ABSTRACT

The ‘age of migration’ (Castles and Miller 2008) poses well-known policy problems for policy-makers everywhere. It presents particular normative and practical challenges to liberal democracies whose legitimacy depends on the participation of their citizens in collective decisions either directly through referendums or, more typically, indirectly through the election of representatives who remain accountable to voters. The legitimacy of liberal democracies also hinges on the equal right of each citizen to stand for election. In modern nation-states, access to the right to vote and the right to get elected to represent one’s peers has traditionally been tied to citizenship status. We know from decades of research that variations in the definition of citizenship and differences in the way individuals acquire the status of full citizens have varied substantially across liberal democracies and have had significant implications for the political involvement of immigrants (Messina 2007: 170-93). It would be a mistake, however, to disregard a further, political dimension of the political incorporation of immigrants: political parties are the main agents of institutionalized democratic participation in liberal democracies. Their patterns of competition and cooperation are strongly influenced by cleavages that crystallized in the late nineteenth and early twentieth centuries. The cleavage structures and voter alignments observed in Lipset and Rokkan’s (1967) seminal study predate the peak period of post-war migration, but still influence – despite their declining capacity to determine the shape of party systems – the way new socio-economic, cultural or political preferences can be expressed. In some cases, immigrants may have neatly fallen on one side or the other in the predominant socio-economic cleavage. In many other cases, immigrants have brought further, cultural or other dimensions to existing party cleavage structures. In a number of cases, immigration itself is believed to have contributed to the development of new cleavages (e.g. a cleavage of cosmopolitanism versus national protectionism; see Kitschelt 1995; Kriesi et al. 2006). The underlying idea of this volume was to present various stages of democratic representation as a process, which is structured by institutions but is ultimately shaped by human agency (and strategy). On a theoretical level it tried to integrate a number of research perspectives in political science and sociology that tend to look at stages of the representative process in isolation from one

another: migration studies, electoral studies, legislative studies and the study of political parties. On an empirical level, the volume expands our knowledge on three collective agents of democratic representation that V.O. Key (1964) famously found to be characteristic of representative democracies: ‘party in the electorate’, ‘party organization’ and ‘party in government’. Based on this logic, the volume started by reviewing empirical evidence on the link between recent immigrants and their descendants as voters on the one hand and the political parties in a number of mature liberal democracies on the other (‘party in the electorate’). In a second step, several studies in this volume focused on party organizations and analysed the extent to which they (1) provide opportunities for immigrants and/or politicians from visible minorities to be nominated for elected office and (2) use the nomination of minority candidates to mobilize voters from such backgrounds. Finally, this volume offers a number of new insights into the behaviour of recent immigrants and their descendants once they have attained elected office (‘party in government’), focusing on their behaviour on the floors of parliamentary chambers, in committees and in relation to their constituents. In the introduction, we outlined the main theoretical proposition, which informs the work for this volume. While recognizing the importance of structural features and policies as factors driving the process of political representation to a significant extent, we proposed to focus on a less well-researched but crucial area: the strategies of the main collective agents of political representation, the political parties. ‘At a very basic level’, we stated in the introduction, ‘increases in ethnic minority representation may reflect growing political engagement of minority citizens and increasing responsiveness to these potential voters among political parties and candidates’. This concluding chapter seeks to map out one way of taking our research forward. We draw on a body of scholarship that models the disciplinary powers of political parties in parliaments and legislatures as solutions to a collectiveaction problem in which the party label is a collective good (Aldrich 1995; Cox 1987; Cox and McCubbins 1993; Kiewiet and McCubbins 1991). The label’s credibility as a signal to voters depends on the extent to which candidates and office holders evaluate the long-term damage to their respective party’s reputation by extensive free-riding, as well as the party’s ability to employ selective incentives to reward loyal behaviour and punish disloyalty. The parties’ main disciplinary instruments are the process of candidate selection and their control over an MP’s career prospects at the parliamentary and governmental level. One of the key implications of such a model for the purposes of this volume is that MPs from visible minorities may have to trade off the benefit of a party label against the costs and/or benefits of demonstrating a certain level of responsiveness vis-à-vis constituents belonging to visible minorities. Based on the literature cited above, Figure 12.1 provides a stylized general model of an evaluation game with four crucial actors: (1) a constituency party, CP, in charge of selecting candidates for elected office; (2) a voter, V, deciding on the candidate or list of candidates presented to him or her by the party; (3) a Member of Parliament, MP, who may or may not belong to an ethnic minority;

and (4) the leadership of the MP’s parliamentary party, PPL. The game starts with the constituency party selecting an aspirant as its official candidate for a parliamentary election, thus conferring the party’s reputation for certain policy positions, responsiveness to minority interests and competence in government on her or him. Whether or not it selects a particular person depends on the aspirant’s reputation, the voters’ expected response to the prospective candidate’s personality and his or her expected behaviour once elected to Parliament. The constituency party selectors cannot be certain of the ‘type’ the chosen candidate constitutes, that is, whether he or she is honest, competent, responsible, loyal or opportunistic. A candidate’s previous involvement in a party or civic association may reduce the nominating body’s uncertainty. Yet, if the person is a recent immigrant, she or he may not have had the opportunity to acquire a longstanding reputation of loyal service to the party. Nomination of a minority candidate may serve as a signal lending credibility to a party’s policy pledges (e.g. claims of an inclusive policy agenda). In a constituency with a strong concentration of voters and activists from a visible minority community, the minority status of the candidate may be advantageous for the party as a whole. In other cases, the nomination of a candidate from a visible minority may be (feared to be) a disadvantage. At the next stage of the game in Figure 12.1, voters have a choice to elect the nominated candidate, or cast their vote for another candidate or party. Their chances to express their preferences over a certain candidate as well as the parties’ policy agendas are constrained by the electoral system (see the section

on Denmark in Chapter 3, and Johannes Bergh and Tor Bjørklund on Norway, Chapter 5, this volume). Again, the choice is influenced by uncertainty about the candidate, this time on the part of the voters. The candidate’s reputation (e.g. as the incumbent) may reduce the voters’ uncertainty. His or her party affiliation may be a relatively credible signal of the candidate’s policy preferences, if the party is able to maintain a degree of discipline and the candidate has a track record of loyalty to the party. The extent to which the candidate’s ethnicity reduces the information asymmetry in the voters’ perceptions depends on some of the contextual factors described in Figure 1.1 (Chapter 1, this volume) (e.g. cultural rights of citizenship or size and spatial concentration of ethnic group). Once in office, the MP chooses a representational strategy. In response to the MP’s observable behaviour, his or her parliamentary party leadership, PPL, can (1) apply party discipline (‘whip’) or (2) allow independent decision-making. When the preferences of a visible-minority MP and his or her party are not identical, the use of both strategies available to the PPL may involve costs as well as benefits. The costs of employing sanctions against a dissenting MP will have to be weighed against the benefits of maintaining a reputation for unity (Laver 1999). The PPL’s expected cost-benefit calculus will be taken into account by the visible-minority MP when choosing her or his representational strategy. The MP may have some leeway vis-à-vis the PPL, if he or she has a strong position in the constituency party or among voters. In some cases, this gives rise to relatively sophisticated individual strategies based on what Necef (2002) called ‘impression management and political entrepreneurship’ in a case study of a Member of the Danish Folketing. Whatever the representational strategy pursued by the MP, at the end of a parliament the members of the constituency party have an opportunity to update their beliefs on his or her suitability and performance, which may influence their decision to re-select or de-select the MP. They are likely to do this, again, in the light of probable voter responses to this process. At the election stage, the voters have their opportunity to judge the MP’s performance, weighing his or her personal record as well as the collective record of his or her party. Thus the game continues. While the diagram in Figure 12.1 resembles a game tree, we have not attached any payoffs to the outcomes. It is more illustrative than exhaustive as far as possible strategies are concerned. What is important here is the theoretical expectation that voters, party organizations and representatives are modelled to use something akin to backward induction to reduce uncertainty about representatives. This will help to better understand the dynamics underlying the entire process of representation as opposed to analyses of merely one stage of this process, and could be analysed as ‘nested games’ or ‘games in multiple arenas’ (Tsebelis 1990). What type of relations would a strategic model of minority representation lead us to predict? And how do predictions based on the stylized model above fare in the light of the evidence assembled in the chapters of this volume? In an expected-utility model of voter participation, minority voters would behave like any other voter: They would turn out to vote, if the product of

the expected benefit of voting and the probability of their vote actually making a difference in attaining this benefit outweighs the cost of voting. In some expected-utility models, authors add a ‘duty term’ to this basic equation, which predicts voter participation to be higher, if voters feel a civic duty to participate (Riker and Ordeshook 1968). The collective chapters on the turnout of voters from minority backgrounds and their party choices do not ‘test’ the fit of such predictions with the data collected in 11 countries (Austria, Belgium, Canada, Denmark, France, Germany, the Netherlands, Norway, Sweden, Spain and the United Kingdom). The country sections generally suggest that the turnout of recent immigrants and their descendants tends to be lower than the turnout levels found among autochthonous voters. While this finding would not be predicted by an expectedutility model of voter participation per se, some of the variations between countries, between different ethnic groups and between voters within such groups are consistent with the model. In a number of countries (Canada, France, Norway and the UK) the gap between minority and majority turnout decreases with increasing levels of socio-economic status and increases with the length of residence. This would be in line with the predictions of an expected-utility model of voter turnout: Citizens with relatively low socio-economic status and relatively little experience as citizens of their country of residence may generally be expected to have lower levels of civic knowledge, perceived efficacy, less information on the differences in expected differential benefits of particular parties in government and possibly a lower ‘duty term’, in Riker and Ordeshook’s (1968) terminology. Some investigations specify alternative mechanisms, which are also generally compatible with the rationalist framework outlined in Figure 12.1. Studies in France demonstrate that political socialization in the parental home (citizenship of parents) plays an important role in determining turnout among secondgeneration immigrants. Studies from Canada and the Netherlands suggest that minority voters with strong personal networks in their respective ethnic communities are more (rather than less) likely to vote. The higher turnout among citizens of Indian origin in the UK is often explained as a result of their strong communal networks in densely populated areas. In a rationalist interpretation, socialization and social capital, memberships of social networks, can reduce information costs, help the dissemination of views about the expected policy benefits of certain parties, provide a context for partisan mobilization strategies and generate intrinsic social rewards or punishment through social disapproval from significant others. Variations in the parties’ efforts in mobilizing ‘the immigrant vote’ – and the nature of the political opportunity structure constraining the parties in their ability to do so – may also be a factor. From the party organizations’ and parliamentary candidates’ vantage point, courting minority voters may involve a high risk for a relatively minor payoff. Sara Claro da Fonseca’s study of German immigrant candidates conjectures that major parties, in particular, are acutely aware of the political risks of politicizing immigration through targeted policies or by promoting immigrant candidates. This may explain why smaller

left-of-centre parties have been successful in attracting relatively large shares of votes from minority communities. While some of the findings on voter turnout may at least be compatible with a strategic framework in which voters, candidates and party leaderships are utility-maximizing actors, the party choices made by those immigrants and members of minorities who do vote in national, regional and local elections are more difficult to accommodate in such a framework. The contributions to Chapter 3 (this volume) show that the loyalty of immigrant and minority voters to particular parties is truly remarkable. In the 2007 French presidential election, for example, 80 per cent of Maghrebin-and African-French voted for the Socialist candidate, Ségolène Royal. Approximately eight out of ten British voters of African or Afro-Caribbean origin voted Labour in the general election of 2005. In the early 2000s, approximately eight out of ten German citizens of Turkish origin preferred either the Social Democrats or the Greens. A similar pattern emerges in nearly all contributions to Chapter 3. Apart from a few idiosyncratic electoral alignments such as the overwhelming and almost unwavering support the German Christian Democrats have enjoyed among ethnic German resettlers from Eastern Europe, the alliance of recent immigrants and their descendants on the one hand and left-of-centre parties on the other has remained remarkably stable over several decades. The surprising finding from a strategic perspective is that the support of immigrants and minority voters for left-of-centre parties seems to be largely independent of policy performance. Messina (2007: 207-8), however, argues that these affinities are not entirely unexpected from a strategic perspective, but are likely to be based on the policies that especially left-of-centre parties have offered to immigrants and their descendants for some time. This may have been reinforced by the early incorporation of immigrants into organizations politically close to left-wing parties (especially trade unions). Whatever the causes of these findings, their consequences are clear: low turnout and strong and stable ties between left-ofcentre parties and ethnic minorities reduce the incentives for candidates to compete for the immigrant vote. Other reasons for the lack of competition for the immigrant and minority vote in many European countries are likely to be found in the structures and norms prevailing within party organizations. Sylvain Brouard and Vincent Tiberj conclude from their experimental study on French voter responses to different types of candidates ‘that diversity candidates are generally accepted, and widely supported by French voters . . . this leaves mainstream parties as the principal barrier to improved political representation for ethnic minorities’. To investigate this further, the second major part of our volume is devoted to this neglected dimension of the process of participation and representation of immigrants and minorities in advanced liberal democracies. In the model in Figure 12.1, the party organizations constitute the central node in a complex multi-level game. Maritta Soininen’s contribution on the barriers to minority candidate nomination in Sweden (Chapter 6) is based on extensive qualitative interviews. She shows that even parties that are ideologically committed to equal opportunities are

struggling to secure the nomination of a proportional number of immigrants on their party lists for elections – unlike their relatively successful efforts to ensure fair representation of women both on Swedish nomination committees and party lists. The reasons for these difficulties resonate in several other chapters on candidate nomination: Whereas efforts to secure a fair share of women on party lists are a norm shared by all Swedish parties (effectively removing the issue of female representation from inter-party electoral competition), this is not the case for immigrants and minorities. The fear of negative responses from majority voters in elections may play a certain part in the parties’ reluctance to nominate many candidates with a minority background. In addition, nomination committees are often concerned about the possibility that an explicit policy of minority nomination may antagonize activists and groups within the parties themselves. Finally, Soininen uncovers a seemingly paradoxical effect of the parties’ shrinking membership base. Like most of their European counterparts, Swedish parties are generally losing members. Rather than improving the chances for immigrants to get nominated under these conditions, however, this decline in the membership base seems to have the opposite effect: As a result of membership losses, some parties have shrunk to a core of loyal but very traditional activists. Immigrants, often younger members who joined their parties more recently, find it relatively difficult to build up the networks and trust necessary for nominations among such older, long-standing members who rely strongly on cues like a credible track record of long-term organizational work in their decisions about candidate nominations. Soininen’s findings are echoed by a number of other contributions to this volume. Interpreting first results from her novel dataset on German candidates with an immigrant background, Sara Claro da Fonseca (Chapter 4) concludes that the two major German parties, in particular, may be reluctant to mobilize immigrants through the nomination of candidates with a minority background, as long as questions of immigration and citizenship remain a matter of intense inter-party electoral competition as well as intra-party contestation. Nevertheless, there are good reasons to believe that the problems identified by Soininen and Fonseca could be overcome through institutional design. In this regard, Bergh and Bjørklund’s chapter on Norway (Chapter 5) demonstrates that the choice of electoral system may be an important factor in overcoming organizational barriers by increasing intra-party competition at the polls. Where preferential voting is permitted under proportional representation (as in Norwegian local elections), the personal attributes of candidates (such as their ethnic background) may become an important additional cue for the voters’ choices. Similar findings were reported on Denmark in Chapter 2, this volume. The low participation of minority voters and the barriers which minority candidates face within the party organizations are crucial factors to explain the disproportionately low ‘presence’ (Phillips 1995) of minority representatives in national parliaments and legislative assemblies. In Part III of this volume, we investigate whether the same strategic (essentially electoral) incentives that work against proportional descriptive representation of immigrants and visible

minorities in liberal democracies are also likely to work against enhanced substantive representation of minority interests through those minority candidates who do successfully master the obstacle course of intra-party competition for nomination and inter-party competition for votes. We do not claim that minority interests could not be articulated and pursued effectively by majority representatives. Studies of the substantive representation of minority interests have had a certain tradition in Congressional research. In their contribution to this volume, Jason Casellas and David Leal (Chapter 8) confirm the findings of Welch and Hibbing (1984) that US Members of the House of Representatives are generally responsive to the share of Latino and African-American voters in their districts. Using data on Congressional voting between 1993 and 2000, they show that the share of Latino and African-American constituents was associated with the voting records of the House of Representatives, whatever the Member’s own ethnicity. This suggests a certain degree of responsiveness of election-seeking Congressional candidates, if the share of minority voters is high. However, the institutional context seems to matter: for reasons Casellas and Leal can only speculate about, there was no statistically significant relationship in Senators’ voting records. Compared to the US, research on the substantive representation of immigrants and minorities in parliamentary systems of government is still in an emerging stage. Given low levels of turnout and high levels of voter loyalty among immigrant voters, MPs (including minority MPs) do not have strong incentives to use Parliament as an arena for the promotion of minority interests. Moreover, many minority MPs consider themselves as representatives of all of their constituents rather than of particular minorities. None the less, the contributions of Karen Bird on Canada, Thomas Saalfeld and Kalliopi Kyriakopoulou on the United Kingdom and Andreas Wüst on Germany do suggest that MPs’ ethnic backgrounds are statistically linked to their behaviour in the respective Chamber. All three contributions on parliamentary behaviour offer some evidence that a significant number of MPs with an immigrant or minority background use some of the ‘spaces’ granted by their parties to turn their personal attributes into political advantage. The precise shape this takes depends on the nature of the political system. Andreas Wüst found that German minority MPs generally respond in a systematic way to the incentive and reward structure provided by the German Bundestag with its emphasis on detailed committee work. Although German MPs with a migratory background are active in all policy areas, there is a clear tendency for many of them to focus on policy areas relevant to immigrants and minorities. In the Canadian and British case, by contrast, there are few rewards for sterling yet largely invisible committee work. Both chapters demonstrate that the personal characteristics of Members of Parliament generally do not override the effect of party discipline on Member votes. Nevertheless, Karen Bird demonstrates how the ethnic background of MPs may influence positions taken in parliamentary debates. Thomas Saalfeld and Kalliopi Kyriakopoulou provide analogous findings with regard to parliamentary questions in the British House

of Commons. The latter chapter also demonstrates the increasing importance of personal websites as a means of sending relatively sophisticated and targeted signals to majority and minority voters. Individually and as a whole, the contributions to this volume demonstrate the fruitfulness of studying the process of democratic representation not only from a structural or institutional, but also from a strategic perspective, which allows potentially to model the entire process of representation from voter preferences – via the parties’ process of candidate selection and nomination – to parliamentary behaviour. Some chapters in this volume are largely designed to consolidate knowledge about behavioural patterns across a larger number of countries (especially Chapters 2 and 3). Others open up new avenues for research on minority representation, which have so far been subject to theoretical speculation rather than rigorous empirical analysis. All of our chapters demonstrate that the study of minority representation offers a fascinating area of scholarly work for students of elections, parties, parliaments and migration.