Liberalising Change or Continued Restrictiveness? Explaining Variation in Citizenship Laws in 11 EU Countries

Liberalising Change or Continued Restrictiveness? Explaining Variation in Citizenship Laws in 11 EU Countries Marc Morjé Howard Associate Professor Department of Government Georgetown University 681 ICC Washington, DC 20036 phone: 202-687-5029 email: mmh@georgetown.edu Biographical Note: Marc Morjé Howard is an Associate Professor in the Department of Government at Georgetown University. His research has focused on immigration and citizenship, civil society, democracy and democratization, and hybrid regimes. In addition to numerous journal articles on these topics, he is the author of The Weakness of Civil Society in Post-Communist Europe (Cambridge University Press, 2003) and Varieties of Citizenship in the European Union (Cambridge University Press, forthcoming). Keywords: citizenship, immigration, public mobilization, far right parties, Europe

Liberalising Change or Continued Restrictiveness? Explaining Variation in Citizenship Laws in 11 EU Countries Abstract: Over the past decade, the empirical study of national citizenship policies has emerged as a fast-growing area of research. This article contributes to that literature by focusing on the question of continuity and change in the 11 historically restrictive countries within the 15 older member states of the European Union. It seeks to explain why six of the countries have liberalised their citizenship policies since the 1980s, whereas five of them have not. The article develops and supports an explanation that focuses on the politics of citizenship. The main finding is that while citizenship liberalisation is more likely to occur with a leftist government, the most important factor is the relative strength of far right parties, which can serve to mobilise latent anti-immigrant public opinion, and thereby trump the pressures for liberalisation.

1 Introduction Over the past decade, the empirical study of national citizenship policies has emerged as a fastgrowing area of research (Aleinikoff and Klusmeyer 2000, 2001, 2002; Hansen and Weil 2001, 2002; Vink 2005; Howard 2005, 2006; Bauböck et al. 2006; Joppke 2008). The boundary of citizenship allows rich states to draw a line that separates its citizens from potential immigrants from poor countries, as well as to create internal distinctions between citizens and foreign residents by associating certain rights and privileges with national citizenship (Brubaker 1992). Despite predictions about the disappearance or decreased importance of national citizenship (Soysal 1994; Sassen 1996, 1998), distinctions between citizens and foreigners remain an essential and enduring feature of modern life whether in terms of politics and elections, welfare state benefits, public-sector employment, social integration, or demographics and pension systems even in the supra-national European Union (Howard 2006). In fact, many EU states have recently been revising and adapting their laws and policies that determine who can become a national citizen, and under what conditions. This article contributes to this burgeoning literature by exploring the question of continuity and change in national citizenship policies. It focuses in particular on the 11 historically restrictive countries within the 15 older member states of the EU, and it seeks to explain why six of the countries (Finland, Germany, Luxembourg, the Netherlands, Portugal, and Sweden) have liberalised their citizenship policies since the 1980s, whereas five of them

2 (Austria, Denmark, Greece, Italy, and Spain) have not. In other words, the four historically liberal countries within the EU-15 (Belgium, France, Ireland, and the U.K.) are excluded from the analysis in this article, for the very simple reason that since their citizenship policies were already liberal, the question of liberalisation does not apply to them in the same way. In order to address this question, I develop an argument that emphasises the politics of citizenship. This political explanation stresses not only whether parties of the left or the right have been in power, but also the extent to which the issue of citizenship became a contentious and politicised issue that received wider public debate and mobilisation. I then consider arguments and hypotheses adapted from the broader literature on comparative immigration that emphasise economic and socio-economic factors (GPD per capita, economic growth, and unemployment), demographics (the number and percentage of immigrants), and public opinion (measuring anti-immigrant sentiment). After presenting all of the arguments, I turn to an empirical analysis that evaluates the bivariate relationships between each of the hypotheses and citizenship liberalisation. The results show that the alternative factors do not work out empirically, but the association between the political factors and citizenship liberalisation is very strong. In short, while citizenship liberalisation is more likely to occur when the left is in power, the most important factor is the relative strength of far right parties, which can serve to mobilise latent anti-immigrant public opinion, and thereby trump the pressures for liberalisation. Although this cross-national analysis remains limited to fairly rigid measures on a common set of factors, it helps to show the broad patterns involved and to explain why some of the restrictive countries liberalised, while others remained restrictive.

3 A New Citizenship Policy Index (CPI) The Citizenship Policy Index (CPI) that constitutes the empirical starting point of this article represents a significant improvement over another version that I published in two previous articles (Howard 2005, 2006). Whereas the earlier CPI was based on my own coding of the citizenship laws on a few key elements of citizenship policy, the new CPI draws on the tremendous research by individual country experts, within a common methodological framework that went into a major project known as NATAC. 1 It also incorporates a very recent development that most studies of comparative citizenship have not yet dealt with systematically, namely the new language and civic integration requirements that a number of countries have mandated as a condition for naturalisation (Joppke 2008; Jacobs and Rea 2007). Like the earlier version, the new CPI consists of a simple addition of the following three factors: 1) whether a country grants jus soli, i.e., whether children of non-citizens who are born in a country s territory can acquire that country s citizenship (either immediately at birth or automatically after a certain number of years of residence); 2) the minimum length of its residency requirement for naturalisation (for both immigrants themselves and immigrant spouses who are married to citizens); and 3) whether or not naturalised immigrants are allowed to hold dual citizenship. Each component is scored on a 0-2 scale, yielding a 0-6 point range for the total index. Although space limitations preclude a full description and presentation of all the scores, the new CPI now includes several more fine-grained distinctions, along with better contextual information, thus yielding a final index that is more precise and accurate. More specifically, on 1 The NATAC project, also called the The acquisition of nationality in EU Member States: rules, practices and quantitative developments, was directed by Rainer Bauböck and funded primarily by the European Commission. Its findings have been published by Amsterdam University Press in two volumes (Bauböck et al. 2006), which provide the most detailed, comprehensive, and systematic analyses of citizenship policies in the EU-15 that exist to date, thus representing a tremendous resource to scholars in this area.

4 jus soli, the new index now allows for more options, in two respects: first, it captures both jus soli at birth and jus soli after birth (whereas the old CPI only measured jus soli at birth), with separate scores for both of these elements of jus soli, along with a final score that is the average of the two; second, for jus soli at birth, it provides several intermediate scores between 0 and 2 points, corresponding to the strictness of the parental residency requirement and whether a country allows so-called double jus soli. The coding of naturalisation requirements is also considerably more precise in the new CPI. It now considers both the requirements for an individual immigrant to become a naturalised citizen (which I refer to as immigrant naturalisation ) and those for an immigrant who is married to a citizen ( spousal naturalisation ). These are the two most common ways for an immigrant to acquire a country s citizenship. In addition, the new score penalises countries that have added civic integration requirements such as mandatory language and civics tests that add restrictions to naturalisation. And it also uses actual naturalisation rates as a check to ensure that the relative inclusiveness or restrictiveness of the policies are also reflected in the extent of people really naturalising, by reducing the scores of countries (to different degrees) with low levels of naturalisation rates. The final component of the CPI, dual citizenship for immigrants, still gives 2 points for countries that have no restrictions on allowing immigrants to retain their prior citizenship upon naturalisation, and 0 points for those who expressly forbid it. But it now gives intermediate scores to countries that have a large number of dual citizenships in practice, or that have created numerous and explicit exceptions to the formal prohibition on dual citizenship, or that do not require proof of renunciation of the prior citizenship, thereby not really enforcing its official ban

5 on dual citizenship. And these scores are then corrected, or reduced, by the same naturalisation rates criterion described above. 2 Overall, the new CPI provides a much more refined and accurate index that allows us to assess and compare the citizenship policies across countries. In order to conduct an analysis of the presence or absence of change, it is important to start with two clearly-defined points in time as benchmarks. One, which I call the 1980s, is admittedly somewhat arbitrary, but that time period can serve as a general proxy for previous laws and traditions. Indeed, in most cases the laws in the 1980s were identical to or closely in line with the historical origins of each country s laws (Hanson and Weil 2001). Using this general label allows us to account for any changes that may have occurred at any time during that decade. 3 The second point in time is the contemporary period, based on the existing laws and practices at the end of 2007. The comparison of the two time periods using the same methodological criteria allows us to explore the causes of both liberalising change and restrictive continuity. Turning to the empirical results of the new CPI, Table 1 shows the country scores for the EU-15 with the countries ordered in terms of their total CPI score, from lowest to highest on each of the three components in the 1980s, and Table 2 presents the same scores for the contemporary period. Each of the tables distinguishes between three groups of countries, depending on whether their citizenship policies can be characterised as restrictive (scores 2 Although technically the reduction should be applied strictly to naturalization requirements, in my view the application of a liberal dual citizenship policy is ultimately meaningless if very few people can actually become citizens. It is therefore important to adjust the scores to reflect the possibility for immigrants to actually benefit from these policies. The result is a more sensitive and meaningful final score on dual citizenship for immigrants. 3 In a sense, therefore, the effective cut-off date for this time period is 1990.

6 between 0 and 1.5) medium (over 1.5 but less than 4) or liberal (4 and above). 4 Both tables show a great deal of variation on each of the three components of the CPI, as the scores span the full spectrum of the 0-2 range on just about all of the components. And the total scores also display wide variation in both time periods as well. [Tables 1 and 2 about here] Table 1 shows that seven countries (Austria, Denmark, Germany, Greece, Italy, Luxembourg, and Spain) could be characterised as historically restrictive in the 1980s, although obviously some were more so than others. Finland, the Netherlands, Portugal, and Sweden were in the medium category, though with different combinations of scores distinguishing the Nordic and non-nordic countries. Finally, the four-country group of Belgium, France, Ireland, and the U.K. constituted the historically liberal category, with quite inclusive scores on all three components of the CPI. Overall, the table indicates that the historical variation in citizenship policies in what later became the EU-15 was very pronounced. This historical pattern was rather rapidly undone, however, over the next decade or two. As Table 2 shows, five of the seven countries (Austria, Denmark, Greece, Italy, and Spain) still remain in the restrictive category, while Germany and Luxembourg moved up to the medium group. At the same time, Finland, the Netherlands, Portugal, and Sweden all moved up from medium to liberal. As a result, a slight majority of the EU-15 can now be considered to have liberal citizenship policies. [Figure 1 about here] Figure 1 compares the two time periods more explicitly, showing the overall CPI scores in both the 1980s and today. The figure shows that most of the countries displayed relatively 4 Note that my use of the word liberal in this article is mainly with reference to the issue of citizenship policies, not to the various meanings and traditions associated with the concept of liberalism. In other words, the categories liberal and restrictive are essentially measures of the inclusiveness of each country s citizenship policies.

7 small changes of under 1 point, but six countries saw their CPI scores liberalise considerably. Luxembourg increased by 1.34 (with more liberalisation planned for the year 2008), the Netherlands by 1.5, Germany by just over 2 points, Portugal and Finland by more than 2.5 points, and Sweden by 3.75 points. The only country that moved considerably in a negative direction was Denmark, which dropped from 1.43 to 0 on account of its lengthening of the residency requirement for both immigrant and spousal naturalisation, as well as new citizenship and language tests. Looking at the changes on aggregate, a comparison of the overall scores from the 1980s to the contemporary period shows that the average score increased from 2.17 in the 1980s to 3.01 today corresponding to a quite large change on average. In other words, within the EU-15 as a whole, the recent liberalisation of citizenship policies is undeniable. And in this sense, it is clear that a relative convergence of citizenship policies has taken place. However, whether this becomes an absolute convergence in the sense of reaching a single, common policy, as with so many other areas of institutional harmonisation within the EU is still very uncertain, and appears unlikely in the foreseeable future (Joppke 1999). Indeed, although the cross-national differences are not nearly as vast as they were a few decades ago, they remain quite wide today. Figure 1 also shows how the EU-15 countries can be grouped into three particular types. The first group, on the bottom of the figure, consists of countries that have displayed restrictive continuity, meaning that their historically restrictive citizenship policies have remained impervious to the liberalising trends of the past decades. The six countries in the second group, liberalising change, have all liberalised their citizenship policies considerably since the 1980s increasing their CPI scores by at least 1 point, and in most cases much more 5 thus 5 Note that Luxembourg only increased by 1.34 points, but since it is scheduled to implement further liberalization in 2008, which should introduce dual citizenship for immigrants, it certainly belongs in this group.

8 moving up to either the medium or liberal category as a result. Finally, the historically liberal countries are at the top, but since by definition they cannot be considered cases of restrictive continuity or liberalising change, they are not included in the analysis that follows. Having established an empirical baseline that shows the extent of continuity and change since the 1980s, we can now turn to causal analysis. Why have some countries liberalised, while others have resisted the pressures of liberalisation 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 actually 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 have they functioned? The Argument: The Politics of Citizenship In order to address these important questions, I introduce a perspective that I call the politics of citizenship. I seek to understand how political actors have navigated the potentially treacherous waters on this volatile issue, how they have dealt with the various pressures for both liberalisation and restrictiveness, and how they have attempted to implement their choices into new laws and policies. Although it is certainly possible that lack of change simply represents a form of institutional inertia where the old policies continue to persist simply because they already existed I argue that elites have pursued conscious strategies and fought open battles, and that these contingent political factors were decisive in shaping citizenship policy change. Since the social science literature on comparative citizenship is still quite recent, there exist few off the shelf theories, hypotheses, and measures that can readily be incorporated into

9 empirical analysis. Instead, one can look to the related and more long-standing literature on immigration to find arguments that have been used to explain policy change. As it turns out, the immigration field is much stronger on providing theoretical reasons for liberalisation than for restrictiveness. Indeed, several arguments have been developed to explain increasing liberalisation, with some emphasising global causes and others stressing domestic factors. On the global level, scholars have stressed economic globalisation (Sassen 1996, 1998), neofunctional economic cooperation (Philip 1994), and new norms of postnational human rights (Soysal 1994), to explain liberalisation. And on the domestic level, scholars have considered interest group politics, whereby organised groups and businesses often exert influence quietly on policymakers to expand immigration (Freeman 1992), 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 (Joppke 1998). Despite the quite different points of emphasis, each of these arguments has the expectation of increasing liberalisation across the countries of the EU. Indeed, scholars have usually applied them with a broad sweep meant to apply to the entire advanced industrialised world, rather than as a differentiated source of variation between countries. In other words, according to these arguments, the entire region of Europe should have liberalised. In some sense, the empirical baseline presented above provides support for these arguments, since most of the countries that changed became more liberal in their attribution of citizenship. But these arguments do not help us account for the countries that did not liberalise. What can explain the lack of liberalisation in some countries? This perspective has been less explored in a literature that primarily focuses on (and predicts) liberalisation. 6 6 For one important exception, see Money (1999).

10 While these latent liberalising pressures are certainly real and influential, they in fact only tell part of the story. For on the other side lies an important latent pressure for restrictiveness: public opinion. Indeed, a common and perhaps unifying feature throughout the EU is the strong hostility to immigrants on the part of many, if not most, Europeans (Schain, Zolberg, and Hossay 2002; Norris 2005). For example, Figure 2 presents the data on the 11 relevant countries 7 from a 2000 Eurobarometer survey, which included the following five anti-immigrant statements, with which respondents could either tend to agree or tend to disagree : The presence of people from these minority groups is a cause of insecurity. The presence of people from these minority groups increases unemployment in [country]. There is a limit to how many people of other races, religions or cultures a society can accept. [Our country] has reached its limits; if there were to be more people belonging to these minority groups we would have problems. People belonging to these minority groups are so different, they can never be fully accepted members of [nationality] society. The figure shows the average percent of respondents who tend to agree with these five statements. As it turns out, over 40 percent of people in all countries, and in most cases substantially over 50 percent, agree with these very harsh statements. 8 What is particularly noteworthy about the figure is that while the scores are certainly not uniform or identical across countries, the variation is fairly limited. Anti-immigrant sentiment seems to be consistently high throughout the region. [Figure 2 about here] As a result, it seems clear that in contrast to the latent pressures for liberalisation identified above anti-immigrant public opinion serves as a latent pressure for restrictiveness. 7 Note that I am still excluding the four historically liberal countries, both because this article focuses on liberalization and because the figure is already very dense. But the results for Belgium, France, Ireland, and the U.K. are not at all different from the other EU countries in terms of their high levels of anti-immigrant sentiment. 8 Data from the 2002 European Social Survey paints a similar picture, showing high levels of anti-immigrant sentiment throughout the EU countries (Sides and Citrin 2007).

11 The question, of course, is whether or, more precisely, where it has an influence. For if political outcomes were a direct implementation of the popular will, liberalisation would not have occurred in most countries, if at all. Indeed, as Douglas Massey (1999, 313) writes, Most citizens [ ] are poorly organised and politically apathetic, leaving immigration policies to be determined quietly by well-financed and better-organised special interests operating through bureaucratic channels. On the other hand, if public opinion were entirely irrelevant, liberalisation might have already occurred everywhere, since the liberalising pressures are often more direct and better organised (Cornelius et al. 2004). The key question, therefore, is how political decision-makers in individual countries respond to the common pressures of liberalisation and restrictiveness in order to implement policy. Figures 3 and 4 attempt to answer this question, and to provide a crystallisation of my argument about the politics of citizenship. Figure 3 recaps the various latent pressures on each side, showing that the increasing demographic change within Europe, the rise of new international norms, the long-standing impact of interest groups, and the role of the courts all provide pressures on political decision-makers to liberalise their citizenship policies. On the other side, of course, public opinion remains staunchly and consistently anti-immigrant. [Figures 3 and 4 about here] So far this only explains pressures that are latent; in other words, they are in the background without necessarily being active or dominant. But is there a mechanism through which we can determine which types of pressures lead to which outcomes? Figure 4 presents my two-part political model for explaining why citizenship liberalisation generally does or does not occur. Part 1 draws on Christian Joppke s argument about the distinctive politics of center-left and center-right governments on citizenship issues.

12 As Joppke (2003) shows, leftist governments are typically in favor of increasing the citizenship rights of immigrants (which he calls de-ethnicisation ), whereas right-of-center governments want to resist such impulses while simultaneously expanding the country s connections to its émigrés (which he calls re-ethnicisation). In other words, the expectation is that citizenship liberalisation will be much more likely to occur if a leftist government is in power. In that case, liberalisation is possible (though certainly not guaranteed), whereas if a center-right government is in place, liberalisation remains unlikely (though not impossible). Joppke s argument is largely supported by the fact that of the six countries that liberalised their citizenship policies, Finland, Germany, Portugal, and Sweden all had left-of-center governments at the time the reform was passed, and the Netherlands had multiple grand coalition governments that included the Social Democrats, whereas only Luxembourg had a center-right government. In other words, one can argue that the first step applies to five of the six cases in question, and that having a leftist government is more of a necessary than a sufficient condition for citizenship liberalisation. The more important part of the argument and the particular contribution of this analysis on Figure 4 is Part 2. It focuses on the latent pressure for restrictiveness provided by public opinion, and asks whether this anti-immigrant sentiment becomes activated politically. In other words, the issue is not whether a center-left or center-right government is in power, but whether the far right is active and mobilised on the issue of immigration and citizenship reform. 9 This can happen 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. If this is the case, then liberalisation will essentially be blocked, since no government (whether on the left or right) will be able to overcome the intensity of this mobilisation, because it has so much resonance in the 9 Note that I use the term far right to refer to what others sometimes call the radical right or extreme right.

13 population. In other words, politicians can easily overlook public opinion when it remains dormant, restricted to private beliefs about an issue; but once that sentiment becomes mobilised politically, politicians can only ignore it at their own peril (Givens and Luetke 2004). If there is no far right mobilisation, however, then citizenship liberalisation is possible as an outcome. In short, the mobilisation of anti-immigrant sentiment can essentially trump the liberalising pressures and a leftist government, and thereby block liberalisation. Operationalisation of the Argument As with many concepts in social science, it is often easier to define a term and spell out an argument in abstract language than it is to assign clear and systematic measures based on existing data from the real world. The activation of anti-immigrant public opinion is difficult to measure using data that can apply to all of the relevant countries. Mobilising events such as referenda or petition campaigns are not only relative rare, but they are also very context-specific, thereby defying a common measure that could be applied across countries. The best systematic operationalisation of this concept captures the electoral support that far right parties i.e., those whose main platform emphasises opposition to immigration and immigrants have received. Although this measure does not cover all aspects of the concept explained above, it does get at the core of it. I therefore incorporate a measure of the average electoral strength of far right parties in national elections in each country between 1992 and 2006. 10 10 For the six countries that liberalized, the average far right scores go up to the year of citizenship liberalization, since the strength of the far right after the change has gone through is no longer relevant to the question of liberalization. Indeed, in some cases, such as the Netherlands, the far right became much stronger after the law was liberalized, precisely in opposition to that law.

14 Before turning to a more causal analysis of the relationship between the factors identified in my political model and the variation in the occurrence of citizenship liberalisation, we need to address some possible alternative arguments that could account for that variation. Alternative Explanations In this section, I briefly discuss some alternative arguments that derive primarily from the literature on immigration. These arguments lead to hypotheses that emphasise economic and socio-economic factors, demographics, and public opinion (Givens and Luedtke 2005). The field of political economy has established many powerful relationships between economic factors and political outcomes. Although these types of arguments have not been applied specifically to citizenship policy, it seems plausible that there would be a meaningful relationship between variation in either economic levels or economic growth and the liberalisation in citizenship laws. The expectation would therefore be that countries with more successful economies as indicated by either higher levels of GDP per capita or by higher levels of annual growth would be more likely to have the extra margin to afford citizenship liberalisation, since they would have more resources available to absorb a greater number of immigrants as new citizens. An important socio-economic factor that often enters into discussions about immigration is the unemployment rate. It is typically assumed that countries with lower unemployment rates can more easily absorb new workers (and citizens) than those who cannot fully employ their own populations already. Since the practice of citizenship acquisition depends in part on the extent to which immigrants live in a given country, it seems logical that there would be a connection between

15 these two factors. In principle, the greater the number of foreigners living in a country, the greater the pressure on that country s policymakers to liberalise the granting of citizenship. I include two measures of this factor: the raw number of immigrants, which of course is much higher in the larger countries; and the percentage of foreigners within the overall resident population. A final alternative explanation incorporates public opinion as an explanatory factor. According to this approach, the policies chosen by elites in power should reflect the views and preferences of their constituents. And countries whose citizens display relatively low levels of anti-immigrant sentiments should be more likely to liberalise their citizenship policies, whereas high levels of xenophobia should be linked to continuing restrictiveness. There are many possible survey measures of anti-immigrant sentiment, and indeed the results can easily shift from one question to the next, depending on how each question is phrased. In the analysis below, I include the composite measures shown above, based on the 2000 Eurobarometer survey. Evaluating the Evidence Having introduced a series of arguments that emphasise a variety of different possible causal factors, we can now turn to the empirical evidence. Since the N of 11 countries is too small for multivariate statistical analysis, we must rely on bivariate relationships between each of the variables and the outcome to be explained, namely the change in CPI scores from the 1980s to today. [Table 3 about here] Table 3 puts all of the competing explanations together with the alternative explanations at the top summarising the expected direction, and showing the correlation and

16 statistical significance for each hypothesis. The table shows the regression results for all seven separate bivariate regressions. It makes clear that several of the alternative arguments actually have the opposite of the predicted sign, and none of them even comes close to statistical significance. In contrast, the support for far right parties has the strongest of the relationships with CPI liberalisation, with the expected sign and a relatively strong correlation of -.48 and given the small sample size a reasonably high level of statistical significance. Figure 5 shows three clear clusters of countries: 1) Austria, Denmark, and Italy, which have very high levels of far right support and did not liberalise their citizenship laws; 2) the six liberalising countries, all of which had low levels of support for far right parties at the time of citizenship liberalisation; 3) and then Greece and Spain, the two countries with weak far right movements yet still no liberalisation. In short, as expected by Part 2 of the political model from Figure 4, liberalisation has only occurred without the presence of a significant far right party or movement. And when such a party flourished, the pressures for liberalisation were effectively blocked. [Figure 5 about here] The cases of Greece and Spain, which have not (yet?) liberalised, help to draw attention to an important feature of my argument about the importance of far right parties. While the presence of a strong anti-immigrant movement seems to be a necessary and sufficient factor that prevents citizenship liberalisation, the absence of the far right is a necessary but not sufficient condition for liberalisation. In other words, given the current constellation of political forces, there is probably a much greater chance for liberalisation in Spain and Greece than there is in Austria or Denmark.

17 Conclusion Although this measure of the mobilisation 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 liberalisation. And public referenda and other forms of social mobilisation, 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. For example, this alternate form of far right pressure was critical to citizenship developments the case of Germany, where anti-immigrant public opinion was activated through a massive petition campaign, rather than a far right party (such parties are closely watched and often banned in Germany), and which led to a restrictive backlash that diluted the extent of citizenship liberalisation in 2000. Moreover, looking for a moment at the historically liberal countries, the case of Ireland highlights the role of public mobilisation outside of party politics. Although Ireland does not have an organised far right movement, proponents of restrictions on citizenship acquisition succeeded in implementing a controversial referendum, which passed overwhelmingly (with 80% 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. 11 This remarkable development shows the tremendous salience of this issue when it becomes publicly mobilised and the result is almost always change in the direction of restrictiveness. The same phenomenon 11 It should be pointed out, however, that while the new law is certainly more restrictive than it was previously, Ireland still grants jus soli, and in fact it is still more liberal than most other countries, such as Germany, which have lengthier residency requirements for the parents of children born in the host country, and which sometimes include employment requirements that many poorer immigrants do not fulfill.

18 has occurred repeatedly in Switzerland (which is not an EU member state, and therefore not included as a case in this study) where voters have consistently rejected referenda that would have liberalised the extremely restrictive Swiss citizenship law. Finally, expanding the analysis further to include other historically liberal countries, we see that Belgium and France both of which have fairly strong far right movements, with 1992-2006 average far right returns of 12.3% and 13%, respectively have seen immigration and citizenship emerge as highly polarising political issues. These two countries have either experienced (in the case of France in the mid-1990s) or are currently experiencing (in the case of Belgium) some tinkering with their citizenship laws as a result of the pressure of the far right. In short, while the evidence above shows that my argument about the impact of far right parties helps to account for whether or not liberalising change has 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. Much more analysis and evidence is needed. But the crossnational comparison in this article provides a valuable starting point for understanding the complex social and political processes involved.

19 References Aleinikoff, T. A. and Klusmeyer, D. (eds) (2000) From Migrants to Citizens: Membership in a Changing World. Washington, DC: Carnegie Endowment for International Peace. Aleinikoff, T. A. and Klusmeyer, D. (eds) (2001) Citizenship Today: Global Perspectives and Practices. Washington, DC: Carnegie Endowment for International Peace. Aleinikoff, T. A. and Klusmeyer, D. (eds) (2002) Citizenship Policies for an Age of Migration. Washington, DC: Carnegie Endowment for International Peace. Bauböck, R., Ersbøll, E., Groenendijk, K., and Waldrauch, H. (eds).(2006) Acquisition and Loss of Nationality. Amsterdam: Amsterdam University Press, 2 volumes. Brubaker, R. (1992) Citizenship and Nationhood in France and Germany. Cambridge: Harvard University Press. Cornelius, W., Tsuda, T., Martin, P.L. and Hollifield, J.F. (eds) (2004) Controlling Immigration: A Global Perspective, 2 nd ed. Stanford: Stanford University Press. Freeman, G. P. (1992) Migration Policy and Politics in the Receiving States, International Migration Review 26(4): 1144-67. Givens, T. and Luedtke, A. (2004) The Politics of European Union Immigration Policy: Institutions, Salience, and Harmonization, The Policy Studies Journal 32(1): 145-65. Givens, T. and Luedtke, A. (2005) European Immigration Policies in Comparative Perspective: Issue Salience, Partisanship and Immigrant Rights, Comparative European Politics 3(1): 1-22. Hansen, R. and Weil, P. (eds) (2001) Towards a European Nationality: Citizenship, Immigration, and Nationality Law in the EU. New York: Palgrave. Hansen, R. and Weil, P. (eds) (2002) Dual Nationality, Social Rights, and Federal Citizenship in the US and Europe: The Reinvention of Citizenship. New York: Berghahn Books. Howard, M.M. (2005) Variation in Dual Citizenship Policies in the Countries of the EU, International Migration Review 39(3): 697-720. Jacobs, D. and Rea, A. (2007) The End of National Models? Integration Courses and Citizenship Trajectories in Europe. Paper presented at the European Union Studies Association, Montréal, 17-19 May. Joppke, C. (1998) Why Liberal States Accept Unwanted Immigration, World Politics 50(2): 266-93. Joppke, C. (1999) How Immigration is Changing Citizenship: A Comparative View, Ethnic and Racial Studies 22(4): 629-652. Joppke, C. (2003) Citizenship between De- and Re-Ethnicization, European Journal of Sociology 44(3): 429-58. Joppke, C. (2008) Comparative Citizenship: A Restrictive Turn in Europe?, Journal of Law & Ethics of Human Rights 2. Massey, D. S. (1999) International Migration at the Dawn of the Twenty-First Century: The Role of the State, Population and Development Review 25(2): 303-22. Money, J. (1999) Fences and Neighbors: The Political Geography of Immigration Control. Ithaca: Cornell University Press. Norris, P. (2005) Radical Right: Voters and Parties in the Electoral Market. New York: Cambridge University Press.

Philip, A.B. (1994) European Union Immigration Policy: Phantom, Fantasy, or Fact, West European Politics, 17(2): 168-191. Sassen, S. (1996) Losing Control? Sovereignty in an Age of Globalization. New York: Columbia University Press. Sassen, S (1998) The de facto Transnationalizing of Immigration Policy, in Joppke, C (ed) Challenge to the Nation-State. Oxford: Oxford University Press. Schain, M., Zolberg, A., and Hossay, P. (eds) (2002) Shadows over Europe: The Development and Impact of the Extreme Right in Western Europe. New York: Palgrave. Sides, J. and Citrin, J. (2007) European Opinion About Immigration: The Role of Identities, Interests and Information, British Journal of Political Science, 37(3): 477-504. Soysal, Y. (1994) Limits of Citizenship: Migrants and Postnational Membership in Europe. Chicago: University of Chicago Press. Vink, M. P. (2005) Limits of European Citizenship: European Integration and Domestic Immigration Policies. Basingstoke: Palgrave Macmillan. 20

Table 1: The Three Main Components of Citizenship Policies for the EU-15 in the 1980s Country Jus Soli (0-2) Naturalization Requirements (0-2) Dual Citizenship for Immigrants (0-2) CPI SCORE (0-6) Restrictive Medium Liberal Germany 0.00 0.00 0.00 0.00 Austria 0.00 0.50 0.00 0.50 Luxembourg 0.00 0.50 0.00 0.50 Italy 0.00 0.72 0.00 0.72 Greece 0.00 0.00 1.25 1.25 Spain 0.50 0.75 0.00 1.25 Denmark 0.00 1.43 0.00 1.43 Finland 0.00 1.72 0.00 1.72 Sweden 0.00 1.72 0.00 1.72 Portugal 0.75 0.00 1.00 1.75 Netherlands 1.50 1.22 0.00 2.72 France 1.50 1.22 1.50 4.22 Ireland 2.00 1.11 1.25 4.36 Belgium 1.50 1.75 1.75 5.00 U. K. 1.75 1.72 2.00 5.47

Table 2: The Three Main Components of Citizenship Policies for the EU-15 Today Country Jus Soli (0-2) Naturalization Requirements (0-2) Dual Citizenship for Immigrants (0-2) CPI SCORE (0-6) Restrictive Med. Liberal Austria 0.00 0.00 0.00 0.00 Denmark 0.00 0.00 0.00 0.00 Greece 0.00 0.00 1.00 1.00 Spain 0.50 0.38 0.50 1.38 Italy 0.00 0.25 1.25 1.50 Luxembourg 1.00 0.84 0.00 1.84 Germany 0.75 0.54 0.75 2.04 Netherlands 1.50 1.22 1.50 4.22 Finland 1.00 1.32 2.00 4.32 Portugal 1.75 1.07 1.50 4.32 Ireland 2.00 1.36 1.50 4.86 France 1.50 1.47 2.00 4.97 U. K. 1.75 1.22 2.00 4.97 Sweden 1.50 1.72 2.00 5.22 Belgium 1.50 2.00 2.00 5.50

Figure 1: Citizenship Policy Index, 1980s and Today Belgium United Kingdom France Ireland Sweden Portugal Finland Netherlands Germany 1980s Today Luxembourg Italy Spain Greece Denmark Austria 0 1 2 3 4 5 6

Figure 2: Average Agreement Rate with Five Anti-Immigrant Statements Greece Germany Netherlands Luxembourg Austria Portugal Denmark Italy Sweden Spain Finland 0 10 20 30 40 50 60 70 80 Average percentage of respondents who "tend to agree" with the statements Source: Eurobarometer 53 (2000)

Figure 3: Latent Pressures on National Citizenship Policies Latent Pressures for Liberalization: Latent Pressures for Restrictiveness: 1) Demographic Transformation (increasing numbers of foreigners, de facto continent of immigration) 2) International Norms (human rights, modernizing to European standards) 3) Interest Groups (businesses, employer associations, immigrant groups) Decision-makers (political elites and political parties in the government) Public Opinion (anti-immigrant and xenophobic sentiments in the population) 4) Courts and Judges (judicial system generally supports the rights of immigrants)

Figure 4: A Two-Part Political Model for Explaining Citizenship Liberalization Part 1: Is a leftist government in power? No Yes Liberalization Unlikely Liberalization Possible Part 2: Does anti-immigrant public opinion become activated, either by a strong far right party or by public mobilization? Yes No Liberalization Unlikely Liberalization Possible

Table 3: Correlations between Country-Level Factors and Citizenship Liberalization Variable Expected direction Correlation Significance N GDP per capita +.02.48 11 Economic Growth + -.03.47 11 Unemployment - -.06.43 11 Number of foreigners + -.04.45 11 Percentage of foreigners + -.12.36 11 Anti-immigrant sentiment - -.27.21 11 Support for far right parties - -.48.067 11

Change in CPI from 1980s to Today Figure 5: The Relationship between Support for Far Right Parties and CPI Change Swe 3.00 Fin Port 2.00 Germ Neth Lux 1.00 Italy 0.00 Spain Greece Den Aust 0.00 5.00 10.00 15.00 20.00 Ele ctoral Support of Fa r Right Partie s