What Is Citizenship?

What exactly is citizenship, and what does it entail? On the most basic level, citizenship bestows upon individuals membership in a national political community.⁷

In liberal democracies, it gives them the right to vote, to run for office, and to participate freely in public activities, while also requiring the obligation of paying taxes and possibly serving in the military. In terms of the larger international community, citizenship serves as what Rogers Brubaker calls “a powerful instrument of social closure,”⁸ in two respects. First, the boundary of citizenship allows rich states to draw a line that separates its citizens from potential immigrants from poor countries. Second, it allows states to create internal boundaries that separate citizens from foreign residents, by associating certain rights and privileges with national citizenship.⁹

Citizenship therefore evokes a fundamental paradox within liberal democracies, namely, what Seyla Benhabib calls “the paradox of democratic legitimacy.”¹⁰

Liberal democracies are based on the universal language of fundamental human rights, along with the free association and participation of “the people,” yet they also delineate clear and enforceable borders and boundaries. This refers not only to territorial borders, but also to the boundaries of political membership. Determining who is included in the concept of “the people” also implies at least an implicit understanding of who is excluded.¹¹ In essence, the paradox is that liberal democracies are “internally inclusive” while remaining “externally exclusive.”¹²

In other words, my understanding and use of citizenship here is as a legal category, rather than as a feature of civil society, social capital, or state-society relations more generally—although, of course, this category has tremendous social ramifications. I focus on the rights that formal citizenship entails, and especially the formal legal requirements for having access to those rights, rather than on the beliefs or practices of citizens. Although this may deviate somewhat from some contemporary discussions that use the term citizenship as a synonym for civic engagement and related concepts, it provides a more focused and grounded definition, while staying true to the theoretical and empirical debates that started with Aristotle.

In a major contribution—one that is both theoretical and empirical—to the study of citizenship, T.H. Marshall developed a model of citizenship based on the experience of industrialization and the emergence of democracy, and his work raises questions that are still relevant for contemporary debates about the future of citizenship.¹³

Marshall argued that the extension of rights and benefits go in a specific historical sequence as democracy develops and expands, starting with basic civil rights (freedom of conscience, protection of property, and some associational liberties), leading eventually to political rights (to vote, hold office, speak and associate freely), and finally culminating in social rights (to form labor unions, and eventually to receive the many social benefits that welfare states provide). The argument is compelling, and it fits the historical experience of the United Kingdom—and to some extent Western Europe in general—quite well.

In recent decades, however, the development and establishment of the welfare state has created a new logic that is quite different from Marshall's historical account. In most liberal democracies today, wide-ranging civil as well as social rights are extended to almost all workers and legal residents, even if they are not citizens, and therefore do not have political rights. In other words, political rights are no longer a prerequisite for social rights. Moreover, in an increasing number of places in both Western Europe and North America, non-citizens are being granted local or regional (but not national) voting rights.¹⁴

At the same time as this sub-national political participation has been expanding, citizens of countries that are members of the supra-national EU can now choose to vote in European elections in their EU country of residence, rather than their country of origin.

While Marshall's historical progression may no longer apply to the contemporary situation, many scholars agree with his normative conclusions concerning social rights as the ultimate priority and objective. But this has, in turn, led some scholars to proclaim the current or impending empirical irrelevance of citizenship in the nation-state. According to this argument, since social rights can now be achieved without political rights, and since an increasing number of political rights are now available on the sub-national and supra-national level, national citizenship no longer matters. As one proponent of this view argues, “when it comes to social services (education, health insurance, welfare, unemployment benefits) citizenship status is of minor importance in the United States and in Western Europe.”¹⁵

In short, this type of argument places great emphasis on the recent emergence of transnational and “postnational” norms based on individual human rights, which undermine the previously dominant system of nation-states.¹⁶

Why National Citizenship Matters

For a number of reasons, however, it is far too early to dismiss the relevance of the nation-state and national citizenship. And this is even the case in the EU—where the broader umbrella of “European citizenship” entitles citizens of any EU member state to have a vast set of rights and privileges across the territory of the union—since EU citizenship is itself strictly derivative of national citizenship. As a result, “third-country nationals” (people who are not citizens of an EU country) still face limitations on their rights and opportunities, and the citizen v. non-citizen distinction therefore remains very important to them and to the society in which they live.

First, the right to vote and run for office in national elections is still extremely consequential in all countries. Non-citizens, even if they are permanent residents and long-time workers, have no opportunity to participate in the democratic process on the national level. And since citizenship, immigration, and asylum policies are generally implemented on the national level, this means that non-citizens are excluded from taking part in decisions that may directly affect their own lives.

Second, despite exaggerated claims that social rights are guaranteed to all, regardless of national citizenship, in many countries non-citizens are still excluded from significant social benefits. For example, five of the nine provinces in Austria do not provide their social assistance programs to people who are not citizens of Austria or another EU country. Many other countries place significant restrictions on the rights granted to new immigrants, particularly third-country nationals, who often receive lower benefits, and in many cases are barred from noncontributory social programs for a certain number of years after arrival.¹⁷

In short, while the modern welfare state undoubtedly provides greater benefits to immigrants than did nation-states at earlier points in history, non-citizens still receive significantly fewer social rights than do people with national citizenship. And for that reason, the citizenship distinction really matters in people's lives.

Third, although citizenship is generally less relevant for most private-sector employment,¹⁸

it is still very important in the allocation of public sector jobs. For example, France only accepts French or EU citizens in railway, postal, and hospital jobs; in Germany, government service employment positions in such areas as public transportation and education are restricted to German or EU citizens; and to cite a non-European case, the U.S. government can restrict such postings as public school teachers, state troopers, and probation officers to American citizens.¹⁹ Within the EU itself, it should be added, while citizens of an EU country automatically have the right to live and work in another EU country, third-country nationals can generally only reside and work in the country into which they have immigrated. In other words, non-citizens face de facto restrictions on their labor mobility that EU citizens do not.

A fourth reason why national citizenship is still quite significant has to do with the eventual integration of immigrants into the host society. Many scholars have argued that immigrants who become naturalized citizens are likely to become much more integrated in their new country than those who remain non-citizen residents, or “denizens.”²⁰

Although more empirical research on these questions is still needed, it is logical to assume that naturalized citizens will tend to have better command of the national language, to experience more loyalty to the new country, to be relatively accepted by their fellow citizens, and of course to enjoy some protections that are only available to citizens, such as the right not to be deported. In other words, while there is considerable variability across groups and countries, citizenship acquisition can serve as a rough measure of integration, and the different possibilities to acquire citizenship will have lasting implications for the long-term integration of immigrants.

A final, and much more practical and policy-oriented, dilemma facing the advanced industrialized world—and EU countries in particular—has to do with demographics. European countries have among the lowest birth rates in the world, and they desperately need more workers in order to prevent their pension systems from collapsing over the coming two decades.²¹

One obvious (though partial) solution to this problem, which has been recognized by scholars and political elites for years, involves increasing levels of immigration and naturalization. Yet the resistance and outright hostility to immigrants has increased noticeably over the past decade, whether measured by public opinion surveys, support for extreme-right parties and candidates, or criminal attacks against foreigners. These two countervailing pressures—the need to incorporate more immigrant workers within a context of an often xenophobic public opposition—will have to be resolved, in one form or another, over time. And political elites will struggle with these contradictory demands, though politicians tend to be more responsive to the short-term nature of the electoral process. The demographic problem, however, is a longer-term one, and a key part of its eventual resolution will depend on how these countries define and enforce their citizenship policies.

In short, whether in terms of politics and elections, welfare state benefits, public-sector employment, social integration, or demographics and pension systems, national citizenship remains an essential and enduring feature of modern life—even in the “supra-national” European Union.

Characterizing and Explaining Historical Variation

Having established the importance of studying citizenship, we can now turn to some important empirical and theoretical questions related to historical variation and relative change over time in the citizenship policies of the established 15 EU members. I focus in particular on the “EU-15” for two main reasons. First, most obviously, the EU-15 have become de facto countries of immigration, with tremendous pressures from would-be immigrants from around the world who would like to benefit from Europe's high standard of living, protection of social rights, and the need for more labor. In other words, the stakes surrounding citizenship in the EU-15 are extremely high, much more so than would be the case in less industrialized countries with few pressures for immigration and citizenship. Second, from a methodological perspective, the EU-15 constitute a relatively coherent entity, consisting of countries that face similar pressures of immigration and globalization within the common framework of the EU and its institutional and juridical “harmonization.” The EU therefore forms a useful “laboratory” for analysis, since it spans a range of variation as wide as that of the entire industrialized world: some countries have been among the most liberal—comparable to Australia, for example—and others have remained staunchly restrictive—along the lines of Japan. In other words, a careful examination of the EU-15 provides for analytically useful contrast and variation within a relatively similar set of cases with tremendous real-world importance, as highly desired destinations, thus allowing for more systematic comparisons than would be possible by looking at the entire world or the European continent.

Unlike many highly studied topics in comparative politics that contain ready-made empirical puzzles to be explained, we first need to establish an “empirical baseline” of citizenship policies across the countries of the EU. Only then can we address the related theoretical questions. This section explores the “varieties of citizenship” in historical perspective, while the following section focuses on the extent of change (and potential convergence) in citizenship policies that has taken place over the last two decades of EU integration. My goal is not necessarily to provide definitive evidence, but to present some suggestive findings that could spark debate on important questions that have so far escaped rigorous comparative analysis.

In order to make broad cross-national comparisons, and through them to draw general theoretical conclusions, it is necessary to move beyond the complex legal and technical specifics of each case, thereby to reach a better understanding of the variation in citizenship policies across countries on aggregate. Such an aggregation procedure will allow us to distinguish between countries that can be considered “liberal,” “medium,” or “restrictive” in their granting of citizenship.²²

I have therefore developed a coding scheme that classifies and scores the citizenship policies of the 15 EU countries based on three main components: 1) whether or not it grants jus soli, i.e., whether children of non-citizens who are born in a country's territory can acquire that country's citizenship; 2) the minimum length of its residency requirement for naturalization; and 3) whether or not naturalized immigrants are allowed to hold dual citizenship.²³ In my view, these are the most important general elements of a country's citizenship policy, even though there are, of course, many other features, conditions, and exceptions in each country's laws.²⁴ They capture the two main modes of citizenship acquisition (by birth and by naturalization), as well as the primary deterrent that can potentially discourage immigrants to naturalize even if they are eligible (dual citizenship).

Based on a detailed analysis of the citizenship laws that were in place in the 1980s,²⁵

Figure 1 presents an aggregate Citizenship Policy Index (CPI), which consists of a scale from 0 to 6, with scores derived from a simple coding scheme registering scores of 2 points for each of the 3 components.²⁶ In most cases the laws in the 1980s were identical to, or closely in line with, the historical origins of each country's laws,²⁷ and in this sense they can serve as a general proxy for previous laws and traditions, thus allowing us to measure historical variation. The figure shows consideration variation in national citizenship policies in the 1980s, with six countries grouped into a restrictive category, five in the medium level, and four in the liberal grouping.

But what explains this historical variation? Even a cursory look at the grouping of countries shows that the variation has little or nothing to do with such standard structural factors as a country's territorial or population size, the configuration of its political institutions, or the performance of its economy. But are there some coherent and systematic factors that explain why some countries developed quite liberal traditions, while others remained very restrictive?

My answer to these questions involves two main (and overlapping) historical factors, which I argue have helped to explain the trajectories that developed in each country. The first is whether or not a country is a former colonial power—on a large scale, outside of Europe, and over a sustained time period. Those countries that were colonial powers generally developed relationships with the outside world and its people. And even though these relationships were usually hierarchical, exploitative, and violent—i.e., not at all based on liberal principles—they created long-term “affinities” between colonizing and colonized countries that still exist today.²⁸

These contacts often have long-lasting and wide-ranging implications, in such areas as security and peacekeeping, for example, but especially in the domain of immigration policy. In short, although racism and xenophobia may be as strong in former colonizers as in countries that were not colonial powers, an unintended—and, for immigrants, beneficial—consequence of colonialism is that people from formerly colonized countries receive special immigration opportunities and rights,²⁹ and they usually have closer cultural and linguistic ties than would immigrants from other parts of the world.³⁰

The second, related factor involves whether a country was democratic in the nineteenth century. Early democratizers were much more likely to develop a more inclusive conception of national identity—based on civic, rather than strictly ethnic, terms. And that identity was therefore more likely to allow foreigners to be included as members of the society. In contrast, countries that only democratized in the twentieth century were more likely to solidify a conception of the national group that excluded people with different appearances, languages, and traditions. This factor is obviously closely related to the first, since most of the early democratizers were also colonial powers that accepted immigrants from their former colonies into their societies. What matters most is whether a country had both a prior experience as a colonial power and was an early democratizer, not just one of the two. And it is in these countries that one would expect to find the most liberal citizenship policies.

How do these arguments map out empirically? Figure 2 presents a basic 2×2 matrix, dividing up the 15 countries into four quadrants, with the key explanatory divide situated between Quadrant I and the three others (indicated by the solid lines. The columns distinguish between countries based on whether or not they used to be colonial powers, and the rows differentiate countries that had become democracies by the end of the nineteenth century.³¹

As expected, the countries that are both former colonial powers and early democratizers (in Quadrant I) are among the most liberal in terms of their citizenship policies. Indeed, Belgium, France, and the U.K. have enduringly liberal citizenship traditions, correlating strongly with their historical experiences.

The only country in Quadrant II—for countries that first democratized by the nineteenth century but have no past history as a colonial power—is Greece. Quadrant III presents the countries that have neither a colonial nor an early democratic experience, and indeed they are almost all at the restrictive end of the citizenship continuum. Austria, Denmark, Finland, and Luxembourg had very low citizenship index scores, and Italy and Sweden were not far ahead. The only disconfirming case here is Ireland, which is actually among the most liberal countries, despite not being classified as having had a colonial past or having been an early democratizer. But this can be explained by the fact that Ireland only became independent from the U.K. in 1922, and its laws and practices have been closely linked to British traditions.³²

Finally, Quadrant IV consists of countries that were colonial powers but were not early democratizers. As expected, Germany and Spain's citizenship policies were extremely restrictive in the 1980s; and Portugal and the Netherlands were somewhat more liberal.

Overall, although it is not a perfect match, this analysis shows that the historical experience of individual countries—both in terms of their past experience as a colonial power and the onset of democratization—corresponds quite closely with their citizenship policies in the 1980s.

Characterizing and Explaining Recent Continuity and Change

These two historical factors seem to explain longstanding variation very well, but what about more recent changes? Over the past several decades, as the EU has been integrating and “harmonizing” in just about every area—from economic to judicial to social issues—has a similar development taken place in the realm of citizenship policy? How have countries changed their policies, and to what extent have they converged?

In order to answer these empirical questions, table 1 compares the CPI score in the 1980s to the current index, and then highlights any change that may have occurred. The results show that ten countries did not change at all; but the five countries that have changed all moved in the positive direction on the scale, towards increasing liberalization.³³

More specifically, the most common change was when countries began to accept dual citizenship for naturalized immigrants, as occurred in Finland (in 2003), the Netherlands (over the course of the 1990s),³⁴ and Sweden (in 2001); in all cases this was a departure from the previous policies. Germany (in 2000) and Luxembourg (in 2001) reduced their residency requirements (from 15 to 8 years and from 10 to 5 years, respectively), resulting in a liberalizing change in their scores. Of all the countries, Germany liberalized the most, as the new law not only reduced the residency requirement, but it also now allows for a form of jus soli, representing an important change from its notorious 1913 law.³⁵

Overall, this empirical analysis of change since the 1980s shows that some, but certainly not all, of the more restrictive countries have changed their citizenship policies significantly. And the cross-national differences are not quite as vast as they were a few decades ago, although they are still very wide.³⁶

In short, there does appear to be a relative convergence taking place (i.e., countries are closer to one another than they used to be, rather than more distant), but it is far from the level of absolute convergence (or “institutional harmonization”) occurring in many other areas and sectors of EU integration,³⁷ and there is clearly not a common EU policy today, or—especially after the major setbacks to the EU Constitution—on the horizon.³⁸ This lasting variation reinforces the point about the importance of earlier national historical legacies, which still seem to matter a great deal, even in the era of globalization and Europeanization.

But we still need to explain why change has or has not occurred across the countries of the EU. Why have some countries liberalized while others have resisted the pressures of liberalization and remained quite restrictive? Can one identify common patterns that apply across countries, despite the national idiosyncrasies that inevitably apply to policymaking? In each case, the decisive actors who determine citizenship policy are, of course, domestic political elites and political parties, but they do not act without constraints. What, then, are the factors that have influenced them, and how?

The existing literature on immigration and citizenship is much stronger on providing theoretical reasons for liberalization than for restrictiveness. Indeed, several arguments have been developed to explain increasing liberalization, with some emphasizing global causes and others stressing domestic factors. On the global level, scholars have stressed economic globalization,³⁹

neo-functional economic cooperation,⁴⁰ and new norms of “postnational” human rights,⁴¹ to explain liberalization. And on the domestic level, scholars who focus on immigration have considered interest group politics, whereby organized groups and businesses often exert influence quietly on policymakers to expand immigration,⁴² or domestic courts and the judicial system in general, which have often sided with immigrants, thereby putting added pressure on political elites to adjust the policies themselves.⁴³ Despite the quite different points of emphasis, each of these theoretical arguments has the expectation of increasing liberalization across the countries of the EU.

The findings shown on table 1 certainly provide support for these arguments, since the five countries that changed all moved in a liberal direction. But what about the countries that did not liberalize? What explains the resistance to liberalization, and how does it play itself out politically? These questions have been less explored in a literature that primarily focuses on (and predicts) liberalization.

Several plausible arguments could be presented, but most do not work out empirically. Structural factors such as rates of economic growth, unemployment, or immigration levels, do not shed any light on the puzzle of why, among the nine most restrictive countries with citizenship index scores in the 0–2 range in the 1980s, Finland, Germany, Luxembourg, and Sweden liberalized, but Austria, Denmark, Greece, Italy, and Spain did not. Nor do more cultural factors, such as the level of racism or xenophobia, hostility to immigrants, or general discontent with the EU—at least as measured by such public opinion surveys as the Eurobarometer or the European Social Survey—help to establish any connection to the liberalization of national citizenship policies.⁴⁴

And differences in political institutions, such as whether countries have parliamentary, presidential, or mixed electoral systems, different minimal percentage requirements to enter parliament, or various electoral rules and practices, explain very little as well.⁴⁵

What is missing from these structural, cultural, and institutional factors is the actual politics of citizenship. How have political actors navigated the potentially treacherous waters on this volatile issue? How have they dealt with the various pressures from interest groups, social movements, and public opinion? How have they made choices, and attempted to implement them politically into new laws and policies? While it is possible that lack of change simply represents a form of institutional inertia—where the old policies continue to persist simply because they already existed—it is also quite likely that elites have pursued conscious strategies and fought open battles, and that these contingent political factors were decisive.

An in-depth analysis of the politics of citizenship across the EU-15 exceeds the bounds of this article, but even a rudimentary examination of the political dynamics within countries will help to explain why some of the restrictive countries liberalized but others did not. And it allows us to draw some more general conclusions about the mobilization of public opinion on issues connected to immigration and citizenship, which generally results in the prevention of liberalization.

One way of exploring the liberalization differential between countries is to consider whether political parties of the left or right were in power. As Christian Joppke has argued, left-of-center governments are typically in favor of increasing the citizenship rights of immigrants (what Joppke calls “de-ethnicization”), whereas right-of-center governments want to resist such impulses, while simultaneously expanding the country's connections to its émigrés (what he calls “re-ethnicization).⁴⁶

This argument certainly applies to Germany, where the 2000 citizenship law clearly resulted from the installation of a new Social Democratic-Green coalition government in 1998, and Joppke also applies it effectively to France, Spain, and Italy. Finland and Sweden also had Social Democratic governments when their citizenship laws were changed, though they had been in power for significant periods prior to this liberalization. Luxembourg, however, had a right-of-center government at the time its new citizenship law was proposed and passed, so it cannot simply be a matter of whether the left is in power. Moreover, several of the countries that did not liberalize their laws also had left-of-center governments at one point or another, which clearly did not result in the expected change. In short, the left-right orientation of the government does seem to be related to the liberalization of citizenship laws, but having a leftist government only accounts for part of the liberalization.

My own interpretation builds on Joppke's left-right dichotomy, but I argue that the political orientation of the right is even more important than the constellation of forces on the left. And the issue is not simply whether a right-of-center government is in power, but whether it is mobilized on the issue of immigration and citizenship reform. Thus, if we accept that the liberalizing pressures mentioned above are influential in the EU-15, the question is what can counteract those forces. My answer is the mobilization of public opinion—which once again is latently hostile to immigrants throughout the EU—either in the form of a successful far right party, a popular movement, or a referendum of some kind on the issue of immigration or citizenship. In other words, the mobilization of anti-immigrant sentiment essentially “trumps” the liberalizing pressures that other scholars have identified.⁴⁷

It may seem counterintuitive to focus on mobilization in the context of European politics, since public involvement has certainly not been a significant feature of EU integration, which has been widely characterized as having a “democratic deficit.”⁴⁸

But as more and more elements of domestic policymaking have moved to the EU level, anti-EU sentiments and actors have been emboldened in their resistance and opposition to EU-level policymaking. And perhaps because they touch on raw nerves that are most closely connected to a country's identity and sovereignty, no issue has been more sensitive, explosive, or politically effective than immigration and citizenship.

How does this argument play out empirically in the EU countries? The mobilization of anti-immigrant sentiment is very difficult to measure, since it can take on different forms. The most obvious and common form is the emergence of a far right party whose main platform emphasizes immigration and national citizenship issues. Figure 3 therefore presents another 2×2 matrix, showing the nine most restrictive countries from the 1980s,⁴⁹

and distinguishing between the strength of far right parties and whether citizenship liberalization occurred. In order to measure the strength of the far right, I incorporate a measure of the average electoral support for far right parties between 1992 and 2003.

As expected, of the four restrictive countries that liberalized their citizenship laws, all are located in Quadrant II, leaving Quadrant I empty. In other words, liberalization only occurred without the presence of a significant far right party or movement. Among the five restrictive countries that did not liberalize their laws at all, the argument about the importance of the far right has some success as well. As expected, in Austria, Denmark, and Italy, which have had strong far right movements for well over a decade, the pressures for liberalization were effectively blocked by their influential far right parties. But Greece and Spain had virtually nonexistent far right movements, yet they also did not liberalize. What explains these exceptions? This question cannot be answered definitively, but I would speculate that the potential for change is still there, and that there is certainly a much greater chance for liberalization in Spain and Greece than there is in Austria, Denmark, or Italy, at least with the current constellation of political forces. In other words, the fact that they have not changed yet does not mean that they will never do so.

Moreover, if we were to consider the six relatively liberal countries in this analysis, Belgium, France, and the Netherlands—all three of which have fairly strong far right movements, with 1992–2003 average far right returns of 11.7, 13, and 6.4 percent, respectively—have seen immigration and citizenship emerge as highly polarizing political issues, and they have either experienced (in the case of France in the mid-1990s) or are currently experiencing (in the case of Belgium and the Netherlands) some tinkering with their citizenship laws as a result of the pressure of the far right. Ireland does not have an organized far right movement, but proponents of restrictions on citizenship acquisition succeeded in implementing a controversial referendum, which passed overwhelmingly (with 80 percent support) in June 2004, to limit the jus soli rights of the children of non-citizens, so that children born on Irish soil can only receive Irish citizenship if at least one of their parents has resided in Ireland or the U.K. for three of the previous four years.⁵⁰

This remarkable development shows the tremendous salience of this issue when it becomes publicly mobilized—and the result is almost always change in the direction of restrictiveness.⁵¹ Finally, Portugal and the U.K. have much weaker far right movements, and their citizenship policies have not changed much in the recent past.

Although this measure of the mobilization of anti-immigrant sentiment is probably the best single indicator available, it only captures part of the larger political story. The reaction of more mainstream conservative parties to the challenge of the far right's message can be just as effective in blocking liberalization. And public referenda and other forms of social mobilization, which are not captured by the far right measure, can in some ways result in more rapid and decisive restrictions than the standard process of elite and party politics.

In other words, while my argument about the impact of the far right helps to account for variation in the type of change that has or has not occurred, it certainly does not provide the final word on the topic of citizenship policies and how they have adjusted to new pressures and circumstances.⁵²

Neither does the argument from the previous section about the importance of historical legacies for understanding existing (and to some extent persisting) cross-national variation. But the goal of this analysis was to provide new sources and categories of evidence, while illustrating the potential fruitfulness of incorporating the empirical study of citizenship into mainstream debates within comparative politics. Hopefully the suggestive evidence and arguments presented above will show the value in renewing our empirical focus on Aristotle's theoretical questions.

Conclusion

Over the past few decades, almost every country in the EU-15 has revisited—though not necessarily revised—its citizenship law. And while the international and domestic pressures to liberalize have been significant, and sometimes decisive, they have sometimes been held in check by the countervailing pressure of a mobilized public opinion that is latently hostile to immigrants. In fact, as the evidence in this article suggests, it appears that when public opinion gets activated politically, with a concrete sponsor or means of expression, liberalization is usually stopped, or an existing law becomes even more restrictive. This was the case in Austria, Denmark, and Italy, where the far right parties have played leading roles, in Ireland, where a restrictive referendum passed overwhelmingly, and in Germany, where an unprecedented petition campaign rapidly stopped the momentum of liberalization. But if, on the other hand elites manage to pass reforms without significant public involvement—as occurred in Finland, Luxembourg, and Sweden—then liberalization will most likely be the outcome.

This brings us to a larger paradox, if not a serious normative problem: in terms of issues dealing with immigration and citizenship, a non-democratic, elite-driven process may lead to more liberal outcomes, whereas genuine popular involvement can result in more restrictive laws and institutions. In other words, proponents of liberal, inclusive outcomes should give more thought to the role of democracy—whether representative democracy that results in the inclusion of far right parties in governments and policymaking, or direct democracy that takes the form of referenda and initiatives—on issues that are prone to populism, xenophobia, and racism.⁵³

The trend is clear, as countries are increasingly relying on referenda and popular initiatives, which advocates of the far right view as being the ultimate expression of “true democracy.” The great challenge, particularly in the EU, will be for elites to surmount the much-criticized “democratic deficit,” while avoiding the trap of populism.⁵⁴ Whatever the longer-term outcome, however, the issue of citizenship will clearly remain an important concept, and one that theoretically-informed but empirically-oriented comparativists should incorporate into their research.