Political participation as transformative reactive mobilization: a qualitative study of voter preferences among Turkish origin residents in the Netherlands

Introduction

A considerable academic literature in migration studies and political science seeks to identify the factors that determine whether persons with a migration background vote in national or local elections in the destination country, or stand as a candidate (see, for example, Martiniello, 2006; Morales and Pilati 2014; Finn, 2020; Soininen & Qvist, 2021). In that literature, immigrants’ interest in and turnout rates at destination country elections are typically treated as indicators of immigrant integration (see, for example, Wright & Bloemraad, 2012). High turnout levels, for example, are seen as evidence that ethnic minorities have acquired a considerable level of knowledge of, and access to, the domestic political system. Furthermore, the overrepresentation of highly educated persons with a migration background among voters and those standing as a candidate is generally seen as evidence for mature immigrant incorporation, which has been facilitated by integration policies, such as through the granting of local-level voting rights to non-citizens, or the offering of naturalization and dual citizenship opportunities. Indeed, it has always been argued in the Netherlands—the country on which the present study focuses—that immigrants’ interest in the domestic political system would increase once they had obtained voting rights (Van Heelsum et al., 2016). The interest to promote immigrants’ political participation in liberal democracies also stems from the institutional need to preserve the legitimacy of domestic political institutions, which requires democratic involvement of different social groups (Kymlicka, 1995).

The political participation among immigrant groups does not unequivocally indicate social inclusion, however, as it may also be a response to perceived social exclusion. Otjes and Krouwel’s (2019), for example, report that perceived discrimination in the Netherlands had the strongest effect on migrants’ voting behaviour among all items investigated. Similarly, Oskooii (2020) shows that increased ethnic discrimination is positively connected to voting behaviour in national elections in the UK in spite of such discrimination potentially reducing voting opportunities for groups with a migration background. Geese (2018) argues that German candidates of immigrant origin tend to receive a significant share of the ‘immigrant vote’ because voters of immigrant origin generally believe that such candidates have personal knowledge of migration-related disadvantage, and are better equipped to fight ethnic discrimination.

A relatively recent political phenomenon, which can be observed in the Netherlands in particular, requires us to pay more attention to how a complex mix of social inclusion and social exclusion may drive specific forms of political participation among immigrant groups: in 2017, votes from about half of the Dutch citizens of Turkish origin—especially from younger residents—helped the newly formed DENK party to gain three seats in Dutch Parliament, which contributed to a historic loss of the traditional party of preference of the Turkish Dutch, the Partij van de Arbeid labour party (PvdA): 20% of the PvdA voters in 2012, voted for DENK in 2017 (Holsteyn, 2018). DENK’s success contributed to an already fragmented Dutch parliament: the newly elected 150 parliamentarians represented no less than 13 different parties. Similar shifts in voting occurred at the local level. While more than 85% of voters of Turkish origin in Amsterdam voted for the PvdA in 2006, only 40% did in 2014 (Vermeulen, 2019). In the 2021 national elections, DENK consolidated its three parliamentary seats.

While DENK does not exclusively cater to residents of Turkish heritage—38% of the Moroccan Dutch also voted for the DENK in 2017 (Vermeulen et al., 2020)—it is strongly rooted in the Turkish immigrant group: a second-generation and a 1.5 generation Turkish immigrant, both members of Parliament for the PvdA between 2012 and 2014, founded the party because they disagreed with the labour party’s views on immigrant integration (a 1.5 generation immigrant immigrated before or during his/her early teens (Rumbaut, 2004). The founders then decided to also involve candidates from other larger minorities, especially the Moroccan-Dutch. Most citizens of Turkish or Moroccan origin are descendants or other family members of the labour migrants that were recruited from Mediterranean countries in the 1960s and early 1970s (Kulu-Glasgow & Leerkes, 2013). While both groups mostly consist of Muslims, and also share a similar migration history, the Moroccan-Dutch are generally seen as a more fragmented, unorganized immigrant group than the Turkish-Dutch (cf. Fennema & Tillie, 1999). In other European countries, such as in Germany,¹ separate political parties have similarly been founded among immigrant groups with Turkish roots, but these parties have, as of yet, not been as successful as DENK in the Netherlands.

The lower electoral threshold in the Netherlands (0.7%) than in Germany (5%) is likely to partly explain DENK’s success. However, it can also be hypothesized that stronger anti-Muslim discourses, as well as related changes in political debates and integration policies in the Netherlands, have alienated groups of Turkish origin from mainstream Dutch parties (also see Vermeulen, 2020). In the Netherlands, a commitment to multiculturalist policies in the 1980s and 1990s was followed by a more assimilationist, or even exclusionist, turn in the 2000s and 2010s. Under Dutch multiculturalism, the Dutch state sought to promote immigrant integration via relatively liberal naturalisation policies and through migrant organizations, which were incorporated in national and local consultation structures. These policies indeed contributed to a steady increase in political participation. In Rotterdam, for example, the turnout among Turkish immigrants in local elections was only 28% in 1994, increasing to 56% in 2006 (Van Heelsum et al., 2016). At the same time, the policies also fostered latent opportunities for group-based organisation outside of mainstream political parties, while the later assimilationist turn possibly contributed to a desire to make use of these opportunities.

In what follows, we shed more theoretical and empirical light on processes of group-based political participation outside of established political parties—or what we will call ‘transformative reactive mobilisation’, drawing upon Portes and Rumbaut’s (2001) theory of reactive ethnicity, and recent additions on reactive religiosity. We theoretically proliferate these theories by applying them to voting behaviour, an empirical phenomenon that they have not, to our knowledge, been applied to yet.² Additionally, we theoretically elaborate these theories by pointing out that reactive mobilisation processes do not merely conserve or re-invent past ethnic identities. They are also future oriented and transformative: the mobilisation of identities around a shared ancestry is also aimed at improving the immigrant group’s societal position, and has the potential to both reinforce and alter ethnic identities, such as by inserting the ethnic group, as a group, more strongly in the polity of the destination country, and by possibly fostering a, to some extent, a transethnic ‘Muslim immigrant’ identity. Against this backdrop, we aim to answer the following research question: How can theories of reactive ethnicity help us understand DENK’s popularity among Dutch citizens with a Turkish background in particular, including those who are born and raised in the Netherlands? The analysis aims to identify the main theoretical mechanisms through which, and the conditions under which, reactive mobilisation processes seem to occur.

For three main reasons, the Turkish-Dutch constitute a strategic group for a study on transformative reactive mobilization. First, the immigrant group experiences a considerable amount of ethnic and/or religious discrimination (Maliepaard et al., 2015) and has been disengaging from traditional parties more than other immigrant groups (Bahçeli, 2018; Vermeulen et al., 2020). Second, the assimilationist turn has been more pronounced in the Netherlands than elsewhere (Bloemraad & Wright, 2014). Third, after Germany and France, the Netherlands hosts Europe’s largest Turkish origin community that, with over 400,000 people, constitutes the largest ethnic minority in the Netherlands.³ Rotterdam, where most of the fieldwork was carried out, is an especially strategic site as the assimilationist turn was more marked there than in other Dutch cities, such as Amsterdam (Scholten, 2013), while the municipality also has the largest number of residents of first- or second-generation Turkish origin in the Netherlands (about 50,000), as well as a relatively well-developed ethnic-religious diaspora infrastructure (Phalet and Ter Val, 2004). Indeed, one of the DENK founders is a former Rotterdam city council member. Transformative reactive mobilization processes, or elements thereof, may also occur in other European countries, Dutch cities, and immigrant groups, but our case selection allows us to see the phenomenon especially clearly. In the discussion, we will reflect on the findings’ generalizability for other contexts and groups.

We do not claim to offer a full explanation of political representation of the immigrant group outside of mainstream political parties. It has been hypothesized, for example, that transnational Turkish diaspora networks, which are partially under the influence of the Turkish state, have also contributed to DENK’s success (Otjes & Krouwel, 2019). Indeed, the Turkish Directorate of Religious Affairs allowed DENK to promote itself in Dutch mosques receiving Turkish state funding. The appearance of the DENK leaders during a Rotterdam rally against the 2016 coup attempt in Turkey may also indicate links with the Turkish state (Bahçeli, 2018). Indeed, by facilitating dual citizenship and diaspora organizations, the Dutch multicultural policies also assisted the mutual engagement between migrants and the polity of the sending countries (Kymlicka, 1995; Ostergaard, 2003). Political parties in Turkey still closely engage with Turkish-origin diaspora with a view to making external Turkish citizens cast their votes in Turkish elections (Mügge et al., 2019). While we do not deny that Turkish diaspora policies may partially explain DENK’s success, we highlight the more domestic aspects of the reactive mobilisation process which are rooted in destination country experiences. It should also be mentioned that DENK, as far as is known, does not receive funding from the Turkish state—in fact, the party supported a 2022 bill that forbids financial gifts from outside of the European Union.

Political mobilization, integration policies and reactive identities

Migrants’ mobilization revolves around obtaining more political, social and economic rights in receiving states, and is determined by both transnational and domestic forces (cf. Koopmans & Statham, 2001; Ostergaard, 2003; Wright & Bloemraad, 2012). Sending countries may assist their citizens in improving their status abroad, and migrants’ incorporation trajectories and receiving countries’ integration policies both codetermine immigrants’ desire to participate politically, and their capabilities. National differences in naturalization requirements, for example, imply differences in voting rights: while countries such as Germany set up high barriers to naturalization, other countries, including the Netherlands, followed a more open-inclusive citizenship regime and granted migrants more opportunities for political claims making via internal and external voting rights, and through group-based consultation structures (Koopmans & Statham, 2001). The higher rate of domestic political participation among second-generation migrants compared to their parents confirms the importance of incorporation for domestic political participation (Mügge et al., 2019; Quintelier, 2009). The second generation tends to be better educated, speaks the destination country language better, and is generally more politically socialized (Ten Otjes & Krouwel, 2019; Teije et al., 2013). Most scholars therefore see immigrants’ knowledge of, and interest in, receiving countries’ political systems as both a consequence and cause of successful integration. It indicates, and reinforces, a sense of membership, belonging and trust in the polity (Huddleston, 2009; Mügge et al., 2019; Wright & Bloemraad, 2012).

Such a view nonetheless risks overlooking important complexities, since immigrants may also desire to participate politically when they experience social exclusion. Interpreting the recent mobilization of the Turkish-Dutch through a ‘reactive ethnicity’ lens (Portes & Rumbaut, 2001) may rightfully add more complexity. In their study on second-generation ethnic identities, Portes and Rumbaut (2001) found that ethnic identities were often paradoxically strengthened in the second generation, when strong expectations to equal social status were frustrated because of perceived discrimination and persistent social disadvantage. When faced with several hardships in achieving upward social mobility and equal social status in the country of destination—ambitions that are common among immigrants orienting themselves to permanent settlement-, persons with a migration background may come to reject the norms and values of the receiving community and acquire a stronger connection with their culture of origin. The second-generation may then develop a desire to assemble around race and ethnicity, display solidarity through daily interactions and activities, and strategically mobilize relations with the majority society in order to protect group interests (Maliepaard et al., 2015).

The notion of reactive ethnicity is in line with earlier insights from American sociology that assimilation—the empirically well-documented phenomenon that immigrant and non-immigrant groups tend to become more similar over time, usually mostly as a result of changes on the part of the immigrant group—may, in some cases, follow a ‘bumpy-line’, rather than a ‘straight line’ (Gans, 1992). In the European context, too, there is some evidence for a ‘paradox of integration’, where those with higher educational levels—e.g., who score higher on indicators of ‘structural integration’—perceive less respect for minorities, and more discrimination (De Vroome et al., 2014).

Recent research has theoretically proliferated the concept of reactive ethnicity by arguing that perceived discrimination may similarly trigger processes of ‘reactive religiosity’, where not just the ethnic identity as such is reinforced, but also its associated religion (Herda, 2018; Maliepaard et al., 2015; Torrekens & Jacobs, [39]). For example, Maliepaard et al. (2015) report that Dutch Muslims, who perceived the most discrimination, were more likely to frequent mosques, which the authors interpret as a response to a perceived negative climate in the Netherlands.

Various studies support the reactive ethnicity hypothesis: perceived discrimination, economic and social marginalization, lack of esteem, denial of social acceptance or group identities increase ethnic identification and consciousness (Platt, 2014; Simonsen, 2021; Stone & Meekyung, 2005). Verkuyten and Yildiz (2007) show that prejudice towards Islam or non-Western minorities among native Dutch culminated in a stronger Turkish and weaker Dutch identity. Çelik (2015), too, documents how Turkish-origin youth were more likely to reject the idea of having German citizenship when perceiving discrimination. The empirical base for religious religiosity hypothesis is more limited: the correlation between religiosity and perceived discrimination could also indicate that strongly religious Muslim Turks simply experience more discrimination than Turks who are less religious. In our view, perceived discrimination does not necessarily increase religiosity, but may indeed give it a more politicized, ‘reactive’ meaning. In case of perceived discrimination, it may begin to symbolize the threatened yet valued minority identity.

By extension, we cannot simply posit an unequivocal positive relationship between inclusive integration policies and immigrant’ political mobilization. As we will show, the Dutch turn to more assimilationist and exclusive policies contributed to a desire among the Turkish-Dutch to engage in the policy process outside of established parties, while the capacities of the Turkish-Dutch were indeed fostered by more inclusive policies, now and in the past. A move away from multiculturalism, which occurred in the second half of the 1990 and the 2000s, triggered oppositional desires, but the legacy of multiculturalism partially explains why political entrepreneurs like the DENK founders could mobilize the immigrant group outside of existing political parties. The next section therefore discusses Turkish immigration to the Netherlands in the context of changing integration discourses and policies, and gives more information about the characteristics of the DENK party.

Turkish immigration, Dutch integration policies, and DENK characteristics

Partially recruited through temporary labour employment agreements, Turks began emigrating to the Netherlands in the early 1960s. After the economic crisis in the 1970s, when recruitment ended, most ‘guest workers’ settled permanently, while immigration continued, mostly through family reunification and formation channels (Kulu-Glasgow & Leerkes, 2013). When unemployment soured in the 1980s, tensions between local Dutch populations and newcomers increased, and racism and discrimination became hot topics. At that time, a 'Minority Policy', the country’s first comprehensive integration policy, was introduced, which argued for ‘integration with preservation of culture’. Ethnic minorities received various cultural rights, and it was attempted to incorporate minority elites into the policy process via migrant organizations. The government also opened up civil service positions, introduced local voting rights, and aimed to reduce barriers to naturalization, so as to provide a space to migrants within existing political structures.

The Dutch tradition of 'pillarization' partially inspired the multicultural policies. In the early twentieth century, each religious group was granted some autonomy, and was promoted to develop separate ‘pillars’, e.g. institutions in areas such as schooling and medical care (cf. Uitermark et al., 2005). By 1997, when the Dutch liberal policies with regard to dual citizenship had become stricter, almost half of all Dutch residents with a first- or second-generation Turkish migration background had acquired Dutch citizenship, usually next to Turkish citizenship. Twenty years later, during the 2017 elections, 82% of the immigrant group was Dutch (while dual nationality is no longer promoted, there are various exceptions that still make it possible).

The multicultural approach offered immigrants a beneficial environment for claims-making and institutional trust, and the Turkish immigrant group even stands out: it organized itself more than other immigrant groups, had relatively high turnout rates in Dutch elections (Fennema & Tillie, 1999), and developed stronger identifications with the host country than, for instance, the Turks in Germany (Ersanilli & Koopmans, 2010).

In the second half of the 1990s, the political discourse on migration and integration in countries such as Germany, France, England and the Netherlands became more restrictive and assimilationist, with even some exclusionist undertones, especially in public debates. The changing discourse mostly occurred because of dissatisfaction with slow improvements in immigrants’ social positioning and persistent segregation, which were blamed on immigrants (Scholten, 2011), and, in some readings, a concern about Islam in relation to national identity and security, a discourse that became stronger following the 9–11 attacks and the ISIS insurgency in Iraq and Syria.

In the 2000s and 2010s, the multicultural approach made way for a more duty-based, individual-level approach to integration: the national government abolished group-based consultation structures, and made residence rights conditional on civic integration exams. Permanent residence status and naturalization became conditional on passing a civic integration exam in the Netherlands; admission to the Netherlands via the family formation channel became conditional on passing a civic integration exam abroad. The Netherlands also introduced various other restrictive measures for family reunification within limits sets by international law, with the explicit aim to limit continued immigration from the former guest-worker countries Turkey and Morocco (Kulu-Glasgow & Leerkes, 2013). While it is still widely believed that newcomers should be able to obtain similar rights as natural citizens, it is now argued that residence rights should be conditioned on proven progress in civic integration, and that group-based structures hampered rather than facilitated integration.

Although multiculturalism was discarded in public discourse, it would be incorrect to say that multicultural policies were fully abandoned, or that multicultural legacies did not create path dependencies persisting until today. For example, immigrant organizations, including religious organizations, remained influential among the Turkish-Dutch—if only because of the constitutional right of association and freedom of religion. Additionally, local authorities in various cities pragmatically preserved certain multiculturalist policies such as by funding ethnic projects under the banner of ‘diversity policy’ (Uitermark et al., 2005). The restrictive turn in admission policies, too, was stronger rhetorically than practically: for example, the Turks were actually exempted from the civic integration exam abroad in 2011 when the Dutch Administrative High Court ruled that the requirement violated the Association Agreement between the EU and Turkey (cf. Kulu-Glasgow & Leerkes, 2013). In spite of these nuances, the tone was set, and a considerable part of the Turkish Dutch increasingly detached itself from the mainstream due to changes in discourse and policies surrounding immigration and Islam (cf. Bahçeli, 2018).

As a result, a considerable part of the Turkish-Dutch reinvented its Turkish national and religious identities. The interethnic (and international) tensions peeked in 2017 when the Turkish-Dutch—once considered the exemplary case of ‘integration with preservation of culture’—organized a massive local protest when the Dutch authorities denied a Turkish minister entry to a political rally in the framework of the Turkish national elections. Twelve demonstrators were arrested, and huge polemics followed, in which Dutch politicians seized the protest as proof of yet another integration failure, while the Turkish-Dutch defended the protests as a reaction to a discriminative environment (Bahçeli, 2018).

DENK is generally not seen as a one-issue party, although it does present itself strongly on theme of discrimination. It’s 2015 election manifesto mentions 5 five main positions: The Netherlands should be (1) ‘tolerant’, (2) ‘social’ (caring), (3) ‘learning’ (innovative), using ‘everyone's talent’, (4) ‘sustainable’ and (5) ‘just’, aimed at promoting ‘international justice’. For example, DENK strives for a national programme to fight Islamophobia and hatred against Muslims, and wants the government to hire specialized police offers against ethnic discrimination. Its financial resources mainly come from regular Dutch governmental subsidies for political parties, membership fees (at the time of writing it has about 3700 members), and some private gifts. Dutch political parties are legally required to report all gifts of at least €4500, and it should be noted that DENK has not reported having received any gifts from outside of the Netherlands. There are no institutionalised links between DENK and the Turkish-Dutch ‘ethnic infrastructure’, but various centrally positioned members, such as the present board members, are owners of successful Turkish-Dutch companies or board members of local Turkish-Dutch mosques.

Research has shown that younger citizens (in 2021, the median age of the DENK voters was 38), citizens of ‘non-Western’ immigrant origin—especially from Turkey and Morocco -, and, less so, female citizens are overrepresented among the DENK voters, while no significant correlations seem to exist with educational level.⁴ Interestingly, DENK is not particularly popular among other larger minorities experiencing ethnic discrimination, such as the ‘Surinamese’, ‘Antilleans’ and ‘Sub-Sahara Africans’, even if there have also been tendencies toward political representation outside of mainstream political parties for these groups.⁵ All in all, these patterns confirm that DENK is relatively popular among second-generation immigrants of Turkish origin in particular, followed by those of Moroccan origin, including the upwardly mobile in these groups (which also include a relatively high number of women).

Research design

The first author is a Turkish national who originates from a conservative Turkish family. As an adult he has lived and worked in various European countries as a spouse in an ‘ethnically mixed’ marriage. The second author is a male Dutch national without a migration background who has conducted ample research on migration policies and the position of immigrant groups in the Netherlands. Arguably, the positionality of the authors as partly insider, partly outsider helped them to get sufficiently ‘involved’ with the research setting—i.e., to get access to the Turkish diaspora organizations and to make sense of the participants’ perspectives—while also helping them to ‘distance’ themselves from the empirical observations, by standing back and interpreting the findings from a theoretical angle (cf. Rojek, 2014).

The first author collected the data during extensive fieldwork between May 2019 and February 2020. It focused on the Delfshaven and Bospolder-Tussendijken neighbourhood in Rotterdam, where between 10 and 15% of the residents are of Turkish origin. The fieldwork focused on perceived discrimination, politics in the Netherlands and Turkey, attitudes about DENK, the 2017 Rotterdam protests, and their interrelationships. On every occasion, the researcher openly exposed his identity, his research, and his affiliation to Erasmus University Rotterdam. He attended prayers in local Turkish Mosques or teahouses where Turks hang out, and dropped by at ‘Turkish’ shops. He regularly visited a local migrant association, the Union of International Democrats (UID), which represents the Sunni-conservative Turkish population and is closely related to DENK, in order to understand migrants’ search for security and solidarity through such organizations.

The author also regularly visited the Utrecht Turk Culture Centre (UTCC), which openly represents nationalist Turks and functions as a political branch of the Turkish Nationalist Party (MHP). Arguably, UID is the most influential Turkish diaspora organization in the Netherlands. It has been criticized for influencing external Turkish voters by organizing events and rallies for Turkish politicians, including the controversial 2017 rally (Frymark, 2017). The data collected include expert interviews, group interviews, informal conservations, and qualitative observations. In total, 8 months of participant observation was carried out in Rotterdam and 2 months in Utrecht.

After the researcher had obtained sufficient research access, 21 more formal face-to-face interviews were carried out with DENK voters and people who were considering voting for DENK in the next elections (see Table 1). Participants, 16 men and 5 women, were between 18 and 73 years old, and had different educational levels: 5 university graduates, 3 university students, 3 with pre-college secondary school degrees, and the rest with primary or other secondary school background. Most respondents were second-generation immigrants (14), 6 were first-generation, 1 was third-generation. All interviews lasted between 15 and 90 min and were conducted in Turkish. These respondents were recruited through UID or UTCC and most interviews (16) were conducted inside the premises of these organizations. To aid with recruitment, and to obtain a better understanding of how perceived discrimination influences political participation, it was also said that we were interested in the 2017 Rotterdam demonstration, which most respondents (17) had joined. Unstructured, unrecorded interviews were used since respondents were uncomfortable participating in ‘official’, recorded interviews. Our interviews thus resembled everyday conversations, allowing participants to open up and share thoughts and feelings.

Table 1 Respondent characteristics (with fictive names)

Although the first author spent quite some time in the two organizations during participant observation, various interviewees were in doubt whether they wanted to be interviewed and what they would say, as they feared that the answers might be used against them. Sharing information about the author’s conservative family background and his affiliation with the university helped to promote trust and rapport; for example, various participants mentioned that they were proud of the first author for succeeding at this academic level. Participants who were initially uncomfortable to speak eventually started to go more deeply into their experiences and became more forthcoming about their political views.

The data were analysed using concepts that inductively emerged from the interview data, and more deductively derived sensitizing concepts from the literature on reactive ethnicity, which was used because of indications that perceived discrimination explain immigrants’ voting behaviour in particular (Otjes & Krouwel’s, 2019) and that the move away from multiculturalism has alienated immigrant groups from mainstream political parties (Vermeulen, 2019). It turned out that various inductively emerging themes (e.g., ‘discrimination’, ‘religion’, ‘values’) could indeed be usefully interpreted using the reactive ethnicity framework. Other emerging themes were more indicative of collective action (e.g., ‘unity’, ‘joint Turkish action’, ‘power’), which led us to speak of ‘reactive mobilisation’. As a final step, we realized that we also needed a term to highlight that reactive mobilisation processes do not merely strengthen pre-existing ethnic and religious identities; they also seem to have relevant transformative aspects by potentially altering the destination country context in which ethnic and religious identities are embedded, and by potentially also changing these identities themselves, such as by creating new group links to the polity of the destination country and by contributing to specific transethnic ‘immigrant-origin’ or Muslim minority identities.

Results

A visualized summary of our main results is offered as a conceptual model in Fig. 1, which will be explained and illustrated in the present section. We intend to present the figure as an outline of the theory of transformative reactive mobilisation, which future research can elaborate and test further. Reactive mobilisation consists of two main mechanisms: (1) the formation of reactive identities under the influence of perceived social exclusion and interethnic tensions in the destination country, and (2) the translation of these identities into reactive mobilisation. The process seems to occur under specific conditions, both regarding the genesis of reactive identities and their translation into group-based political participation. For both ‘steps’ a relatively mature, yet hampered degree of incorporation of the immigrant group seems to be required. Such incorporation increases the likelihood that exclusion is perceived as discrimination, and facilitates the translation of reactive identities into (group-based) political mobilization. The presence of ‘established’ diaspora organizations strengthens reactive identities but also enhances opportunities for the immigrant group to mobilize politically outside of established political parties.

Fig. 1

The transformative reactive mobilization framework

Perceived discrimination

All participants whom we talked with reported perceived discrimination in the Netherlands, and in Europe more generally, in relation to their ancestry and religion. The generally believed that Muslims are now much more seen as a threat to Dutch culture and national security than before, and various respondents even claimed that European states are deliberately trying to eradicate, or marginalize, Islam. The complaints centred on Dutch politicians and the police; the attitudes towards the general Dutch population were less negative or even positive.

The heightened sense of discrimination seemed to be a relatively recent instinct. One older first-generation immigrant, for instance, explained that he remembered that it wasn’t a big deal in the Netherlands if you were Muslim. Incidents like 9/11, the London and Madrid bombings, the ISIS insurgencies, and the political use of these incidents, were felt to have contributed to an assumed connection between Islam and terror. Participants noted that right-wing politicians openly try to create tensions between assumed Dutch and Islamic values. As Turks represent the country’s biggest Muslim population, they were bound to become implicated in these tensions.

In the beginning, they [the native Dutch] didn’t even care if you are religious. We were praying in the streets. But everything changed in the last 15 years. It is like, why are you still religious when religion is so dangerous. Especially politicians, they are looking for excuses to blame you for your religion (Ahmet, mechanic, first-generation).

We asked participants how these changing perceptions of Muslims had influenced their lives, leading to different answers that were mostly connected with rights and perceived life chances. Some participants complained about increased police surveillance in neighbourhoods; others felt deprived from their rights as Turks because of the stricter policies with regard to dual citizenship; some claimed that recent policies, such as stricter admission and naturalization policies had created unsurpassable barriers for migrant populations.

They do not see any more that different cultures are good; like Turkish wedding is good or Turkish food is good. They believe that different culture is dangerous. They want everyone to follow their life style (Sevim, secretary, second- generation).

The participants also complained about labour market discrimination. For instance, in an informal chat with five Rotterdammers, one person mentioned that he might advance in his job because someone had left a position. The others immediately responded, ‘as a Turk, he would never get it’. Such affirmation by others perceiving similar social hostilities notably served to prompt for ethnic identity. Similarly, a young participant shared a story on how his teacher questioned him about his leisure activities in class, which he thought was extremely disrespectful and discriminatory. A bus driver claimed that he found himself being scheduled for early hours all the time. Others heard these stories as the room was small, and the research attracted some attention, leading people to share similar stories, which reminded them of the allegedly negative environment, and created a stronger sense of a shared social identity.

While these observations supported the reactive ethnicity framework, a few young Turks experienced some measure of discrimination, but still seemed to be oriented to mainstream assimilation. They contended that some measure of discrimination is ‘inevitable’, and tried to adopt ‘a positive attitude’. For instance, Fatih, a second-generation university student who works in his dad’s store, claimed that native Dutch people do accept Turks, but that you have to be ‘hardworking’, ‘have good intentions’, and should ‘not mix religion with work’: ‘If you do not show any sign of religious behaviour, they [the native Dutch] are fine’. Others, especially older migrants, countered that such strategies would not protect him from discrimination. For example, Mustafa, 54, who was temporarily unemployed, explained that his son ‘doesn’t pray’, ‘speaks perfect Dutch’, and ‘even goes on holiday to Greece not Turkey’, yet still encountered discrimination: he was convinced that his son, still in middle management, should have moved into a better position in the city hall by now.

All respondents strongly believed that discrimination had become worse after 2017. In that year, about 1000 Turks protested in Rotterdam because the Dutch authorities had denied a Turkish minister entry by air who wanted to campaign for a Turkish referendum, and then expelled a substitute Turkish minister who came from Germany by car in spite of the Dutch ban. While the protests were peacefully at first, the police arrested 12 protesters because of demolition and public violence.

The respondents claimed that increased interethnic and institutional tensions made it harder to find a job or to get in touch with Dutch authorities such as the police. They specifically commented that Turks now receive the lowest level of recognition in the country: If they understand you are a Turk, for example from the headscarf in the picture, they will put your CV at the bottom of the others (Halit, first-generation, retired).

Whereas all participants experienced ethnic and religious discrimination, they perceived different measures of it, and held different beliefs on how to respond to it. As was mentioned, some participants accepted it; others fought against it; still others, especially first-generation migrants, believed that their issues would be solved once they returned to Turkey. The second-generation immigrants in particular believed that becoming politically strong by voting in Dutch elections, was a desirable response to the increased securitization of Islam and the more assimilationist discourses and policies.

Formation of reactive and religious ethnicity as political behaviour

Relatively settled minority groups are more prone to developing reactive identities than recent arrivals (Maliepaard et al., 2015: 2638), and the Turkish-Dutch are no exception. They represent an upwardly mobile, yet disadvantaged, ‘mature’ immigrant group that is characterized by notable levels of formal citizenship, increased educational attainment and, relatedly, an increasingly strong claim to equal social status in the destination country. Perceived discrimination seemed to be especially likely in case of what could be called ‘integration inconsistencies’, i.e., when improvements in social positioning in terms of citizenship acquisition and educational attainment are not matched up with status improvements in other respects, such as income and social prestige. Additionally, the maturation of the immigrant group makes people more aware of environmental hostilities since they speak the official language and can better track the news, where they also encounter anti-immigration statements by far-right parties.

While respondents attributed perceived discrimination to real and perceived changes in discourse and policies, we should not overlook these stronger claims to equal membership, and the opportunity structures that a significant measure of settlement and integration unlock. For example, the participants demanded a right to gather for rallies to protests, because, being Dutch citizens, they felt entitled to the same civil rights as other protesters, whom the Dutch authorities allegedly do tolerate. As such double standards are, in their view, merely related to ethnicity, the only solution most respondents saw is to use the voting rights that were directly or indirectly acquired under the Dutch naturalization policies.

Before, there was Geert Wilders, now there is another one [Forum for Democracy Party]. We need to do something because Dutch people vote for them and they grow bigger. We must have people who can speak for us (Emre, second-generation, call centre).

Evidently, incorporation in the destination country also facilitates political participation beyond formal citizenship. Second-generation migrants, for instance, also clearly had a better understanding of the Dutch political system than first-generation migrants. They closely followed Dutch and Turkish politics, and saw political participation as the appropriate response to social problems, assuming that if enough of the Turkish-Dutch would vote, they could bring about change. Possibly because of their higher education, they were generally optimistic about the democratic system and its ‘transformative’ promise:

The current [exclusive] politics of immigration only focus on Turkish immigrants in this country. Our status, our lives are at the centre of populist nationalist politicians. We need to be more politically active and we need to tell our story to everyone (Murat, second-generation, social worker).

First-generation respondents, by contrast, were more pessimistic about the immigrant groups’ problems being solved in the Netherlands, as Dutch politicians supposedly do nothing about the far right and ‘even adopt their discourses’. They were more inclined to believe that leaving the Netherlands after retiring would be an easier solution.

The influence of perceived discrimination on voting behaviour thus seemed to be especially strong for the second generation, which is in line with the age composition of the DENK voters (Van der Meer et al., 2017) and theories about reactive ethnicity. DENK was clearly seen as the best advocate of immigrant communities and a counterforce to anti-immigrant discourses and policies. For instance, in an activity that we observed just before the 2019 European elections, the DENK leader invited people to vote for their candidate. Many young people mentioned that their parents used to vote for regular Dutch parties such as the labour party, but complained that these parties, even if they were relatively liberal on immigration and immigrant rights, could not represent the Turkish community like DENK, which allegedly understands ‘the life of an immigrant’ better.

Participant: Of course, I will [vote for DENK]. We need people also in the European Parliament who can defend us.

Interviewer: There are some leftist candidates who can defend your rights as well.

Participant: Yes, but I don’t think they represent us; we need people who are immigrants as well, because they understand the life of an immigrant (Mesut, second-generation, student).

Most participants were Muslims, and Islam has become a popular subject in Dutch politics in relation to ethnic minorities. Far right parties in particular, try to portray Islam as backward and dangerous. While there were nationalist-secular Turks among the DENK supporters, we also met respondents who strongly associated their ethnic identity with religious identity, and then tried to create a stronger bond with other Muslim groups, such as the Moroccan-Dutch. For instance, one participant interpreted the Rotterdam incidents as a fight against Islam, while others argued that assimilative and exclusionist tendencies are directed to any Muslim, not just Turks.

Participant: What happened there (Rotterdam) is that Dutch authorities tried to humiliate the Muslims. If we were from European countries, they would allow us to protest.

Interviewer: Were they any Muslims from other communities?

Participant: Yes of course, many brothers and sisters from other Muslim countries joined us that night (Erdal, First-generation, Freelance).

Both reactive ethnicity and religiosity, which jointly originated from perceived discrimination, thus progressed into a desire to use the acquired political rights to participate politically against perceived hostilities in Dutch politics and Dutch society. The assimilative turn also seemed to contribute to a desire to be represented by politicians of Turkish and/or Muslim origin, thus partially explaining the strong support for an immigrant-led party. And yet, the assimilative turn and mature settlement and acquired political rights do not, in them self, provide a sufficient explanation of transformative reactive mobilization outside of established parties: we also need to consider the role of diaspora organizations.

Diaspora organizations as a vehicle for transformative reactive mobilization

The diaspora organizations helped to construct reactive social identities and facilitated the translation of these identities into group-based political mobilization outside of established political parties. Somewhat ironically, these capabilities were partially developed during the heyday of multiculturalism, when national or local Dutch authorities actively supported immigrant organizations.

The role of the organizations in facilitating transformative reactive mobilization could be well observed in relation to the 2017 protests when UID and UTCC recruited a large part of the protesters. For example, one participant from Utrecht explained that he got a phone call to gather at UTCC to go to Rotterdam and attend the protests, as ‘the Dutch police was beating the Turks’. He mentioned that he had to be there ‘for his brothers and sisters’. Another interviewee pointed out that he did not join the initial protests but went to Rotterdam the day after to protest against the police violence.

Participant: Incidents showed us why we should be together. They [Dutch police] were violent and I am telling you, they will be more and more against us in the future (second-generation respondent)

Interviewer: Why do you think so?

Participant: Because they have the support from politicians and the public.

Attendance is a sign of social belonging (Maliepaard & Phalet, 2012), and we could indeed observe that the diaspora organizations facilitated in-group solidarity, both during and after the protests. People increasingly began to visit the organizations in order to be together and talk with like-minded people. Sharing and hearing stories about the incidents over and over again, and mutual speculation about what would happen to the arrestees, strengthened reactive identities and helped to lay the ground for additional collective action. While no participant mentioned that diaspora organization officials told them directly to vote for DENK, the organizations did advise them against engaging in ill-considered individual actions and instead join activities of the diaspora organizations and cast their vote in elections, preferably electing a party that the diaspora organizations support, such as DENK.

The diaspora organizations also added a transnational dimension to the reactive identity process by fostering solidarity with Turks and/or Muslims in disadvantaged positions around the world. For instance, one of the activities in Utrecht was the Uighurs Solidarity Nights against Chinese Oppression. After the event, the first author was invited to stay for informal talks and listened how people agreed on the type of financial support they could deliver to ‘other Turks who are in trouble’.

Interestingly, the potential to strengthen reactive identities and translate them into group-based mobilization seemed to require a substantial level of incorporation in the country of destination as well, which is partially facilitated by inclusive elements of the country of destination, now and in the past, including the multicultural integration policies: migrants had more opportunities to socialize around identity issues when they had access to facilities to pray and could spend time in the main public rooms where they could follow news on television and on newspapers. Both first- and second-generation migrants visited the centres, albeit for different reasons. First-generation migrants mostly visited the organizations in order to find solidarity by fulfilling their longing for Turkey:

We come here because this is like a small village in Turkey for us. Of course, we share when we have problems but I like to come here to be with our own people (Rasim, first-generation, estate agency).

Second-generation migrants primarily made use of the opportunities to pray, study or develop a project, and made more instrumental use of facilities of the organizations, such as libraries. They nonetheless formed groups in these organizations like ‘UID Youth’ and were assigned tasks in the election campaign, such as visiting residents to remind them to vote. These assignments and formations involved them in the group-based mobilizations and also increased their attachment to the organizations.

Discussion and conclusion

Existing theories predominantly see immigrant political participation as an indicator of integration and belonging in the destination country, which inclusive policies facilitate, especially in the multicultural model. However, our study indicates that changes towards a more assimilative or even exclusionist discourse and policies do not necessarily limit immigrants’ political participation; on the contrary, they seem to have increased the desire among the Turkish-Dutch to participate politically, and do so outside of mainstream political parties. It is mostly the ability of the immigrant group to participate that has a more unequivocal positive relationship with host state incorporation and inclusive policies. In the present case, the ability of the Turkish-Dutch to translate reactive identities into reactive mobilization outside of established political parties is also, somewhat ironically, related to the heritage of multiculturalism, which helped to institute immigrant organizations among larger immigrant groups like the Turkish-Dutch. By acting as a source of social support and solidarity, diaspora organizations facilitated the translation of individual discontent into a reactive social identity, and helped to channel discontent into collective action.

In the alternative framework used here, the political participation of immigrants and their native-born descendants is not necessarily an indicator of social cohesion and trust in the polity of the destination country, even if the more educated second-generation in particular certainly have certainly not lost trust in the political process. It may also reflect social tension and problems or perceived injustice. Much in line with notions of reactive ethnicity and religiosity such ‘reactive mobilization’ seems to require a complex mix of inclusion and exclusion.

While earlier studies have already identified that paradox of inclusion and exclusion, the present analysis also points to another, related, paradox, to which the existing studies have not paid enough attention. On the face of it, reactive identities are backward-looking responses on the part of immigrant groups to cope with persistent disadvantage, leading persons of immigrant origin to strengthen ethnic and religious identities that are rooted in the past. However, in some cases—especially if there is a collective action component—, the immigrant group is likely to do so for forward-looking purposes, namely in the belief that it will improve the position of the immigrant group in the country of settlement. Basically, for Dutch Turks, voting behaviour becomes a way of gaining more power in order to obtain more rights and to be treated equally, and address a perceived lack of tolerance in Dutch society. Ethnic and religious identities may then also change in the process, such as by tying these identities more strongly to the polity in the destination country, or by fostering specific transethnic identities (e.g., a shared Muslim or ‘immigrant-origin’ identity). Existing studies do not seem to sufficiently capture that transformative, future-oriented, political element.

A combination of reactive identity formation and opportunity structures seems to be required to actually translate reactive identities into group-based mobilization. Reactive identities typically develop when an upwardly mobile immigrant group, with a stronger claim to equality, perceives persistent social disadvantage. Opportunities to translate such identities into group-based mobilization depend on various factors such as the size and maturity of the immigrant group, but also on inclusive elements in the destination country’s integration policies, now and in the past. Institutional characteristics of the Dutch political system further add to these political opportunities: the Dutch House of Representatives follows the principle of proportional representation and has an electoral threshold of only 0.67%.

Our fieldwork here focused on Rotterdam and, less so, Utrecht, and mostly concerned ethnographic work among conservative-nationalist residents of Turkish ancestry who, in one way or the other, had connections with diaspora organizations or other elements of the Turkish ‘ethnic infrastructure’. Men were also overrepresented due to the structures of the diaspora organizations and because of cultural and religious barriers, which set limitations to speaking with the male researcher who conducted the fieldwork. Evidently, the findings may therefore not be completely ‘representative’ for all residents of (territorial) Turkish ancestry in the Netherlands, which also include Alevi, liberals and Kurds, and for other immigrant groups, including the Moroccan Dutch. However, DENK’s relative popularity among the Turkish-Dutch and Moroccan-Dutch voters more generally, especially among younger voters, and the finding that perceived discrimination is a crucial determinant of immigrant voting in the Netherlands (Otjes and Krouwel’s (2019), does suggest that similar mechanisms operate among other (sub)groups. It should also be re-emphasized that statistical representativeness was not our aim; we wanted to conduct a strategic case study with a view to conceptual development within the reactive ethnicity framework.

Future research should elaborate on the role of the sending state and international relations in the reactive mobilization process. Here, we highlight factors in the destination country, but a fuller analysis could pay more attention to transnational dimensions by elaborating on the role of the Turkish State and the tensions in the relations between the Netherlands and EU Member States such as the Netherlands. Additionally, it seems fruitful to elaborate on the interaction between relatively domestic and relatively transnational factors, which may reinforce each other in a kind of ‘elective affinity’. There is some evidence, for example, that perceived discrimination strengthens migrants’ identifications with the home country (Snel, Hart and Bochove 2016), which may increase the opportunities of political actors in the home country to influence diasporas and fuel perceived discrimination and/or facilitate mobilisation processes. At the same time, it is clear that the diaspora politics of the Turkish State do not sufficiently explain the success of the mobilisation process, also given DENK’s popularity among Moroccan-Dutch voters in particular.

Future research could examine whether similar processes occur among migrant communities and ethnic minorities outside of the Netherlands. While the partial abandonment of multicultural policies is somewhat specific for the Netherlands, the assimilationist turn in political discourse is not. The desire to engage in group-based mobilization is therefore likely to also be present among immigrant groups elsewhere, especially among Muslims. The degree to which the transformative reactive mobilization process actually unfolds will depend on different contextual factors, including immigrant group size, degree of ethnic organization, and degree of naturalization. Higher electoral thresholds in other countries are likely to hamper the opportunities of immigrant groups to be politically represented in parliament outside of mainstream political parties, or may limit such representation to cities where immigrant groups represent a larger percentage of the electorate. In such cases, a larger part of the underlying discontent may well manifest itself outside of formal political systems.