bar Ass'n Section of Taxation, Report of the Task Force on International Tax Reform, 59 TAX LAW. 649, 660-61, 680-89 (2006); Harry Grubert & Rosanne Altshuler, Corporate Taxes in the World Economy: Reforming the Taxation of Cross-Border Income, in FUNDAMENTAL TAX REFORM: ISSUES, CHOICES, AND IMPLICATIONS 319, 320 (John W. Diamond & George R. Zodrow eds., 2008) ("[N]one of these benchmarks is satisfactory because the argument supporting them usually takes place within very simple models."); see also Kane, supra note 87, at 60-66 (finding empirical evidence inconclusive); Stephen E. Shay, Ownership Neutrality and Practical Complications, 62 TAX L. REV. 317, 323-24 (2009) (same).
(91.) Michael P. Devereux, Clemens Fuest & Ben Lockwood, The Taxation of Foreign Profits (CEN, CON and All That): A Unified View (June 27, 2011) (unpublished manuscript), http://www.sbs.ox.ac.uk/centres/tax/symposia/Documents/Lockwood_MD_%20CF %20final%20v3.pdf.
(92.) Mason & Knoll, supra note 33, at 1074.
(93.) Id. at 1086 n.195 ("[W]e do not argue that a competitive neutrality interpretation of nondiscrimination would do a better job of promoting economic welfare or any specific notion of the good, justice, or fairness than other possible interpretations that might be imposed on the member states.") In a recent article, Ruth Mason herself acknowledged that "more empirical evidence is needed to guide states' choice between the competing mobility benchmarks," which she describes as bearing on "where taxpayers work, how much they work, who works which job, and where workers reside." Mason, supra note 46, at 1559.
(94.) Mason & Knoll, supra note 33, at 1086 n.195, 1098.
(95.) Id. at 1040.
(96.) Id. at 1087.
(97.) Id. at 1099-1105.
(98.) Id. at l060-72.
(99.) Id. at 1040.
(100.) Id. at 1053. While here we take this definition as given for the purpose of understanding the logic of the authors' argument, we note that the vagueness of the definition leaves us somewhat unsure about its relationship to the authors' definitions of locational and leisure neutrality and to the relationship of leisure to competitive neutrality. See our discussion supra Part II.
(101.) See Mason & Knoll, supra note 33, at 1046 n.121.
(102.) Id. at 1060-72.
(103.) Id. at 1061-62.
(105.) Id. at 1063-64, 1069-71.
(106.) Id. at 1067 n.163.
(107.) At one point, the authors declare that competitive neutrality "requires" ratios that would satisfy proposition 1, but that is only a further definition or assumption, proving nothing. Id. at 1061 n.154. Footnote 163 does some of the work of proposition 3 by using algebraic reasoning to show that certain tax structures do not disturb the ratio of ratios. The careful reader of The Yale Law Journal will, however, see immediately that proposition 3 cannot lead to the conclusion above without some showing that propositions 1 and 2 are true.
(108.) Id. at 1062-69.
(109.) Id. at 1069.
(110.) A profit-maximizing German firm will compare how much Gunther and Francoise will each produce for the firm with how much each will cost the firm in salary. According to Table 4, Gunther will produce 100 if he works in Germany. In order to attract Gunther to Germany, the German firm must meet (or slightly exceed) the after-tax income Gunther would retain if he took the competing job in France. Gunther's after-tax income in France is 40. The German firm would therefore have to pay him 55-56 for Gunther to retain 40 after paying the 20% German source tax (11.11) and the 10% German residence tax under the ideal deduction (4-45) postulated in Table 4. The profit to the firm would thus be 44.44 (or slightly less) if it hired Gunther. On the other hand, Francoise will produce 150 for the German firm. In order to attract her to Germany, the German firm must meet (or slightly exceed) the after-tax income she would retain if she took the competing job in France. Her after-tax income in France is 64. The German firm would therefore have to pay her 100 for her to retain 64 after paying the 20% German source tax (20) and the 20% French residence tax under the ideal deduction (16). The profit to the firm would thus be 50 (or slightly less) if it hired Francoise. A profit-maximizing firm would hire Francoise (for profits of 50) rather than Gunther (for profits of 44.44). Mason and Knoll conclude the opposite, saying that the German firm would employ the less productive Gunther, rather than the more productive Francoise.
(111.) Mason & Knoll, supra note 33, at 1069.
(112.) See supra note 110.
(113.) Mason & Knoll, supra note 33, at 1062.
(114.) For readers who want more, we propose a final thought experiment to test whether the authors have demonstrated what they claim in this section. Consider the very simple Table 1 (on page 1062), which shows results for Francoise and Gunther in the absence of taxes. We understand the conclusion that Francoise will work in Germany because of her higher productivity there. Now let us introduce a very low nonuniform source tax in Germany that applies only to foreign workers. Assume the rate is 0.1%. Such a tax will, of course, affect the relevant ratios and the ratio of ratios. The authors insist at many points that such a nonuniform tax will therefore change the sorting of jobs. Id. at 1068-70. (Indeed, they also assert that no one else has ever before made this point. Id. at 1069 n.165.) But as far as we can tell from the facts presented in Table 1, Francoise will keep working in Germany because her higher productivity there makes her more valuable to her employer than the less productive Gunther, even though the ratios have changed. Perhaps the mysterious sentence quoted in the text accompanying note 113 would change the result, but the authors provide the reader no reason to think that the unspecified assumptions made in that sentence are plausible. Perhaps an assumption that the firm in Table 1 does not maximize profits would also change the result, but the authors provide the reader no reason to make such a strange assumption.
(115.) Id. at 1060. The idea of efficiency gains from worldwide taxation with unlimited tax credits was originated by Peggy Brewer Richman (later Musgrave) in the 1960s. See PEGGY BREWER RICHMAN, TAXATION OF FOREIGN INVESTMENT INCOME: AN ECONOMIC ANALYSIS (1963); PEGGY B. MUSGRAVE, UNITED STATES TAXATION OF FOREIGN INVESTMENT INCOME: ISSUES AND ARGUMENTS (1969).
(116.) Mason & Knoll, supra note 33, at 1060. Mason and Knoll note in footnote 152 that eliminating source-country taxation would also satisfy competitive neutrality, but they rule this out as unrealistic because no country forgoes source taxation. They do not indicate why they regard unlimited foreign tax credits as realistic even though no country has such a credit.
(117.) Id. at 1064.
(118.) See, e.g., PETER ANDREW HARRIS, CORPORATE/SHAREHOLDER INCOME TAXATION AND ALLOCATING TAXING RIGHTS BETWEEN COUNTRIES: A COMPARISON OF IMPUTATION SYSTEMS 447-50 (1996) (discussing the "composite tax principle").
(119.) Mason & Knoll, supra note 33, at 1075.
(120.) See supra Part III.
(121.) See, e.g., Commission Staff Working Paper: Company Taxation in the Internal Market, COM (2001) 582 final (Oct. 23, 2001), available at http://ec.europa.eu/taxation_customs/ resources/documents/company_tax_study_en.pdf (finding rate differences to be the most important corporate tax distortion within the EU); Kleven et al., Taxation and International Migration of Superstars, supra note 70 (finding that the level of top tax rates has a large impact on the migration decisions of EU soccer players); Razin & Sadka, supra note 80 (considering effects of tax rates net of benefits in home and host countries).
(122.) Mason & Knoll, supra note 33, at 1038. They recognize, of course, that many other factors are crucial to these decisions; indeed, they list thirteen others, including wage differentials, which are particularly large among EU member states. One estimate is that wages in the first group of Eastern European countries admitted to the Union ranged from 8% to 23% of those in Western European member states, such as Germany. Sinn, supra note 80, at 299. In 2003, hourly wages in the accession countries were only 14% of those in Germany. Hans-Werner Sinn, EU Enlargement, Migration and the New Constitution, 50 CESIFO ECON. STUD. 685, 686 (2004). Mason and Knoll assert, without any analysis or evidence, that adding these factors would not "dramatically" change their results. Mason & Knoll, supra note 33, at 1038 n.98.
(123.) Mason & Knoll, supra note 33, at 1029.
(124.) See supra notes 53-55 and accompanying text.
(125.) According to the definition of the ideal deduction method, a taxpayer who receives an amount of income (I) retains after paying source taxes ([t.sub.s]) and residence taxes ([t.sub.r]) the following amount: I-[t.sub.s]I-[t.sub.r](I-[t.sub.s]I), which is equivalent to I-([t.sub.s]+[t.sub.r]-[t.sub.r][t.sub.s])I. Interested readers can confirm that in Tables 3 and 4 (on pages 1066 and 1069), the taxpayers in each case retain (1-[t.sub.t]) of their income where [t.sub.t]=[t.sub.s]+[t.sub.r]-[t.sub.s][t.sub.r]..
(126.) Mason & Knoll, supra note 33, at 1069.
(127.) Mason and Knoll assert in note 249 that the ideal deduction method is consistent with graduated residence and flat-rate source taxes. Id. at 1103 n.249.
(128.) Mason and Knoll simply assume the existence of a "total tax" rate without explaining how it is calculated. Id. at 1067 n.163.
(129.) Id. at 1075-76.
(131.) A second problem created for the ideal deduction method by graduated rates is the exact scope of the authors' requirement that the source tax be "uniform," by which they mean that "source taxes are imposed at the same rate and upon the same base for both resident and nonresident workers." Id. at 1068 n.164. Once again, Mason and Knoll do not consider the implications of graduated rates for this limitation. Recall that the ideal deduction method includes the possibility of simply exempting foreign income (so there are no residence taxes), which is said to be the simplest version of the method. Id. at 1065 n.161. If we take the article's statement of the uniformity limitation seriously, we do not see how, for example, a country that chose to exempt foreign income (therefore levying only a source tax) could both (a) tax individual income produced by its residents within its borders at graduated rates, taking into account spousal income or children's circumstances, and (b) apply flat tax rates to income earned within its borders by nonresidents. Suppose a French and German resident both work in France, which does not tax foreign income, but applies graduated rates to the combined income earned in France by a French husband and wife. France has no way of knowing how much income the German worker's spouse earns from working in Germany or anywhere else. How can it possibly impose a tax on the German worker "at the same rate and upon the same base" it uses for the French couple?
This is a well-known problem in international income tax design. One possible solution is that all taxpayers reveal to every country in which they earn any income the amount of income they earn in any other country, with the total income then divided among all the countries on a fractional basis. See generally Kees van Raad, Nonresidents-Personal Allowances, Deductions of Personal Expenses and Tax Rates, 2 WORLD TAX J. 154 (2010) (describing "fractional taxation"). As Mason and Knoll indicate elsewhere in their article, that possibility is not, however, practical. Mason & Knoll, supra note 33, at 1027 n.56. When discussing their ideal deduction, the authors avoid addressing these kinds of issues, because they use only proportional rates in the examples and tables used to explain their ideal deduction. If they mean to prohibit a country from exempting foreign income and coupling graduated-rate source taxation of residents with fiat-rate source taxation of nonresidents, such a prohibition is not evident.
(132.) Mason & Knoll, supra note 33, at 1103-04.
(133.) Id. at 1103 n.249.
(134.) Id. at 1075-76. Mason and Knoll define a uniform source tax as a tax applied by the source country "to all workers with income from its territory, regardless of the workers' residence." Id. at 1045. A uniform residence tax is a tax applied by a country "on the same basis to all residents, regardless of the source of their income." Id. at 1047.
(135.) In their discussion of how the courts should apply the recommended standards, id, at 107685, the authors once again fail to work through the consequences of their analysis for a world in which individual income is taxed at progressive rates, taking into account family circumstances. Consider again the case we put forth supra note 131: A French and German resident both work in France, which exempts foreign income, but applies graduated rates to the combined income earned in France by a French husband and wife. In accordance with standard international practice, France decides to tax the German resident working in France using a fiat rate because it does not know her economic situation outside France. Assume that the European Court of Justice is to apply the partial competitive neutrality construction of nondiscrimination favored by Mason and Knoll. The judicial standard to be applied is that France must "assess uniform source taxes or uniform residence taxes, or both." Id. at 1078. As France has no residence tax, its source tax must be uniform, which means that if France wants to tax the German working in France, it must do so at the same rate and using the same tax base it applies to the French couple. But as we indicated above, that is an impossible task for France, because it does not know how much the German citizen's spouse earns in Germany or anywhere else. Above, we were discussing an analytical standard proposed by the authors. Here we are discussing the constitutional requirement they derive from that standard. For Mason and Knoll, it would apparently violate the European Treaties or the U.S. Constitution if a state enacted an individual income tax that exempted foreign income and combined graduated-rate source taxation of residents with flat-rate source taxation of nonresidents. We express no view as to whether such a tax system would be sensible, but we certainly do not think it should be constitutionally prohibited.
(136.) R. G. Lipsey & Kelvin Lancaster, The General Theory of Second Best, 24 REV. ECON. STUD. 11, 11 (1956). In Mason & Knoll, supra note 33, at 1099 n.235, the authors recognize the difficulty of making definitive statements about welfare in the presence of other distortions, but do not indicate how that affects their analysis.
(137.) The ECJ case law is confusing and has changed over time regarding when discriminatory taxation of a cross-border transaction in one member state will be upheld if offset by tax advantages in the other member state. Englisch, supra note 45, at 16-17. The ECJ has made it clear, however, that it will not impose a most-favored-nation requirement in bilateral tax treaties within Europe. See Case C-376/03, D. v. Inspecteur van de Belastingdienst/ Particulieren/Ondernemingen Buitenland te Heerlen, 2005 E.C.R. I-5821, paras. 49-63. See also the discussion of "reverse discrimination," infra notes 143-146 and accompanying text.
(138.) Mason & Knoll, supra note 33, at 1087-97.
(139.) Mason and Knoll make two other arguments that the ECJ's jurisprudence is consistent with competitive neutrality, but not locational or leisure neutrality. First, they argue that the idea that discriminatory taxes harm particular parties aligns only with competitive neutrality. Id. at 1092-93. They say that violations of locational neutrality do not create winners and losers (even relative winners and losers), but harm everyone. They apparently have in mind the idea that high taxation of capital in one country could be shifted to workers there (resulting in lower wages) and cause capital to flow to lower-rate countries (reducing the marginal pretax rate of return for capital in those countries). Id. While we agree that the incidence of any tax must take into account taxpayer reactions, there is no reason to believe that those reactions would mean that all taxpayers are harmed to the same extent by violations of locational neutrality. Suppose, for example, residence country R decides to tax foreign income significantly more heavily than domestic income. We think that an R company that engages primarily in foreign commerce will be disadvantaged more than an R company that engages primarily in domestic commerce. It is certainly true that the ensuing taxpayer adjustments (e.g., more capital remains at home in R, perhaps reducing the marginal rate of return to all R companies) may affect all taxpayers. But the same is true of taxpayer adjustments in response to violations of competitive neutrality. Suppose higher taxes in R cause an R company to lose a bidding contest to acquire a company in source country S (where there is no local tax) to a competing company from low-tax country L, when the R company would have outbid the L company in the absence of taxation due to the R company's greater expertise in the S company business. We agree that the R company is disadvantaged, but that is not the end of the story. As in the previous case, there will also be taxpayer adjustments here that need to be taken into account. Workers at the S company may have lower wages because L management is less efficient than the R company management would have been. All investors in R may earn a lower rate of return on their investment, because the R company capital remains at home, increasing the supply of capital there. Without a fully specified model of the relevant relationships, there is simply no logical basis for claiming that only competitive non-neutralities harm particular parties.
Second, Mason and Knoll argue that the two labor income decisions analyzed in their article provide "anecdotal evidence that the ECJ does not interpret the nondiscrimination principle to require locational neutrality or leisure neutrality," whereas those decisions could be reconciled with competitive neutrality. Id. at 1093. The authors do not put much weight on this argument, a judgment we share, particularly since the holdings of the two decisions are strongly criticized elsewhere in the article. See id. at 1093-96 (criticizing the holding of Case C-279/93, Finanzamt Koln-Altstadt v. Schumacker, 1995 E.C.R. I-225, that personal benefits must be available at least "once somewhere" as inconsistent with competitive neutrality); id. at 1094 n.218 (characterizing the facts in Case C-385/00, De Groot v. Staatssecretaris van Financien, 2002 E.C.R. I-11819, as violating both locational and leisure neutrality, but not competitive neutrality).
(140.) See, e.g., Sayde, supra note 20, at 401-03 (discussing the acceptance of reverse discrimination in the EU).
(141.) See id. at 401.
(142.) Mason & Knoll, supra note 33, at 1086 (emphasis added). Mason and Knoll also express the typical concern with protectionism, id. at 1022, and with nonresidents or residents with foreign source income receiving "worse tax treatment" than residents, id. at l029 (emphasis added).
(143.) MADURO, supra note 11, at 131.
(144.) See Kleven et al., Taxation and International Migration of Superstars, supra note 70, at 2.
(145.) See, e.g., Jim Shanahan, Is It Time for Your Country To Consider the "Patent Box"?, PRICEWATERHOUSECOOPERS LLP 4 (May 23, 2011), available at http://download.pwc .com/ie/pubs/2011_is_it time_for_your_country_to_consider_the_patent_box.pdf. Countries that have implemented a patent box regime include Belgium, France, Hungary, Ireland, Luxembourg, the Netherlands, and Spain. The United Kingdom has announced its intention to adopt a patent box regime effective in 20l3. Id.
(146.) For a description of the Danish system and its success in attracting highly skilled or highly compensated foreigners to Denmark, see Kleven et al., Taxation and International Migration of Top Earners, supra note 70. The Danish regime, which taxes residents' income based on their country of origin and the duration of residence, is obviously a nonuniform residence tax in Mason and Knoll's lexicon.
(147.) Compare Allied Stores of Ohio, Inc. v. Bowers, 358 U.S. 522 (1959) (sustaining an exemption for nonresident but not resident corporations from the state's property tax for merchandise stored in a local warehouse), with Wheeling Steel Corp. v. Glander, 337 U.S. 562 (1949) (invalidating a tax on accounts receivable owned by foreign corporations while exempting similar accounts receivable owned by residents and domestic corporations).
(148.) 1 JEROME R. HELLERSTEIN & WALTER HELLERSTEIN, STATE TAXATION [paragraph] 3.05[a] (3d ed. 2004).
(149.) See, e.g., Bos. Stock Exch. v. State Tax Comm'n, 429 U.S. 318 (1977). For a comprehensive discussion of the case law, see generally Walter Hellerstein & Dan T. Coenen, Commerce Clause Restraints on State Business Development Incentives, 81 CORNELL L. REV. 789 (1996).
(150.) Bos. Stock Exch., 429 U.S. at 336. The Court's most recent encounter with state tax incentives designed to encourage local economic development involved review of the controversial decision of the U.S. Court of Appeals for the Sixth Circuit in Cuno v. DaimlerChrysler, Inc., 386 F.3d 738 (6th Cir. 2004), which struck down Ohio's income tax credit for new in-state investment on the ground that it discriminated against interstate commerce in violation of the Commerce Clause, but at the same time sustained the state's personal property tax exemption for new in-state investment over Commerce Clause objections. The U.S. Supreme Court vacated the portion of the decision striking down the income tax credit on the ground that the plaintiffs, taxpayers who objected to the "corporate welfare" the state was providing to DaimlerChrysler, lacked standing to pursue their claims in federal court. DaimlerChrysler Corp. v. Cuno, 547 U.S. 332 (2006). At the same time, the Court denied the taxpayer's petition for certiorari from the portion of the opinion that sustained the tax exemption. Cuno v. DaimlerChrysler, Inc., 547 U.S. 1147 (2006). While the Court did not reach the merits of either claim, its disposition of the issues, at a minimum, reflects the continuing incoherence of the law in this area.
(151.) Mason & Knoll, supra note 33, at 1089.
(152.) Id. at 1098.
(153.) Id. at 1094.
(154.) Id. at 2095.
(155.) Graetz & Warren, supra note 1, at 1216-19.
(156.) Mason & Knoll, supra note 33, at 1096.
(157.) We do not, for example, know of any ECJ case law that would support the conclusion that the court would strike down an income tax that exempted foreign income and combined graduated-rate source taxation of residents with flat-rate source taxation of nonresidents, even though that system would apparently violate competitive neutrality, as defined by Mason and Knoll. See supra note 135.
(158.) Mason & Knoll, supra note 33, at 1106-15. For a brief discussion of a few of the differences between the ECJ's tax discrimination jurisprudence and that of the U.S. Supreme Court, see, for example, Mason, supra note 46, at 1614-20 ("[T]he ECJ and the U.S. Supreme Court have come to somewhat different conclusions about what the tax nondiscrimination principles of the EC Treaty and the U.S. Constitution mean....").
(159.) See Mason & Knoll, supra note 33, at 1108-10; see also Paul v. Virginia, 75 U.S. (8 Wall.) 168, 177 (1869) (concluding that "[c]orporations are not citizens within the meaning" of the Privileges and Immunities Clause); 1 HELLERSTEIN & HELLERSTEIN, supra note 148, [paragraph][paragraph] 4.12-.13.
(160.) Or. Waste Sys., Inc. v. Dep't of Envtl. Quality, 511 U.S. 93, 99 (1994). See generally 1 HELLERSTEIN & HELLERSTEIN, supra note 148, [paragraph] 4.14 (discussing case law).
(161.) Mason & Knoll, supra note 33, at 1111.
(162.) Id. at 1052 n.136, 1105-06.
(163.) Graetz &Warren, supra note 15, at 1622; Graetz &Warren, supra note 1, at 1219.
(164.) See supra notes 15-20 and accompanying text.
(165.) These requirements are summarized at Mason and Knoll, supra note 33, at 1074 tbl.5.
(166.) Id. at 1060-67.
(167.) That discrimination is shown in Mason and Knoll's Tables 3 and A by the higher pretax wage in Germany. Id. at 1066 tbls.3 & A. As the authors indicate, enough workers have shifted their place of employment to lower-taxed France in response to the tax differential to achieve the equality of after-source-tax wages shown in the tables. Id. at 1065 n.161; see also id. at 1052 n.134.
(168.) Id. at 1105-06.
(169.) Id. at 1072 n.171.
(170.) Id. at 1074.
(171.) Id. at 1104-05.
(172.) See Graetz & Warren, supra note 1, at 1220.
(173.) See id. at 1220-21 nn.116-18.
(174.) Id. at 1220 n.116 (citing, inter alia, Wolfgang Schon, Tax Competition in Europe- The Legal Perspective, 9 EC TAX REV. 90, 97-99 (2000)).
(175.) The definition used by Mason and Knoll for uniform source and residence taxes is set forth supra note 134.
(176.) Mason & Knoll, supra note 33, at 1105.
(178.) See, e.g., Case C-470/04, N. v. Inspecteur van de Belastingdienst Oost/Kantoor Almelo, 2006 E.C.R. I-7409 (exit taxes); Case C-446/03, Marks & Spencer plc v. Halsey, 2005 E.C.R. I-10837 (deductibility by parent company of losses of subsidiary located in another member state depends on treatment in other state); Case C-319/02, In re Manninen, 2004 E.C.R. I-7477 (stating that the dividend credit in the country of a shareholder depends on the treatment in the country of the dividend-paying corporation). For other examples, see Graetz & Warren, supra note 1, at 1221 n.120.
(179.) Englisch, supra note 45, at 16. Once both countries are considered, it becomes clear that Mason and Knoll's approach would not create the "level tax playing field" between domestic and cross-border transactions they claim to supply, Mason and Knoll, supra note 33, at 1021, 1096, because, as they admit, there would still be a "drag" on cross-border transactions, id. at 1114.
(180.) Id. at 1014, 1115.
(181.) Cf. Case C-147/08, Romer v. Freie und Hansestadt Hamburg (May 10, 2011), http://curia.europa.eu/juris/document/document.jsf?text=&docid=80921 &pageIndex=0&doclang=EN&mode=doc&dir=&occ=first&part=1&cid=364923 (holding that extension of public employee benefits to different-sex married couples but not to same-sex registered partners violates an EU directive prohibiting discrimination on the basis of sexual orientation). Regarding a same-sex marriage recognized in one member state, the European Commission's website contains the following statement: "In principle, your marriage is guaranteed to be recognised in all other EU countries--but this does not fully apply to same-sex marriages." Couples: Marriage, Your EUR., http://ec.europa.eu/youreurope/ citizens/family/couple/marriage/index_en.htm (last updated Aug. 2011) (emphasis omitted).
(182.) Ruth Mason herself recently recognized the impracticality of implementing the norm that she and Michael Knoll are now urging the ECJ and U.S. Supreme Court to impose on the EU member states and U.S. states. Mason, supra note 46, at 1581 ("[N]either LIN ["labor import neutrality," here labeled "leisure neutrality"] nor LON ["labor ownership neutrality," here labeled "competitive neutrality"] represents a practical policy goal at this time.").
AUTHORS. Michael J. Graetz is the Isidor and Seville Sulzbacher Professor of Law, Columbia Law School. He is grateful to the Milton Handler Research Fund at Columbia Law School and the European University Institute for support of this Response. Alvin C. Warren is the Ropes & Gray Professor of Law, Harvard Law School. The authors would like to thank Mihir Desai, Daniel Halperin, James Hines, Louis Kaplow, Alvin Klevorick, Peter Merrill, Alex Raskolnikov, and Stephen Shay for helpful comments.
|Printer friendly Cite/link Email Feedback|
|Author:||Graetz, Michael J.; Warren, Alvin C., Jr.|
|Publication:||Yale Law Journal|
|Date:||Mar 1, 2012|
|Previous Article:||The future of the Voting Rights Act: lessons from the history of school (re-)segregation.|
|Next Article:||The Twenty-Sixth Amendment enforcement power.|