Earning citizenship. Economic criteria for naturalisation in nine EU countries

Abstract

In several European countries, access to citizenship via naturalisation is conditional upon the payment of substantial fees or upon proof of a certain degree of economic self-sufficiency. Thus, if not acquired by birth, citizenship as political membership depends on the economic performance of an individual. In this article, economic criteria for naturalisation (income, employment, welfare benefit requirements and naturalisation fees) were scrutinised in nine EU countries (Austria, Belgium, Denmark, Finland, Germany, Ireland, Netherlands, Sweden and the United Kingdom) over the period from 1985 to 2014. After describing their empirical variation, the article investigates in which contexts they are more or less strongly developed. The study finds that naturalisation fees in particular have increased over time. Differences in economic criteria for naturalisation can be explained by the long-term power of far-right parties and by immigration rates, while other institutional, structural and political factors have only very limited explanatory power.

Keywords:

• Citizenship
• naturalisation
• policy analysis
• economic inequality

1. Introduction

In order to obtain the citizenship of their country of residence, immigrants need to meet several criteria for naturalisation. In several European countries, these criteria include a certain degree of economic self-sufficiency and the payment of substantial fees in the naturalisation process. Since electoral participation at the national level tends to be a privilege of citizens of the respective country (Arrighi and Bauböck 2017), an electoral exclusion of resident foreigners can be a consequence of insufficient economic resources. In 2012, Joachim Stern estimated that about 60–70% of all Austrian female blue-collar workers would not be able to meet the income requirements for naturalisation in Austrian citizenship law (Stern 2012, 60). Consequently, if an ‘average’ female blue-collar worker had not acquired Austrian citizenship by descent – being born to Austrian parents – she would be excluded from citizenship and the rights that come with it, including voting rights. Although some scholars argue that focusing on economic resources as a criterion for access to citizenship is a widespread trend in western Europe (van Houdt, Suvarierol, and Schinkel 2011), little is known in a comparative perspective on how economic criteria for naturalisation have evolved.

In this paper, I investigate to what extent economic resources are relevant for becoming part of the legitimising basis of a democratic polity and the institutional and structural context of these criteria. Covering nine EU countries (AT, BE, DK, FI, DE, IE, NL, SE, UK) over the period of 1985–2014, I show that economic requirements for naturalisation changed only infrequently, while naturalisation fees increased substantially over time. In order to describe their development over time and allow a comparison of their relative strength across countries, data on economic requirements and fees in the naturalisation process are merged into an index of economic criteria for naturalisation. In a second step, the development of economic criteria for naturalisation is explained using time-series, cross-section regression models. The findings suggest that the strength of economic criteria for naturalisation can be explained by the long-term strength of far-right parties and changes in immigration patterns.

2. Economic status and citizenship in mobile societies

While many aspects of citizenship and naturalisation policies have received considerable academic attention, economic criteria for naturalisation remain fairly uncharted in contemporary studies of citizenship policy development. Koopmans, Michalowski, and Waibel (2012) cover information on welfare dependency as potential barrier for naturalisation in their ICRI index. Janoski (2010) includes naturalisation fees in his BNI index and Goodman (2014) covers costs for civic integration requirements in her CIVIX index. Healy and Reichel (2013) extracted information on economic requirements from the MIPEX index (Huddleston et al. 2011) and explored their cross-national variation in the EU in 2010. In sum, these studies include only some economic criteria as a component of citizenship policy indices. However, studies focusing on one or just a few countries suggest that such criteria became more elaborate recently and have strong exclusionary effects (van Houdt, Suvarierol, and Schinkel 2011; Reichel and Perchinig 2015; Stern 2012). Since there are no studies providing a comprehensive measurement, let alone a comparative analysis of the development of economic criteria for naturalisation, current knowledge on economic criteria for naturalisation is vague at best. To account for this research gap, the study merges information on income, employment and welfare benefits as criteria for naturalisation, plus various fees in the naturalisation process, to an index covering nine countries over 30 years.

2.1. Hypotheses on the development of economic criteria for naturalisation

Since many comparative studies aim to provide broader, comprehensive perspectives on citizenship and immigrant rights (Helbling 2013) and do not focus on specific details – such as civic integration or economic requirements – their explanatory approaches may not be applicable to the study of distinct aspects of citizenship policies per se. Consequently, while some institutional settings may be connected to the development of inclusive or exclusive policies towards immigrants in general, they do not necessarily imply consequences for the presence or absence of economic criteria for naturalisation. Furthermore, the sample of nine EU countries is homogeneous in many respects. Many institutional factors, therefore, which may account for variation in citizenship policies in a global perspective (Koopmans and Michalowski 2017), cannot be applied for this sample in a meaningful way. Focusing on economic criteria for naturalisation in the given sample of countries and years, the following factors might be associated with different degrees to which access to citizenship is conditional upon economic status:

Studies of citizenship regimes reveal that citizenship policies can be classified in several distinct regimes with regard to their relative openness towards immigrant populations (Howard 2009; Koopmans, Michalowski, and Waibel 2012; Vink and Bauböck 2013). Assuming that citizenship regimes distinguished by their openness towards resident foreigners are consistently inclusive or exclusive, I expected economic criteria for naturalisation to be more strongly developed in contexts in which citizenship is generally difficult to acquire for immigrants.

H1: Economic criteria for naturalisation are more strongly developed in contexts of exclusive citizenship regimes towards immigrants.

Although it is evident that citizenship policies were amended in western Europe in response to changing migration patterns, there is no clear evidence that the inclusiveness of citizenship policies towards immigrants is indeed shaped by higher or lower levels of immigrant inflows (Howard 2009; Janoski 2010). Since naturalisation is conditional upon meeting residence requirements (in this sample ranging from 3 to 10 years), policy-makers may not respond to changes in immigration rates immediately, but rather when naturalisation becomes viable for the newly immigrated after several years (Reichel and Perchinig 2015; Vink, Prokic-Breuer, and Dronkers 2013). In this perspective, any effect of immigration on economic criteria for naturalisation is expected to be delayed.

H2: Economic criteria for naturalisation are more strongly developed when immigration rates were high in the recent past.

As economic criteria for naturalisation are often framed by far-right parties as a measure to protect the welfare state from immigration (Stadlmair 2017a), we may assume that such a rationale may resonate more in highly developed welfare states. In this view, a more comprehensive and redistributive welfare state may be deemed worthier of protection than a welfare state offering fewer benefits to its members.

H3: Economic criteria for naturalisation are more strongly developed in contexts of comprehensive public welfare systems.

Many recent studies focus on power resources of political actors as a source for variation in citizenship and immigration policies (Akkerman 2012; Bale 2008; Howard 2009; Sredanovic 2016). The assumption is that actors of the political left are ‘agents of de-ethnicisation’, who aim to extend rights to immigrant populations, while the actors of the political rights are ‘agents of re-ethnicisation’, who try to maintain rights exclusively for the ethnic community (Joppke 2003). Primarily looking at governments, right-wing governments are expected to pursue exclusive policies towards immigrants and left-wing governments are expected to pursue inclusive policies (Sredanovic 2016). Focusing on economic criteria in particular, a case study on Austrian political parties supports the assumption that left-wing parties oppose economic criteria for naturalisation, while parties of the political right support them (Stadlmair 2017a).

H4: Economic criteria for naturalisation are less strongly developed when left-wing governments are in power.

While this argument is widely accepted, recent studies have highlighted the distinct relevance of far-right, populist parties for the development of citizenship and immigration policies. Since these parties are typically fierce opponents of immigration, their electoral successes in recent decades, including participation in several governments, may very well result in exclusionist policies towards immigrant populations. In his seminal study on citizenship policy change in Europe, Howard (2009) argues that when far-right parties successfully mobilised anti-immigrant opinion, a general liberalisation of citizenship policies did not take place. In contexts of already liberal citizenship regimes, far-right parties may cause restrictive turns in citizenship policies (Goodman and Howard 2013). Similar approaches were taken in several studies, some of which confirm the effect of far-right power (Abou-Chadi 2016; Howard 2009; Koopmans, Michalowski, and Waibel 2012), but others do not (Akkerman 2012; Sredanovic 2016). In order to identify possible conditions for the success of far-right parties in terms of policy output, two possible specifications are investigated in this paper: First, the effect of far-right power may be delayed, for several reasons. As far-right parties have been shown to be organised rather weakly, they may need some time to consolidate and effectively shape policies (Akkerman 2012). Furthermore, far-right parties may face a cordon sanitaire of other parties, hence their policy preferences may be adopted by other parties only when they constitute a long-term competitor for office, which is unlikely to occur immediately after their first electoral success (Bale et al. 2010; Mudde 2013). Second, the policy success of far-right parties may depend on the salience of ‘their’ issues. In this sense, they may be more likely to successfully tighten economic criteria for naturalisation in contexts of increasing immigration and comprehensive welfare systems, since economic criteria for naturalisation can be connected well with these issues in welfare-chauvinist narratives (Stadlmair 2017a). Consequently, four hypotheses on the (conditional) effect of far-right parties on economic criteria for naturalisation are tested:

H5a: Economic criteria for naturalisation are more strongly developed when far-right parties are strong at present.

H5b: Economic criteria for naturalisation are more strongly developed when far-right parties exhibit long-term strength.

H5c: Economic criteria for naturalisation are more strongly developed when far-right parties are strong and public social spending is high.

H5d: Economic criteria for naturalisation are more strongly developed when far-right parties are strong and immigration rates are high.

Lastly, economic development may shape immigration and citizenship policies (Cornelius, Martin, and Hollifield 1994). As research on economic development and immigration tends to focus on migration policies, possible consequences for citizenship policies are quite unclear. However, economic development should be controlled for in statistical models aimed at explaining the development of economic criteria for naturalisation.

3. Data and methods

Economic resources as a requirement for naturalisation may take three principal forms: the requirement to participate in the formal economy, to have an income, or not to draw certain welfare benefits (Jeffers, Honohan, and Bauböck 2012; Stadlmair 2014). In order to measure the relative strength of these requirements, six indicators on their legal format, thresholds, duration and exemptions are combined into an index ranging from 0 (no requirement) to 100 (most difficult requirement) (see Appendix 1): (1) The legal format distinguishes between explicit requirements mentioned in citizenship laws and mere procedural requirements (Huddleston 2013, i.e. the need to provide information on one’s income in a naturalisation application, even though there is no formal requirement mentioned in the citizenship laws). (2) Applicants may need to meet a certain income threshold to naturalise, or the assessment of what is deemed a sufficient income can be left to the discretion of the authorities processing naturalisation applications. (3) The time period for which an economic requirement needs to be met may differ. (4) Authorities may be able to disregard a requirement to avoid social hardship. For instance, such exemptions may be made for applicants with disabilities or who lost their employment through no fault of their own (Stadlmair 2014). (5) Second generation immigrants may be explicitly exempted from certain requirements in the naturalisation process, or may obtain the citizenship of their country of birth via ius soli or a distinct facilitated acquisition mode (Jeffers, Honohan, and Bauböck 2012). (6) Spouses of citizens and citizens of certain kin-states may also be exempted from some naturalisation requirements. These six indicators vary over time and across countries and can give a meaningful account of differences in economic requirements for naturalisation. Each indicator measuring the strength of economic requirements has three coding options. The index score for each observation (country_year) is measured by taking the mean of the six indicators. To be sure, economic requirements are not only relevant for naturalisation, but also for obtaining other residence titles, in particular for permanent residence (Healy and Reichel 2013; van Houdt, Suvarierol, and Schinkel 2011). If such a permanent residence title is necessary for naturalisation and is conditional upon economic requirements, this would constitute an indirect economic requirement for naturalisation. Such indirect requirements cannot be covered in the index in a meaningful way, but are briefly discussed in Section 4.

Apart from requirements, fees may constitute an economic obstacle for accessing citizenship. For the purpose of investigating costs in the naturalisation process over longer periods and across countries, only general expenses in the naturalisation process, which are independent from an applicant’s individual condition, can be considered. These expenses are measured with five indicators: (1) Fees for different stages in the naturalisation process, such as fees for submitting an application, fees for receiving a naturalisation certificate and fees for specific parts of the naturalisation process (such as citizenship tests or conferment ceremonies, Goodman 2014) are summarised and – if necessary – converted to EUR. Subsequently, the relative purchasing power in a country at a given point in time is controlled by applying inflation rates and purchasing power parities. These fees are then rescaled to a continuous indicator ranging from 0 (no fees) to 1 (maximum sum of fees in the sample). (2) Language skill certificates are often required and may incur additional costs, which are included as a dummy indicator (0 = no certification necessary). Since these certificates usually need to be obtained at market prices, they cannot be added as concrete EUR amounts to the sum of fees in a meaningful way. Hence, they are measured as a dummy indicator and subsequently added in the index. Since some categories of immigrants may need to pay lower fees for acquiring the citizenship of their country of residence, the measurement of naturalisation fees includes an indicator on possible exemptions or reductions for (3) second generation immigrants, (4) spouses of citizens and (5) kin-citizens. These five indicators are subsequently summarised to a weighted index, in which the total fees make up 70%, language skill certificates and exemptions/reductions for the second generation 10%, and exemptions/reductions for spouses and kin-citizens 5% of the index (see Appendix 1). This weighting strategy is driven by the intention of maintaining the substantial differences in the total fees over time and across countries. The analyses presented below include three dependent variables: an index of economic requirements for naturalisation (henceforth ERN index), an index of naturalisation fees (fee index) and an index of fees and economic requirements (overall ECN index).

Focusing on policy output, not on policy outcomes (Helbling 2013), citizenship laws and government decrees specifying the interpretation of citizenship laws are used as data to describe economic criteria for naturalisation. These documents are taken from government sites and from the Global Citizenship Observatory database (http://globalcit.eu/). To strengthen the internal validity of the study, I have contrasted the information from the EUDO database with information from comparative legal studies, such as de Groot (1989), Nascimbene (1996) and Davy (2001). In addition, GLOBALCIT country experts were consulted to discuss preliminary findings (see Acknowledgements).

The study covers nine EU member states – Austria, Belgium, Denmark, Finland, Germany, Ireland, the Netherlands, Sweden and the United Kingdom – over the period from 1985 to 2014. This selection of countries was driven by the aim of tracing the development of economic criteria for naturalisation over a longer time-frame. While it might strengthen the argument if more countries were included, such a course would meet some important obstacles: Especially in Southern and Eastern European countries, citizenship laws and decrees were – and still are – held in a very vague language, hampering a meaningful comparison of different policies (see e.g. Christopoulos 2009; Martín-Pérez and Moreno-Fuentes 2012; Zincone 2010). The selected timeframe starts with 1985, as Western European countries started to amend their citizenship laws by the mid-1980s in response to immigration based on labour migration programmes, family reunification and asylum migration (Howard 2009).

In order to show in which context economic criteria for naturalisation are more or less strongly developed, cross-section time-series data were analysed using multivariate regression models covering nine countries from 1990 to 2014 (225 country_years). Importantly, the intention was not to investigate processes of policy change per se, but to provide a general account of the political, institutional and structural context in which economic criteria are in place, which requires a quantitative analysis covering a longer time span and several countries (Wenzelburger, Jäckle, and König 2014). I use panel-corrected standard errors to account for the times-series cross-sectional data structure. In a simple test for non-stationarity, using a fixed-effects model with the overall ECN index as dependent variable and independent variable with a lag of one year (Wenzelburger, Jäckle, and König 2014, 132, 142), the beta-coefficient is .92, but the 95% confidence interval does not include 0. The serial autocorrelation of the dependent variable is thus substantial, but the dependent variable is not non-stationary. As a remedy, the regression models were specified to control correlated errors, not only across panels, but also over time (Prais–Winsten specification).

Concerning the independent variables used in the regressions, citizenship policy inclusiveness is measured with an index covering residence requirements, ius soli, dual citizenship policies, other naturalisation requirements and procedural safeguards (for details see Stadlmair 2017b), for which data were taken from MIPEX for the period of 2007 to 2014, complemented by own research for earlier years. The development level of welfare states is measured as the percentage of public welfare spending as share of the GDP, taking OECD data. Immigration rates are obtained from OECD data, taking the annual share of foreign citizen inflows per 1000 inhabitants. Furthermore, GDP per capita (OECD) and an index of economic globalisation (Dreher 2006) are included as controls in the regression models. Power resources of political parties are measured as voter shares (far-right party strength) and share of government members (left-wing government) using data from the Comparative Political Data Set (Armingeon et al. 2017). The coverage and descriptive statistics of the variables used for the regression analyses are available in Appendix 2.

4. Descriptive findings

Given that naturalisation tends to be a discretionary process, the requirements applicants have to meet in practice are sometimes not explicitly laid out in citizenship laws (Huddleston 2013). Of the nine countries covered here, Austria, Finland and Germany had explicit economic requirements throughout the period from 1985 to 2014, in Denmark such a requirement was introduced in 2006, and in Belgium in 2013. The concrete requirements vary across these countries, ranging from income requirements in Austria and Germany, a requirement not to draw welfare benefits in Denmark, an employment requirement in Belgium, to a simple declaration of the origins of one’s livelihood in Finland. Evidence of the economic situation is also required in Ireland, where naturalisation is a discretionary procedure. Even though there is no official requirement concerning economic status, applicants for naturalisation in Ireland must provide a documentation of their economic situation. In cases of a dependency on public funds or unemployment, the application may be rejected (Becker and Cosgrave 2013). Lastly, requirements for permanent residence may constitute an indirect economic requirement for naturalisation: In the United Kingdom, holding the Indefinite Leave to Remain permit is a necessary condition for naturalisation, for which many foreign residents need to comply with demanding income requirements (Gower 2012; Wray 2013). Here, access to citizenship is conditional upon resources through the back door of residence requirements.

Apart from these differences in the legal format, the difficulty of economic requirements is determined by their thresholds and duration. While there is no threshold in terms of income or welfare benefits imposed in the Belgian (since 2013) and Finnish requirements (since 2003), in Austria, Denmark and Germany, applicants may not draw means-tested welfare benefits or prove an income at a level precluding the reception of such benefits. Since 2009, the Austrian citizenship law stipulates a higher income requirement, where regular expenses – such as rent, alimony or mortgages – have to be deducted before an assessment of the income level, leading de facto to a higher threshold (Stern 2012). The duration ranges from a documentation of the current economic conduct when applying for naturalisation (FI) to an assessment of the applicant’s economic situation over several years, such as in Austria or Denmark.

In addition to the difficulty of the requirement, the level of the index depends on the capacity of citizenship authorities to waive the requirement in cases of undue social hardship. As such, the Austrian income requirement included an exemption, if the applicant was in an economically dependent situation through no fault of his/her own, until this provision was abolished in 2006. After a ruling of the constitutional court that the absence of such a provision would constitute a discrimination of disabled people, a narrower exemption, focusing specifically on applicants with disabilities, was introduced in 2013 (Stadlmair 2017a). A similar provision is also in place in Belgium, where disabled applicants do not need to meet the employment requirement. In Ireland and Germany, the assessment of the applicant’s economic situation is a discretionary decision and may also provide some leeway for exemptions to avoid hardship. In Denmark (since 2006) and Finland (until 2003), there were no provisions allowing for social exemptions from economic requirements for naturalisation.

Second generation immigrants typically have access to citizenship through procedures other than ordinary naturalisation or are explicitly exempted from economic requirements. Only in Austria and in Denmark, from 2006 to 2013, and in Germany until 1990, birth or socialisation in the respective country did not imply an exemption from economic requirements. Lastly, in Belgium since 2013, spouses of citizens were exempted from economic requirements, and in Denmark and Finland, other Nordic citizens may acquire citizenship without proving economic self-sufficiency. Based on the indicators on the level and scope of economic requirements for naturalisation, Figure 1 displays an aggregated index on the strength of economic requirements.

Figure 1. Economic criteria for naturalisation in nine European countries, 1985–2014.

The number of countries imposing economic requirements for naturalisation slightly increased in the past decade from four to six out of nine (Figure 1 and Appendix 3). Furthermore, the requirements often become more detailed over time, leading to complex legal provisions, such as in Austria, Belgium, Denmark and Germany. However, this development should not be interpreted just as a move towards stronger requirements, but also as a move towards less administrative discretion. While provisions in the 1990s often left many issues open to interpretation, current provisions on economic status as criterion for naturalisation are formulated in fairly specific ways (Stadlmair 2014). This observation fits a trend of legal entitlements to citizenship instead of discretionary decision-making (Bauböck and Perchinig 2006, 468). To some extent this implies more particularistic approaches towards citizenship acquisition, where different categories of applicants need to comply with different requirements.

Fees in the (ordinary) naturalisation process increased from 1985 to 2014 in every country except Belgium. Controlled for inflation and cross-national purchasing power inequalities, this trend softens, but the overall tendency remains the same: Naturalisation is becoming an increasingly costly process, with an average of EUR 653 in fees to be paid in 2014. However, this development occurred in the second half of the covered time frame (Figure 1): While average fees increased only slightly from 1985 to 2000 – especially when controlled for inflation – they have more than doubled since then, from EUR 307 to EUR 653 (Appendix 3). These changes cannot be attributed to inflation only rather they constitute substantial de facto increases in naturalisation fees. Furthermore, fees serve as a criterion for distinguishing more from less desirable applicants for naturalisation: Foreign residents of the second generation pay less, in some cases also citizens of certain kin-states. Spouses tended to be treated preferentially, but such provisions had been abolished in most cases by 2014.

While fees for civic integration programmes may be justified as measures to cover additional costs for the authorities from these programmes, naturalisation fees clearly have a regulative function (Grohs, Knill, and Tosun 2013): There is only very little evidence that changes in fees or their cross-national variation may correspond to the complexity of a naturalisation process: (1) There are discretionary procedures (i.e. in Austria and Ireland) as well as procedures involving a legal entitlement to naturalisation (i.e. in the Netherlands), where high fees are imposed. (2) Naturalisation fees tend to be rather strongly developed in contexts of otherwise inclusive citizenship regimes towards immigrants (Figure 2). (3) There are only very few cases in which fees were amended as a result of changes in naturalisation procedures: In Finland naturalisation fees were reduced in 2014 from EUR 440 to EUR 400 for applicants submitting their applications for naturalisation online. In Belgium the citizenship policy reform of 2000 brought a very simple naturalisation procedure and an abolishment of fees (Foblets and Loones 2006). In 2013, together with a more demanding naturalisation procedure, naturalisation fees were reintroduced. Much more frequent are changes in naturalisation fees which are conducted independently from a reform of naturalisation procedures, suggesting a rationale of fees as incremental tightening of access to citizenship (i.e. in the Netherlands and the United Kingdom), as an instrument to differentiate between more or less desirable applicants, or as a simple consequence of taxation reforms (Stadlmair 2014).

Figure 2. Main regression results for economic requirements, fees and the overall ECN index.

5. Explanatory findings

The first hypothesis stipulates that economic criteria for naturalisation are more strongly developed in a context of exclusive citizenship policies. In a basic regression model controlling for time (decades), there is no significant effect of citizenship policy inclusiveness on economic requirements and a positive significant effect on naturalisation fees (models 1–3, Appendix 4). Contrary to the expectations, naturalisation fees are even more strongly developed in inclusive citizenship regimes. This effect remains significant in full models including all independent variables (Figure 2). Thus, the first hypothesis must be rejected: Economic criteria for naturalisation are not stronger in exclusive citizenship regimes, naturalisation fees are even stronger in inclusive regimes.

Immigration rates only have an immediate statistically significant effect on naturalisation fees, but not on economic requirements and the overall economic criteria index. Since citizenship is a highly path-dependent policy area and becomes relevant for newly immigrated foreigners only after several years (residence requirements), it is more plausible to expect a delayed, long-term effect on policies, not immediate policy changes (see models 4–7, Appendix 4). When we lag immigration rates and take long-term averages (in Figure 2: 5-year averages, lagged by 3 years), they have a significant effect on all three dependent variables: Naturalisation fees and economic requirements for naturalisation are more strongly developed in contexts of high and long-term immigration, leading to a confirmation of hypothesis 2. This effect remains significant when including lagged naturalisation rates as independent variable, suggesting that not past naturalisations, but prospective naturalisations drive increases in economic criteria for naturalisation (see model 15, Appendix 4). However, since the variable on immigration rates includes all immigrant inflows, regardless of the countries of origin or types of immigration, there is a caveat to this finding: There is an effect of long-term immigration on economic criteria for naturalisation, but it remains unclear which immigrant inflows are of particular relevance. It is not possible to make more specific inferences due to the unavailability of reliable data of particular types of immigration in the selected period of analysis.

Different levels of public welfare spending do not have statistically significant effects on economic requirements or naturalisation fees, even when we lag the variable. Economic criteria for naturalisation thus do not correspond to how well developed the welfare system is, leading to a rejection of hypothesis 3.

Looking at the effects of power resources of political parties, left governments do not have the expected effect on economic criteria for naturalisation: There is no significant effect of left governments on any of the three dependent variables. Even if we lag the variable on left governments by 1–3 years, the effect remains insignificant. Rescaling the variable from an interval (percentage of left government members) to an ordinal scale that distinguishes between non-left, mixed and left governments leads to the same result. The same holds true for a counter-perspective: Right governments neither have a significant effect on economic requirements, nor on naturalisation fees. Hence, hypothesis 4 must be rejected.

In contrast to government composition, the strength of far-right parties does have an effect on economic criteria for naturalisation, but only in a long-term perspective: Far-right power, measured as their annual voter share, returns no significant effects on the dependent variables (model 8, Appendix 4). If their voter share is averaged over several years, the coefficients are significant (and positive) for naturalisation fees and for economic requirements, as well as for the overall economic criteria index (Figure 2 displays 10-year averages). I interpret these results as an effect of the long-term presence of far-right parties on policy output, while their short-term presence does not have a clear effect on the development of economic criteria for naturalisation. Hence, hypothesis 5a cannot be confirmed and hypothesis 5b can be confirmed.

GDP per capita has a significant and positive effect on the overall economic criteria index, but not on its two components: Neither naturalisation fees nor economic requirements for naturalisation are significantly more or less strongly developed over different levels of economic development. The effect on the overall index becomes insignificant when lagging GDP per capita for more than 3 years. For economic globalisation, there is a significant negative effect on economic requirements and the overall index, indicating that the economic requirements for naturalisation are more strongly developed in contexts of lower economic globalisation, but not naturalisation fees. This effect also becomes insignificant when the independent variable is lagged.

5.1. Interaction effects

From the repertoire of issues typically addressed by far-right parties, increasing immigration and welfare system change appear as plausible contexts in which far-right mobilisation may focus on the economic implications of immigration and so contribute to more strongly developed naturalisation criteria (Mudde 2013; Stadlmair 2017a). Figure 3 shows the conditional effect plots of regression models, including the respective interaction effects, and distinguishes between their long-term electoral strength (first row) and their current voter share (second row). While there is no (conditional) effect of current far-right power on economic criteria for naturalisation, the effect of long-term far-right party power varies over different levels of public welfare spending and immigration rates. For public welfare spending, the very modest differences do not point in the expected direction, but indicate a stronger effect of far-right parties in contexts of small welfare systems and a weaker effect in larger welfare systems. However, the differences are very small and the confidence intervals tend to overlap. Consequently, hypothesis 5c can be rejected.

Figure 3. Interaction effects of immigration and welfare spending with far-right power.

The interaction effect between long-term far-right power and immigration rates is more pronounced and points in the expected direction: The effect of far-right parties’ long-term strength on economic criteria for naturalisation is stronger in a context of high immigration and weaker in a context of low immigration. In a (hypothetical) context of a long-term immigration rate of zero, the predicted effect of far-right party power becomes insignificant. Thus hypothesis 5d, stipulating a conditional effect of immigration and far-right power can be confirmed when focusing on their long-term voter shares.

5.2. Robustness checks

In order to account for the time-series, cross-section structure of the data, the regression models were specified to control correlated errors across panels and over time (panel-corrected standard errors and Prais–Winsten specification). Despite these specifications, a serial correlation of residuals is evident in the residuals displayed in Figure 4. In the terminology of Wenzelburger, Jäckle, and König (2014, 133), this constitutes a non-dynamic model focusing on long-term trends instead of short-term policy changes.

Figure 4. Robustness checks.

The strongest outliers are observations in Belgium, Finland and Sweden. In the case of Finland from 1994 to 2002, the level of economic criteria for naturalisation is underestimated, the actual values are higher than those predicted in the regression model. In a context of little immigration and a rather weak far-right party at the time, we would expect very weakly developed economic criteria for naturalisation. In 2003, the former requirement not to draw welfare benefits was softened. Since then, applicants have only had to provide evidence of a legal source of income, including welfare benefits. According to Fagerlund and Brander (2013, 12), this reform was barely contested and passed by consensus of all major parties in parliament. In Belgium, there were no economic criteria for naturalisation in place from 2000 to 2012, which is in line with the Belgian government’s view of naturalisation as an instrument for integration and not its consequence (Foblets and Loones 2006). When the entire naturalisation procedure was reformed in 2012/2013, the government introduced a requirement to show participation in the formal economy as well as a naturalisation fees of EUR 150, which fits the estimates of the regression model. For Sweden from 2011, the regression model also predicts stronger economic criteria for naturalisation than those observed, since far-right parties are increasingly strong and there are steady immigrant inflows.

6. Discussion and conclusion

Many scholars noticed a recent shift towards economic status as primary condition for accessing many rights, such as family reunion, or various forms of a regularised residence status (Chauvin, Garcés-Mascareñas, and Kraler 2013; van Houdt, Suvarierol, and Schinkel 2011). However, a view on citizenship as something that must literally be earned is not entirely new: Requirements referring to economic resources were already present in naturalisation policies in the 1980s and amendments of these requirements are rare. They were introduced in Belgium, strengthened in Austria, Denmark and Ireland, and softened in Germany and Finland. At best, there is a modest convergence, but in Sweden, the Netherlands and the United Kingdom, direct economic requirements for naturalisation were absent throughout the period of 1985 to 2014. The development of fees in the naturalisation process is a different matter: They were amended more frequently, and add up to a substantial increase towards the end of the period. By 2014, four out of the nine countries imposed fees of more than EUR 1000 for ordinary naturalisation (AT, IE, NL, UK). The lowest fees in 2014 were imposed in Belgium (EUR 150). Furthermore, all countries except Ireland and Sweden require a formal language skill certification for naturalisation, which may lead to further expenses. Since the increases in naturalisation fees took place towards the end of the period of analysis, they may well be attributed to an agenda of austerity measures (van Houdt, Suvarierol, and Schinkel 2011; Morris 2016). However, this assumption cannot be tested empirically with a sample of nine countries, where austerity measures might not vary substantially. In a similar vein, recent studies on naturalisation took up the argument that in times of economic crisis, citizenship status may become more relevant also for intra-EU migrants, but the respective empirical findings are inconclusive (Alarian 2017; Graeber 2016). To conclude, contrary to suggestions of other studies on immigrant rights and economic resources, a trend towards ‘neoliberal communitarian citizenship’ (van Houdt, Suvarierol, and Schinkel 2011) is driven by the increase and spread of fees in the naturalisation process, not by formal economic requirements.

The sample of countries covered in this analysis raises the question how generalisable the findings are and of possible overestimation. The deliberate exclusion of Southern and Eastern European countries implies that the findings may not be transposed on other contexts indiscriminately. Given the (un-)availability of necessary and adequate data, as many years as possible were covered within the selected countries. The results reveal substantial variation for the explanatory factors and underline the validity of regression analyses. Nonetheless, the robustness checks show that some observations fit the explanatory findings better than others, such as the policy development in Belgium. The regression analyses suggest that there is no one pattern of institutional, structural and political contexts in which economic criteria are strongly developed. The results are equivocal for several reasons: First of all, fees and requirements may be functionally equivalent in terms of preventing non-wealthy immigrants from naturalisation, but they may have different origins in terms of their structural, institutional and political context. As such, naturalisation fees are particularly high in contexts of otherwise inclusive citizenship regimes, but economic requirements are not related to the inclusiveness or exclusiveness of citizenship regimes in general. Second, economic requirements are path-dependent policy instruments, where differences across countries are more pronounced than differences over time, which limits the explanatory power of time-series regression analyses. There are, however, two factors with a strong effect on economic requirements and on naturalisation fees: immigration and the long-term strength of far-right parties. Increasing immigration rates do not lead to immediate changes in citizenship policies. It takes some time until economic criteria for naturalisation are tightened, but then the effect of immigration is independent of the party spectrum of the government or citizenship regime.

Arguing that the presence and power of far-right parties may contribute to restrictive policies towards immigrants is not a new argument in citizenship and immigration policy research (Howard 2009), but was called into doubt by several recent studies. The main finding of these studies is that the mainstream right may also politicise immigration and take tough stances on immigrant rights (Bale 2008; Meyer and Rosenberger 2015; Sredanovic 2016), which suggests that the effect of far-right parties gaining power and pushing anti-immigrant policies may be overestimated. The findings presented above suggest that far-right parties contribute to a distinct focus on economic performance as criterion for naturalisation, but only when they have been strong over a longer period of time. Plausible reasons for this delayed, long-term effect may be their earlier organisational weakness, a cordon sanitaire by other parties and the long-term adaption of their policy preferences by other parties (Bale et al. 2010; Mudde 2013). There are three instances in this sample in which far-right parties participated in government or supported a minority government. In Austria and Denmark, they successfully transformed many of their citizenship policy preferences into legislation, which led to a tightening of economic criteria (Stadlmair 2014, 2017a). In the Netherlands, the PVV-supported government also introduced a bill including income requirements for naturalisation in spring 2012, but the bill did not pass until the government fell in autumn 2012 (van Oers, de Hart, and Groenendijk 2013, 42). Furthermore, the impact of far-right parties on the development of economic criteria for naturalisation is stronger in a context of high immigration levels. The factors immigration and far-right power may thus have some impact on economic criteria for naturalisation on their own, but their effect is much stronger when high levels of immigration and a strong far-right party are present.

Taking economic resources as a criterion for accessing political rights via citizenship has implications for the quality of democracy, which is measured inter alia by the inclusiveness of democratic polities (Blatter, Schmid, and Blättler 2017). In citizenship regimes that do not offer birthplace citizenship to their resident alien populations, economic criteria may substantively limit the inclusiveness of democracies. How such economic criteria affect the lives of immigrant populations and their inclusion in consolidated democracies should be subject to further research: There are some plausible indications that economic criteria influence naturalisation outcomes (Reichel and Perchinig 2015; Stadlmair 2017b) and there is a growing literature on the positive effects of naturalisation on labour market participation and socio-cultural integration (Ersanilli and Koopmans 2010; OECD 2011). How naturalisation matters in a context where access to citizenship is conditional upon economic status and which differences in life chances this entails, remains a task for further research, for which this study can provide a basis.

Acknowledgements

I would like to thank Hilkka Becker, Sampo Brander, Dimitris Christopoulos, Betty de Hart, Eva Ersbøll, Anuscheh Farahat, Alberto Martín-Pérez, Patrick Wautelet, and Helena Wray from the Global Citizenship Observatory for contributing their country expertise to the data of this study. Thanks also go to Djordje Sredanovic, Grete Brochmann, Sam Schmid and Jelena Džankić and the anonymous reviewers of the Journal of Contemporary European Studies for their comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the author.

Appendix 1.

Coding scheme and index construction

Table

Coding scheme for economic requirements for naturalisation
Indicator	Coding options
	(0)	(0.5)	(1)
Legal format	No requirement	Documentation required	Explicit requirement
Threshold	No threshold	No welfare benefits or income threshold at minimum standard of living	Income threshold beyond minimum standard of living
Duration	Only at naturalisation	One to three years	Three years or more
Exemption on social grounds	Explicit exemption	Discretionary decision	No exemptions
Second generation exempted	Exemption or ius soli		No exemption or ius soli
Spouses or kin-citizens exempted	Both are exempted	One category is exempted	No exemption

The six indicators on economic requirements for naturalisation are summarised to an index of economic requirements for naturalisation by taking their mean values, multiplied by 100. This index of their mean ranges from 0 to 100, with intervals of 8.3, the value of 100 indicating strongly developed economic requirements:

$$Ec.req.index=\frac{format+threshold+duration+soc.ex.+sec.gen.+spo.kin.}{6}\times 100$$

Table

Coding scheme for costs in the naturalisation process
Indicator	Coding options
	(0)	(0.5)	(1)
Level of costs for naturalisation (continuous variable)	No fee		Highest fee in sample (AT2011: 1314)
Second generation	No fee	Lower fee	Standard fee
Spouses	No fee	Lower fee	Standard fee
Kin-citizens	No fee	Lower fee	Standard fee
Language skills	No certificate required		Certificate required

Different costs in the naturalisation process (application, certificate and other costs) are summarised and multiplied with inflation rates and purchasing power parities. The total fees are transformed into a continuous variable ranging from 0 to 1, with 1 the highest amount found in the sample (Austria in 2011, value of 1314), which is labelled level of naturalisation fees. In a second step, exemptions or lower fees for the second generation, spouses and kin-citizens, and the indicator on language certificates are added:

$$Feeindex=\frac{7\times level+sec.gen.+0.5\times spo.+0.5\times kin.+lang.cert.}{10}$$

With the weighted additive index of naturalisation fees, the level of fees makes up 70% of the index, exemptions or lower fees 20%, and a necessity to submit a language certificate 10%. This weighting strategy is of necessity somewhat arbitrary, but aims to include some information on other costs and possible exemptions without levelling down the absolute amount of fees, which is the main interest here.

An overall index on economic criteria for naturalisation is obtained by taking the mean of the economic requirements and the naturalisation fee indices.

Appendix 2.

Independent variables

Table

Variable	Mean	Min	Max	SD	Observations (1990–2014)	Brief description
Index of economic requirements (ERN index)	23.9	0	100	27.8	225	See Appendix 1
Index of naturalisation fees (fee index)	38.7	0	90	24.0	225	See Appendix 1
Index of economic criteria (ECN index)	31.3	0	95	22.2	225	Mean of ERN index and fee index
Citizenship policy inclusiveness	58.4	44.8	79.2	8.8	223	Index of citizenship policies, covering ius soli, residence, civic integration and good conduct requirements and dual citizenship provisions; data and definitions from Stadlmair (2017b); missing data: BE 1990–1991
Immigration rate (5-year average)	6.4	0.8	19.0	3.3	217	5-year average immigration rate: inflows of foreign citizens / population in 1000s; own calculation based on OECD data; missing data: IE 1990–1997; net migration rate for AT 1990–1995
Naturalisation rate	3.6	0.4	11.4	2.2	213	Annual naturalisations/foreign resident population in 100s; Own calculation based on EUROSTAT data; missing data: AT 1990–1994, IE 1990–1996
Economic globalisation	85.4	60.8	97.3	6.5	225	KOF index of economic globalisation (Dreher 2006), version of 2017;
GDP per capita	36.8	21.2	48.6	5.8	225	GDP per capita in 1000 USD (OECD)
Public social spending	23.8	12.6	34.2	4.1	225	Public social spending in % of the GDP (OECD)
Left-wing government	38.0	0	100	34.1	225	Share of leftist cabinet members; data and definition from the Comparative Political Data Set (Armingeon et al. 2017)
Far-right votes (10-year average)	5.6	0	28.2	7.6	225	10-year average electoral share of far-right parties; author’s calculation; data and definition from the Comparative Political Data Set (Armingeon et al. 2017)

Appendix 3.

Descriptive findings

Table

Economic requirements for naturalisation in nine EU countries, 1985–2014
Country	Legal format	Type/threshold of requirement	Duration	Exemptions	Exempted categories
Austria	Explicit	Income (until 2006)	≤1 year (until 2006)	Yes (until 2006)	None
		=Minimum standard (2006–2009)		No (2006–2013)
			≥3 years (since 2006)	Yes (since 2013)
		>Minimum standard (since 2009)
Belgium	–	–	–	–	–
	Explicit (since 2013)	Economic activity	1–3 years	Yes	Spouses, 2nd generation
Denmark	–	–	–	–	–
	Explicit (since 2006)	No welfare benefits	≥3 years (2006–2013)	No	Kin-nationals (since 2006)
			1–3 years (since 2013)		2nd generation (since 2014)
Finland	Explicit	No welfare benefits (until 2003)	≤1 year	No (until 2003)	Kin-nationals, 2nd generation
		Legal income (since 2003)		Yes (since 2003)
Germany	Explicit	=Minimum standard	≤1 year	Discretion (until 2000)	None (until 1990)
					2nd generation (since 1990)
				Yes (since 2000)
Ireland	Documentation	–	≤1 year (until 2011)	Discretion	2nd generation
			≥3 years (since 2011)
Netherlands	–	–	–	–	–
Sweden	–	–	–	–	–
United Kingdom	–	–	–	–	–

Table

Fees in the naturalisation process in nine EU countries, 1985–2014
Country	Year	Application (EUR)	Certificate (EUR)	Other (EUR)	Language certificate	Overall (EUR)	Controlled for inflation & PPP		Main fee for 2nd generation
								Kin-nationals		Spouses
Austria	1985	29	509	58^a	no	596	1107	Standard	Standard	Lower
	2000	44	727	153^a		923	1215
	2014	110	977	150^a	yes	1237	1217	Lower	Lower
Belgium	1985	397			no	397	734	Standard	None	None
	2000					0	0	–	–	–
	2014	150			yes	150	147	Standard	Standard	Standard
Denmark	1985				no	0	0	–	–	–
	2000	134				134	176	None	None	Standard
	2014	134		93^c	yes	227	223
Finland	1985	26			no	26	47	Lower	Lower	Standard
	2000	336				336	391
	2014	400			yes	400	358
Germany	1985	51^b			no	51	75	Standard	Standard	Lower
	2000	256			yes	256	312		None	Standard
	2014	255		25^c		280	273
Ireland	1985		190		no	190	383	Standard	None	Lower
	2000		635			635	847
	2014	175	950			1125	1062			Standard
Netherlands	1985	136			no	136	237	Standard	Lower	Standard
	2000	227				227	300
	2014	829		230^c	yes	1059	1041
Sweden	1985	32^d			no	32	62	Standard	Lower	Standard
	2000	59				59	63
	2014	161				161	142
United Kingdom	1985	13	195		no	208	417	Lower	Lower	Lower
	2000	195				195	269
	2014	1073		169^ce	yes	1242	1255			Standard
Average	1985	76	99	6	0/9 yes	182	340	2/8 ≠standard	6/8 ≠standard	5/8 ≠standard
	2000	139	151	17	1/9 yes	307	397	3/8 ≠standard	7/8 ≠standard	3/8 ≠standard
	2014	365	214	74	7/9 yes	653	635	4/9 ≠standard	8/9 ≠standard	1/9 ≠standard

Notes: all amounts in EUR; ^avarying fees at sub-national level, amount for Vienna; ^bminimum fee depending on the applicant’s income, the fee can amount to EUR 2500; ^cfees for citizenship/integration test; ^destimate, taken from Janoski (2010); ^efee for citizenship ceremony.

Appendix 4.

Regression models

Table

Model	1	2	3	4	5	6	7
Dependent variable	ERN index	Fee index	ECN index	ECN index	ECN index	ECN index	ECN index
Citizenship policy inclusiveness	−0.236	0.389**	−0.088	−0.006	0.000	−0.002	0.010
	(.183)	(.121)	(.117)	(.126)	(.131)	(.138)	(.151)
Immigration rate				−0.093
				(.226)
Immigration rate (3-year average)					0.108
					(.308)
Immigration rate (5-year average)						0.502
						(.414)
Immigration rate (5-year average, 3-year lag)							2.044**
							(.448)
Left-wing government				−0.003	−0.002	−0.003	−0.003
				(.019)	(.020)	(.021)	(.023)
Far-right votes (10-year average)				1.917**	1.899**	1.885**	1.694**
				(.349)	(.359)	(.353)	(.382)
Economic globalisation				−0.403	−0.455*	−0.487*	−0.605*
				(.219)	(.225)	(.234)	(.251)
GDP per capita				0.379	0.434	0.461	0.755*
				(.318)	(.323)	(.337)	(.345)
Public social spending				−0.081	0.018	−0.041	−0.398
				(.351)	(.362)	(.378)	(.467)
Decade 1995–2004	0.712	0.791	1.064	−0.218	0.135	0.078	−1.505
	(2.993)	(2.399)	(2.121)	(2.092)	(2.184)	(2.245)	(2.180)
Decade 2005–2014	2.216	2.268	3.244	1.265	1.173	0.83	−1.786
	(4.231)	(3.268)	(2.834)	(2.882)	(2.909)	(2.983)	(2.805)
Constant	39.81*	11.02	32.67**	44.71	42.88	44.65	46.75
	(15.82)	(7.373)	(8.328)	(25.06)	(25.87)	(26.66)	(28.43)
R²	0.059	0.170	0.184	0.314	0.313	0.341	0.405
N	223	223	223	219	217	215	202

Note: Standard errors in brackets; *p < 0.05, **p < 0.01; panel-corrected standard errors and Prais–Winsten specification, regression models in bold are displayed in Figure 2.

Table

Model	8	9	10	13	14	15
Dependent variable	ECN index	ERN index	Fee index	ECN index	ECN index	ECN index
Citizenship policy inclusiveness	−0.124	−0.268	0.387**	0.003	0.000	0.031
	(.155)	(.232)	(.138)	(.151)	(.153)	(0.153)
Immigration rate (5-year average, 3-year lag)	1.744**	1.556*	2.304**	2.079**	1.997**	1.912**
	(.435)	(.700)	(.425)	(.467)	(.516)	(0.474)
Left-wing government	−0.009	0.009	−0.021	−0.003	0.000	−0.010
	(.0225)	(.037)	(.026)	(.023)	(.024)	(0.023)
Far-right votes	0.208
	(.148)
Far-right votes (10-year average)		1.703**	1.294**	3.076	1.843**	1.672**
		(.551)	(.400)	(2.912)	(.589)	(0.458)
Immigration rate (5 year av 3 year lag) X far-right votes (10 year av)					0.011
					(.067)
Public social spending X far-right votes (10 year av)				−0.0524
				(.111)
Economic globalisation	−0.489*	−0.951*	−0.0601	−0.605*	−0.677**	−0.435
	(.237)	(.417)	(.247)	(.251)	(.249)	(0.251)
GDP per capita	1.256**	1.294	0.145	0.732*	0.731*	0.720
	(.361)	(.742)	(.468)	(.347)	(.358)	(0.378)
Public social spending	0.337	−0.297	−0.642	−0.277	−0.404	−0.249
	(.483)	(.845)	(.511)	(.462)	(.489)	(0.480)
Lagged naturalisation rate						−0.546 (0.393)
Decade 1995–2004	−0.663	−1.444	−1.473	−1.476	−1.675	−0.633
	(2.189)	(3.833)	(2.518)	(2.192)	(2.184)	(1.912)
Decade 2005–2014	−1.147	−2.529	−1.458	−1.785	−2.101	−0.835
	(2.822)	(5.101)	(3.270)	(2.817)	(2.789)	(2.603)
Constant	15.22	66.8	12	44.79	54.43	28.64
	(25.48)	(52.4)	(30.97)	(28.31)	(28.13)	(29.74)
R²	0.349	0.301	0.247	0.407	0.533	0.337
N	202	202	202	202	202	198

Note: Standard errors in brackets; *p < 0.05, **p < 0.01; panel-corrected standard errors and Prais–Winsten specification, regression models in bold are displayed in Figure 2.