Introduction
The legislative branch in democratic government is widely believed to be the facilitator of restrictive immigration policies.Footnote 1 Electoral accountability to the public potentially creates a need to take restrictive positions that appeal to voters with nativist attitudes.Footnote 2 Ellermann (Reference Ellermann2021) surmises this tension in liberal democracies, “Given that public opinion on immigration typically favors restriction, popular mobilization at the polls or during referendum campaigns will increase pressure on policy makers to restrict immigration” (p. 21). Concurrently, legislatures increasingly feature policymakers with direct experiences with immigration themselves, a trend likely to intensify as global migration continues to grow.Footnote 3 In the United States, for example, nearly 15% of the 118
${{\rm{\;}}^{{\rm{th}}}}$
Congress are immigrants or children of immigrants.Footnote 4 How does the experience of immigration affect the behavior of elites and how they interact with political institutions? In this paper, we study the dynamics of immigration politics through the lens of migrant legislators in the United States, the United Kingdom, and Canada.
Although conventional wisdom suggests that legislatures are often the driver of restrictive immigration policy, there is surprisingly little scholarship on elite decision-making in the face of personal migration experiences. Existing literature finds deep divides in public opinion between immigrants and native-born citizens (Just and Anderson Reference Just and Anderson2015), but little work empirically tests these differences among political elites. Contemporary scholars of migration have emphasized the differences in rules and norms particularly between democratic systems, attributing differential outcomes to these deviations. For instance, these theoretical models often compare an electorally insulated committee system in the United States and an empowered executive branch in Canada to explain variation in responses to changing international norms on migration (e.g., Triadafilopoulos Reference Triadafilopoulos2010).
In this article, we expand on the literature on social identity and political behavior, arguing that legislators with personal experiences with immigration face structural barriers when advocating for permissive immigration policies. To begin, we present stylized facts about roll-call voting behaviors in Canada, the United Kingdom, and the United States. Through qualitative interviews, we isolate four specific factors that affect the efficacy of immigrant legislators—political party sanction, electoral penalties, sociotropic concerns about the health of political parties, and a sense of duty to immigrant constituents. In the contemporary politics of Canada, the United Kingdom, and the United States, these mechanisms pose unique career obstacles to migrant legislators, while mitigating movement on immigration legislation. Therefore, we find that personal identity actually drives legislators to conform and heed the whip more closely than their native-born counterparts. On the other hand, native-born legislators face “compensatory pressures” from growing migrant constituencies, which paradoxically push legislators without experiences with immigration towards accommodating preferences more than their foreign-born counterparts.
We highlight three particular contributions of our paper. First, to our knowledge, we present the first comprehensive exploration of the role of migration in legislative decision-making, adding to a robust scholarship on the role of elite social identities and their impact on governance (Hutchings Reference Hutchings1998; Tate Reference Tate2003; Burden Reference Burden2007; Carnes Reference Carnes2012; Juenke and Preuhs Reference Juenke and Preuhs2012; Butler and Broockman Reference Butler and Broockman2011; Griffin and Newman Reference Griffin and Newman2007; Broockman Reference Broockman2013). Second, while scholars have highlighted the gatekeeping powers of political parties in electoral arenas as a mechanism by which immigrants are underrepresented in legislatures (Dancygier et al. Reference Dancygier, Lindgren, Oskarsson and Vernby2015), we argue that institutional powers within legislative institutions are also a factor in suppressing the expression of inclusive immigration preferences. Third, we introduce a mechanism by which legislatures can facilitate the passage of restrictive immigration policies or maintain restrictive status quos, adding to literature that more generally speaks to the role of parties in legislative chambers (Cox and McCubbins Reference Cox and McCubbins2005; Aldrich Reference Aldrich1995; Poole and Rosenthal Reference Poole and Rosenthal1997) and specifically discusses migration policy (Ellermann Reference Ellermann2021; Triadafilopoulos Reference Triadafilopoulos2010; Freeman Reference Freeman2006).
This paper draws on a unique mix of methodologies and data sources. To begin, we present analysis of 60,000 roll-call votes on immigration bills in the U.S. Congress, the UK, and Canadian Parliaments between 1997 and 2014, establishing elite voting patterns across party and demographic characteristics. We then present insights from 25 interviews with legislators from the United States, the United Kingdom, and Canada. These interviews were conducted in 2015 and 2016. We test the hypothesis that legislators act and vote strategically, as opposed to sincerely, in anticipation of political party discipline. The data support this inference, building on work that relies on observational data to predict immigration policy outcomes (Wong Reference Wong2017). Our results show how foreign-born legislators are solidaristic with fellow immigrants but, to the extent they have more permissive policy attitudes, they are constrained by legislative institutions and the sense that they must compensate for public perceptions that their foreign origins bias their decision-making. We also explore variation by the proportion of foreign-born in electorates, political party, and bill type. Combined, our findings shed light on an under-explored aspect of immigrant incorporation—their political representation—and reveal a mechanism by which legislators are limited in the expression of their personal political beliefs. We conclude with a discussion of the implications of these gatekeeping powers by party leaders in democratic legislatures.
Legislatures and Immigrants
The notion that the legislative branch is the most accountable to the public is one that is as old as political science itself. Indeed, the foundational premise of representative democracy lies in the premise that constituents hold their representatives accountable for their legislative decisions (Ansolabehere and Kuriwaki Reference Ansolabehere and Kuriwaki2022; Mansbridge Reference Mansbridge2003). In the context of immigration, however, this has often meant the expression and representation of nativist attitudes. The history of immigration policy in many democratic countries is one that is rife with discrimination and restriction. In the United States, for example, the quota system of the early twentieth century and Asiatic Barred Zones exemplify this propensity as does more modern discourse on border securitization and discriminatory admissions policies (Zolberg Reference Zolberg2006).
In contemporary social science, the characterization of the legislative branch as the most easily swayed by nativist preferences is the dominant one. Ellermann (Reference Ellermann2021) dubs the ability to withstand political pressure as “insulation,” arguing that the executive branch and courts serve as champions for more permissive immigration policymaking in comparison with the restrictionist tendencies of the electoral arena. Triadafilopoulos (Reference Triadafilopoulos2010) similarly centers the legislature in a discussion about the differential outcomes between the United States and Canadian immigration policy. Other frameworks incorporate the role of economic benefits and costs, primarily analyzing the role of organized interests (Freeman Reference Freeman2006; Wilson Reference Wilson1989; Wilson Reference Wilson1973).
However, extant scholarship overlooks the rise in legislators that have themselves experienced immigration. Immigration policy provides an important test of political representation. It is an issue that has not only divided legislatures but has also created divisions within political parties and coalitions (e.g., Tichenor Reference Tichenor2009, Anonymous Reference 1). And while many white legislators can trace their ancestry to a foreign origin, particularly in immigrant settler states like the United States and Canada, these same legislators are often generations removed from the immigrant experience, and thus their empathy (or lack thereof) for the plight of contemporary immigrants varies widely. Yet, we know little about the role that the immigrant background of legislators plays in their decisions when it comes to immigration policymaking across countries and institutional regimes, particularly under the twin pressures of constituent demands and legislative caucuses. Addressing this question is the task of our research here.
Immigrants as Legislators
Since Pitkin (Reference Pitkin1967)’s work on The Concept of Representation, scholars have sought to identify how groups, often women and minorities, are represented in legislatures. Indeed, a robust scholarship exists that estimates the effect of shared social identity on governance. Scholars have found evidence that descriptive representation can produce substantive outcomes for Latinos in the United States (e.g., Bratton Reference Bratton2006; Casellas and Leal Reference Casellas and Leal2013; Minta Reference Minta2009; Preuhs Reference Preuhs2006; Santos and Huerta Reference Santos, Huerta and Menifeld2001; Welch and Hibbing Reference Welch and Hibbing1984; Wilson Reference Wilson2010; Hero and Preuhs Reference Hero and Preuhs2009). Similar effects have been found for Black people in an American context (Grose Reference Grose2005; Mansbridge Reference Mansbridge1999; Minta Reference Minta2009; Owens Reference Owens2005). In Canada and the United Kingdom, racial minority MPs tend to address “minority issues” more frequently than non-minority MPs (Bird Reference Bird, Bird, Saalfeld and Wust2011; Saalfeld and Bischof Reference Saalfeld and Bischof2013).
Scholars have also found that diverse constituencies can improve substantive representation for racial and ethnic minorities (Casellas and Leal Reference Casellas and Leal2013; Bird Reference Bird, Bird, Saalfeld and Wust2011; Celis and Erzeel Reference Celis and Erzeel2013; McLendon, Mokher and Flores Reference McLendon, Mokher and Flores2011; Saalfeld and Bischof Reference Saalfeld and Bischof2013; Saalfeld Reference Saalfeld2011; Boushey and Luedtke Reference Boushey and Luedtke2011; Chavez and Provine Reference Chavez and Provine2009; Bloemraad Reference Bloemraad2013; Biles and Tolley Reference Biles, Tolley and Andrew2008; Wilson Reference Wilson2010; Juenke and Preuhs Reference Juenke and Preuhs2012; Griffin and Newman Reference Griffin and Newman2007; Saggar and Geddes Reference Saggar and Geddes2000; Griffin Reference Griffin2014). Moreover, descriptive representation can benefit minorities indirectly. Voters can relate more to legislators who “look like them,” leading to more trust, and thus enhancing political incorporation (Menifield Reference Menifield2001, 6). It can also facilitate communication between constituents and representatives, legitimize the ability of a previously disadvantaged group to lead, and improve the credibility of government in minority communities (Mansbridge Reference Mansbridge1999, 628).
Building off this robust scholarship, there are a number of reasons why the experience of immigration may result in differential preferences on immigration policy. Although there are a litany of theories about why migration occurs, political economists highlight the costs and risks associated with migration (Massey Reference Massey2009; Kimberlin Reference Kimberlin2009). Being a migrant is often associated with an uprooting of political, economic, and social stability, while relocating to environments that are entirely unaccepting. Using the case of Latin American immigrants, Jones-Correa (Reference Jones-Correa1998) highlights gendered differences in adapting to the United States, but emphasizes the significant role of migration on political socialization. Other studies highlight how immigrants develop a greater sense of linked fate and group consciousness than future generations given their shared experiences (Sanchez and Masuoka Reference Sanchez and Masuoka2010). While Dancygier and Saunders (Reference Dancygier and Saunders2006) find little difference in redistributive attitudes between migrants and host country natives, immigration attitudes are found to differ significantly between the two groups (Just and Anderson Reference Just and Anderson2015). In addition to the shared identity of migration, classic theories of cognitive dissonance avoidance may apply in this context (Just and Anderson Reference Just and Anderson2015). In particular, people of migrant origins may identify with and empathetically evaluate new migrants.
In addition to shared experience, legislators with experiences of immigration may feel a heightened level of responsibility to immigrant populations. Classical perspectives in Congressional scholarship, in particular, highlight the genuine sense of duty that legislators feel towards their co-ethnic constituents (Fenno Reference Fenno2003). Further scholarship finds that Black legislators, for example, create concrete improvements in policy outcomes for their co-ethnic constituents, such as better constituent services (Broockman Reference Broockman2013). Altogether, these shared experiences, psychological processes, and sense of responsibility towards fellow immigrants suggest that foreign-born legislators are more likely than their native-born legislators to support immigration-related bills that advance the interests of immigrants.
We also note that Pitkin (Reference Pitkin1967) also distinguishes between descriptive and substantive representation, where the former refers to the similarities between the represented and elected officials and the latter refers to whether representation takes place “in place” of those communities. We believe this paper contributes to this understanding by bridging this gap, by not only focusing on the differential voting behaviors of foreign-born representatives but also the differences by institutional design.
Political Parties as Moderators
In the study of legislative institutions, the role of political parties is paramount to understanding legislative action. We draw on two theories, in particular. First, the theory of conditional party government (henceforth CPG) states that “intraparty homogeneity of preferences and interparty divergence” will determine the willingness of legislators to delegate parliamentary powers to party leaders (Rohde Reference Rohde2013). CPG theory relies on the notion that the institutional structure of a majoritarian legislature is rooted in the goals and preferences of their constituent members. Building on this foundation, Cox and McCubbins (Reference Cox and McCubbins2005) propose their procedural cartel model of legislatures, which similarly asserts the importance of intraparty homogeneity and interparty divergence in the delegation of administrative powers to party leaders. The procedural cartel theory of legislatures adds to CPG by highlighting the positive agenda control powers of party leaders, as opposed to negative punishment powers. Unlike alternative theories of legislative bargaining that focus on individual preferences (Krehbiel Reference Krehbiel1998; Baron and Ferejohn Reference Baron and Ferejohn1989), these theories particularly center political parties and their leaders in the analysis of legislative action.
We believe the focus on political parties is an apt foundation in the analysis of immigration policy. Partisan divides on immigration attitudes has remained a subject of interest to social scientists (e.g., Johnston, Newman and Velez Reference Johnston, Newman and Velez2015; Hout and Maggio Reference Hout and Maggio2021; Wright and Levy Reference Wright and Levy2020; Hainmueller and Hiscox Reference Hainmueller and Hiscox2010). Drawing on the CPG tradition, we argue that political parties affect immigration in two particular ways. First, political parties exercise direct control of the appointment process to key legislative and executive posts, wielding deviations from and incentivizing adherence to the party line. In the United Kingdom and Canada, the governing coalition holds the power in the appointment of Cabinet members. In the United States, the majority party, through their caucus, decides appointments to key committee seats and also supports campaign fundraising and endorsement. Deviation from the party line may involve direct sanctioning by party leaders in the form of removal from key positions in subsequent legislative sessions or significant harm to the future career prospects of a given member. Other models highlight the information costs of deviating from the party brand (Snyder and Ting Reference Snyder and Ting2002). Second, political parties can incentivize discipline in legislative action in the form of norms, fomenting a sense of duty to the organization. As per the procedural cartel theory, legislators care not only about their own re-election (Mayhew Reference Mayhew1974) but also about the electoral fortunes of their parties (Cox and McCubbins Reference Cox and McCubbins2005). In such majoritarian institutions, legislative action is nigh impossible without a sufficient critical mass or coalition (Baron and Ferejohn Reference Baron and Ferejohn1989).
Moreover, we believe that the personal nature of immigration makes it a policy arena that is best suited to explore the tensions between identity and party sanction. As mentioned, the dominant theoretical understanding of parties in legislatures emphasizes the lack of defections from the party line. However, it is unclear how the trend of rising diversity in foreign-born representation would affect these party dynamics. Dancygier (Reference Dancygier2017), for example, argues that party diversification is strategic, particularly emphasizing the electoral dynamics that motivate leadership to include racial minorities in their coalition. Bélanger and Godbout (Reference Bélanger and Godbout2010) further by exploring the dynamics of coalition government, highlighting changes to governing and minority coalitions in response to similar electoral connections. This is particularly concerning given that legislators themselves are navigating a constantly changing public opinion environment on immigration, which may put additional pressures on representatives (Segovia and Defever Reference Segovia and Defever2010). These studies focus on the role that parties play in inviting racial minorities into their membership, but theoretical expectations of how foreign-born legislators balance this tension between personal identity and their behavior are unclear.
The comparative study of immigration policymaking, while emphasizing institutional differences between different legislatures, oftentimes neglects the differences in the power of political parties in a comparative framework. A canonical literature in American politics describes the political party from the early twentieth century to the modern era as an incredibly weak organization compared to parliamentary systems such as those in the United Kingdom and Canada (Shepsle Reference Shepsle2002; Mayhew Reference Mayhew1991; Tsebelis Reference Tsebelis1995). The study of Congressional institutions highlights factors, such as the domination of seniority in the delegation of key roles (Hall and Shepsle Reference Hall and Shepsle2014), in the maintenance of a weak party government. Although the United States features weak parties, the United Kingdom and Canada feature strong parties. Indeed, the threat of sanction and direct removal from party rolls are credible and motivate political behavior directly.
Additionally, the role of political parties affects the level to which individual legislators are accountable to their constituencies. Dancygier (Reference Dancygier2017) theorizes about these dilemmas faced by political parties in their recruitment of racial minorities, highlighting the role that national constituencies play in the calculus of party leaders to diversify their coalitions. The ability of political parties to coordinate such strategies in the United States is doubtful, however. Political parties have much less of an ability to punish; rather, legislators are far more incentivized to respond to constituent pressures. Indeed, if legislators are single-minded seekers of re-election (Mayhew Reference Mayhew1974), we would expect that immigrant legislators are able to not only shirk the party line but also respond to their particular constituencies in their voting behavior.
All in all, the preeminent role of political parties within government, and the subsequent variation in their power, will result in vastly different legislative behavior in immigrant legislators. We hypothesize that the threat of party sanctioning will limit the ability of immigrant legislators to express their positive immigration preferences in strong party states, such as the United Kingdom and Canada, while allowing for more independence for legislators with weak parties, such as in the United States. However, we offer one caveat to this prediction. The relatively smaller effect of party politics on individual legislative action of American legislators does not mean the party is absent—quite the opposite. Instead, we argue that members of Congress actively monitor the ability of political parties to sanction. In the current state of party governance, we argue this threat is simply not credible, when compared to that of parliamentary legislatures.
With specific expectations on the role of immigration experience on individual legislative behavior, and political parties in a comparative context, we are now ready to present our mixed-method empirical strategy and data.
Data and Empirical Strategy
Quantitative Data
To analyze whether foreign-born legislators vote systematically differently from native-born legislators on immigration-related legislation, we first compiled a list of all immigration-related legislation voted on in Canada from 2007 to 2014, in the UK from 1997 to 2014, and in the United States from 2005 to 2014. The case countries were selected because they are immigrant-receiving, liberal democracies with legislatures that contain numbers of foreign-born officials sufficient to support our analysis. They offer variation in institutional regimes—Westminster and Presidential systems—and among the Westminster systems, variable procedures. The examined period was one featuring a great deal of immigration-related legislation before strong partisan views of immigration dominated left and right-wing parties.
The dependent variable in our analysis is votes on immigration bills. Each bill was hand-categorized and then coded based on its valence-permissive or restrictive.Footnote 5 Multiple coders analyzed all bills independently to ensure inter-coder reliability.Footnote 6 The outcome is therefore roll-call votes on immigration policies, which broadly follow after the third reading of the bill. We also disaggregate permissive and restrictive bills based on the content of the bill. We employ this more manual method for categorizing immigration bills, given their general complexity and the difficulty of fitting a multidimensional scaling procedure with subsets of roll-call data (Clinton Reference Clinton2012).
Table 1 highlights the types of bills in our sample, ranging from those that pertain to admissions to those on border security. Our categorization scheme reveals important variation across countries. For instance, while the American legislature has considered mainly interior enforcement policies, the Parliaments of Canada and the United Kingdom have focused much more heavily on refugee and asylum policy and border security respectively.
Bills in the sample by type
Our primary independent variable of interest is foreign-born identity. We compiled original data on the place of birth for each legislator who cast at least one vote in our sample of immigration-related legislation. We consulted primary sources first, focusing on the official biographies of legislators, and used secondary sources to fill in gaps. We also coded whether each legislator in our database was a member of a visible minority-determined on the basis of their declared race, ethnicity, skin tone, and/or surname. However, given that visible minority is nearly collinear with foreign-born, we omit it from analysis.
We start with the country with the highest percentage of foreign-born MPs during the period under study. From 2007 to 2014, 509 Canadian legislators cast a combined 13,054 yes or no votes on 47 different pieces of immigration-related legislation. Of these 509 legislators, 73 legislators, or 14.3 percent, were born outside of Canada. Among these 73 foreign-born MPs, 25 or 34.2 percent are from the Liberal Party, 24 or 32.9 percent are from the Conservative Party, and 19 or 26.0 percent are from the New Democratic Party (NDP).Footnote 7 Moreover, 42 of the 73 (57.5 percent) are visible minorities.
We turn next to the country with the second highest percentage of foreign-born MPs during the period under study. From 1997 to 2014, 1,119 UK legislators cast a combined 26,023 yes or no votes on 61 different pieces of immigration-related legislation. Of these 1,119 legislators, 49 legislators, or 4.4 percent, were born outside of the UK. Among these 49 foreign-born MPs, 26 or 53.1 percent are from the Labour Party, 21 or 42.9 percent are from the Conservative Party, and 3 or 6.1 percent are from the Liberal Democrats.Footnote 8 Moreover, 17 or 34.7 percent are visible minorities.
The country with the largest foreign-born population, surprisingly, has the fewest number of foreign-born legislators among the countries under study. From 2005 to 2014, 741 U.S. legislators cast a combined 14,153 votes on 34 different pieces of immigration-related legislation. Of these 741 legislators, 17 legislators, or 2.3 percent, were born outside of the United States. Among these 17 foreign-born legislators, 10 or 58.8 percent are Democrats, and 7 or 42.2 percent are Republicans. Just fewer than half, 7 or 47.1 percent, are visible minorities.
Using these two measures, we fit the following OLS model for each country:
$${Y_{it}} = {\alpha _t} + \beta \times {\rm{Foreign}} - {\rm{Bor}}{{\rm{n}}_{it}} + {\rm{\Omega }} \times {{\bf{X}}_{it}} + {\varepsilon _{it}}$$
where the dependent variable
${Y_{it}}$
represents a legislator’s particular vote on permissive or restrictive immigration bills and
$i$
indexes the legislator. Foreign-born is the binary variable that identifies place of birth for each legislator.
$\beta $
represents our primary quantity of interest, defined as the differences in voting propensity attributable to the experience of immigration. On restrictive bills, a negative value of
$\beta $
would indicate that there is a negative association between foreign-born status and an affirmative vote on such legislation. In short, foreign-born legislators are less likely to vote for restrictions to migration.
In our main specification, we include a fixed effect for legislative session. Given that Foreign-born status is essentially static, it is not possible to include a fixed effect for each legislator to uncover within-member changes. The aforementioned fixed effect captures differences in voting patterns within legislative session. Therefore, the specification gives us an average estimate of the effect of Foreign-born within a given legislative session.
We also include control variables, denoted by
${X_{it}}$
, that are commonly found in roll-call studies. These include other demographic information, such as age and gender. Additionally, institutional information, such as a member’s tenure in office and political party affiliation, is included. Finally, we also include controls for district characteristics, such as percent foreign-born, and changes in percent foreign-born from previous year for specifications in the United States.
Given that foreign-born status is static, we cluster our standard errors by legislator. Although the literature recommends estimating clustered standard errors by a variety of bootstrap methods (Cameron and Miller Reference Cameron and Miller2015), the large number of clusters gives us confidence that Liang-Zeger standard errors are sufficient in this setting.
Qualitative Interview Data
This article draws on 25 in-depth interviews of former and current members of the United States House of Representatives, the Canadian House of Commons, and the British House of Commons. Using official avenues in each case country, we requested interviews with all current and past, foreign-born MPs for whom contact details were available, along with a selection of native-born MPs who represent districts with high numbers of immigrants and people with recent immigrant origins. We go back twenty years in order to construct our sampling frame, under the logic that legislatures have only recently started to diversify to include foreign-born legislators. Those interviewed responded to our requests and agreed to set time aside to speak. All interviews were administered in-person in Washington, Ottawa, and London between 2015 and 2016—with the exception of four subjects who asked to be interviewed by phone—according to a pre-set but unseen topic guide. We explicitly did not send the topic guide or questions in advance of interviews, in order to facilitate candidness in interviews. Moreover, we were wary of priming or altering the responses of legislators.
Due to the public nature of parliamentary positions and the non-sensitive nature of our questions, interviews were on the record unless a Member requested that a portion of their comments be unattributed. We took notes by hand and conducted questioning in a conversational style to facilitate the most candid responses.
We pursue this qualitative approach to account for the subtleties of legislative behavior, lobbying, and identity construction that cannot be captured by quantitative data alone. Interview questions ascertain the conditions under which preferences of immigrant constituencies are considered and acted upon by representatives. They address the subjectivities of legislative decisions and relationships, rather than impressions of particular minority groups. Interviews also seek to solicit the salience and significance of representatives’ immigrant origins where applicable. They serve as a window into the perceptions and self-understandings of foreign-born and native officials but also Members’ perceptions of different competing pressures. Although not fully abductive in nature (Ternullo Reference Ternullo2022; Timmermans and Tavory Reference Timmermans and Tavory2012), the semi-structured approach of the interviews allows us to both test previous theories of social identity and come to novel conclusions about this process.
Results
Voting Records on Bills
We begin with results from our analysis of roll-call votes from the three legislatures. How does the credible threat of party sanction and electoral backlash affect legislator voting behaviors? Tables 2 and 3 present our main results from our fixed-effect specification.
Foreign-born identity on permissive immigration policy
+ p < .1, * p < .05, ** p < .01, *** p < .001.
The dependent variable is a binary variable indicating a ‘yes’ or ‘no’ vote on immigration legislation. This table reports OLS models with standard errors clustered at the legislator level. All models include a battery of legislator, district, and legislature-level covariates, including district foreign-born proportion, a binary indicator for party, gender, age, tenure in office, and NOMINATE ideology scores. A full table of results can be found in Appendix B.1.
Foreign-born identity on restrictive immigration policy
+ p < .1, * p < .05, ** p < .01, *** p < .001.
The dependent variable is a binary variable indicating a ‘yes’ or ‘no’ vote on immigration legislation. This table reports OLS models with standard errors clustered at the legislator level. All models include a battery of legislator, district, and legislature-level covariates, including district foreign-born proportion, a binary indicator for party, gender, age, tenure in office, and NOMINATE ideology scores. Full table of results can be found in Appendix B.2.
The coefficient estimates on Foreign-Born are statistically significant and in the expected direction for the United States. We find that foreign-born legislators are approximately 12.1 percent more likely and 14.3 percent less likely to vote for permissive and restrictive immigration policies respectively. This association persists even after controlling for political party and the share of district constituents who are foreign-born. These results follow from our hypotheses on the relative lack of credible party sanction in the United States, when compared to the strong party systems of Canada and the United Kingdom. Moreover, it provides quantitative support for statements made during our interviews from members of Congress, who identified a responsibility towards immigrants and their co-ethnic constituents. We consider this evidence that weak party discipline results in increased leeway to express permissive immigration preferences and respond to the foreign-born constituents in each district.
On the other hand, the coefficient estimates on Foreign-Born are null in the case of Canada. We find a null relationship between foreign-born identity and voting on both permissive and restrictive immigration policies. These results follow from our hypotheses about the threat of party sanction and support the conclusions drawn from our qualitative data about the credible threat of punishment.
In the United Kingdom, we find that foreign-born legislators actually vote more for restrictive immigration policy and less for permissive immigration policy. Similar to Canada, we believe these results stem from the system of party sanction that exists in the two parliamentary Westminster style governments. As stated by a number of our interviewees, careers are made by toeing a party line with both advancement and current positions in jeopardy from deviations. However, unlike Canada, the United Kingdom features significant effects in the opposite direction. We interpret these results as evidence that not only do party sanctions factor significantly in the decision-making of immigrant legislators but also force them to compensate in their votes.
Given these differential results in the UK, we explore the effects by specific policy type. Figure B.1 highlights our results in the United Kingdom disaggregated by policy type. On permissive admissions policies, foreign-born identity has a statistically significant negative effect. The coefficient on refugee and asylum policy is negative, but is statistically insignificant at the p < .05 level. We take this moderator as suggestive evidence, owing mainly to the sample size of refugee and asylum policy. On the other hand, both restrictive admissions and refugee policy are statistically distinguishable from zero, indicating that foreign-born identity is associated with decreased probabilities in voting for such legislation.
The quantitative section is intended to give attention to broad associations in foreign-born identity and voting behavior. Certainly, the results here are in line with theoretical expectations of legislator voting behavior. That is, foreign-born representatives vote more permissively and more restrictively on immigration legislation differentially based on the strength of party sanction. The results from the OLS regressions highlight these patterns, but given concerns about omitted variable bias, we caution against a causal interpretation of these estimates. Moreover, there are key limitations to the roll-call analysis in testing the theoretical expectations of the party and descriptive representation literature. For instance, given the near-linear relationship between party and foreign-born identity, our quantitative analysis cannot adjudicate between the differential party behaviors of pro v.s. anti-immigrant parties, in part due to the power issues particularly for foreign-born members of legislatures that are affiliated with anti-immigrant parties. As such, we now turn to our qualitative analysis to not only explore the mechanisms behind the differential point estimates but also more directly address the tension between identity and political party.
Legislative Behavior and Identity
In our qualitative work, we highlight three key findings. First, we show how the strength of the political party affects both rank-and-file members and leaders in each faction, while highlighting that the institutional contexts also induce variation in legislative action of immigrant legislators. Second, we differentiate between the punitive and reward structure of the political party, introducing a typology for understanding legislator behavior in immigration politics. Finally, we highlight the role of electoral pressures that affect legislators across the board, which along with party pressure limits the ability of immigrant legislators to advocate for migrants.
Party and Whipping Institutions
In the Westminster system, grave consequences await legislators who defy the Party Whip. For junior cabinet members, the ramifications of shirking the party line are not only understood but also especially real. A Tanzania-born Conservative Party parliamentary secretary and MP from Calgary, at the time of our interviews, Deepak Obhrai articulates this sentiment, “During our meetings in the caucus, we do speak our mind. And there have been issues when I speak up. The majority of people who are anti-immigrant are ignorant. They don’t know what being an immigrant is about. It’s easier for you to change the system from the inside. And [previously] I was very vocal in the [now defunct] Reform Party caucus, and you could see their [platform] move in that direction. I did not agree with the Foreign Criminals Act, and for that I was put in the penalty box. I went against the government—I abstained—and all my colleagues knew why. But I was not allowed to answer questions in the House for three days—the penalty box.”
Obhrai’s comments underlie a common trend in parliamentary governments. Depending on Party conventions, junior ministers who stray face setbacks in their privileges, status, and career prospects. Some are asked to resign from leadership. Accordingly, British and Canadian respondents consistently cited the rigidity of the Westminster system, when we inquired about their voting records. Chungsen Leung, a Taiwanese-born MP from Toronto and Parliamentary Secretary for Multiculturalism at the time, says, “For example, if these were free votes, I would have voted differently. But I give up a little bit of my freedom when I was appointed as Parliamentary Secretary. That’s the way the cookie crumbles. It’s a matter of duty.”
This dynamic, while pervasive, does have exceptions when it comes to bills with particularistic benefits to individual districts. In the Westminster system, candidates therefore can only distinguish themselves on free votes when they are not constrained by the Party Whips. “On government legislation, we don’t really have free votes,” says Liberal John McCallum, former chief economist for the Royal Bank of Canada, Minister for Immigration, Refugees, and Citizenship and, at the time of the interview, an MP for the eastern suburbs of Toronto. “And very rarely, a Member will vote against their party or be deliberately absent. There are only free votes on private members’ bills, but they’re not totally free. Among Liberals, they are, but people still feel awkward if they vote against the party. […] There’s not much boat-rocking that goes on.” Here, McCallum illuminates a somewhat symbiotic relationship between the governing coalition and the minority party. Indeed, what he dubs “private bills” give legislators some leeway in expressing their true policy preferences, but they are only able to act sincerely on bills introduced by the opposing party.Footnote 9
Obhrai also shed light on motivations beyond the threat of credible punishment, that of a duty to his political party. He continues, “People need to understand that you must work inside the system. I’ve been here for 18 years. If you don’t have discipline as a party, multiple views come in. You work within the system.” Obhrai’s sentiments as a backbencher apply also to that of party elites. Leung articulates, “I have had to be loyal, and some of [my votes] have not been what I would call ‘pro-immigration’. But sometimes you have to say that ‘I’m a Canadian first’, and I have to follow the mainstream.” Critical in Leung’s assessment and previous comments is the invocation of national identity and the need to appeal to as broad of an electorate as possible, which hints at a responsibility he feels as a party leader.
In the American context, however, the threat of party sanction is less credible. Legislators have much greater liberty to cast rebel votes that reflect their personal beliefs or adhere to those of constituents. Party Whips initially merely count votes and only press legislators to fall into line when voting outcomes are uncertain. While American parties incentivize legislators’ conformity with leadership positions and committee chairs, they are far more lenient and legislators are expected to adhere to the bespoke preferences expressed by their constituents. “On immigration, I bucked the caucus many times,” says Ciro Rodriguez, a Democrat who represented districts in south and southwestern Texas for six terms. “And they accused me of selling out. I’m not one of these guys who won their district with 90% of the vote. And sometimes I would go off-message; that’s what got me in trouble.” However, it is precisely through such relatively independent voting that candidates may accommodate their constituents’ preferences.
Identity and Re-election
While the party looms large in the minds of legislators, they are also acutely cognizant of their own racial identities. In the context of native-born members of Congress and Parliament, this oftentimes means symbolically representing their constituents. “It’s about learning the people, the culture, the language,” says Kirsty Duncan, a native-born, Liberal MP who represents one of Canada’s most diverse ridings around Toronto’s Pearson International Airport. “I can get by in Punjabi, Hindi, Urdu, Somali and Tamil. I have to. From the time I got elected, I’ve been learning. We have a number of interns, and every day the whole office had to learn a sentence in the different languages.” Duncan says that her goal is to give a speech entirely in Punjabi. None of the Indian-origin representatives we interviewed expressed an interest in learning any Mandarin, Tamil, or Ukrainian, despite the presence of such language speakers in their constituencies.
Precisely because they do not descriptively represent their constituents, native-born white MPs sense a greater need to demonstrate the ways they may symbolically represent such interests. Sensitive to accusations of “white privilege,” they reported exerting special effort to combat the image that they benefited from establishment cronyism, to demonstrate their acquaintance with diverse constituencies and their qualification to represent them. “I work them hard; I show up at the mosque between elections; not just at elections,” says Liberal John McKay, a sixth-generation Canadian who represents the Toronto suburb of Scarborough. “I have some feel for their issues. So ironically, it’s easier for me to be the face representing their community to the wider world saying things like, ‘Relax, they’re not terrorists’, and, ‘The burqa is not going to kill you’. So I bring something that someone from their community cannot bring. I see my role as a bridging job between populations and colleagues. I’m everybody’s second choice. But you can replace me in a heartbeat. If you’re not disciplined to the constituent, you’re not going to be around very long. If I had been a representative for Halliburton [rural Ontario], which is not diverse at all, I would not be as sympathetic as I have been.” While McKay therefore acknowledges that he calculates the way he accommodates political attitudes in his riding, he also notes his ability to rebut prejudices between and about immigrant communities.
Native-born legislators are also wary of the electoral wage paid by ethnic affinities, and seek out ways to overachieve. However, John McCallum has found that running as an immigrant-friendly white candidate has allowed him to solicit the support of multiple ethnic groups and transcend their boundaries. “If I had to run as a WASP, you would have hordes of ethnic communities lining up behind their candidate, and it would have been a problem,” says McCallum. “I was a parachute candidate, and once I was chosen, I won large majorities. […] But also, a white guy like me is kind of neutral. If you’re Chinese, you might not want to vote for someone Indian. But I also make efforts to get to know the different communities. I’m least successful with the Caucasian community in terms electoral support, fundraising, and prominence. My wife is Chinese, and we didn’t get married for political reasons, but it didn’t hurt. […] Most of the time my Conservative opponent is a non-white person.” Implicitly, this opposition tendency reinforces native-born MP’s sense that their ethnicity is their vulnerability, leading to further pressures to compensate. Although McCallum’s comments may be specific to his representational style, they dovetail with other work on the role that racial affinity towards racial minorities plays in parliamentary electoral politics (Besco Reference Besco2015).
Among foreign-born representatives, compensatory pressures manifest as a sense of obligation to exhibit their independence from their symbolic affinities, and to exhibit their independent qualification as a legislator. “I am elected by my constituents, and privileged to be a member of Government,” says Shailesh Vara, an Indian-born Conservative MP who represents Northwest Cambridgeshire in the British House of Commons. “But I have a duty to do what’s right rather than to do what’s driven by emotion.” The Calgary Conservative Deepak Obhrai says, “In the initial stages when you come out for office, electoral campaigns are based on—I hate to say this—ethnic and racial lines. Across all groups. People like to see their own. I had to reach out to build my base. The way I was successful was to become involved in the community associations to overcome ethnic and racial lines. In my riding there is no Indian community. Throughout my career, I’ve been challenged on those ethnic and racial lines. I don’t talk about my foreign background. I just try to be policy-conscious.” As a contrast, native-born legislators who represent diverse districts are unsure whether being “policy-conscious” would be sufficient.
A number of foreign-born respondents described experiences in which their promotion or qualifications were discounted because of their ethnic affinities. Chungsen Leung says, “When I first ran in 2000, some constituents said they want to make sure I speak the language. My campaign manager was quick to the draw and said, ‘Mr. Leung speaks five languages; which did you want to check?’ You know, you cannot run away from your skin color, the color of your hair, your eyes. […] You do feel a glass ceiling because you’re a visible minority. You’re never quite inside the inner cabinet, the boardroom seat where people make the big decisions. That is true in [the Conservative] caucus as well.” Indian-born British Labour MP Virendra Sharma believes that such discounting also emerges within caucuses. “Inside the Party group meetings, you can become a specialist in a field, like how to retain the support of [South] Asian and Black communities,” he said. “The disadvantage is that you become compartmentalized. You’re good at race issues, rather than mainstream issues. These are earmarked roles for minority communities. Why not health care or local government?” (e.g., Saggar and Geddes Reference Saggar and Geddes2000)
A South African-born Labourite, Peter Hain explained that foreign-born British MPs not only feel pressure to adhere to British traditions; they are also asked to adapt to the customs of the British class system. “Class is very dominant in the Labour Party as well, despite starting off as the voice of working-class people in Parliament. That’s part of Labour’s DNA. Being outside the class system, I didn’t feel the need to defer to working-class people because of some vestigial guilt. I have also seen all kinds of deferrals to members of the aristocracy and the upper classes.” While Hain says that he has resisted the pressure of class integration, he also says that he subsequently did not have party leadership aspirations.
Wary of being discounted by virtue of their ethnic affinities in diverse districts, foreign-born MPs sense a greater need to demonstrate loyalty to their party and country before their countries of origin—a qualification with which native-born MPs did not concern themselves. “I don’t bring it up,” says Devinder Shory, an Indian-born Conservative MP for Calgary. “I have not felt ignored once I showed I wanted to be integrated.” Ciro Rodriguez, the former Democratic Congressman from Texas says that he never “played up” his Mexican heritage, despite being the only Mexican-born Member of Congress. “You play up things that can help you, and I never thought that it would help me. I did put it on my website for a while, but then took it down. In Texas, people like my wife don’t like being asked “When did you come over?” The [Tejanos] will say “‘Hell, I didn’t cross the river; the river crossed me.”
Still, this compensatory pressure is balanced with senses of obligation to foreign-born Member’s ethnic groups, particularly in the United States. Ahn “Joseph” Cao, was a one-term Republican congressman from Louisiana and the United States’ first Vietnamese-born Member of Congress. “I had two bodies of constituents,” he says. “One were the people I represented in the second district [of Louisiana]. And the other are Vietnamese people from all over the United States. They looked at me as a connection to the government. Whenever there was an issue of importance to Vietnamese people—whether or not they were part of my district—I was asked to get involved by other members and by the Vietnamese community. So you had 2.5 million Vietnamese Americans looking to a first-term congressman to solve all their problems.”
In diverse districts, therefore, foreign- and native-born legislators often want precisely what each can’t have—the credentials that accompany each other’s identity as an immigrant or as a native. Accordingly, each reports modifying behavior to compensate for their perceived inadequacies. For foreign-born legislators, they see this process of triangulation as a means of qualifying for governmental authority despite their appearances and accents. For native-born legislators, they see this process as a means of reassuring diverse constituents that they understand their experiences, attitudes, and can be trusted to translate them to power. In the end, the effect of electing foreigners is neutralized.
Table 4 summarizes the themes that emerged in our qualitative work. In particular, we find support for our hypotheses on the punitive role of party politics, electoral backlash, and duty to party and immigrant communities.
Factors for immigrant legislators
Discussion
In the case of foreign-born legislators, our analysis highlights the deeply personal and informed consideration that legislators of immigrant backgrounds undertake when voting in this policy space. However, we also uncover direct barriers to their ability to act on these preferences. Our paper makes three primary contributions. First, we highlight the political party within legislatures as a key actor in suppressing advocacy by immigrant legislators through the threat of sanction and a cultivated sense of duty to the party itself. Our quantitative and qualitative results point to differences between Westminster parliamentary systems, where the party rules, and the American system, where the party is historically weak. In particular, our analysis argues that political parties not only have gatekeeping power in the selection of candidates Dancygier (Reference Dancygier2017) but also within the legislature itself through key influence over executive positions and career advancement.
However, our contributions are not limited to adding to this literature on parties and sincere preferences, as we find that legislators with personal experiences with migration are driven by these institutions to conform and heed the whip more closely than their native-born counterparts. As such, the accountability pressures from growing migrant constituencies are more likely to influence native-born legislators to become greater champions for immigrant rights and break from their parties on this pivotal issue.
Third, our results emphasize a key barrier for legislators to act and invoke their social identities in government. A bastion of research on race in political science is the connection between elite racial identity and representation of racial minorities (Hutchings Reference Hutchings1998; Tate Reference Tate2003; Burden Reference Burden2007; Carnes Reference Carnes2012; Juenke and Preuhs Reference Juenke and Preuhs2012; Butler and Broockman Reference Butler and Broockman2011; Griffin and Newman Reference Griffin and Newman2007; Broockman Reference Broockman2013). However, the literature often comes to differing conclusions about the role of elite social identities and their impact on governance. Our findings offer nuance to this question by highlighting the power of the environments in which legislators operate. In introducing institutions in the study of this question, we connect the study of immigration policymaking to that of organizational legislative studies (Cox and McCubbins Reference Cox and McCubbins2005; Aldrich Reference Aldrich1995; Poole and Rosenthal Reference Poole and Rosenthal1997).
We recognize the limitations of our analysis here. First, our conclusions are only as good as our data. The phenomenon of migrant legislators is a particularly nascent one, in part due to the relatively recent liberalization of immigration policy in Canada, the United Kingdom, and the United States. While our evidence here offers unique insight into the lives of an understudied proportion of the political elite, more work is required to explore the intersection between social identity and governance in a more general way. We believe that the integration of institutions into the study of immigration enriches our understanding of not only the politics of migration but also the nature of legislatures in democratic governance.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/rep.2025.10013.
Acknowledgements
We thank Paddy Torsney and Ratna Omidvar for facilitating interviews with several subjects. The paper received valuable feedback and guidance from Jason Casellas, David FitzGerald, Matthew Wright, Anna Boucher, Jennifer Victor, and Albert Fang. The authors are also deeply grateful for the time each legislator shared with us when we conducted the interviews. Across party lines and three countries, they recognized the importance of this work and wished to contribute their voice and perspective. We thank Anna Coleman, Miguel Higuera, and Aaron Roper for excellent RA work.
