Jump to ContentJump to Main Navigation
Rescaling the European StateThe Making of Territory and the Rise of the Meso$

Michael Keating

Print publication date: 2013

Print ISBN-13: 9780199691562

Published to Oxford Scholarship Online: January 2014

DOI: 10.1093/acprof:oso/9780199691562.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2022. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. Subscriber: null; date: 17 May 2022



(p.154) 8 Norms
Rescaling the European State

Michael Keating

Oxford University Press

Abstract and Keywords

Rescaling poses two normative questions. The first concerns social solidarity. It has been argued that the weakening of the nation-state threatens solidarity and distribution because of the loss of affective solidarity and the ability of privileged groups to engage in partial exit. The first argument rests upon the assumption that states and nations are congruent so that territorial specificity necessarily weakens community. Yet it may be that solidarity lost at one level may be gained at another. Where the sub-state community is more solidaristic, decentralizaton may enhance solidarity. The second argument is more serious. Rescaling to the european level has not been accompanied by mechanisms for redistribution. On the other hand, evidence for a race to the bottom has proved elusive. Rather, different solidarity regimes are emerging around certain services, while state-level redistribution remains strong. The second normative question concerns self-determination. This has been a conundrum in political theory, because of the difficulty of defining the people who have the right, and the lack of alignment of people and territory. Modern understandings of both groups and territory see these as socially constructed and changeable. In some ways this complicates matters further. On the other hand, a more flexible concept of both opens up new ways for their mutual adaptation and for effective.

Keywords:   solidarity, welfare, self-government, autonomy

Normative implications of rescaling

The nation-state has borne a heavy analytical burden. It sets the boundaries of political authority (the state), defines identity and culture (nation), delineates social systems, and bounds recognizably distinct economies. All of these are underpinned by the normative principles of sovereignty, nationality, and solidarity. Rescaling puts all these normative principles in question, or at least shifts their locus to new spaces. Two issues in particular are reframed. Solidarity is no longer exclusively framed within the nation-state context but arises at multiple levels. Self-determination claims have proliferated and often strengthened but are also detached from the classic state form. The resulting questions do not have straightforward answers or, arguably, any ‘answer’ at all, but are worked out in new forms of compromise.


The rationale for redistribution in general is threefold. First, there are abstract questions of justice, on which opinions differ, some arguing only for the relief of poverty, others for greater equality. Whichever criterion is used, the idea is, in principle, universal, extending to all humanity, but in practice limited to specific communities. The second rationale is economic and practical. Economic performance can be enhanced through investment in human capital, bringing into the workforce excluded social groups and individuals. It has been argued that greater interpersonal equality in itself is associated with enhanced economic performance (Wilkinson and Picket, 2010). We have noted how regional policies in the post-war state were justified by the need to improve allocative efficiency and overcome market failures. There is a utilitarian argument that it makes economic sense to pool coverage for risks such as unemployment or illness across larger areas and populations. The third rationale is political. States redistribute income to maintain the loyalty of (p.155) social groups with a potential for disruption of the economic and social equilibrium or party politics. Bridging the spheres of economics and politics was the ‘Keynesian welfare state’, which linked a technique of economic management to a system of social provision. Business was assured of social stability and public goods in exchange for taxation to support welfare. Labour could balance gains in the ‘social wage’ against the constraints of capitalism.

In principle, social solidarity operates among individual humans, providing a specified minimum for everyone. In practice, however, solidarity and redistribution are rarely based on the individual alone but operationalized by particular status, such as age, family responsibility, or labour-market position. Many countries have established universal and equal access to education and health services but even these are sometimes segmented, while cash benefits are usually tied to age, family, or occupational status. Gender and age have become more salient in defining common interests and needs, either because they are proxies for individual conditions or because they have been theoretically linked to the generation and reproduction of inequalities. The use of territory as a basis for defining needs and the idea of territorial solidarity and distribution are more controversial. There may be a public interest in maintaining territorial communities for cultural or historical reasons and people may have a legitimate interest in wanting to remain where they are without having to move in order to enjoy equal access to employment and services (Storper, 2011b). On the other hand, territorial redistribution may contradict inter-personal equity or non-territorial group-based solidarity. If the emphasis is on equalizing across territories, then rich people in poor regions may benefit over poor people in rich regions. If welfare systems are territorially differentiated, then similarly-endowed individuals may have different entitlements in different places.

In practice, however, welfare systems are always territorially differentiated; it is just that this territory has historically that of the nation-state (Clarke, 2004). T. H. Marshall (1992), in the aftermath of the Second World War, famously formulated the concept of social rights as a third set of rights, following legal and political rights. Such rights posed a challenge to the market economy as the basis for allocating resources, although at the same time they could serve to legitimate it. Like other rights, they could help to legitimate the political order, which in turn provided the institutional basis to realize them. Contending social groups could be brought to share a common political and social space and integrated into the constitutional order. Such reasoning assumes the normative primacy of the nation-state, although Marshall himself took this for granted and did not consider the role of welfare in integrating the constituent nations of the United Kingdom into a common political space. Other observers have more explicitly justified the use of the nation-state to define welfare entitlements. David Miller (1995) has argued that there are principled reasons for privileging solidarity to fellow nationals. Others have (p.156) merely argued that solidarity is in fact stronger where underpinned by national identity or that the state level is wider and thus more universal than that of sub-state territories. Welfare state scholars have described it as part of the broader process of boundary building, in which a number of distinct systems were enclosed in the same territorial borders, integrating national space while differentiating it from the outside (Rokkan, 1980, 1989; Keating, 1988, 1998; Bartolini, 2005; McEwen and Moreno, 2005). As the economic system, the welfare state, governmental institutions, political representation, and cultural reproduction coincided they locked social classes into the same space and provided an underpinning for positive-sum social compromises.

Welfare states tended to be centralized and federalism may work against the construction of welfare states, by providing veto points, frustrating territorial transfers, and discouraging broader solidarities. Where federal countries did develop welfare states, this was at the expense of federal differentiation, as in Germany or (to a lesser extent) Canada (Banting, 2006). Theories of federalism have usually argued that redistribution should be handled at the federal level in order to avoid negative externalities and market distortions, including welfare migration, as well as to ensure the unity of the market. Social democratic parties supported centralization, arguing that only unitary government could redistribute resources and command the authority to confront the market. Shared affective identities, promoted by state-building elites, reinforced the tendency to unity and centralization.

Rescaling has put these economic, political, and normative assumptions in question. It has served to construct and reinforce the notion of territorial interests, alongside class, age, gender, and other collective interests. Claims that the nation-state is the highest level of solidarity are discredited by the emergence of wider spatial scales, including the European Union, where similar normative and instrumental arguments about solidarity can be made. The ideological and normative supremacy of the ‘nation’ is called into question by the demystification of the nation-state and the emergence of rival nation-building projects within and across states. Spatial Keynesianism can no longer convincingly bridge the fields of territorial economics and welfare, when wealthy regions see neither a cultural rationale for territorial transfers (based on shared identity) nor an economic one (in that the money comes back in the form of orders for their products). Territorial competition appears inherently at odds with inter-territorial solidarity. Market promotion has migrated to sub-state levels, while market regulation is increasingly supranational. Market correction, in the form of the welfare state, remains largely a state responsibility and has resisted Europeanization. These different functional systems thus no longer complement each other at the same level and within the same institutions (Bartolini, 2005). Moreover, distribution can no longer be confined to explicit income transfers since economic strategies and (p.157) many services are implicitly redistributive, so putting the old spatial division of tasks in question (Chapter 7).

It is often hypothesized that this form of rescaling will generate a ‘race to the bottom’ as states, regions, and cities seek to sustain competitiveness by cutting welfare standards. States will offload social welfare costs by decentralizing to regions and localities and these in turn will be forced to cut in the interest of attracting and retaining footloose investors. National solidarity will give way to regional egotism, boosted by the re-emerging politics of identity. Within regions, those classes and sectors that are most internationally competitive will have an advantage, able to defect from social compromises and venue-shop among spatial scales for the most favourable treatment. The argument is a staple of American urban political economy, in which cities are said to be forced to keep social provision to a minimum in the race for competitive advantage (Peterson, 1985). As the social inequalities increase so does opposition, squeezing city governments between external competitive pressures and internal social demands (Keating, 1991; Kantor, 1995). There is consequently a threat to both interpersonal and territorial solidarity.

Yet it is a mistake to believe that if social solidarity weakens at the level of the ‘nation-state’ this necessarily represents an absolute decline in solidarity. Modernist writers tended too easily to elide nationality and citizenship in a defence of the homogeneous and unitary nation-state; and then assume that this represents a more universal form of order. This takes us back to the modernist contradiction noted in Chapter 1, that the nation-state is seen both as a bounded community and a form of universal normative order. In fact, state-wide solidarity is merely one form of particularism among others. Segall (2007), for example, takes me to task for suggesting that social solidarity could be located at the ‘sub-national’1 level of Scotland rather than at the level of the (British) political community. In Scotland, the term ‘national’ commonly refers to the Scottish nation as well (confusingly) as to the state level, and both are political communities which can equally well claim to be an appropriate basis for solidarity. The error is compounded where the territorial community or ‘stateless nation’ is defined as a closed ethnic group, while the nation-state is presented as an ethnically-neutral or universal space (Keating, 2001a). Then any devolution of social responsibilities to the sub-state level becomes by definition an infringement of citizenship in the name of ethnic particularism or discrimination.2

(p.158) The Commission on Scottish Devolution (2008) established by the unionist parties in response to the Scottish Government’s proposals for independence, makes a lot of social citizenship in its interim report, but assumes that this in itself is an argument in favour of the United Kingdom. A less prejudicial way of proceeding would be to ask just what forms of citizenship and nationality are developing at the state, sub-state, and trans-state levels and how these might be consistent with normative conceptions of solidarity and social citizenship. This requires that we cease unthinkingly to identify state, nation, and political community, and treat these as distinct concepts, related contingently and differently in different places. It is thus incoherent to argue that the Norwegian nation is an appropriate unit for the expression of social solidarity but the Scottish nation is not, or that Canada is appropriate but Quebec not, merely because Norway and Canada happen to be states recognized in the international order.

On the other hand, we cannot assume that the old forms of solidarity will be re-established at new territorial levels. Welfare states have rested upon shared social citizenship, on assumptions about nationality and national solidarity, and on institutions and arguments about efficiency, trade-offs and social compromises. As these are more difficult to sustain at the level of the nation-state because of the effects of boundary transformation, territorial unbundling, and economic competition, then it follows a fortiori that they cannot be re-established in exactly the same way at the sub-state level. Yet new forms of social solidarity may be viable at new levels. If the devolved territory is more solidaristic than the state as a whole, then social solidarity and redistribution may be greater (Noël, 1999). Where regions and stateless nations have a strong sense of shared identity, this might replace state-level identity as the focus for solidarity. Just as states have used welfare provision to build national identity, citizenship, and solidarity, so governments in devolved territories engaged in region building or ‘stateless nation building’, may do the same (Keating, 1997; McEwen and Moreno, 2005; Béland and Lecours, 2008). New forms of citizenship could emerge at new institutional levels and there is some evidence that solidarities might be stronger among people with stronger sub-state territorial attachments (Berg, 2008). So the repeated concerns by the Institute for Public Policy Research3 and the Commission on Scottish Devolution (2008) that devolution will somehow undermine social solidarity are (p.159) puzzling, since the welfare settlement appears more under threat at the centre, while the devolved territories have defended it; a similar argument is made about Quebec. Nor is it the case that smaller units are necessarily less able to engage in distribution because they have a smaller resource base; the most solidaristic states in Europe tend to be smaller. Where boundaries have been drawn to exclude lower income people and so hinder redistribution, as has happened in metropolitan areas (Chapter 5), then solidarity is curtailed but this does not necessarily apply to all regional levels. That is, rather, an empirical question.

This still leaves the question of inter-territorial solidarity and the level at which it should be addressed (Van Parijs, 2004). Interpersonal and inter-territorial solidarity may be connected to the degree that those pressing for less inter-territorial redistribution are also less inclined to share resources internally, but this is by no means necessarily the case. On the contrary, neo-liberals wanting to roll back the state have often been rather centralist constitutionally while social democrats over recent decades have become less hostile to decentralization. In principle, territorial solidarity can be reconciled with decentralization by providing block grants, which devolved governments are free to spend according to their own priorities, and indeed this has been the tendency over recent years. This poses issues both of principle and of practice since it is necessary to decide what should be equalized and how it is to be defined and measured. Systems of fiscal equalization typically have a resources element, in which funds are transferred from richer to poorer regions, and a needs element, which gives more to needier places. Richer can be defined in a number of ways. GDP per capita is one, but does not necessarily reflect incomes. Other measures are disposable incomes, fixed wealth, or tax base. Defining need is even more contentious. In principle it refers to the differing costs of providing the same level of welfare in different conditions, particularly where there is social stress and deprivation but the judgements involved are often subjective and political. So the higher cost of providing services in two languages may, for one person, be an indicator of need but, for another, a matter of political choice. Measures of both resources and needs are sensitive to the way boundaries are drawn, to include or exclude major production sites or households. In practice, the boundaries are drawn around politically-defined communities, which might not capture the geography of wealth and need. As a result, equalization is managed through sequential compromises rather than by framing a single overarching rationale.

(p.160) Even if equalization can be agreed, devolution does mean that individuals will be treated differently according to the area they live within, as devolved governments make different policy choices. Such differences are often criticized in the United Kingdom as a ‘post code lottery’ implying that they are arbitrary. This would be true it they resulted merely from the malfunctioning of the administrative system, as does happen in some centralized states. Where it is the result of local political decision it is an anomaly and an injustice only to those who see social citizenship as working purely at the level of the whole state but not above or below it.4 Critics of devolution in the United Kingdom, Spain, Italy, and Germany have insisted that devolution should not affect unitary social citizenship, interpreting this to mean that there should not be differences in entitlements to any particular service (d’Aloia, 2003). Such a principle, however, is difficult to sustain either on the grounds of normative universalism (which would point to global equity) or on the grounds of communitaristic solidarity (which could operate at various levels). It is another example of the statist blind that still affects such large areas of social science.

Upward rescaling to the European level is affecting systems of social solidarity at both state and regional levels as the dissociation of economic regulation from social policy may have allowed for forms of ‘partial exit’ from national solidarity (Bartolini, 2005). The design of European institutions, including the European Court of Justice, has built in a strong preference for market solutions, which may undermine national systems of labour protection and favour the marketization of public services (Sharpf, 2009; Mény, 2013). Monetary integration in the Euro zone, with its concomitant centralized control over deficits and debt, further constrains national welfare systems, and penetrates all the way down the scale to the local level. Given the weakness of affective solidarity at the European level, provision for territorial transfers is minimal, among either states or regions. On the other hand, national welfare compromises have generally held against pressures for a race to the bottom, as have their regional variants. While there is no European-level welfare state, there is still a recognition of the value of social welfare and some commitment to securing it through mutual recognition and the Open Method of Coordination, a form of horizontal rather than vertical integration. The result is that welfare is less institutionalized than market competition but not entirely absent and that the battle for a social Europe is still being fought.

So multiple levels of social solidarity and welfare are emerging, none of them as secure as the classic post-war national welfare state, while the scale at which the issue of solidarity might most appropriately be addressed is a politically-contested question. There may be a trade-off between solidarity at (p.161) one territorial level and solidarity at another, and solidarity at different levels may take different forms, be thicker or thinner, and rely on different policy instruments. Governments have often sought to reconcile these dilemmas and the tension between competitive regionalism and social solidarity by arguing that, in the long run, there is no conflict since growth will benefit everyone. Another strategy involves a discursive turn, playing on the meaning of competition. So the British Government in the early 2000s, for example, proclaimed the goal of regional policy as being to make all regions more competitive, at the same time as reducing regional disparities. The European Commission has coined the term ‘territorial cohesion’, denying that this (as represented by the Structural Funds) is a redistributive policy and insisting that it is about improving regional competitiveness. The problem with these formulations is that competition is by definition a relational term and that it is not possible for all regions simultaneously to become more competitive, as opposed to more productive. In practice, the Structural Funds are able, politically, to serve as a vehicle for inter-territorial redistribution while sailing under the flag of economic orthodoxy.

The evidence

Territorial distribution has become an increasingly salient issue as devolution creates institutional actors and the notion of territorial interest is strengthened. There has been a ‘revolt of the rich’ as politicians in Catalonia, northern Italy, Flanders, London, and the southern and western German Länder argue that they cannot afford to finance transfers to poor regions, given their own need to compete internationally. Redistribution occurs in two ways, by the disproportionate impact of central spending in certain regions, and by explicit intergovernmental transfers under fiscal equalization. Arguments in Belgium have centred on the former, notably within the social security system, which Flemish interests argue favours Wallonia because of its social structure and antiquated industrial model.5 As decentralization proceeds, however, the redistribution becomes more transparent as it takes the form of intergovernmental transfers. In Catalonia there is a consensus across the political parties, the right and the left, nationalists and non-nationalists, and business and labour, that they pay too much into the Spanish treasury and get too little in return. There is less consensus on the exact solution but all point to the need for more fiscal autonomy, less equalization, and, ideally, a system in which Catalonia collects its own taxes and passes on a (p.162) sum to Madrid for common services, as happens in the Basque Country. In northern Italy the Lega Nord has made capital out of transfers to the south and the need for ‘fiscal federalism’. The southern German Länder have long complained about the scale of transfers to the north and then to the east. In the United Kingdom the main complaints have focused on the alleged favourable treatment of Scotland via the Barnett Formula—although this is in fact a mechanism for convergence in spending levels.

While actors in prosperous regions consistently criticize the scale of transfers, however, they rarely question the principle, since the idea of inter-territorial solidarity still carries resonance. Instead they argue that transfers should be more transparent and explicit, although this may be disingenuous, since explicit transfers may be more difficult to sustain. They also call for limitation on territorial redistribution, such that prosperous regions do not end up with less money per capita to spend, as has happened in Germany and Spain. Politicians in richer regions also argue that excessive transfers are not in the interests of the poor regions themselves, as it stifles initiative and traps them into dependency.

Decentralization, moreover, may actually lead to a reduction in inequalities, as they become more visible and politically salient. Since devolution, the Barnett formula has, for the first time, begun to have the convergent effects which it always logically entailed but which had been postponed by various forms of by-pass. Decentralization has, if anything, produced more even spending in France (Dupuy and Le Galès, 2006) and Spain (Gallego and Subirats, 2011).

There are strong institutional obstacles to dismantling both systems of state-wide social protection and mechanisms of fiscal equalization. As shown in Chapter 9, the lack of agreement on what territorial equity would look like has frustrated reform effort, while constitutional mechanisms have given losers veto powers, Switzerland being an exception. Even where formal consent from territorial governments is not required, it has proved difficult to shift the balance of transfers. When formulas are changed, there are usually provisions inserted to dampen the effect and compensate losers.

Territorial solidarity is also sustained by popular opinion, even where it supports autonomy (Chapter 4). Surveys in decentralized systems regularly show support, even in richer regions, for the principle of sharing wealth territorially (Henderson et al., 2013). A German survey (Bertelsmann, 2008) showed 85 per cent of the population in favour of uniform tax rates across the country and majority support in favour of equal living conditions. Seventy-four per cent wanted to keep fiscal equalization, although this figure was higher in the east and lower in Bavaria (65 per cent) and Baden-Württemberg (58 per cent). Inter-regional competition was accepted by about half of citizens but this issue did not really separate the rich and poor Länder. Saxony-Anhalt had both the highest percentage favouring competition and the highest supporting solidarity. In Belgium, where economic advantage has been one factor (p.163) in driving the move to autonomy, there was a plurality of Flemish voters in favour of decentralizing social security in 2003 but not a majority (37 per cent in favour and 31 per cent opposed). Opinion on whether individuals should be treated the same wherever they live is more complicated to interpret. The devolution or federalism paradox, already noted, arises from the fact that substantial numbers of people support the idea of more autonomy for their own region but also think that services should be the same in all regions of the state.

There is scant evidence, then, that regional social solidarity has displaced state-wide solidarity. In wealthy regions and those with strong identities (and especially where these are combined) there is less support for territorial redistribution but it is rarely opposed outright. Regions of strong identity show some tendency towards internal territorial solidarity and the idea that social policies and entitlements might differ from one part of the state to another, and welfare has been used for the purposes of polity building. It appears that solidarity has been rescaling in complex ways, with new territorial solidarities emerging alongside continued state-wide attachments. If there is little evidence of a race to the bottom, there are some signs of a race to the top in welfare provision, although this is constrained by the institutional limitations on the scope for redistributive policies at the regional level.

Differences in the organization of public services, access, universalism or selectivism, and charging policies have also been noted (Chapter 7). While advocates of NPM would argue that it is all about efficiency, there are distributive implications both for service recipients and for employees. It can be argued, however, that competition and selectivity help those who are better equipped to work the system, namely the middle classes, and that dividing the system may undermine solidarity by converting citizens into consumers (Crouch, 2003; Needham, 2003). Privatization and contracting out is also often associated with a deterioration of working conditions for those in the public services. The individualization of public services could, moreover, serve to undermine the very territorial social solidarity that underpins the use of welfare as a means of region building.

It is difficult to make an overall judgement on whether providing universal services free at the point of use is more progressive or redistributive than charging or means-testing, as that depends on who is using them most and who is paying. It is argued that university fees are progressive because graduates have higher earnings and should not be subsidized by taxpayers who did not have the chance of higher education. On the other hand, it is argued that the older generation, who enjoyed free education, rising living standards, and a buoyant property market, are using fees to offload burdens onto the younger generation. It is safer to argue that they represent a different choice in distributive policies and the use of resources. Underlying these are conceptions of solidarity and desert rather than the principle of solidarity itself. These (p.164) result from local cultures and preferences and political competition, as well as the political advantage enjoyed by particular groups within various arenas, and practices of collective and individual clientelism. Even in 1980s France there was evidence of distinct choices being made among favoured groups in social policy, with the aged usually well cared for compared with other groups (Tymen and Nogues, 1988). It is clear, then, that the detailed meaning of social citizenship is being reshaped at new territorial levels, even as broad notions of state-wide solidarity survive. Solidarity is not dead but is rescaling, with new issues arising in multiple arenas and at multiple scales, from the European (where it is poorly developed) to the local.

Self-determination and autonomy

One of the most difficult questions in normative theory is that of self-determination of peoples. In its classic formulation, this has four elements. First, there is a group of people who are the holders of self-determination rights. Second, is a territory over which they claim control. Third, is the idea of sovereignty, or the ultimate right to legal authority, recognizing no superior claims, nor those of constituent parts of the territory. Fourth, is effective control, in the sense of the capacity to manage functional systems at the level of the territory concerned. These are, of course, the classical elements of statehood. The dilemma of self-determination as a doctrine has been that these elements have rarely aligned themselves in the past. Rescaling means that they are even less in alignment in the present.

Defining the people has never been easy. Jennings (1956) famously declared that the doctrine of self-determination of peoples made no sense because, before the people could self-determine, somebody would have to determine who the people are. Limiting the right to nations does not help, since there is no objective definition of a nation and the claim for self-determination is itself usually one of the defining features of nationality, as opposed to ethnicity or regionalism. At the other extreme, permissive theories under which any group of people can constitute themselves into a self-determining community (Beran, 1998) are too broad and risk trivializing the idea of self-determination or encouraging opportunistic behaviour. Remedial theories argue that groups have rights to self-determination if they have suffered oppression or discrimination. This looks too narrow (Moore, 2001) as it would exclude groups to which we would normally concede the right, such as the people of Scotland, and could be an invitation to stage confrontations and provoke repression. It would also, inconveniently, de-legitimate existing states that have not suffered alien oppression; it is a sign of the statist trap in social science that most (p.165) scholars have ignored this point, taking the legitimacy of existing states for granted.

Matters are further complicated by the constructivist turn, since most social scientists now see groups as socially constructed, malleable, and contested. Even if we are able to define the people, they do not always correspond to a fixed territory. Procrustean attempts to overcome the problem by aligning territory with groups have proved ethically difficult and practically fraught. Partitions (adjusting the territory to the people) have been ineffective because of the inevitable existence of minorities within the new divisions, however drawn. It is not just that existing minorities will be trapped in new units but that, once group identity becomes the basis for political claims, there is an incentive for new minorities to form. The alternative, moving the people to adjust them to the territory, has proved equally ineffective, since it requires policy makers to make often arbitrary decisions on who belongs to which group and has been morally discredited as ‘ethnic cleansing’. Moreover, territory as well as identity is often part of the very claim that is being made so that nationalists will not happily give up parts of their putative homeland, even where their own group is a minority.

The doctrine of sovereignty is similarly difficult to apply, since there are often competing and overlapping sovereignty claims to the same territory. It is for this reason that self-determination has often been seen as a form of zero-sum politics, not amenable to the compromises of ‘normal’ political exchange. The functional issue, the ability actually to manage public policy matters, is also problematic, since the self-determining unit, even where it can be agreed, may be too small or lacking in resources and policy capacity. Rescaling may have altered matters here, in that, as we have noted, smaller units may be more viable, but this does not eliminate the problem entirely.

One answer to this is to delink self-government from territory altogether by providing for non-territorial government. The old Austro-Hungarian socialist idea of National Cultural Autonomy (NCA) provides for territorial self-government where it is possible and, where it is not, the rights of communities to organize matters on a non-territorial basis (Nimni, 2005). In this case, however, the boundaries of self-governing units are even more difficult to draw as they now encompass groups, which must be defined according to some principle, whether ascriptive or voluntary. The risk of exclusion is even greater than in the case of territorial autonomy since, in the latter case, at least everyone within the geographical limits can be considered as a member of the community. Public choice advocates, starting from different premises, based on individual choice rather than community, come to similar conclusions, and recommend unbundling competences and attributing them to different territorial and non-territorial jurisdictions (Frey and Eichenberger, 1999). In this case, the problem of defining community is by-passed as it is assumed that citizens can voluntarily come together to provide public goods.

(p.166) It is not, however, easy to de-territorialize public policy. Even purely cultural matters such as language require territorial proximity and territorial institutions, as noted in Chapter 3. Education is largely delivered in territorially-based institutions. Moreover, in a modern society it is difficult to separate out public policy into discrete spheres, which can be dealt with by different agencies. Economic policy interacts with social policy and environmental policy, while cultural issues are a component in all of them. Claims for self-determination are not just based on the desire to control specific policy items, but are part of a drive to redraw the boundaries of political communities or design a distinct project of society. In many parts of western Europe there has been a merging of demands based on recognition of national distinctiveness with the themes of the new regionalist political economy (Keating, 2004).

The constructivist approach to defining groups and the rescaling approach to territory together risk complicating matters still further. On the other hand, the very malleability of the concepts of peoples and territory can open up new approaches to managing the issue of self-determination, although not providing a definitive solution. Rather, the very fluidity of the key concepts allows new ways of reconciling them. Since self-government demands conflict and overlap, it is more appropriate to see them as claims rather than absolute rights (which would imply that they trump other claims). Such claims are both ontological, to the effect that the group or people exists as a unit and has some collective consciousness, and normative, that they have a right to self-government. Peoples are formed in specific circumstances, giving rise to more or less plausible ontological claims. Territories are not fixed and rigidly bounded. Self-government itself can be unpacked and delinked from the claim to sovereign statehood, which is itself a problematic concept in the modern world. The functional components of self-government can also be unbundled and recombined in different ways, while recognizing that there is no perfect or ‘correct’ level for assigning any given competence. This provides a set of criteria for the management of self-determination claims without reducing them to a single, ultimate, argument. The application of the criteria depends on the context and is ultimately a matter of political judgement and accommodation.

Rethinking peoplehood

Peoples are neither primordially given, nor are they constructed on a whim or as a result of political manipulation. Rather, they are formed over time as a result of social and political mobilization. It may not be possible to resolve the ontological issue through a scientific definition of a nation but scholars like Buchanan (2003) too easily deconstruct the idea of nation or (to use more (p.167) neutral vocabulary) autonomy-seeking political community. It is true that ethnogenesis is partly the product of manipulation by power-seeking elites, but this is not all there is to the matter. Groups may be constructed, but do take on a life of their own as sociological realities. Mobilizing to achieve self-government is not costless and if mass opinion has adopted a particular form of identity and aspiration to self-government, this is to be taken seriously. In this way, we can dismiss the national aspirations being articulated in the name of Padania (the imaginary nation of Italy’s Lega Nord) not because it is ‘artificial’ but because it has not established itself within its own territorial constituency.6 This leaves open the possibility of new communities of self-government being built in the future, as well as the revival of others.

Some authors have sought to resolve competing claims by resort to history (Kymlicka, 1995) but this is slippery ground. We know that history is a battlefield, prone to revisionisms and counter-revisionisms and that, in arguing about what happened in the past, we are often merely rehearsing arguments about the present. Historic rights have been dismissed as unworkable and irrelevant, based upon fanciful ethno-histories and carrying no obvious normative implications (Blanco Valdés, 1995, 2005). So we cannot resolve differences among groups on the basis of such criteria or who arrived first in the territory or under what circumstances, as Kymlicka (1995) has proposed. Israelis and Palestinians have very different stories, as do Catholics and Protestants in Northern Ireland. It is also true that the specific rights contained in historic charters often relate to matters of no current relevance.

Historic rights might be rescued, however, in two ways. Herrero de Miñon (1998) has characterized them as rights to be, rather than rights to have. That is, they address the ontological question and help delineate the entities (specifically territories) that might have certain rights, although the precise content of those rights can evolve over time. Secondly, historic rights do not have to refer to specific agreements at a definite point in time, as in the largely futile debates about original intent or efforts to read the minds of various founding fathers. Rather they refer to a more complex history of constitutionalism and constitutional understanding over time. So the debate about the ‘two nations’ theory of Canada is not about what the men gathered in Charlottetown in 1867 thought, but about how Canada has functioned and been understood since. There are similar debates about constitutionalism and constitutional practice in Spain and the United Kingdom. Talk about historic rights has been dismissed as inherently conservative, since it seeks literally to conserve the past, while the interests so defended are also conservative and privileged. Yet such rights have in recent years been harnessed by progressive groups to contest the power of states and multinational corporations, while traditional (p.168) moral economies have questioned the premises of market liberalism. Finally, it is often ignored that the main principle of legitimacy underpinning nation-states themselves is their historic existence; efforts to endow themselves with a founding or constitutional moment based on popular consent are usually as fictional as those of their opponents.

In some cases, historic rights are specified in written constitutional documents, of ancient or recent vintage. In other cases, they are enshrined in conventions, so that most observers now consider that the existence of the Scottish Parliament is a binding constitutional principle, notwithstanding its recent creation and the clause in the relevant legislation asserting continued Westminster sovereignty. Historic rights, referring mainly to the Basque fueros are included in the Spanish constitution, albeit only in an appendix. This allows the Basques to consider them fundamental and pre-constitutional law, while Spanish jurists insist that they are subordinate to the constitution. In both readings, they are effectively entrenched. In other cases, rights are subject to constitutional interpretation and re-interpretation (Tierney, 2004), which does not provide for a definitive resolution but moves the debate onto constitutional politics, in which it is not always necessary to share a belief in the foundations of the constitutional order in order to reach practical compromises.

It is important to distinguish this conception of historic rights and practices from claims that a specific ethnic group has exclusive rights to a territory, in which it may or may not presently live. Constitutional rights are vested rather in institutions, with the question of who belongs to the political community as a separate issue. The difference is illustrated in the history of Basque nationalism, which started out in defence of historic territorial rights, transmuted into an ethnic nationalism based on essentialist and primordial conceptions of Basqueness, then returned by the late twentieth century to a focus on historic territorial rights, which in principle can include everyone regardless of ancestry.7 It can also cope with inward migration of people into the territory. A similar story might be told about Quebec.

Territorial claims

The territorial element of self-determination claims itself has a number of components, including sovereignty, ownership, settlement, and access (Buchanan and Moore, 2003). Places may also have a symbolic importance. In some cases it may be possible to align these meanings of territory, as in the (p.169) ideal-type nation-state. Yet, while existing nation-states are often taken for granted, or have used their infrastructural power to forge a unity of the various components, claims for new states face a steeper challenge. The external boundaries are contested and there may be internal territorialized minorities. In some cases, there is a core territory but dispute about how far it extends. So Basque self-government demands have been realized within the autonomous community under the Spanish constitution, but there is a broader concept of Euskal Herria, which includes Navarre and three provinces in France. The autonomous community of Catalonia comprises a historically institutionalized unit but the concept of the Països Catalans, based on linguistic and cultural criteria, is much wider. Brussels is territorially surrounded by Flanders and claimed by Flemish nationalists, but is predominantly French-speaking. The French region of Brittany is smaller than the various historical Brittanys but where it should end is disputed. As one moves further east into Europe, the drawing of lines becomes increasingly fraught.

Then there are communities that share a territory but are politically divided. Sometimes these include national minorities, that is, groups which have a political affinity with the majority in a neighbouring state. In Northern Ireland, the community is divided on allegiance to two neighbouring states. Euskal Herria reaches across two states. Central and eastern Europe contains many apparently stranded national minorities. Such situations are intractable within the traditional view of territory as closed space belonging to one integrated national-institutional community or another. Shifting the boundaries only creates more minorities. More fundamentally, it does not address the complexities of identities in these cases, where people rarely identify simply and uniquely with their kin-state. Sometimes it is not even clear what the kin state is. For example, it is often thought that the kin state for the German speakers of South Tyrol is Austria, which has indeed assumed this role over the post-war decades. Yet the region has never been part of an Austrian nation-state, which itself has only existed since the 1950s. Before the Second World War, their reference point would more likely be Germany, as it was for Austria itself. Nowadays there is a Tyrolean identity, linked in various ways to Austria but not sharing the Austrian nation-building experience. The national minorities of central and eastern Europe have varying degrees of allegiance to their putative kin states and their mobilization is often the consequence of political entrepreneurship (Brubaker, 1996).

One response might be non-territorial forms of accommodation but this, as already noted, is not enough. Demands do not merely concern the right to practice a culture or speak a language but extend to the boundaries of political community and broader political identities. Yet a more open conception of territory can allow us to unpack its various dimensions and allow different territorial expressions to each. An elaborate version of this is the Northern Ireland settlement, which allows people to choose their own imagined (p.170) community, be it Northern Ireland, the island of Ireland, or the United Kingdom. In the Basque Country, territorial autonomy has been achieved in three provinces but there is a provision allowing the autonomous community to merge with Navarre, in contrast to the general prohibition on autonomous communities merging or federating. There are complex ways in which the imagined community can be extended into France, taking into account that Basqueness itself is felt very differently in the various component territories (Keating and Bray, 2013). The point about these various arrangements is that they are not territorial in the traditional sense, establishing bounded spaces, but neither are they non-territorial. Rather they employ territory in complex ways to allow institution-building, functional capacity and identity all to be accommodated. This is not to say that they always work well. Belgium, with its complex constitutional arrangements, is riddled with veto points and the electoral and party systems tend to entrench community divisions. It is not clear how the Northern Ireland settlement will develop including, as it does, two opposed visions about the telos. The Dayton arrangement for Bosnia and Herzegovina incorporates consociational elements which give vetoes to rent-seeking elites. Yet, properly designed, they allow for evolving relationships among territory, function, and identity, rather than fixing them for all time.

Sharing of territory is also possible through cross-border cooperation and extending certain citizenship rights across borders. The success of such arrangements does not depend on eliminating borders or de-territorializing politics and institutions. On the contrary, it requires the mutual acceptance of the state border as a condition for transcending it through free movement, cultural cooperation, and economic exchange. Where extending citizenship rights across the border is seen as a nationalist move, it is seen as a threat to the neighbouring state. This was the fate of Hungarian status laws, seen as part of a move by Hungarian nationalists to ‘reverse Trianon’.8 In the Northern Ireland settlement, on the other hand, people enjoy reciprocal rights and cooperation with the understanding that the border can be changed only with the consent of the majority according to a specified procedure.

The claims of indigenous peoples may appear to be a case apart since they are based on ancestry. Yet while these claims are linked to ethnicity they do not generally invoke the monopolistic conception of territory that underpins western theories both of sovereignty and of property ownership but rest on conceptions of community and its relationship with space. This has permitted other ways of unbundling territory, with mixtures of territorial autonomy, extra-territorial forms of community membership, and bargaining about natural resources. So the normative debate can be moved on from a single principle, such as who got there first, towards a balancing of different (p.171) considerations. The real problem perhaps lies in the way that indigeneity has come to be seen as a trump card, encouraging groups to make claims about being there first and thus for special consideration. If indigeneity is defined precisely, it refers to a specific context of peoples displaced or dominated by (normally western) imperial expansion and cannot be generalized.

Post-sovereignty and autonomy

As long as sovereignty is seen in its traditional sense, as an absolute claim to jurisdiction over a territory, then sovereignty claims are liable to be non-negotiable. New understandings of sovereignty, more complex and less monistic than the so-called Westphalian paradigm, however, provide more scope for compromise. Sovereignty, in this new sense, is divisible and there are multiple sources of it. These include states but also international regimes, which become sources of law in themselves, historic rights vested in territorial communities, and federal arrangements providing for a constitutionalized division of authority. Such ‘post-sovereignty’ understandings do not deny that sovereignty exists, or that there are multiple sovereignty claims, but do point away from the monist understandings, linked to the nation-state, which have prevailed since the nineteenth century (MacCormick, 1999; Keating, 2001a). Sovereignty in this sense ceases to be an absolute or something that an entity has and becomes embedded in relationships with other holders of authority.

Such understandings of sovereignty in legal theory can be married to new understandings of self-government in political theory under the rubric of autonomy (Lapidoth, 1997). Unlike the traditional ideas of sovereignty, rights, nationality, or territory, autonomy can never be seen as an absolute, something that one either does or does not have (Keating and Gagnon, 2012). On the contrary, it is inherently relational in the sense that one is autonomous in relation to some other person or entity. In the past, the concept has been used in relation to the state, with territories or groups having some degree of freedom within the parameters set from above. Nowadays, the concept is more multidimensional. Groups, institutions, or territories have multiple dependencies or inter-dependencies, with the state, supranational and transnational institutions, other entities at the same level, the economic environment, and interest groups.

This allows for an unbundling of territory and for its re-articulation in new and complex forms. We have already noted the emergence of new institutional forms within European regions and at the European level and the progressive politicization of these institutions. On the other hand, the new spaces are loosely bounded, take on different political, economic and social meanings at (p.172) the same time, and their external borders differ for different functions and purposes. Given all this, territorial autonomy takes on a new meaning. It does not create spatial monopolies of authority identity or capacity but involves bundles of competences, which may have different territorial reaches. Imagined territories can also be unbundled and recognition of the symbolic or historic importance of particular spaces does not necessarily entail sovereignty, possession, or settlement. The spatial imaginations of people and groups may vary and different functional and political systems may conform to different spatial scales and boundaries, or escape boundaries altogether, as with global production chains. The flexible and constructivist meanings of both group and territory then allow new ways for their mutual adjustment.


Earlier chapters have stressed that the size and shape of governing units cannot be derived, either causally or normatively, from purely functional considerations. Yet the functional viability of territorial units is a relevant criterion in discussions of self-government. New regionalist theories have pointed to the constitution of governing capacity at new spatial levels, and many nationalist movements have adopted new regionalist arguments to underpin autonomist claims. This is not to say that national self-determination claims can be reduced to functional regionalism. They are essentially normative, seeking to reconstitute the bounds of the political community and involve claims for recognition as well as power. Functional regionalism, indeed, has been challenged by political and social movements seeking to politicize regional space and to introduce themes of community and belonging. Yet, bringing these two concepts together, there is scope for polity building without going for the constitution of a new nation-state but recognizing the constraints of international interdependence. Autonomy becomes meaningful only when one has power to do something, drawing on the classical distinction between ‘power over’ and ‘power to’. This allows us to bring the question of governing capacity, and ideas from the study of public policy, back into the debate about self-government.

Territorial autonomy in this sense is about the creation of new space for culture, economics, solidarity, and politics but never in a complete or perfect way. These are loosely-bounded spaces, which overlap with and interpenetrate other spaces. Nation building and region building in the modern era consist in constructing such spaces and in securing a degree of complementarity among the various cultural, political, social, and economic dimensions of the project. Models of regional government similarly vary from the functionally-specific, as in France, where regions are a component in national planning and (p.173) resource allocation, to the general purpose, as in Scotland or the Belgian regions where (at least in principle) the powers reserved to the state are enumerated, with the residual powers going to the devolved level. At the limit, the difference between strong devolved regions and weak independent states disappears, in spite of the formal sovereignty claims of the latter. This provides room for negotiation and compromise between independence-seeking and unionist forces, as we see in many places.

The right to autonomy

Statehood in its traditional sense is an inappropriate response to most problems of group accommodation, since one cannot fit hard state boundaries to fluid groups, and the state itself is changing. In any case, few groups have actually demanded a state for themselves and often only as a last resort, after other proposals have been frustrated. Nationalists in Scotland, Wales, Catalonia, the Basque Country, Flanders, and elsewhere have regularly advanced claims for self-government and recognition. When these are not fully conceded, they turn to independence as an alternative but, when it comes to defining what independence means, revert to what look like federal or at least confederal proposals (Keating, 2012b). Full sovereignty seems ever more elusive.

Given this elusiveness of the final aim, the doctrine of self-determination might better be seen as a theory for defining the subject rather than the object, and the right for that subject to negotiate its position within the wider state and international order. Secession is only an extreme case and not a first resort (Bauböck, 2005). It is paradoxical and unfortunate that theoretical reasoning on the claim to autonomy has been poorly developed compared with the literature on the right to secession, when the former is usually to be preferred as less destabilizing. It is often argued that prudent states should concede autonomy as a means of conflict management and diffusion, but rarely that groups have a right to it. One reason is no doubt that it is difficult to enshrine rights of internal self-determination in international law but that is not an insuperable obstacle, given the increasing penetration of international human rights law into domestic systems. There has also been resistance in the political arena. A politician like Margaret Thatcher (1993) was prepared to concede that Scotland had a right to secede from the United Kingdom but no right to autonomy within it.9 More recently unionists in the United Kingdom have (p.174) insisted that the only question that can be put to the Scottish people in a referendum is that of separation or remaining in the United Kingdom, terms laid down by the centre. After the NATO intervention in Kosovo, the international community (or most of it) was prepared to recognize Kosovo’s independence but there was no effort to enforce the more obvious remedy to Milosevic’s take-over, which was to restore the old autonomy regime.

So there is no one foundational principle that could legitimate a right of territorial autonomy or provide a basis for its design. Rather, in individual cases, different principles must be brought to bear and balanced. This requires a contextual analysis rather than a preformed grid to be imposed on cases (Coppieters and Sakwa, 2003). This form of constitutional practice flourishes better in liberal societies where constitutionalism is recognized and differences resolved through political means. In other societies, autonomy and power-sharing have had to be imposed from outside, or at least with external help, often in the course of peace processes. Although the international community has been reluctant to adopt a doctrine of autonomy as a right it has often introduced it in such circumstances.

All this might appear to critics as indeterminate, giving scope for picking criteria to suit, with no means of resolving differences. In another paper (Keating, 2013), I have sought to work through cases applying the various criteria and coming out with more simple or complex formulas. Scotland lies at one end of a spectrum, with a historic understanding, a present demand for autonomy, an inclusive principle of membership, clear boundaries, and viable institutions. It may therefore be seen as a self-determining entity able to negotiate its own autonomy within the wider system of interdependencies in which it is encompassed. At the other extreme is Republika Srpska, with no historic right and a claim for autonomy based purely on ethnicity, itself the product of ethnic cleansing. The borders represent little more than a cease-fire line. The constitutional document recognizes autonomy, which should therefore be respected, against demands from the Bosniak side to dismantle it, but does not constitute the Republika Srpska as a self-determining entity. Northern Ireland and the Basque Country are more complex, with a mix of principles at work, suggesting more complex solutions, based on territorial autonomy with an open conception of territory and an external projection.

Taking an open and more constructivist perspective on territory may also help meet a common objection against both territorial and non-territorial autonomy, that it freezes group identities and reifies what may be the result of (p.175) conjectural power politics or manipulation. The nation-state provided one answer to the conundrum of reconciling identities, institutions of government, and systems of representation but it itself was an effort to freeze a more fluid reality. My analysis has sought to show how we can reframe the issue by seeking autonomy as a relational concept, so that all polities are self-governing but interdependent; and by seeing autonomous territories as more or less loosely-bounded.

This perspective does not provide a solution to the nationalities question. Arguments will continue over which arena should be the primary focus for political identity and defining the political community. There will be pressures to privilege one level or another as the main locus for interest intermediation, social solidarity, and regulation, as we have seen. Self-government arguments will not go away and in many ways have increased as a consequence of rescaling, which has called into question the supremacy of the nation-state. What this approach does is to provide ways of bringing the arguments within the scope of ‘normal’ politics rather than seeing them as a zero-sum game or a clash of absolute and non-negotiable claims.

Territorial politics

So neither the question of social solidarity nor that of autonomy has a definitive answer, to be derived from first principles. The traditional nation-state provided a rather deceptive short cut, by supplying a ready-made unit whose normative justification was given rather than demonstrated. So solidarity could be presented as a universal and uniform idea, as long as it was confined within state boundaries. Self-determination movements could be hedged around with numerous restrictions, as long as the nation-state itself did not have to justify its own existence. Functional and political rescaling, together with the loss of normative hegemony on the part of the nation-state, have put these assumptions in question. Solidarity and self-determination then become matters of political contestation, in which multiple, and not always consistent, normative considerations can be brought to bear. Normative analysis therefore needs to appraise individual cases against these criteria and reach a judgement rather than pressing all cases into a single procrustean framework.


(1) I did not use the term ‘sub-national’ in the article in question.

(2) Segall (2007) seems to equate policy differentiation by territorial governments within compound states to providing different levels of public services to Arabs and Jews in Israel, a completely different matter. A better analogy for the latter would be the treatment by the Northern Ireland Stormont regime (1922–72) of Catholics, a disempowered non-territorial minority.

(3) In the prospectus for their 2009 project on devolution and policy making the IPPR team stated:

However, greater policy divergence across the constituent parts of the UK, particularly in public services, also raises difficult and as yet unresolved questions about the nature and content of UK-wide citizenship, about the level of policy differences within the UK that are acceptable and unacceptable and whether divergence matters in terms of ensuring a degree of national [sic] solidarity.

As all the divergences in Scotland and Wales had involved higher or more universal levels of service, the state-wide parameters they suggested could only lead to a leveling down of provision. My own paper, pointing this out and questioning the link between state-wide uniformity and welfare was quietly dropped from the collection.

(4) So the idea may be valid when applied to the health service in England but not to differences between England and Scotland.

(5) They mention less often that, in the medium and long term, Flanders, given its demography, will gain from the centralized public pension system.

(6) Nor has it been a complete failure, in which case we would not have to mention it at all.

(7) This is not to deny the sociological reality of an ethnic element in Basque national identity, but to point to the changing political demands.

(8) The peace treaty, at the end of the First World War, that deprived Hungary of a large part of its territory.

(9) Her extraordinary statement is worth quoting in full:

As a nation, they (the Scots) have an undoubted right to national self-determination; thus far they have exercised that right by joining and remaining in the Union. Should they determine on independence no English party or politician would stand in their way, however much we might regret their departure. What the Scots (nor indeed the English) cannot do, however, is to insist on their own terms for remaining in the Union, regardless of the views of others…it cannot claim devolution as a right of nationhood within the Union. (Thatcher, 1993: 624)

Of course, Thatcher, in the name of the English majority, was herself dictating the terms of the Union here.