2. The democratic polity as territorial enterprise and as club enterprise

A theory of taxation must have as its conceptual basis a theory of the state. How the state is conceived provides the criterion against which potential alternative theories of taxation are to be judged. The democratic polity, the principal subject of the present inquiry, is conceived here, following Rawls's (Reference Rawls1971: 84) well-known definition, as ‘a cooperative venture for mutual advantage’,Footnote ⁵ or, as I prefer to call it, a citizens co-operative. This label is meant to underscore two things. First, that the democratic polity is, like any private-law co-operative (Genossenschaft),Footnote ⁶ a member-governed organization, i.e. an organization in which decision-making authority resides ultimately with its members. Second, that, again like in any private-law co-operative, individuals voluntarily participate only if they expect to benefit by their membership in the joint enterprise.Footnote ⁷

This concept of the democratic state as a cooperative enterprise is at the core of James M. Buchanan's public finance theory. Buchanan's starting point is the contrast between two ‘opposing theories of the state’ (Reference Buchanan and Buchanan1960 [Reference Buchanan and Buchanan1949]: 8), an ‘“organismic” theory’ in which the state is ‘conceived as a single organic unit’ and an ‘individualistic theory’ in which ‘the state is represented as the sum of its members acting in a collective capacity’ (ibid.).Footnote ⁸ According to the individualistic assumptions, which he considers ‘the only appropriate ones for democratically organized societies’ (ibid.: 4), taxation can be viewed, so Buchanan (Reference Buchanan1984: 103) argues, ‘as the cost side of an inclusive fiscal exchange process’. Taxes are conceived ‘as “prices” that persons pay for the benefits provided by collectively financed goods and services made available to them by the government’ (ibid.).Footnote ⁹ From such perspective, Buchanan (ibid.) adds, ‘the limits to taxation are those determined by the preferences of the citizens themselves for collectively provided goods’.

In a democratic polity, a citizens co-operative, the authority to decide and act on behalf of the organized citizenry is delegated to an executive organ, the government.Footnote ¹⁰ For analytical purposes, the functions that a democratic government performs can be classified along various dimensions.Footnote ¹¹ The distinction between governmental functions that I want to draw attention to is, as noted above, between government's roles as territorial enterprise and as club enterprise.

In both its roles a democratic government acts on behalf of the citizenry, and its mandate is to serve interests that the members of the citizens co-operative share in common. The difference between its two roles lies in the nature of the goods it provides in either capacity, namely ‘territorial goods’ on the one hand and ‘club goods’ on the other. As ‘territorial goods’, I classify goods and services governments provide in their capacity as territorial enterprises. Their characteristic feature is that they are available to anybody who resides and/or operates within the territorial boundaries of the respective state.Footnote ¹² In other words, they are available to all jurisdiction users, citizens and non-citizens alike. By contrast, the rubric ‘club goods’ comprises goods and services that a government provides in its capacity as club enterprise. Access to these goods and services is limited to members of the respective citizens co-operative.

Figure 1 illustrates the relations between individuals in their different capacities, as citizens–members and as jurisdiction users, to the democratic state and to government in its two roles.

Figure 1. Government's dual role.

As jurisdiction users, individuals, citizens and aliens alike, are subject to the power of a state by virtue of their residing or operating within the state's territory. Accordingly, they can escape the respective government's authority by leaving the territory. In contrast to aliens, though, citizens are subject to their government's authority in a dual capacity, as jurisdiction users and as members of the polity. In the former capacity, they are free to decide whether they wish to use their home-state's territory for their private purposes or whether they prefer to pursue certain purposes – e.g. residence, employment, financial investment or business activities – in foreign jurisdictions. In other words, in their private capacity citizens are essentially in the same kind of relation to their home state as other jurisdiction users who, as non-citizens, use the state's territory for their private purposes. In their capacity as jurisdiction users, they can evade their home-state's authority by taking their private business elsewhere. By contrast, in their capacity as citizens–members they are subject to the rules that define the conditions of membership in their respective polity as long as they maintain their citizenship. They remain subject to the state's authority as a club enterprise, whether they reside in the state's territory or not, and can evade it only by terminating their membership in the citizens co-operative.Footnote ¹³

The remaining sections examine the implications that follow from the distinction between government's two roles for the theory of competitive federalism.

3. Competitive federalism and the two roles of government

In his classic contribution to the theory of competitive federalism, Charles Tiebout took issue with the received Musgrave–Samuelson doctrine ‘that no “market-type” solution exists to determine the level of expenditures on public goods’ (Reference Tiebout1956: 416).Footnote ¹⁴ The core problem, so Tiebout argued, concerns ‘the mechanism by which consumer-voters register their preferences for public goods’ (ibid.). If such a mechanism existed, he notes, ‘the appropriate benefit tax could be determined’ (ibid.). Finding the current method of solving the problem through the political mechanism – namely, combining ‘expenditure wants of a “typical voter” with an ability-to-pay principle on the revenue side’ (ibid.)Footnote ¹⁵ – ‘unsatisfactory’, Tiebout posed the question whether a set of social institutions might be determined that would ‘force the voter to reveal his preferences’ (ibid.) and would allow ‘to tax him accordingly’ (ibid.: 418). As solution, he suggested a regime of competitive federalism in which ‘consumer-voters’ are free to choose among multiple local governments offering different combinations of public goods and tax burdens. The working principles of such a regime he described as follows.

The fact I want to draw attention to is that in Tiebout's account local governmentsFootnote ¹⁷ are considered only in their role as territorial enterprises, as ‘localities’ in a ‘spatial economy’. Their role as club enterprises is ignored. When the ‘consumer-voter’ is said to move ‘to that community whose local government best satisfies his set of preferences’ (ibid.: 418), it is his preferences for territorial or location goods that he can thus register.Footnote ¹⁸ The choice between ‘locations’ does, however, not provide a mechanism by which individuals could register, as citizens–members of a polity, their preferences for club goods that different governments might offer in their capacity as club enterprises. Accordingly, Tiebout's claim that in the competitive regime he describes expenditures for community services ‘approximate the proper level’ (ibid.) applies to territorial or location goods only, not to club goods that a government provides exclusively for their citizens–members.

Nearly, two decades before Tiebout's contribution F.A. Hayek had published a paper on ‘The Economic Conditions of Interstate Federalism’ (Reference Hayek and Hayek1948 [Reference Hayek and Hayek1939]) in which he looked at competitive federalism as an institutional device to limit the power of government. In Hayek's account, as in Tiebout's, the competing sub-units in a federal system are exclusively considered in their capacity as territorial enterprises. In the regime Hayek envisaged, the federal government's principal task is to assure free mobility between the states of the federation by doing ‘away with impediments as to the movement of men, goods, and capital’ (ibid.: 255), thus creating ‘one single market’ (ibid.: 258). The resulting competition for mobile labor and capital would lead, as Hayek put it in a later contribution, to

Though Hayek speaks of competition for ‘citizens’, the competition he describes is in fact not between governments as club enterprises for citizens–members but only between governments as territorial enterprises for jurisdiction users.

The exclusive focus on governments’ role as territorial enterprises that Hayek's and Tiebout's contributions share is indeed a quite common feature in theories of competitive federalism. Explicitly endorsing Hayek's account, Weingast advocates a ‘market-preserving federalism’ in which sub-national governments have primary regulatory control over their economies while the federal government's task is to secure a common market by preventing sub-units from using their regulatory power to erect trade barriers (Reference Weingast, Ménard and Shirley2008: 155). The ‘induced competition among lower units in the federal structure’, Weingast (Reference Weingast1995: 5) concludes, implies ‘that policy choices will be disciplined by the ability of resources to move between jurisdictions’ and that only those policies will survive ‘that citizens are willing to pay for’ (ibid.). As he puts it:

Again, though Weingast speaks of ‘citizens’, the competition he describes does not permit citizens to register what kinds of club goods they may wish government to provide and for which they are ‘willing to pay’. It only allows individuals to register with their ‘location decisions’ their preferences for location goods.

In his approach to public finance, James Buchanan has put particular emphasis on the role of competition among governments as a supplement to constitutional constraints, which, as he argues, ‘may not offer sufficient protection against the exploitation of citizens through the agencies of government’ (Reference Buchanan2001 [Reference Buchanan1995]: 69).Footnote ²⁰ Like Hayek and Weingast, Buchanan suggests a federal structure in which the central government – though severely ‘restricted in its own domain of action’ (Reference Buchanan2001 [Reference Buchanan1995/96]: 70) – is sufficiently strong ‘to enforce economic freedom and openness over the whole of the territory’ (ibid.), while the ‘remaining political power is residually assigned to the several “state” units’ (ibid.). Under such a federal structure, Buchanan (Reference Buchanan2000 [Reference Buchanan1979]: 264) supposes, the ‘right of citizens to migrate freely, to vote with their feet or with their mobile resources, will limit the extent to which their demands for governmentally provided goods and services can be ignored by governmental units’. In emphasizing the ‘prospects for exit . . . as constraint on political control over the individual’, he notes, the ‘theory of competitive federalism is congenial to economists in particular because it is simply an extension of the principles of the market economy to the organization of the political structure’ (Reference Buchanan2001 [Reference Buchanan1995/96]: 80).Footnote ²¹

Again, even though Buchanan speaks of ‘citizens’ and their ‘demands for governmentally provided goods and services’ he looks in effect, like Hayek and Weingast, at individuals only as jurisdiction users who, if they do ‘not like the results of state or local political action, may move to another area or another location’ (Buchanan and Flowers, Reference Buchanan and Flowers1987; Dercks, Reference Dercks1996: 385).Footnote ²²

4. Federalism and two kinds of exit

As the above summaries of Tiebout's, Hayek's, Weingast's and Buchanan's accounts of federal competition illustrate, with their exclusive focus on territorial mobility they ignore the fact that democratic governments serve in the two noted roles and that, accordingly, we must distinguish between two different meanings of ‘exit’. Exit may mean, on the one hand, that individuals, in their capacity as jurisdiction users, leave a government's territorial domain, and it may mean, on the other hand, that they give up membership in their citizens co-operative and seek citizenship in some other polity.

Exiting as private law subject from a jurisdiction is about ‘location decisions’ (Weingast, Reference Weingast1995: 5), exiting as citizen from a polity is about ‘membership decisions’. As private law subjects individuals can split up their various activities – where to take residence, where to invest, where to work, etc. – between different ‘territorial enterprises’ according to how attractive they are with regard to their respective purposes. They do not give up their status as citizens–members, however, when they exit from their home polity's territory in one or more of these dimensions. In the extreme, they may take all their private activities elsewhere while maintaining their membership status in the polity. Reversely, they may exit as citizens–members from a polity while remaining with some – or even all – of their private engagements within its territorial boundaries.

The calculus of advantage on which individuals base their exit decisions will be characteristically different for location choices as opposed to membership choices. As jurisdiction users individuals can, as noted, distribute their various private activities across different (accessible) sovereign territories. They can decide separately for each type of activity which location to choose. By contrast, membership in a polity comes with an inclusive bundle of rights and duties, and the relevant comparison is between the inclusive bundles that different (accessible) citizens co-operatives offer. The flexibility and easy reversibility that individuals enjoy as jurisdiction users in their location decisions is typically absent in their choice of citizenship–membership in a citizens co-operative.

In the Tiebout–Hayek–Weingast–Buchanan approaches to competitive federalism, the distinction between the two kinds of ‘exit’ is simply ignored. When these authors speak of ‘consumer-voters’, ‘taxpayers’ or ‘citizens’ they refer in fact only to individuals in their capacity as customers of territorial enterprises, moving in a ‘spatial economy’ between jurisdictions. The issue of how individuals may register the preferences they harbor in their capacity as citizens–members of a polity remains undiscussed. To be sure, the failure to distinguish between individuals as jurisdiction users and individuals as citizens–members of a polity is not limited to the cited authors.Footnote ²³ It appears to be indeed a quite common feature in the literature on competitive federalism.Footnote ²⁴ The likely reason for such one-sided focus on government's role as territorial enterprise is that authors writing on the subject tacitly take a federal system like that of the United States as the standard model of a competitive federalism.

In the federal system of the United States, the primary citizenship is in the Union. US citizens acquire membership status in local communities and states by virtue of taking residence in the respective territories.Footnote ²⁵ This means that only the federal government can exercise effective authority over who obtains citizenship status.Footnote ²⁶ Given the provision that US citizens can freely choose their place of residence within the territory of the US, local communities and states can only passively register but not actively control who acquires membership status in the polity. The US rules for assigning citizenship – locating primary citizenship at the federal level while citizenship in sub-units is a matter of individuals’ residential choice – are, to be sure, a quite common feature in other federal systems, such as e.g. in Germany. They are, however, neither a necessary nor a universal feature. They are a matter of constitutional choice. A counterexample is, for instance, Switzerland where the primary citizenship is at the level of local communes and cantons, while citizenship in the Swiss nation derives therefrom.Footnote ²⁷ Another, more significant counterexample is the European Union to which I shall return below.Footnote ²⁸

What I want to draw attention to is the fact that the constitutions of federal systems – how competencies are assigned to the respective levels of government, and how the rules of citizenship are defined – can be differently designed, and that the differences they exhibit may have a significant effect on how the competitive dynamics among sub-units unfold. A general theory of competitive federalism must account for this fact. If a US-type federal system is, expressly or tacitly, taken as the standard model, it follows naturally that the focus is exclusively on territorial mobility as the principal driving force in competitive federalism, while the role of federal sub-units as club enterprises is lost sight of. Where membership in federal sub-units can be acquired simply by individuals’ unilateral residential choice, the ensuing competition among these sub-units will critically affect what kinds of public services they will be able to provide. It will in effect reduce them to their role as territorial enterprises,Footnote ²⁹ incapacitating them in their role as club enterprises. Given their inability to control admission to and exit from the ‘club’, they will not be able to provide sustainably genuine club goods because such goods will be subject to adverse selection.Footnote ³⁰ This is in particular so for the good ‘redistribution as social insurance’ that will be the subject of the next section.

5. Competitive federalism and redistribution

It is quite common to assert that redistributive policies are subject to adverse selection when governments have to compete for a mobile tax base.Footnote ³¹ R.A. Musgrave, to cite a particularly prominent source, has put it in these terms:

Summarizing his reading of the literature on the subject, he states:

Like many others writing on the subject, Musgrave apparently regards as an unquestionable fact that redistribution is a task of government, a task furthermore that cannot be carried out at the level of competing federal sub-units, but must be assigned to the central government. Yet, a theory that conceives of a democratic polity as a ‘cooperative venture for mutual advantage’ cannot treat these matters as unquestionable facts. It must provide arguments for why redistribution may qualify as a good in which the members of a citizens co-operative share a common interest. Furthermore, it must provide arguments for why citizens in a federal system may share a common interest in assigning the redistribution task to a particular level in the federal hierarchy.

There are, in fact, two kinds of arguments for why the members of a citizens co-operative may want their government to engage in redistributive policies. On the one hand, citizens may share a common interest in avoiding potential negative effects of great inequality in their home jurisdiction. Redistribution that is so motivated takes on properties of a ‘territorial good’. To the extent that it helps to make the jurisdiction a more attractive place for private purposes, jurisdiction users can be made to co-finance its production.

On the other hand, citizens may share a common interest in redistribution as a mutual insurance arrangement, covering the members of the citizens co-operative in case of need. It is the second variety that I shall focus on here.Footnote ³² That uncertainties about one's own – and one's childrenFootnote ³³ – future income-earning prospects can provide prudential reasons for citizens to agree to a regime of redistributive insurance has often been observed.Footnote ³⁴ As e.g. Sinn (Reference Sinn1997: 258) has put it:

This is the reason, Sinn concludes, why redistribution ‘as insurance . . . may be welcomed by all citizens before destiny has lifted its veil of ignorance’ (ibid.: 259).Footnote ³⁵ Yet, so he argues, despite the benefits that welfare state provisions can generate for citizens, governments may not be able to provide them because of the ‘increased difficulties of carrying out redistributive policies . . . if the factors of production are internationally mobile’ (Sinn, Reference Sinn1994: 90).Footnote ³⁶ If a ‘country's borders are opened and both capital and labor can freely migrate across them’, he posits (Sinn, Reference Sinn1997: 262), this liberty ‘will affect insurance through redistributive taxes since the government loses its power to enforce the payment of taxes’. This is, according to Sinn (ibid.: 264), so because the ‘good risks leave the insurance state’, creating an adverse-selection problem that destroys the viability of the insurance arrangement.

As noted before, claims about the effects of inter-governmental competition on governments’ ability to provide club goods such as social insurance cannot be assessed in the absence of a clear distinction between governments’ roles as territorial and as club enterprises. Governments’ power to tax in both their roles depends on their ability to make access to the services they provide contingent on the payment of required tax contributions. As territorial enterprises, governments can charge the taxes to finance their services as, in effect, user charges. They can make the permission to take advantage of the benefits their jurisdiction offers contingent on payments made by jurisdiction users. In a competitive environment, the taxes jurisdiction users can be made to pay will tend to take on the character of prices, differentiated according to the respective uses made (as residents, investors, tourists, etc.).

The situation is categorically different for the services governments provide in their capacity as club enterprises. Club goods like social insurance cannot be provided in the same manner as territorial goods. By contrast to user charges, the taxes governments need to raise to finance their services as club enterprises must be charged in the form of membership dues.Footnote ³⁷ They are payments for the option to partake in the bundle of services that the club provides for its members, not payments charged separately according to the actual use made of the respective services. Membership dues must be paid as long as one maintains one's membership status, independent of the extent to which one makes use of the options it offers, even if, as may well happen in the case of social insurance, the option never materializes.Footnote ³⁸

In order for governments to be able to charge taxes – such as redistributive taxes – for their services as club enterprises they must be able to control the conditions under which new members are admitted to the citizens co-operative. Nation-states typically have this ability. They have the authority to decide whom they admit as citizen–member. Contrary to what authors like Sinn suggest, the need to compete for a mobile tax base does not per se undermine states’ ability to charge redistributive taxes. Where states lack this ability competition can only be a proximate, not the ultimate reason. The ultimate reason most surely is their failure to distinguish adequately between the taxes they can charge as territorial enterprises, and the taxes they must charge in order to function as club enterprises. Accordingly, and in light of what has been said above about the two reasons for redistribution, Sinn's (Reference Sinn1994: 101) argument that ‘mobile factors cannot be taxed for redistributive purposes’ needs to be specified. To the extent that redistributive policies make the jurisdiction a more attractive place for jurisdiction users, the latter can be made to co-finance its production. By exiting from the jurisdiction, they can avoid the respective tax, but must also forgo the corresponding advantages the jurisdiction offers. By contrast, to the extent that redistributive taxes serve to finance the club good ‘social insurance’, a good from which jurisdiction users as such do not benefit, the latter can avoid paying the tax by exiting the jurisdiction without corresponding sacrifice.Footnote ³⁹ Redistributive taxes to finance the club good ‘social insurance’ must be collected as membership dues that can only be evaded by giving up one's membership status, thereby sacrificing one's claim to insurance coverage.Footnote ⁴⁰

In federal systems, the ability of lower level governments to function as club enterprises depends, as already noted, on the respective citizenship rules. If, as in US-type federal systems, primary citizenship is at the union level while citizenship in states and local communities is a matter of individuals’ unilateral residential choice, lower level governments are in effect, as I have argued, reduced to their role as territorial enterprises. It is for this reason, not because ‘redistribution is intrinsically a national policy’ (Stigler, Reference Stigler and Phelps1965: 172),Footnote ⁴¹ that in such cases the power to collect redistributive taxes will be assigned to the central government.

Notice that to deny that redistribution is ‘intrinsically’ a central government's task is not meant to deny that there may well be prudential reasons for citizens in a federal system to choose to put the central government in charge of redistributive policies. In fact, the historical trend towards a centralization of redistributive policies that had formerly been the domain of local communitiesFootnote ⁴² has its apparent causes in the difficulties the local provision of social insurance faced in an increasingly mobile world.Footnote ⁴³ Yet, specifying prudential reasons why citizens may want to centralize redistribution as social insurance is quite different from simply positing that redistribution requires ‘intrinsically’ its centralization. To keep this difference in mind is important when one looks at a federal system such as the European Union.

6. The case of the EU: governments’ dual role and ‘redistribution policy’

It is in line with his above-cited claims about the effects of inter-governmental competition when Sinn (Reference Sinn1994: 97) charges that the EU's Single Market Program with its four liberties (free movement of goods, services, labor and capital) must lead to ‘the breakdown of national redistribution schemes under institutional competition’.Footnote ⁴⁴ He concludes that because of the ‘impossibility of redistribution policy’ on part of the member states, tax rates will ‘have to be harmonized across countries or chosen by a central agency’ (ibid.: 99).Footnote ⁴⁵

In terms of what has been argued above, what Sinn fails to take into account in his ‘impossibility’ claim is that in assessing the effects of competition on redistributive policies one needs to distinguish between the member states’ two roles.Footnote ⁴⁶ It is simply not the case that ‘(s)tates could establish redistribution schemes only when borders are closed, not when they are open’ (Sinn, Reference Sinn1990: 10). There is a categorical difference between ‘open borders’ in the sense of allowing free migration of jurisdiction users across national boundaries and ‘open borders’ in the sense of free admission to the services, in particular redistribution as insurance, governments provide as club enterprises for their citizens.Footnote ⁴⁷ The EU's Single Market Program is about the first, not the second. The commitment to the common market's four liberties requires member states in their capacity as territorial enterprises to allow citizens of all EU member states to operate, as jurisdiction users, within their respective territories. Yet, this commitment does not require them in their capacity as club enterprises to grant citizens from other EU member states free access to the club goods they provide for their own citizens. In the EU primary citizenship is with the member states, while the EU citizenship derives therefrom.Footnote ⁴⁸ Different from the US-type federalism, EU citizens cannot acquire citizenship of a member state simply by taking residence therein. Accordingly, as providers of club goods like social insurance member states can avoid problems of adverse selection as long as they enforce an appropriate taxation regime, a regime that carefully distinguishes between the state's two roles. Interestingly, if only passingly, Sinn (Reference Sinn1994: 100) acknowledges this fact when he notes that redistribution can work with free migration if ‘a strict nationality principle for redistributive taxation is applied’.Footnote ⁴⁹

Adverse selection undermining independent national redistribution policies within the EU is not a consequence of free migration per se. It may be a consequence, though, of EU rules that prevent national governments from exercising effective control over whom they admit to the services they want to provide as club enterprises for their citizens–members. A tendency to adopt such rules, and for existing rules to be so interpreted, can indeed be observed in the EU, caused, as I posit, by a failure to distinguish properly between the member states’ two roles. The failure is particularly visible in some interpretations of the non-discrimination principle that is at the core of the Single Market Program. As noted before, the principle requires national governments in their capacity as territorial enterprises to treat citizens of all EU member states equally in their capacity as jurisdiction users, prohibiting them from granting privileges to their own citizens in their capacity as jurisdiction users.

As a defining characteristic of the common market, the non-discrimination principle does per se not apply to member states in their capacity as club enterprises, and there is no intrinsic reason why a principle that is constitutive of the common market should be extended to the club goods the individual member states provide for their citizens. Yet, ambiguities about what the non-discrimination principle implies for member states’ redistributive policies have been created by certain EU regulationsFootnote ⁵⁰ and directivesFootnote ⁵¹ as well as by rulings of the European Court of Justice.Footnote ⁵² One of the ECJ's recent rulings may serve as an illustration.

In the Case C-333/13, the ECJ had to decide whether the Sozialgericht Leipzig, Germany, had justly denied benefits according to the German system of basic provision (Grundsicherung) to a Romanian mother and her son, who had taken residence with the mother's sister in Leipzig.Footnote ⁵³ The ‘Judgment of the Court (Grand Chamber)’ of November 11, 2014, covers 20 pages to answer a question that could have been answered in one sentence if the Court had been guided by the unambiguous understanding that the non-discrimination principle binds the member states in their role as territorial enterprises but not in their activities as club enterprises. If guided by such an understanding the Court could have simply stated that the right to take residence in any member state does not imply the right to be admitted to the club goods the respective state provides for its citizens.Footnote ⁵⁴ Instead, the judgment draws repeatedly on the argument, incidental to the principal issue, that persons exercising their right to move and reside within the EU should not ‘become an unreasonable burden on the social assistance system of the host Member State’. Pondering various directives and regulations pertinent to the issue the Court arrives at the conclusion that they

I should emphasize that by insisting on a clear distinction between the EU member states’ roles as territorial and as club enterprises I do not mean to rule out that citizens in Europe, through their representatives, may choose, if they wish, to extend the non-discrimination principle to such matters as redistributive policies. The point I want to make rather is that such extension should be the subject of explicit and well-considered constitutional choice. It should not creep in tacitly, due to a mistaken interpretation of what the common market's non-discrimination principle requires.