To vote or not to vote? Migrant electoral (dis)engagement in an enlarged Europe

External voting by nonresident citizens has become an important feature of contemporary democratic politics. However, compared to the average voter in domestic elections, we still know signiﬁcantly less about migrants’ motivations to vote or not. Whereas analyses of external voting patterns offer insights into the results of external voting compared to origin populations, there is a lacuna of knowledge about why migrants choose to vote, or not, when they have the right to do so. This article seeks to address this gap by building a framework rooted in both the electoral studies literature and on the growing body of knowledge on external voting within migration studies. We consider migrant voters’ desire, mobilization, and ability to vote, and map the locus of all factors—either in the country of residence, country of origin, or within transnational political space. We explore evidence from 80 in-depth interviews, collected January–May 2020, with four groups of intra-European migrants—Romanian and Polish residing in Norway and Spain—to map the determinants of external voting. Our research generates three insights which challenge or nuance extant research on external voting. We show how migrants’ motivations to vote depend not only on residence and origin contexts but also on subjective factors and perceptions of the legitimacy of external voting. This article complements existing macrolevel studies of voting determinants with an in-depth qualitative microperspective and generates hypotheses that can be further tested in large-n as well as cross-regional comparisons.


Introduction
The number of states that enfranchise their nonresident citizens in national elections, that is, allow for external voting is rising rapidly (Navarro, Morales and Gratschew 2007). Although external voting remains a matter of heated debate (Bauböck 2005;Ló pez-Guerra 2005;Rubio-Marin 2006;Lappin 2016), as of 2020, almost all European countries allow their nationals to cast ballots outside their territories (European Commission 2020). Following the post-Enlargement  wave of migration to other European Union (EU) member states, Central and Eastern European countries faced the expansion of existing émigré communities and the emergence of new ones (Anon 2010). In many cases, these are quite populous, and in principle their vote can make a difference in national elections (Hutcheson and Arrighi 2015;Lesi nska 2018).
Despite the globally expanding émigré enfranchisement and the clear potential impact of migrant votes on politics in some regions, like within the EU, comparative politics, but also migration studies to an extent, still pay only limited attention to external voting (but see Collyer 2014;Escobar, Arana and McCann 2015;Lafleur and Sánchez-Domínguez 2015;Goldberg and Lanz 2019). Most studies of migrant electoral behavior concern the way immigrants vote in the elections of their countries of residence, even though there are still significant barriers to immigrant enfranchisement. The disproportionate attention paid to immigrant voting versus external voting is explained, for example, by data availability and the interest within receiving societies.
In comparison to migrant voting in countries of residence, external voting in country of origin elections receives less attention, and research is dominated by national case studies of either specific countries of origin or countries of residence (Itzigsohn and Villacrés 2008;Boccagni 2011;Lafleur and Chelius 2011;Leal, Lee and McCann 2012;Escobar, Arana and McCann 2015;Lafleur and Sánchez-Domínguez 2015;Mencütek 2015; Lesi nska 2018; Goldberg and Lanz 2019;Mügge et al. 2019;Finn 2020;Sevi et al. 2020). Existing comparative research focuses on explaining the emergence and horizontal diffusion of external voting rights, that is-why do sending countries grant expatriates the right to vote in the first place (Rhodes and Harutyunyan 2010;Lafleur 2011;Collyer 2014;Hartmann 2015;Lisi et al. 2015). Another issue attracting attention has been the impact of party mobilization and party activity abroad (Lazzari 2019;Burgess and Tyburski 2020;Kernalegenn and van Haute 2020). Meanwhile, analyses of the factors that can account for turnout and results beyond a single country of origin or residence remain scarce (Chaudhary 2018;Ciornei and Østergaard-Nielsen 2020;Pallister 2020).
Several studies, like Lafleur and Sánchez-Domínguez (2015), explore the determinants of external voting but knowledge on why migrants vote the way they do is still limited. However, even a passing look at the voting outcomes in different national elections and the number of votes cast abroad reveals an additional puzzle: a grand majority emigré voters do not participate in elections (Hutcheson and Arrighi 2015;Ciornei and Østergaard-Nielsen 2020). This suggests that beyond asking why migrants vote the way they do, we also need to ask the more fundamental question: why they vote, or not, in the first place, and inquire into the determinants of turnout in external voting.
To shed light on this, we draw on the classic studies on political participation, which have inquired into fundamental questions regarding domestic elections for decades. What we are interested in is, according to Ferejohn and Fiorina (1974: 525-emphasis added) 'the most basic political decision in a democratic system: to vote or not to vote'. Given enfranchisement of nonresident voters, we are interested in three sets of factors. First, do they desire to vote-what are their internal motivations to stay home or make their way to the polling station? Second, are they encouraged to vote-that is, what are the outside elements of mobilization? And finally-can migrants vote?-that is, are there administrative and logistical factors that impact their access to the ballot box? We expand the usual scope of electoral studies by introducing the transnational dimension and discussing the way these factors interact with their locus-country of residence, country of origin, or the transnational positioning of migrant voters between the two.
To explore this, we look at migrant communities from states that impose no limits on passive voting rights for migrants, eligibility is not an issue, and which are among the largest senders of intra-EU migrants in absolute terms-Poland and Romania. We conduct eight semi-structured interviews with members of émigré communities in two countries of residence: Spain and Norway. While our study is exploratory, not strictly comparative, we seek to have variance in the studied sample of voters, particularly between those that never vote, those that vote sporadically, and those that show up for every election.
Our contribution is threefold. First, the factors influencing external voting as a particular political practice which our exploratory research identifies can be used as testable hypotheses in further research using broader, representative surveys of migrant populations. Second, we generate three insights that challenge or nuance extant research on external voting, and third, offer a framework that can be used to inform policy development aimed at increasing external voter turnout. We also contribute to expanding the geographic scope of the research on migrant political engagement and transnational politics, which is dominated by studies of South-North migration (Lafleur and Chelius 2011;Leal, Lee and McCann 2012;Palop-García and Pedroza 2017), by exploring intra-European dynamics. Our study complements the robust literature on diasporas engagement in homeland politics, which to date has primarily focused on nonelectoral forms of 'political remittances ' (King and Melvin 2000;Waldinger 2014;Koinova 2021). Finally, our focus on Central-Eastern European transnational politics is relevant both in the context of post-Communist transitions and the recent populist turn and democratic backsliding in the region (Stanley 2017).

The determinants of transnational political participation: A theoretical framework
The issue can be approached from two ends-macro and micro. Macrolevel studies are more geared to answering why a certain share of the eligible population votes. Such studies try to identify the contextual factors that correlate with aggregate voter turnout; for example, that trust in democratic institutions or levels of membership in political parties are positively correlated with average turnout across elections (Riker and Ordeshook 1968;Silberman and Durden 1975;Settle and Abrams 1976). The macroapproach helps explain variance between different societies and states or between elections, but it does not tell us why some people vote, while others do not.
Microlevel factors can predict individual voters' behavior with a significant degree of certainty. Following a comprehensive and systematic review of voter turnout studies, Smets and van Ham (2013) identify a set of independent statistically significant variables predicting voter participation: age and age squared, education, residential mobility, region, media exposure, mobilization (partisan and nonpartisan), vote in previous election, party identification, political interest, and political knowledge.
However, for many years political scientists were also trying to explain the decision to participate. Anthony Downs (1957) put forth a canonical rational choice 'economic theory of democracy', explaining voting preferences in terms of utility maximalization. However, his hard rationalist approach only highlighted the apparent 'paradox' of voting: since casting a ballot is costly (takes time and effort), and the impact of each vote negligible, instrumentally rational citizens would not be expected to vote at all-because it is irrational (Tullock 1972). And yet they do. Or do not. But why that is the case is beyond parsimonious rational choice models, and Downs himself 'left the decision to vote somewhat up in the air, resorting to extra theoretic factors' (Ferejohn and Fiorina 1974: 525). To account for the supposed 'paradox' at the heart of democratic politics, rationalist scholars introduced other forms of benefit that selfishly motivated voters can achieve, for example, informing others about their political preferences (Glazer 1987), proposed 'alternative rationality criteria' around the act of voting (Ferejohn and Fiorina 1974), or varying probability considerations (Owen and Grofman 1984).
But is the fact that millions of people routinely exercise their basic democratic right really a 'paradox'? Empirical research among actual voters, probing their motivations and preferences, renders a picture which is very detached from the rational choice expectations-and neither irrational nor paradoxical, but instead a matter of civic voluntarism (Verba et al. 1993;Schlozman, Verba and Brady 1995). It can be approached from a psychological perspective, as a reflection of personal values (Caprara, Vecchione and Schwartz 2012) but also an affirmation of democratic rights, as well as a demonstration and means of creating social pressure (Dellavigna et al. 2017). Considering the broader question of the (lack of) political participation, Brady, Verba and Scholzman (1995: 217) suggest that people do not actively take part in politics 'because they can't, because they don't want to, or because nobody asked'. Can we think of similar intuitive reasons for not voting?
Drawing on insights from this rich literature we distinguish three sets of factors that account for why people vote (or not): because they want to, because they are mobilized, and because they can (Compare Brady, Verba and Schlozman 1995;Harder and Krosnick 2008). The first set of factors brings forward the voters' motivation and desire, reflecting different social, cultural, and psychological factors, which relate to voters' internal 'push' to vote. Second, mobilization stands for the 'pull' of the sociopolitical environment and others who try to convince the voter to participate and whether enough information about the importance, need, procedure, and choice available in elections is accessible. Third, the ability to vote-which consists of formal eligibility, implied by enfranchisement, as well as access, meaning the process of registration and technicalities of voting itself.

Toward a configuration of factors driving external voting
Having established the three dimensions that condition voter turnout in 'regular' elections, we should bear in mind that external voting has additional features, which can influence the decision to vote or not to vote (cf. Finn 2020). Leal et al. (2012: 540), who have studied the participation of Mexican migrants in external voting, note that the roots of migrants transnational participation include factors familiar to students of domestic elections, 'as well as dynamics uniquely relevant to immigrant communities'. External voters' electoral behavior, according to Lafleur and Sánchez-Domínguez (2015), can be partly explained by the traditional models of electoral studies-the ideal typical 'social voter', 'ideological voter', and 'interest-driven voter'-but there are also uniquely transnational voters.
Emigrants voting in country-of-origin elections can be influenced by domestic country of origin factors, much like voters 'at home', but in different ways and to a different extent. Waldinger, Soehl and Lim (2012) explore the way in which both political attitudes brought from the premigration context and the new transnational positionality in which migrants find themselves shape their political behavior. Meanwhile, their decisions and preferences are also shaped by the context of the country of residence (Goldberg and Lanz 2019). Last but not the least, they can be influenced by their 'in-between' positionality, that is, the very fact that they are migrants which impacts their sociopolitical status both in the country of residence and of origin, can create a unique political agenda for which external voting is an opportunity of expression (cf. Waldinger 2014; Erdal 2020) and can be influenced by diaspora entrepreneurs (Koinova 2021). Hence, Escobar, Arana and McCann (2015) emphasize the importance of the 'context of settlement' as well as the 'context of exit', and analyzing the voting behavior of Colombians in the USA and Europe list motivational, institutional, and migratory factors. Similarly, scholars researching the Turkish diaspora note that participation expresses the links with the origin country, but is contingent on settlement contexts (Sevi et al. 2020), and that external voting turnout can be partly explained with the symbolic dimension of citizenship and the desire to participate directly in homeland politics (Mencütek 2015).
Building on a combination of the broad and rich literature in electoral studies, which provides a set of determinants of voter participation that can be seen as 'universal', and combining it with the growing body of research on external voting in migration studies, which emphasizes voters' 'positionality' and settlement/origin/transnational entanglements, we propose the following framework (Table 1), which offers a systematic configuration of factors driving external voting. The three rows display different sets of factors influencing voters' decisions to partake in elections, while the three columns help identify where specific factors are 'nested'-here/there, or in-between.

Methods and data
Our study focuses on the EU and the European Economic Area (EEA), where the nearly universal emigrant enfranchisement allows us to explore individual-level determinants of voting, not obscured by legal inability to participate. We selected Polish and Romanian migrant communities due to their size and the scale of migration (over 2 and 3 million emigrants, respectively). These two groups are very diverse in terms of demographics and specific motivations for migrating, and importantly for our purpose, the perceived importance of external voting in these two countries is high, as they both have a legacy of émigré political involvement (Burean and Popp 2015; Lesi nska 2019). However, as Supplementary Appendix 3 shows, the actual impact of the diaspora vote varies between countries and elections. Furthermore, both Poles and Romanians are among 5 largest migrant groups in at least 10 Western European countries, although they show regional variation, with Romanians' greater presence in Southern European states, and the Poles' stronger representation in Northern Europe.
Pairing Spain and Norway was motivated by the variation in electoral results in the last parliamentary elections occurring before the research was designed (see Supplementary Appendix 3). In the watershed Polish elections of 2015, Poles in Norway displayed an even stronger support for populist and far-right parties than did voters 'at home', while emigrants in Spain voted for left and liberal forces. Meanwhile, in 2016 Romanians in Spain supported the conservative PNL, while those in Norway favored the liberal Save Romania Union (USR) party. Spain and Norway display a symmetrical pattern, with Romanians as the largest migrant group (over 600,000) and Poles in seventh place (ca. 100,000) in Spain, and a reverse situation in Norway, where the Polish diaspora of ca. 100,000 is the largest and the ca. 15,000 Romanians are seventh. Although Norway is not an EU member state, the fact that it is part of the EEA and the Schengen Zone is meant that the country was the destination of the largest influx of postaccession migration from Central and Eastern Europe in the Nordic region, and migrants enjoy the same external voting rights as they do in any other European country of residence.
Within the countries of residence, we focused on two cities and their metropolitan regions-Barcelona and Oslo. Both are large centers of immigration and both are 'global cities', which see a varied influx of migrants of all ages, classes, and professions. These residence settings vary in many ways; however, our study was not designed to explicitly compare two migrant nationalities or cities of settlement but is exploratory and instead seeks To investigate individual migrants' desire, mobilization, and ability to vote in countryof-origin elections, we followed a qualitative approach, based on in-depth semi-structured interviews. The interview guide was designed to cover different aspects of individuals migration stories, foregrounding themes related to political participation and broader political and socioeconomic perspectives on both contexts of origin and residence (see Supplementary Appendix 1). We also collected systematic background information about all participants, including basic demographic data, educational and professional trajectories, and history of civic and political engagement. The most recent elections at the time of interviews: parliamentary elections in Poland in October 2019 and the Romanian presidential elections in November 2019 were the initial reference point.
Our team comprised seven researchers, with a diverse background in the social sciences, from anthropology, through human geography and migration studies to political science, with necessary language proficiency in Polish and Romanian, including some researchers residing in both Barcelona and Oslo. The interview guide was developed and tested jointly by the team, while the interviews were conducted by four researchers, and the entire team convened twice during the interviewing process. Interviews were conducted primarily in person between January and April 2020; however, the COVID-19 pandemic and the resulting lockdown necessitated the use of telematic means (e.g. Skype, telephone) for some interviews. The interviewees were recruited using a diverse set of entry-points, including the researchers' extended networks, social media, migrant associations, and religious institutions catering to migrants, as well as a limited employment of snowballing from one interviewee to another. All interviews were conducted respectively in Polish and Romanian and lasted, on average, 1 h, though some were briefer and a few longer. They were subsequently translated and transcribed in English. The entire dataset was coded using a joint codebook, with 8 main topics and 72 themes, developed to make the most of the exploratory potential, using NVivo software for qualitative analysis. The software allowed us to explore specific themes and keywords across the 80 interviews.
The data set comprises 80 semi-structured in-depth interviews with migrants living in Oslo and in Barcelona, originating from Poland and Romania (see Supplementary Appendix 2). We interviewed roughly equal numbers of women and men from each origin context in each city. Only a small minority of our interviewees were dual citizens (of their countries of origin and residence), while virtually all held only country of origin citizenship. Most of our interviewees were part of the postaccession migration from Poland and Romania: about a third had lived in Oslo and Barcelona for less than five years at the time of interview, while a third had lived abroad for more than 11 years.

Analysis
The basic prerequisite to even consider voting in national elections is eligibility, founded in enfranchisement and citizenship. However, having the right to vote does not necessarily equate in actual voting behavior. Therefore, we start by investigating the question of migrants' desire to vote. We subsequently move on to examining in the mobilization to vote. We then turn to the ability to vote, which in our analysis focuses on access, the practical dimension translating eligibility into the actual prospect for migrants to vote from abroad. Combined, these three dimensions-desire, mobilization, and ability to vote-as we show, help explain the low turnout among migrants with the right to vote from abroad.

Desire to vote
A key reason for not voting is, unsurprisingly, a lack of interest in politics, which is true for migrants and nonmigrants alike.
I'm not interested in politics. Here, in Spain, still, I listen to something, I read something, but when it comes to Polish politics. . . I don't even know, is Duda still the president?
(Weronika, Pol/Barcelona, female, 38 years) For some respondents, it was not so much a lack of interest in politics but rather of homeland matters which are 'pushed down somewhere to a lower priority' (Michał, Pol/ Barcelona, male, 34), even among those who emigrated fairly recently. However, not voting externally does not necessarily go hand in hand with disinterest. Szymon, who emigrated 10 years earlier, reported voting in local elections in Oslo, and in national elections in Poland before migrating, but emphasized that even following politics 'back home' is not enough of a reason to vote externally: 'I am interested in it like I'm interested in a football match [. . .] But I can't influence this game. I'm not there. I don't lose the game' (Pol/Oslo, male, 45).
The above quote points to a critical factor influencing migrant's desire to vote: considerations of transnational positionality, in relation to democratic processes within nationstates. Rationalist theories of voter behavior emphasize the cost-benefit calculation between the effort needed to participate versus expected returns-either in the form of perceived influence on the political landscape or demonstrating one's partisan stance. Many of our interviewees, however, questioned the legitimacy of their own voting rights in the country of origin-resulting in not voting. We find that convictions about two clashing notions of citizenship-based on residence or on felt national identity-affect the desire to vote. Przemek, who emigrated in 2008, expressed the residence-based view categorically: This argument has a strong moral underpinning and expresses a logic of appropriateness that seems incompatible with purely rationalist accounts. Questioning migrants right 'to decide for those people that live in Poland' (Kasia, Pol/Barcelona, female, 34, emigrated four years ago) or Romania was both a justification for interviewees who did not vote themselves, as well an explanation of the behavior of the nonvoting majority by those interviewees who did vote.
Some of those who voted from abroad accepted the general premise of residence-based voting rights, but still nuanced this. Lack of voting rights in national elections in the country of residence was sometimes cited as a reason to partake in external voting: This highlights the open-ended character of migration, especially in the context of intra-EU mobility, which undermines the idea of a clear distinction between residents and nonresidents and where the elusive nature of residency contributes to complicating (formal) citizenship. Meanwhile Daniela casts voting not merely as a privilege, but as a duty and a matter of responsibility: As long as it is our right to go in the street and complain [. . .] then I should also have my part of the responsibility. If we talk about politics from the comfort of our couch then yes, it is acceptable not to vote. But if we want to go to the streets and protest then the vote should be mandatory. I personally always did everything I could to go and vote.
(Daniela, Rom/Oslo, female, 40) The notion of democratic responsibility links to another factor influencing the desire to vote: the urge to affect political realities in the country of origin. Karolina, who traveled some 100 km to vote in the last election explains: 'I voted to stop the [ruling populist party] from winning' (Pol/Barcelona, female, 46). Although some of the interviewees reported disillusionment with the impact of voting, many among those that vote regularly emphasized that they perceive this act to change their homeland for the better, to remit the experiences of life in more established democracies back home, and in terms of elections-to help those left behind by choosing candidates and parties the migrants perceive as better.

Mobilization to vote
The decision to vote can be influenced by the environment in which a person lives. Voter mobilization can be direct or indirect, focused on the turnout or on rallying a partisan electorate. For emigrants, direct mobilization implies targeted campaigns focused on diasporas. But mobilization can just as well be more horizontal, voter-to-voter, not merely top-down vertical, between political institutions and emigrants. Both the Romanian and the Polish state have recently launched campaigns focusing on émigré voter turnout, sometimes as part of broader identity-building actions among diasporas, as Monica describes: The Romanian state plays on the nostalgia for home [. . .] It's a narrative that looks good, the mobilization, but there is no truth, no real motivation behind it.
(Monica, Rom/Barcelona, female, 47) The chairman of the Polish Senate oversaw the launching of a similar campaign, entitled 'We are 60 million' (TVP Info 2019), but the electoral mobilization around this campaign was generally perceived to be weak and unnoticeable. 'I vote because it's my civic duty, but I don't remember any campaign'-Patryk (Pol/Barcelona, male, 23) claimed. Paulina seconded 'Very weak appeals. . . they didn't reach me a lot. Rather the appeal of other Poles who live here than the ones of the government' and interpreted this governmental action as partisan campaigning: 'But maybe I'm not a group they were targeting cause I'm not their voter, from their point of view, that appeal shouldn't have reached me' (Pol/Oslo, female, 29). Another Polish migrant in Norway reflected a common reaction among our interviewees: The government does something in that direction? I didn't notice that. I think they do not like that [. . .] the majority of people would vote against the government. So, they'd rather they didn't vote. [. . .] We mobilized ourselves.
(Julia, Pol/Oslo, female, 29) This is linked to a broader reticence toward official institutions in the country of origin, which was noticeable among many interviewees, most clearly expressed by an academic with a long migration track record: I was part of the League of Romanian Students Abroad [trying] to bring students together and I thought that was really nice until I bumped into the servants of the Romanian government, their ambassador to be exact, and when I heard his speech, I didn't like it one bit.
(Mihaela, Rom/Oslo, female, 30) Among Romanian migrants, mobilization activities of homeland political parties were also visible and underlined. As underscored by Marius: 'new parties [try to] increase the political weight of the diaspora by recruiting people from the diaspora', to the extent that for one of the recently established parties, USR, 'the diaspora branch was the largest in the party'. In Marius' eyes, for political parties: the diaspora is a large reservoir of votes that they want to tap, and not necessarily because of interests related to what these people want, but rather that here are millions of votes that matter, which often make the difference in Romania (Marius, Rom/Barcelona, male, 42) In the 2019 elections, the Polish opposition parties also saw migrants as a potentially important constituency, seeking to 'mobilize the Polish community abroad to vote, but especially not in favor of the current ruling party' according to Dawid (Pol/Barcelona, male, 38). However, the main mobilization campaign, supported by the opposition but framed in nonpartizan terms (i.e. 'go vote'), was not adjusted to the migrant communities-they were only reached through (social)media and debates in the country of origin.
The sense of being treated instrumentally, or purposefully left out, very quickly made interviewees shift to noticing that rather than an object of political mobilization, their Paradoxically, this sense of marginalization had a mobilizing effect on some migrant voters, feeling they needed to express their disappointment by participating in elections and supporting a political alternative that might give them a platform in the future: 'One of the reasons why the diaspora politically mobilized itself was the fact that they've begun talking about the diaspora [which had been] nonexistent in the Romanian political discourse' (Maria, Rom/Barcelona, female, 40). The key unfulfilled demand was that of giving migrants a voice as a specific constituency with a specific set of issues. In the absence of institutionalized mobilization, migrants take matters into their own hands: However, two opposing views were clearly discernible. On the one hand, many recent migrants were either describing themselves or perceiving others as living in 'a parallel reality', that is, physically inhabiting the country of residence, but mentally remaining anchored in the country of origin, as Adam and Marius illustrate: This leads to emphasizing country of residence mobilization, even if the most active diaspora entrepreneurs were themselves socialized into their roles before migrating, as Marius attests: 'The politically active diaspora is generally the recent date diaspora, [these people] have not fundamentally transformed after they moved to Spain' (Rom/ Barcelona,male,42). Although the question of where political preferences are formed-'here', 'there' or 'in-between'-crucial for understanding the dynamics of political remittances, cannot be answered with our material, we can see traces of mobilization through all possible channels.

Ability to vote
Although in our cases, the eligibility to vote was universal, the actual access was conditioned by the voter registration procedures and voting methods available. This varies across countries (of origin) and also across elections, as regulations change (Ostling 2019). For instance, Romania introduced postal voting prior to the 2019 elections, while Poland suspended postal ballots before 2015 and reintroduced this as an option for the 2020 presidential elections held during the COVID-19. Also, while Romania maintains an automatic voter registration system for in-person voting (postal votes require registration three months in advance), Polish citizens voting abroad need to register personally, by phone, letter, or email no later than three days before the election. Overall, all the respondents who voted (and some of those who in the end did not) described registration procedures as straightforward. In some cases, however, even this relatively small hurdle is enough to discourage the less motivated, as Daniel describes: 'A friend persuaded me [to vote], but I was late for the registration [. . .] 48 hours before' (Pol/Barcelona, male, 40). The reliability of the registration system can be an issue, and even those who desire to vote were motivated to register and show up, end up not voting, as Adrian shares: we didn't vote because the organization here abroad sucked. You had to wait many hours outside and I didn't even receive the confirmation that I can vote. Other people received it a few days after the registration. I didn't [receive it] for several months.
(Adrian, Pol/Barcelona, male, 31) Accessing the polling station (or the option of postal and proxy voting) is a second hurdle-even if registration is automatic. Due to a limited number of polling stations, many migrants may not even consider casting their ballot, as they would not be able to go there on election day. As Corina's comment reveals, there is a further dimension to what might be considered a reasonable time investment, or not: [It]j 'was really far, it would take me an hour to get to the voting station in Oslo' (Rom/Oslo, female, 48). Meanwhile, Zofia underscores a reality that is familiar to many migrants living outside the major cities where polling stations would typically be located: I know that people who do not live in big cities have the problem that they would have to drive somewhere in order to vote. In Norway the distances are so long that maybe it's not possible.
(Zofia, Pol/Oslo, female, 33) In some cases, respondents recalled bottom-up initiatives set up to mitigate some of these problems, for example, a Romanian man who taxied people to the polling station, but overall, the logistics of voting abroad is a very important factor limiting turnout. Among interviewees, access was impacted by working on the day of elections (e.g. waiters, cooks), especially when voting is organized on one day only, others explained that the lack of mental time to think about politics and the election was something that limited not just their ability, but the desire to vote at all, as reflected by Tomek: 'I have never voted. Never in my whole life [. . .] I just don't have the time [. . .] I work 7 days a week' (Pol/Barcelona,male,35). This sense of politics as beyond my lifeworld may overlap with questions of interest, which we discussed in the above section on desire. Meanwhile, it is pertinent to also consider the practical elements which enable space for interest in some people's everyday lives, and not for all. However, given that a voter desires to vote, is mobilized, got across the hurdles of registration, and gets to the polling station on election day, a final hurdle remains, which is related to the organization of external voting on elections day. Queues in front of embassies and consulates are the typical image illustrating media coverage of emigrant voting, as Mihaela recalls: People started to get irritated [. . .] I remember I went to get lunch, and I thought everything is ok, that I will just be two more hours in line, and I will go home. I ended up there for 9 hours before they closed the ballot boxes. That was like 10 minutes before people who were already in the embassy could have voted. People got violent; I didn't like that.
(Mihaela, Rom/Oslo, female, 30) An opinion often expressed was that inadequate voting organization was a conscious government strategy to discourage diaspora voting, also among Polish migrants. Among our interviewees, most Polish migrants waited more than an hour (often 2-3 h), while some did not experience queues at all (depending on the election and the polling station they went to). In elections before November 2019, most Romanians waited 3-12 h (sometimes without voting in the end). However, Romanian interviewees reported visible improvements in the last election (November 2019), as Mihai and M ad alin note: You'd go in, scan your ID, you'd get the voting card, and vote. I liked it a lot that they were using the tablets to work with the voting.
(Mihai, Rom/Oslo, male, 27) These were the first real elections [. . .] I came to vote and solved everything in 2 minutes. I do not claim to say that I have to vote in 2 minutes necessarily, but one thing is to wait for 5, 10 minutes, maybe half an hour and another is to wait 10 hours and then stay out.
(M ad alin, Rom/Barcelona, female, 32)For this election, postal voting was also introduced, and that was also evaluated as an improvement, even though the instructions were confusing, and some voters later learned that their votes did not count. Several interviewees suggested that the improvements were an indirect result of the mobilization of Romanian migrants. Indeed, postal voting, opening/asking for new polling stations, and more than one day to vote were all postulates raised by migrants and politicians alike following the earlier electoral failures. Similarly, the 2019 election also saw a mobilization of Polish migrants who used the available regulation to set up polling stations independent of consular institutions, reaching record numbers. However, some Polish migrants complained that the tradition of holding elections on one day (Sunday) influences turnout as some are simply not able to vote.
Meanwhile, migrants adopt individual strategies to cope with queues to vote. Some come early in the morning, others go with friends to queue together, others go in the afternoon expecting queues to have faded, and others again try to get inside the voting station before 9 pm in the hope that those inside will be allowed to vote even after the official end of elections.
Large queues can have contrasting impacts: while some were turned away and did not manage to cast their ballot, others felt encouraged by standing in queues, like Lorena who felt 'more motivated when [she] saw what is happening there' (Rom/Oslo, female, 33). Despite these claims, it is hard to deny that long queues have a voter suppression effect, as Maria and Oliwia attest: 'I've voted twice [but] the fact that these barriers exist [. . .], that you have to wait in line, it dissuaded me. And I know more cases' (Maria, Rom/ Barcelona, female, 40); 'If someone doesn't have time or is at work or has kids then they won't vote even if they wanted to ' (Oliwia,Pol/Barcelona,female,38).
Assuming, however, that the dedicated voter registers reaches both the polling station and the voting booth, a final hurdle remains. This pertains to her ability to express her political views through the election in relation to the issue of external voting districts. Clearly how external votes count in national elections matters, typically organized either as integrated within local voting districts based on former place of residence or based on a separate constituency for external votes. Romanian voters in parliamentary elections elect four members of the lower chamber of parliament and two senators (although their votes also add to the national results of the parties they support). Conversely, Polish migrants vote for candidates on the Warsaw-Centre list (with their votes also adding to the overall results). Among interviewees from Poland, this problem is raised by some, including Julia who says: 'I'm from Silesia, but I had to vote for the Warsaw list. It's unfair. That we all have to vote and decide for the people from Warsaw!' (Pol/Oslo,female,29). Martyna also thought that while perhaps not a major issue it 'could be a negative factor [. . .], that could dissuade people from smaller municipalities, [. . .] who are more closely tied to the local authorities' (Pol/Oslo, female, 35) from voting from abroad. None of the Romanian interviewees commented on this issue, primarily because they were asked to reflect on the last elections at the time-that is, presidential. Taken together, these various technicalities of external voting are fundamental for voters' access and therefore the ability to participate in elections.

Discussion
Our interviews paint a complex picture, suggesting that the final outcome in terms of a person's electoral behavior-voting or abstaining in a particular election-is the result of an interplay of many factors. At times, instrumentally rational calculations of one's vote, the moral legitimacy of deciding for country-of-origin residents, lack of perceived political representation, as well as mundane questions such as the lack of time to vote, can be intertwined in a way that does not allow to answer convincingly why a certain person decided not to vote: Sincerely I did not intend to vote because I did not want to pass my day up in that gigantic queue. And my other reason was that I did not know who to vote for [. . .] Although, I don't know if this one vote of mine would change anything. But I did not want to decide for those people that live in Poland.
-Did you vote when you lived in Poland? Yes, always (Kasia, Pol/Barcelona, female, 34) The desire to vote appears to hang on two key factors-an interest in politics, particularly in the country of origin, and a conception of democratic legitimacy of external voting, which conditions the urge to influence homeland politics. Mobilization is a mechanism working through both official and unofficial, country of origin and country of residence channels, and their relative importance may vary, although in our cases voter-to-voter mobilization appears most influential, because it is adjusted to migrants' transnational positionality, and can help overcome, for example, access problems specific to external voting. Finally, ability to vote, which is a prerequisite of the entire process, also depends on access factors and on voting methods available in each election, registration procedures, as well as the country of residence context (including its geography and the location of polling stations). However, while the possibility of proxy and postal voting is important, its impact on the turnout should not be overstated. In our cases, the Romanian postal vote in 2019 was negligible, while the record high turnout occurred at external voting ballot boxes. Meanwhile, voter turnout among Polish migrants almost doubled between 2015 and 2019 even though the possibility of voting by mail was removed.
Instead of searching for monocausal explanations or even for the most important factor among many, our exploratory analysis helps to map the determinants of external voting, whose relative importance can be further tested on representative samples. Table 2 below organizes our findings along the framework presented in Section 2 and Table 1. Importantly, the same factor can have a different impact depending on the combination with others; for example, long queues are predominantly seen as discouraging from voting, but those interviewees who saw voting as an expression of citizenship or treated their vote as an act of protest-saw queues as mobilizing. Some of the factors we identify can be seen as 'objective', for example, distance to polling station, or lack of postal voting, while others are visibly 'subjective'. While the former is easier to address and use in quantitative analyses, we should not forget that a not inconsiderable weight in shaping the decision to vote or not relies on such subjective elements.

Conclusion: To vote or not to vote externally?
External voting has become an important feature of contemporary democratic politics. However, compared with the average voter in domestic elections, we still know markedly less about migrants' motivations to vote or indeed why migrants do not vote when they are enfranchised. This article sought to address this gap by building a framework rooted in both the electoral studies literature and on the growing body of knowledge on external voting within migration studies.
We explored qualitative evidence from 80 in-depth interviews with four groups of intra-European migrants to map the determinants of external voting. We looked at two groups of migrants from Central and Eastern Europe, Poles and Romanians, currently among the largest emigrant societies in Europe. Although economic concerns are the dominant motivation for the postaccession migration from those countries, its scale and the diversity of the migrants makes these groups valuable cases for studying transnational politics. While not strictly representative, our interview samples included people with different migratory trajectories, education levels, political views, and degrees of political engagement, allowing for analytical generalizations. Apart from the broad mapping of determinants, our research generates three insights which challenge or nuance extant research on external voting.
First, voting behavior is not a binary variable in people's experience. Many migrants may vote once, but not in each election necessarily. Few people participate in all elections, but among permanent nonvoters there are also those who were regularly participating in elections before migration. This suggests that the dominant determining factors in country-of-origin election turnout do not have to have the same pulling power among migrants. Furthermore, the volatility of external voting is much greater than in domestic voting, for the above stated reasons and due to migrant mobility, and large-n research on external voting should be wary of election-to-election comparisons.
Second, both political dynamics in the country of settlement and origin matter, but they do so in varying ways, for example, discrimination versus following news. Our interviewees were very often expressing frustration over their lack of voting rights in national elections in the country of residence (as noncitizens of these nation-states), and for many, external voting in country-of-origin elections acts as political compensation (rather than seeking naturalization), driven by a desire for political participation and expression. However, many migrants perceive themselves and their compatriots as living in a 'parallel reality', physically residing abroad, but mentally anchored 'at home'. This latter feature can be particularly strong for intra-European migration, which is in principle more fluid, and possibly particularly strong in the cases we analyze, but the consequences of this dual dynamic are first and foremost that transnational politics has to be treated as a realm, which is different from domestic politics whether in the country of residence or of origin.
Third, an important determinant of the decision to vote or not is not 'rational' in the instrumental sense, but value-rational with a moral logic of legitimacy. Migrants reflect on whether diasporas should have the right to vote, and the residence-based principle clashes with the citizenship-based voting rights not just between migrants, but in the heads and hearts of individuals, who are weighing their moral right to take part in democratic politics from afar.
Without doubt, nonresident external voting is costlier in terms of time, effort, attention, and sometimes also financially than voting 'at home', and visibly lower turnouts among nonresident eligible citizens should not be surprising. However, if increasing external voter turnout is a state's policy goal, there are a number of 'objective' elements that can be put in place to reduce costs, boost migrants' ability to participate, increase their motivation, and potentially-impact their desire to vote. Notes 1. In the Analysis section, all interview quotes are anonymized and cited using pseudonyms, where we denote whether this was a migrant from Poland or Romania and living in Barcelona or Oslo. Full interview transcripts as well as the coding remain with the authors. Please see Supplementary Appendix 2 for an overview of interviewees.

Supplementary data
Supplementary data is available at Migration Studies online.