3.1 The internal conditions and its implementation

First, the spouse is, along with minor children, the person to whom family reunification is most obviously granted although the Directive does not define the meaning of spouse in Article 4. According to national legislation, the question of what constitutes a migrant family member is unclear because it is not the same for EU citizens and national families. Migration law requires a sponsor’s spouse to have reached a minimum age before he or she can join him or her (Table 1), and requires non-marital separation in legal terms to support their family reunification.

Another requirement related to the family reunification of spouses is the duration of the relationship, e.g. Cyprus requires a couple to have married at least one year before submission of the application. Moreover, a formal marriage does not mean that couples have formed a genuine relationship. A legal marriage contract is a symbol and not the substance of the relationship, and some marriages might be shams ( ^{Wray y Hutton, 2014, p. 209} ). Family migration must be based on a genuine and continuing marriage or partnership. Legal rigour is employed with respect to marriage dissolution in the case of second or subsequent marriages, which sets the economic rights to maintain the spouse, dependent relatives and common children in good faith. This obligation requires the observation of a carefully safeguarded system that prohibits polygamous marriages, governs forced marriage in compliance with a minimum age and controls for repudiation effects. In general, European states have adopted a prohibitive approach that seeks to impose limits on two separate levels: civil and criminal ( ^{Shah, Foblets & Rohe, 2014, p. 143} ; ^{Buchler, 2011} ). For instance, on the criminal level, polygamy and forced marriage are targeted through legislation that makes multiple marriage an offence bearing the potential to trigger penal sanctions. This approach is taken, but a brief glance at the evolution of domestic law and jurisprudence in Europe reveals differences and a lack of legal security in each case. For instance, as demonstrated in France in 1992, for example, the Cour d'Appel de Versailles refused social security benefits to the second wife of a Muslim husband and in 1988 the Cour d'Appel d’Aix-en-Provence similarly denied a Muslim woman her alimony on the ground that she was the second co-wife and that polygamy was considered contrary to “ordre public” ( ^{Shah et al., 2014, p. 32} ). Unless an Islamic ceremony was conducted in the spouses’ country of citizenship, the marriage is considered to have some legal validity under French law as long as it does not violate “ordre public” ( ^{Fournier, 2004, p. 15-26} ).

Similar to France, under German law the recognition of polygamous marriages means in practice that Muslim women can obtain social security benefits, such as inheritance, custody rights and child support payments ( ^{Shah, 2003, p. 382} ). Spain’s approach to polygamy in its immigration context has not been altogether different from German’s recent position and partial recognition is either granted in the context of claims for matrimonial relief or pension schemes in order to protect the interest of the widows. It has helped that in Spanish law each surviving divorced spouse is entitled to the deceased's pension in proportion to the respective periods of their marriage with the deceased. However, in 2003 the Supreme Court of Justice of Catalonia has held that only the first spouse should be assigned the claimed pension. In the United Kingdom, the traditionally liberal approach to the immigration of the wives of polygamists was considerably restricted by the introduction of the Immigration Act, 1988 ( ^{Shah, 2003, p. 383} ). This legislation imposed an effective ban on the admission of a wife when another wife or widow of the same man had already been admitted to the country ( ^{Shah, 2003, p. 391} ). Similar challenges have been brought against polygamy restrictions in immigration policies in the Netherlands, where a female resident in the country who acquired Dutch nationality and then married a man in Morocco who already had a wife was refused a divorce because her marriage was considered to be so connected to the Netherlands that it breached Dutch public policy and therefore could not be recognised ( ^{Campbell, Bala, Duvall-Antonacopoulos, Macrae, Paetsch, Bailey, Baines, Amani & Kaufman, 2005} ).

There is a similar approach to the recognition of divorces pronounced in other jurisdictions, provided that there are no objections to this on “ordre public” grounds. However, multicultural controversies arise when a number of legal systems in the Islamic world permit a husband to divorce his wife by simply pronouncing the word talaq three times. Courts in some European countries take into account the circumstances surrounding the repudiation and whether the divorce resulting from it is in fact unacceptable to the values of the country in question, while other jurisdictions simply note, irrespective of the circumstances of the individual case, that the institution of divorce by repudiation contravenes the principle of equality between the sexes or procedural principles ( ^{Kruninger, 2015} ). For instance, the French attitude implies denial of recognition of any repudiation-based divorce, whether or not it was established after legal proceedings in the case of sufficient proximity of the spouses involved with the French legal order. In the Netherlands, a prominent argument in the discussion is the principle of favor divortii and the interests of the wife in recognition of the repudiation, in order to prevent atypical legal relationships, which is the rationale behind the Dutch attitude towards repudiation-based divorces ( ^{Kruininger, 2015, p. 217} ). The revision to Belgian Law in 2005 provides against the recognition of Muslim talaq except in cases where there is no proximity to Belgium and the same law states that an act abroad recognizing the decision of the husband to dissolve the marriage may not be recognised in Belgium, unless the women benefits from the same right, effectively recognising talaq. Despite the ambivalence about the effects of talaq pronounced in non-European countries, Spanish and German courts now tend to recognize effects in a somewhat analogous manner through a kind of legal transfer to assimilate Islamic repudiation into divorce, but only if there is the possibility of divorce for women ( ^{Quiñones, 2013} ).

Unmarried and same-sex partners are grouped together in immigration rules and separated from married partners when the marital relationship is not simply a formal legal status. For instance, with respect to same-sex partners, different European Member States (e.g. Austria, Belgium, Denmark, Finland, Ireland, the Netherlands, Spain, Sweden and the UK) have decided to extend the right to family reunification to both registered and unmarried same-sex-partners, while three of them restrict the possibility to registered partnerships only, thus excluding unmarried partners in de facto cohabitation (e.g. Czech Republic, Germany, Luxembourg). Undoubtedly, the possible inclusion of this type of non-marital relationship is has been fixed but there are potential problems for the determination of common elements regarding the legal framework of de-facto union registration. Consequently unmarried partners (i.e. registered, non-registered or cohabiting partners, depending on the country) are considered family members. However, there are some countries that impose additional conditions with respect to the characteristics of a relationship (Table 1). In addition, European family reunification rules assume that it is not discriminatory to grant family reunification rights to the spouse of the sponsor without extending the same rights to the unmarried partner of the sponsor, even when the country of origin of the individuals concerned does not allow for two people of the same sex to marry.

Second, under the exclusion criteria for nuclear family configuration there is false scope regarding the status of children. Family admission seeks to assimilate an applicant’s dependents, including those born outside of marriage or a registered domestic union. Of course, this only refers to minors and they must be unmarried in order to be eligible for family reunification. Therefore, family reunification rules exclude older sons or daughters, regardless of whether they are dependent or not. There are no provisions to raise the age to 21 –as has been explicitly explained only in the case of the children of European nationals– but there are further restrictions as regards the maximum age of children and some countries establish different levels for minors: under 16 years; up to 18 years and older than 18 years in cases of extraordinary hardship. However, a more controversial question is the requirement to show evidence of family relationship and age, even implicitly through DNA testing ( ^{European Commission, 2008} ; ^{European Migration Network, 2018} ; ^{La Spina, 2012, p. 65} ; ^{McIntosh, 1988, p. 113} ). Indeed, as stipulated by the European Court of Human Rights, and as one may often recall, biological paternity does not in itself imply ipso iure the existence of a family relationship between parent and child, and instead is linked to other types of indications of family life. In short, since the Court of Strasbourg opted for the presumption of family life if there are legal or blood ties between parents and children, grandparents and grandchildren and sometimes between siblings, it seems clear that the family is not a purely biological construction if, moreover, there is no universal definition of family to which one can appeal ( ^{McIntosh, 1988, p. 104} ).

Nowadays, it is not enough for a child to be single and of the appropriate age in order to determine whether they can be a beneficiary of family reunification. The family relationship must be proved and, for instance, in order to renew a residence permit in Portugal, the sponsor must justify the need for the family reunification of adult children, and must certify compulsory schooling during the stay in the host society. Together with age, disability also allows children to apply for family reunification when they are objectively unable to attend to their own needs due to their condition. However, this legal requirement is more concerned with the idea of care, given the particular situation of dependence derived from the delicate condition of the child, and less with the actual family relationship. This is possible when they are not the applicant’s children, but the applicant is their legal guardian and holds representative powers that do not oppose the principles of law. In fact, there are basic problems with different legal and cultural adoptions, e.g. kafala, an institution in Islamic law that allows the legal adoption of a child by a person other than their biological parents. Many member States, such as Spain, do not consider kafala to be a real relationship or adoption in accordance with domestic rules ( ^{Pascouau, 2011, p. 168} ; ^{Böcker & Strik, 2011} ). However, if biological and adopted children are to be treated in the same way, it is reasonable to believe that this demand is a disproportionate form of discrimination in relation to the object of these different conditions for family reunification. Article 10.1 of the Convention on the Rights of the Child, referring to the obligations stipulated in Article 9.1 points out that “applications by a child or his or her parents to enter or leave a State Party for the purpose of family reunification shall be dealt with by States’ Parties in a positive, humane and expeditious manner”.

A third example of singularity and exclusion is the family reunification of ascendants. Because of the boundaries of the nuclear family, this kind of reunification is governed by a matter of dependence and kinship that creates a legal minefield in terms of migrant grandmothers/mothers. In fact, the legitimate discretion of the Member States following the guidelines of Directive 2003/86/EC enforces several strict limitations on age and degree of relationship, but in terms of the latter, the determinations generally coincide. In European immigration law, this condition explicitly operates in relation to the first-degree relatives of the applicant and their spouse, and is similar to the provisions for children of any domestic partners of the sponsor. An element of dependency is required and the immigration laws of destination countries set a minimum age for parents of 65 years (e.g. Czech Republic or Netherlands) or pensionable age, and alone or single (e.g. Netherlands, Estonia, Czech Republic and Italy). Meanwhile, families are dependent when they prove to have transferred funds or incurred the family’s welfare costs. In this case, this age requirement is a maximum age, in addition to the need for reasons for this family to enter. Thus, it is indirectly presumed that the family life of ascendants with their adult children is not common, so this type of extended family should be justified for humanitarian reasons, such as illness or disability, but immigration authorities could accept these reasons or not under terms of exceptional dependency. The possibility of sharing family life with ascendants of over sixty-five years is only allowed as a rare exception for economic reasons whereby the sponsor must prove sufficient financial income, as is the case in Spain and Italy, for instance ( ^{Bonizzoni, 2009, p. 93} ). Rarely, additional persons are admitted by a family reunification scheme, such as the sponsor’s siblings (Hungary) or a person who the sponsor has parental custody or care over (e.g. Hungary, Czech Republic, Portugal, Estonia, Finland).

These limitations greatly oppose the principle of non-discrimination because despite section 4.2 a) of Directive 2003/86/EC, Member States have the option of not authorizing the entry and residence of an applicant’s ascendant or spouse. The limitation on the exercise of family reunification for ascendants and additional relatives negatively affects the protection of family rights under Article 8 ECHR ( ^{Cholewinski, 2002} ), the Directive and, above all, the direct and indirect configuration of the model and the development of a family dynamic.

3.2 The external conditions and its implementation

Directive 2003/86/EC establishes a set of external conditions that Member States may require of applicants for family reunification, but with room for discretion. For instance, Article 7 of Directive allows Member States to require a sponsor to prove that he/she has appropriate accommodation; medical insurance; stable and regular resources; and complies with integration measures. Moreover, another sensitive question related to “dependent conception” is the access to employment for family members. The Article 14.2 of the directive leaves important margins of discretion to Member States in order to decide according to national law the conditions under which family members shall exercise an employed or self-employed activity.

First, the EU Directive requires the mandatory condition of a residence period, in general two years before the application. A non-lawfully residing third-country national being a sponsor for family reunification purposes often depends on their immigration status in the Member State. Austria, Denmark, Estonia, Germany and Latvia require a permanent residence permit in order to be eligible for family reunification. In most EU countries, however, it is enough for the potential sponsor to hold a temporary residence permit. Cyprus and France also require a temporary permit that is valid for at least one year. Other EU Member States grant different reunification rights according to the type of residence permit. For instance, reunification with parents or grandparents is not possible for persons that have no temporary residence permit and consequently only holders of a permanent residence permit are entitled to bring in family members in a broader sense. In general, holders of temporary residence permits can bring in members of their nuclear family. In addition to the legal status of the sponsor, Article 8 of Directive 2003/86/EU entitles Member States to establish a period of legal residence before the sponsor is entitled to be joined by his/her family. The term of legal residence required varies between Member States from 12 months (Ireland, Luxembourg, Slovenie and Netherland) or more but may not exceed two years (Cyprus, Hungary, Letona, Latvia, Malta, Poland) or three years in Austria ( ^{EMN, 2016, p. 27} ). In addition to the provision of adequate housing and stable resources or sufficient health insurance, quality housing is required, which is not generalized to all families in the host society and social welfare network.

In any case, legislators require financial inputs in the year following the date of the application. The sponsor has to declare possible changes in resources over the previous months. And they also require the family reunification applicant not to obtain this income from the healthcare or welfare system, although the income of other household members can be taken into account. The amount of resources defined by law or regulation is only a reference and therefore the immediate consequence of not meeting the income condition may not mean automatic rejection of the application. Moreover, “needs” can vary greatly depending on individuals, as in the Chakroun case, whereby a 60 year-old couple does not have the same needs as a 30 year-old couple ( ^{European Parliament v. Council of the European Union, 2006; Rhimou Chakroun, 2010} ). The Court has framed the margins of manoeuvre of member states. For instance, a yearly schooling allowance granted to the sponsor due to their level of resources may not lead to rejection. However, given the difficulty to achieve sufficient income under economic crisis, Portugal has adapted rules on family reunification when an applicant involuntary ends up unemployed and their level of resources is diminished. Finally, accommodation is a pre-condition for family reunification. This requirement can either be formulated very broadly as adequate accommodation, i.e. for the family or precisely expressed by the number of square metres required per family member or what is considered normal for a comparable family in the same region in order to meet general health and safety standards ( ^{Minister van Biuitenlandse Zaken c. K and A, 2015} ).

Second, although integration and fraud control seem to suggest very different problems, they are often interrelated and target the same immigrant groups. Both arguments lead to extended scrutiny of all migrant families, opening up questions as to how suited migrant family lives are to the dominant family norms ( ^{Mullaly, 2011} ). Suspicion of widespread fraud or false declarations of parenthood are often based on gender and ethnicity-based stereotypes, and on a lack of understanding of different family forms or even the outcomes of discrimination, which are simply used as arguments regarding the integration issue. For instance, Article 4, paragraph 5 of Directive 2003/86/EC states that “in order to better ensure integration and to prevent forced marriages, Member States may require the sponsor and his/her spouse to be of a minimum age, which is at least 21 years old, before the spouse is able to join him/her”. Accurate information must be obtained on the scale of fraud or forced marriages before introducing restrictive policy measures to combat these phenomena ( ^{La Spina, 2012, p. 39} ). In fact, a number of Member States are developing policies or amending legislation in order to better tackle the misuse of the right to family reunification, although an ^{EMN study (2012)} provides evidence that marriages of convenience do occur but it is not yet possible to fully quantify this across all Member States.

Third, on an EU level, the European Pact on Immigration and Asylum in 2008 and the Stockholm Programme accomplished the widespread transfer of nationalistic integration schemes into common EU immigration policy in what concerns, most crucially, the rule of law and non-discrimination (DOUE 2010/C 115/01). Since 2003, Northern and Central Member States have heavily encouraged a shift to civic integrationism or illiberal social policy in legal and political terms due to failed multiculturalism. Four member states (Austria in 2011, Germany and France in 2007 and the Netherlands in 2006) use these measures as a condition for admission into their territory, requiring family members to pass language tests, tests on the knowledge of the host society or to sign a contract obliging them to take civic and, if needed, language courses before entry in the territory of the member state.

In France, the integration process is pursued upon arrival. The content of integration measures may differ between these countries but their effects are similar in the sense that they play on the legal status of family members i.e. on his/her entry in the territory and/or his/her capacity to remain in the territory and the access to the labour market for reunited families during a year at most. For instance, a majority of Member States prefer submitting the members of the family to the requirement of a work permit: e.g. Austria, Belgium, Czech Republic, Estonia, France, Hungary, Ireland, Latvia, Poland, Romania, Spain, Sweden are in that case, this latter, Member State applies legislation prohibiting any kind of discrimination on the labour market. Hence, most Member States choose to limit or set more (Austria, Greece and Slovakia) or less strict conditions to the access to labour (Finland, Estonia, Lithuania), the flexibility of the directive frame letting them a great margin of action for that purpose.

However, other Member States require family members to undertake certain formal obligations only upon entry, such as taking integration (mainly language) courses ( ^{Böcker and Strik, 2011} ; ^{Strik, de Hart y Nissen, 2013} ), although this new civic integration paradigm presents inherent contradictions with regard to equal treatment, non-discrimination, effectiveness, proportionality and social inclusion in each regional and local area of Europe (III/IV Annual Immigration and Asylum Report COM (2012) 250 final; COM (2013) 422 final, COM (2014) 288 final). In the coming decade, the European Union will be increasingly influenced by the huge impact of the economic crisis, and mandatory integration programs to ensure selective immigration, restrictive control and access to state welfare arrangements will be implemented. However, in Southern European countries, these integration conditions/measures could become an unlimited “national” tool to control the non-national “inside” the nation-state and even abroad⁸. In this regard, Koopmans argues that Sweden, Belgium and the Netherlands, which have combined multicultural policies with a strong welfare state, have had relatively poor integration outcomes. And many countries that have either more restrictive or assimilationist integration policies (Germany, Austria, Switzerland, France) or a relatively austere welfare state (the United Kingdom) have achieved better integration results regarding employment, spatial segregation and incarceration ( ^{Koopmans, 2010, p. 20-25} ). Whereas in Belgium, the requirement is only applicable after arrival in the country and due to constitutional constraints, integration measures are only applicable in the Flemish region of the country.

No comparable tests have been conducted in Spain, Italy or Sweden. For example, in Sweden, there has been little civic conditioning in family migration policy, though this situation may currently be changing due to the large influx of asylum migrants ( ^{Bech et al., 2017, p. 3} ) or in Spain does not formally employ integration measures. After all, rules adopted in 2011 may pave the way to an increased role of integration issues in Spanish migration law. Hence, in order to renew a residence permit, and when other requirements are not fulfilled, the foreigner’s efforts to integrate may be evaluated ( ^{Pascouau, 2011} ; ^{EMN, 2016} ; ^{EMN, 2018} ).