Political party ideology and immigration policy reform: an empirical enquiry

Katharina Natter; Mathias Czaika; Hein de Haas

doi:10.1080/2474736X.2020.1735255

Political party ideology and immigration policy reform: an empirical enquiry

Natter, Katharina; Czaika, Mathias; de Haas, Hein 2020-01-01 00:00:00 POLITICAL RESEARCH EXCHANGE 2020, VOL. 2, 1735255 https://doi.org/10.1080/2474736X.2020.1735255 RESEARCH ARTICLE Political party ideology and immigration policy reform: an empirical enquiry a b c Katharina Natter , Mathias Czaika and Hein de Haas a b Institute of Political Science, University of Leiden, Leiden, Netherlands; Department for Migration and Globalization, Danube University Krems, Krems, Austria; Department of Sociology, University of Amsterdam, Amsterdam, Netherlands ABSTRACT ARTICLE HISTORY Received 29 November 2018 What drives the restrictiveness of immigration reforms? To what Accepted 8 January 2020 extent does the political ideology of parties in government and parliament matter? Drawing on immigration policy data oﬀering KEYWORDS unprecedented historical and geographical coverage, we analyse Immigration policy; political the drivers of immigration reforms in 21 Western immigration parties; migration; policy countries between 1970 and 2012. Our results show that there is determinants; policy no robust eﬀect of the political ideology of governments and diﬀusion parliaments on the overall restrictiveness of immigration reforms. Partisan eﬀects are limited to certain migration policy areas, primarily to integration policies, and to certain migrant groups, particularly asylum seekers and undocumented migrants. In contrast, political party ideology does not fundamentally shape decisions on the core of immigration regimes, such as entry policies or policies towards labour and family migrants. Our ﬁndings also showcase the importance of international policy diﬀusion and of trade-oﬀs between reforms in diﬀerent policy areas. Overall, the analysis highlights that although immigration is subject to heated debates in the public sphere and extensive political bargaining, the actual policies enacted seem primarily driven by factors such as economic growth, social welfare protection and the structure of political systems that are largely independent of the political ideology of parties in power. Introduction It is widely assumed that left-wing governments tend to be pro-immigration and that right-wing governments are generally in favour of restrictive immigration policies. This assumption can be questioned on two theoretical grounds: First, left- and right-wing parties tend to be internally divided on immigration. Typically, right-wing economic market liberalists and left-wing cosmopolitan-humanitarian streams are thought to be pro-immigration, while right-wing cultural conservatives and left-wing economic protec- tionists are expected to adopt more restrictive policy positions (Perlmutter 1996). As a con- sequence, we can expect party positions to vary according to migrant categories and CONTACT Katharina Natter [email protected] © 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (http:// creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 K. NATTER ET AL. policy issues at stake. For instance, right-wing parties may favour labour immigration, while left-wing parties may see immigration of labour migrant as harming the interests of native workers. Second, the need to compromise and bargain at diﬀerent levels of policymaking – between the executive and the legislative, between parties of a coalition government, or between governments and important lobbies – may drive governments to water down initial ideological positions. This can go both ways: The inﬂuence of business lobbies may lead parties with a relatively ‘tough’ immigration rhetoric to ultimately adopt less stringent policies, while parties with liberal positions on immigration might decide to abandon some of their policy proposals in an electoral strategy to not lose voters worried about high immigration. For ideological and strategic reasons, we can thus expect party positions on immigra- tion to be systemically incoherent, varying according to the policy issue at stake and leading to the adoption of ‘mixed’ policy packages that are characterized by trade-oﬀs and compromises reﬂecting the diversity of interests within and between parties. This begs the question whether, ultimately, there is a diﬀerence in immigration policy reforms enacted under left- or right-wing dominated governments or parliaments. It also shows the need to assess the overall importance of partisan eﬀects compared to other immigration policy determinants, such as the inﬂuence of economic factors. In other words: To what extent does the political orientation of parties in power aﬀect changes in immigration policy restrictiveness? A number of prior studies have provided initial insights into the inﬂuence of party orien- tations on speciﬁc immigration policies (for a review, see: Bale 2008; Meyers 2000, 1257– 1261). Such studies generally yielded mixed results and were often limited to one country or a relatively short time period. There are two main exceptions: Givens and Luedtke (2005) studied immigration laws in France, Germany, and the UK over the 1990–2002 period and showed that partisanship does not signiﬁcantly aﬀect the restrictiveness of immigration laws, but is a signiﬁcant predictor for the restrictiveness of integration laws. More recently, Abou-Chadi (2016) analysed immigration policies of nine European countries, Australia and Canada between 1980 and 2006, showing that left-wing governments are more likely to pass liberal immigration reforms, but only if they are in control of both executive and legislative powers. However, the overall lack of pertinent data on immigration policy has limited the ability of researchers to conduct systematic, quantitative comparisons of the eﬀects of party politics on immigration policy restrictiveness over longer time periods and a large number of countries. In particular, it has obstructed an analysis of partisan eﬀects on diﬀerent areas of migration policy or across diﬀerent migrant groups. Given the selective nature of immigration policies (de Haas, Natter, and Vezzoli 2018), such diﬀerentiation is however crucial. In addition, a one-sided focus on the ideological orien- tation of political parties may easily overestimate their importance in shaping immigra- tion policies. A proper assessment of partisan eﬀects can only be achieved if we do so in conjunction with analysing other plausible drivers of immigration reform, such as econ- omic growth, unemployment levels, the structure of political system or the international diﬀusion of policy practices. This paper seeks to address these limitations and expands the empirical and theoretical scope of research on this issue (1) by investigating how partisan eﬀects shape diﬀerent types POLITICAL RESEARCH EXCHANGE 3 of immigration policies in distinct and potentially heterogeneous ways and (2) by embedding the analysis of partisan eﬀects within a broader assessment of structural determinants of immigration policy reform. To do so, this paper draws on the DEMIG POLICY database and investigates the drivers of immigration policy reform in 21 Western immigration countries over the 1970–2012 period. Party politics and immigration policy reform Since the 1990s, immigration has become an increasingly important topic into most Euro- pean and North American elections. Confronted with the rise of far-right populist parties, parties both at the left and the right of the political spectrum have sought to win over voters – or at least not to alienate their traditional constituencies – through promising tougher border controls, as well as stricter access for migrants to labour markets or welfare provisions (Alonso and da Fonseca 2011; Davis 2012). Restrictive discourses on immigration are thus not a prerogative of the right. Nonetheless, political parties continue to diﬀer signiﬁcantly in their discourses on migration, leaving the impression that right- wing parties are more restrictive towards immigration than left-wing parties. For instance, while the conservative Christian Democratic Union (CDU) and far-right Alternative for Germany (AfD) in Germany opposed dual nationality and the granting of local political rights to foreigners in the run-up to the 2017 parliamentary election, the Social Democratic Party (SPD), the Green party, the left Die Linke and also the liberal Free Democratic Party (FDP) supported such measures. Campaigning for the 2015 UK parliamentary elections, most parties promised to cap legal immigration, albeit to diﬀerent degrees: Labour suggested to only cap the number of non-EU workers, the Conservatives announced it would keep annual net immigration within the ‘tens of thousands’, while the far-right UK Independency Party (UKIP) suggested a cap of 50,000 highly skilled workers per year together with a complete ban on low-skilled workers for a 5-year period. A cursory look at the history of immigration policymaking, however, gives reason to question the idea that right-wing parties are more ‘anti-immigration’ than left-wing parties when it comes to the actually enacted immigration policy reforms. For instance, under the inﬂuence of industry lobbies, right-wing and centrist parties in the 1950s and 1960s often favoured (‘guest-worker’) labour immigration from Mediterranean countries to Western Europe. Right-wing parties were also more lenient to enshrine family reuniﬁca- tion rights for ideological reasons related to conservative family values (Bonjour 2011). Left-wing parties, on the other hand, were often more wary towards immigration, because this was seen as undermining the position of native workers and trade unions. At the same time, they advocated socio-economic and citizenship rights of already- entered migrants to defend their rights as workers and to prevent that their presence would cause a downward pressure on wages and employment conditions of native workers (Haus 1999). We, therefore, cannot simply classify the left as ‘pro’ and the right as ‘anti’ immigration. In fact, the migration issue does not neatly cut across the left–right spectrum (Massey 1999, 313; Odmalm 2011, 1076–1077; Schain 2008, 468). Sciortino (2000, 225) argued that migration divides each party internally between those close to the party’s economic tradition and those close to its socio-cultural tradition. Perlmutter (1996, 378) emphasized that left-wing parties have to accommodate conﬂicts between unions who traditionally 4 K. NATTER ET AL. Table 1. Internal party divisions on migration. Left Right Restrictive migration policy reform Dimension Economic tradition Socio-cultural tradition Ideology Market protectionism Value conservatism Actor Labour unions Cultural conservatives Liberal migration policy reform Dimension Socio-cultural tradition Economic tradition Ideology International solidarity Market liberalism Actor Liberal and ethnic groups Employer lobbies favour restrictive policies and human rights advocates or ethnic groups lobbying for more open policies; while right-wing parties tend to be divided between employers favouring immigration and cultural conservatives asking for immigration restrictions. Analyses of Spain (Wutts 1998) or the United States (Tichenor 2002, 2008; Zolberg 2006) have shown the prevalence of such ‘strange bedfellow’ coalitions between businesses and human rights advocates or between labour unions and cultural conservatives that cut right across the typical partisan cleavages. Table 1 summarizes these internal divisions. Several scholars have also argued that because of such potential internal conﬂicts and the electoral risk of their public exposure, mainstream parties have long tried to avoid public debate on migration (Perlmutter 1996; Triadaﬁlopoulos and Zaslove 2006). This ties in with the classic argument by Freeman (1995) that migration policies are often made through client politics in closed political arenas, which can only be inﬂuenced by highly organized groups. Particularly since the end of the Cold War in 1989 and the demise of the perceived ‘communist threat’, however, the political salience of immigration in the public sphere has increased (Freeman 2013; Geddes 2003; Sciortino and Colombo 2004; Van Der Valk 2003). Politicians and the media have frequently presented immigra- tion as a threat to employment and the welfare state, to social cohesion and cultural integ- rity. Since the attacks on the Twin Towers and the Pentagon in the United States in 2001, politicians have also linked migration to terrorism and have cast immigrants as an overall security threat. The concomitant rise of far-right anti-immigration parties has further shifted immigration out of client politics and into the public domain (Davis 2012; Givens and Luedtke 2005, 7; Perlmutter 1996, 377–378; Triadaﬁlopoulos and Zaslove 2006, 172). The increasing politicization of migration may thus have diminished the inﬂuence of business lobbies and civil society groups while increasing the role of party politics on immigration policymaking. Yet, immigration policymaking is not an exclusively domestic aﬀair. National govern- ments closely monitor policy developments in neighbouring or partner countries, facilitat- ing the spread of policy practices regardless of the party in power (Cornelius et al. 2004; Meyers 2002). For instance, if several countries introduce policies to attract the high- skilled, this may motivate other governments to introduce similar policies as part of an international competition for talent (Czaika 2018). The other way around, the introduction of more restrictive asylum policies by some countries may motivate other governments to also restrict asylum out of fear of a large-scale increase in asylum applications (Hatton and Williamson 2004). It seems plausible that globalization and the formation of regional gov- erning systems such as the European Union have increased the importance of inter- national policy diﬀusion (Block and Bonjour 2013; Lavenex 2001). The politicization of immigration and the growing importance of international policy diﬀusion suggest that parties’ impact on policy may have changed over time, but the POLITICAL RESEARCH EXCHANGE 5 two trends seem to work in diﬀerent directions. On the one hand, the growing political salience of migration and the need of mainstream parties to position themselves more clearly on the political spectrum suggests an increasing polarization of party ideologies over time. On the other hand, the growing importance of supra-national policymaking and the concomitant convergence of policy trends can be expected to decrease party inﬂuences on migration policymaking. To further complicate the picture, immigration policies are not homogenous but typi- cally consist of ‘mixed bags’ of policy measures targeting diﬀerent migrant groups and policy areas in distinct ways. Large-scale immigration reforms are rarely purely restrictive or liberalizing, but generally comprise measures that grant certain entry and post-entry (integration) rights to certain groups while closing opportunities for others (de Haas, Natter, and Vezzoli 2018). We know for instance that entry and post-entry rights of most labour and family migrants have generally increased since 1945 partly under the inﬂuence of human rights law and shifting ideologies (Bonjour 2011), while border con- trols and policies targeting unauthorized migrants and asylum seekers have generally become more restrictive (de Haas, Natter, and Vezzoli 2018). In this vein, Money (1999, 37) found that immigration control (entry) and immigrant integration (post-entry) policies follow diﬀerent political logics: While integration splits political parties neatly between the left and the right, pro- and anti-migration control forces exist both within the left and the right. This shows that migration policies can be contradictory and incoherent ‘by design’, because they are inﬂuenced by diﬀerent ideologies and interests, and because they target various immigrant groups through selective immigration policies. Immigration policymak- ing may, therefore, become subject of bartering, in which political parties compromise to guarantee the immigration of their ‘favourite’ immigrant group. This suggests that the restrictiveness of immigration policy reform enacted by a speciﬁc political party is likely to vary according to the type of policy (border controls, entry or inte- gration) and in terms of the migrant categories targeted (such as high- and low-skilled labour migrants, family members, or asylum seekers). We could, for instance, expect that left-wing parties favour more liberal policies towards undocumented migrants and refugees because values such as international solidarity and universal human rights are central to their party ideology, while right-wing parties favour immigration of both lower- and higher-skilled workers because of the inﬂuence of business lobbies. Similarly, left-wing parties are likely to favour integration policies that grant labour, welfare and citi- zenship rights to settled migrants, while right-wing parties may oppose such policies to maintain a more malleable, ﬂexible and ‘returnable’ immigrant workforce. Data and methodology To test these hypotheses and gain more systematic and comprehensive insights into the role of political party ideology in shaping the restrictiveness of immigration policy reform, this paper investigates 21 Western immigration countries over the 1970–2012 period, drawing on the DEMIG POLICY database. The database deﬁnes migration policies as the ‘rules (i.e. laws, regulations and measures) that national states deﬁne and [enact] with the objective of aﬀecting the volume, origin, direction, and internal composition of […] migration’ (Czaika and de Haas 2013, 489). For the 21 countries included in this paper, DEMIG POLICY recorded more than 2600 migration policy changes over the 6 K. NATTER ET AL. 1970–2012 period. The database records changes in ‘policies on paper’ and does therefore not capture the leverage bureaucrats (Eule 2014; Infantino 2010) and private actors (Gute- kunst 2015) have in the implementation of policy measures. Taking this into account would be a future valuable contribution to the literature. Because migration policy reforms generally consist of diﬀerent (and potentially con- tradictory) measures, DEMIG POLICY disaggregates policy packages by breaking them down into individual policy measures. Each policy measure was coded to identify the relevant policy area (border and land control; legal entry and stay; integration and post-entry rights; and exit and return regulations) and the migrant category targeted (e.g. high-skilled workers, low-skilled workers, family members, undocumented migrants, asylum seekers). This enables the analysis of policymaking trade-oﬀs between diﬀerent policy areas or migrant groups. For more information on the concep- tual and methodological underpinnings of DEMIG POLICY, see de Haas, Natter, and Vezzoli (2015). The dependent variable is the change in restrictiveness introduced by an immigration policy measure enacted in a speciﬁc country in a given year. Each policy measure was coded as a change towards more restriction (+1) or less restriction (−1), according to whether it made the existing legal framework more or less restrictive, that is, whether it increased or decreased the rights granted to the migrant group targeted by that policy change. This focus on changes in restrictiveness allows to explore the drivers of immigra- tion policy reform regardless of the level of absolute openness or restrictiveness of a coun- try’s immigration regime. To add nuance, we also weighted changes in restrictiveness according to their magnitude as ‘ﬁne-tuning’, ‘minor’, ‘mid-level’ or ‘major change’, leading to a 9-point change in restrictiveness scale from −4to 4. To provide a robustness check, all analyses in this paper were performed using both the weighted and unweighted variables. With regards to independent variables, our main variable of interest is the political orientation of parties in power. We have consistently run all model speciﬁcations using two alternative measures of parties in power: government composition and parliament composition. First, we used a variable from the Comparative Political Data Set (CPDS) that measures the composition of governments since 1960 through the percentage of cabinet posts held by each party in government (Armingeon et al. 2017). The percentage of each party in government is weighted by information on its right-left ideological pos- ition as provided by the Comparative Manifestos Project (CMP) dataset, which codes the party programmes of all electorally relevant parties in 54 countries since 1945 (Volkens et al. 2017). High (low) scores of this government composition variable represent govern- ments dominated by right-(left-)wing political ideologies. Second, we constructed a parlia- ment composition variable in a similar way, weighting the right-left position of parties represented in Parliament (provided by the CMP dataset) by the number of parliamentary seats they hold, with positive (negative) numbers indicating a dominance of right-(left- )wing parties in parliament (see Appendix Table A1 for descriptive statistics). The two vari- ables are signiﬁcantly correlated (0.622) as the government composition partly reﬂects the composition of the parliament, depending on the electoral system. There are exceptions, such as in the case of Austria over the 1990s and 2000s, where the government was a coalition between social-democrats and conservatives, but the far-right had a strong voice in the parliament. POLITICAL RESEARCH EXCHANGE 7 We also explored the role of international policy emulation and diﬀusion in explaining national immigration policy reform (Shipan and Volden 2008) by constructing a so-called spatial dependence variable (Neumayer and Plümper 2010). This variable captures immi- gration policy changes in all other countries k in the dataset and tests the inﬂuence of (weighted) trends in general immigration policy trends on national immigration policy- making y of country i at time t according to: it y = m v y + bX + 1 . (1) it ik kt it it The weighting matrix v proxies the relative connectivity between N (=45) countries i ik and N countries k using linguistic distance as connectivity variable (Neumayer and Plümper 2010). The spatial autoregression parameter m identiﬁes the eﬀect of the spatially lagged dependent variable on policy changes y . International policy diﬀusion it can impact national policymaking both towards more and less restriction: One the one hand, the increased role of supra-national institutions may limit the inﬂuence of national governments and party preferences on immigration restrictions. On the other hand, inter- national policy diﬀusion can also trigger a ‘race to the bottom’ towards progressively dis- mantling migrant rights, particularly for lower-skilled migrants or, alternatively, a race to attract the ‘best and brightest’, particularly with regards to higher-skilled migrants. To account for the structural context within which political parties operate and migration policies are made, we introduced four sets set of control variables measuring (1) demographic, (2) economic, (3) political and (4) welfare factors that may aﬀect the restrictiveness of immigration policies. First, demographic indicators include population ageing (measured through the share of the population over 65 years) and recent immigra- tion levels (measures through immigration rates as a percentage of total population drawn from the DEMIG TOTAL database (DEMIG 2015)) to capture the demand for migrant labour as well as perceptions of immigration. While it has been regularly argued that ageing societies create a higher demand for migrant labour, high levels of recent immigration may create political pressures to tighten entry policies. Second, we included two economic cycle variables (one-year lagged GDP per capita growth and one-year lagged change in unemployment rates) and two variables measuring economic structure (trade openness measured by the annual trade volume as percentage of GDP, and the strength of trade unions measured by net union membership as a pro- portion of employed wage and salary earners (Armingeon et al. 2017)). The economic cycle control variables were included based on the expectation that growing economies with low unemployment create more demand for migrant labour and that this may increase the power of economic lobbies favouring liberal immigration policies. With regards to economic structure variables, we can expect open economies to have more liberal immigration policies while trade union power may push policies in a more restric- tive direction. Third, we also expect political system variables to shape the inﬂuence of political party ideology on immigration policy reform. As Schmidt (1996) has shown, party preferences tend to be watered down in political systems that have a lot of inbuilt room for nego- tiation. We would, therefore, expect that both presidential systems (measured through a binary variable capturing presidentialism) and federal systems (measured through the absent, weak or strong levels of federalism) enact overall more liberal changes because 8 K. NATTER ET AL. of the need to negotiate laws over multiple decision-making structures. On the other hand, the more fragmented the party landscape, the more restrictive we would expect immigration reforms to be, given the better representation of fringe parties – both on the left and the right side of the political spectrum – with more radical views on immigra- tion. We expect that institutional fragmentation (measured through the fractionalization of 2 2 the party system, index by EF = 1/ v , where v captures the squared share of votes i=1 it it for party i in an election at time t (Laakso and Taagepera 1979)) increases the tendency of parties in power to introduce restrictive policies. Lastly, we included two variables to capture the structure of a country’s welfare system (measured by social security spending in the percentage of GDP and the strictness of employment protection) in the expectation that immigration policies will be more restric- tive in countries where migrants have access to strong welfare state arrangements. Next to these four sets of control variables, we also account for possible structural changes in general immigration policy trends (i) after the fall of the Berlin Wall in 1989 and (ii) after the 9/11 attacks in the United States in 2001. By introducing two time dummies, we test to what extent immigration policies have been aﬀected by the end of communism and the start of the so-called ‘war on terror’ which is often associated to the increasing ‘securitization’ of migration. An overview of all variables used can be found in Table A1 (Appendix). The empirical analysis is divided in a descriptive and multivariate part. The descriptive analysis in Section ‘Disaggregating immigration policy reform’ compares trends in migration policy restrictiveness under left- and right-wing governments. To account for the increasingly targeted and selective nature of immigration policies (de Haas, Natter, and Vezzoli 2018), we zoom in on immigration policy changes targetting diﬀerent policy areas and migrant groups. Sections ‘Putting partisan eﬀects in context’ and ‘Disen- tangling partisan eﬀects’ use multivariate regression analyses to investigate the extent to which the restrictiveness of immigration policy reform is driven by the political orientation of parties in power. The analyses are performed separately by policy areas and migrant groups targeted, as well as for the three diﬀerent time periods (1970–1988, 1989–2000, 2001–2012). Looking at changes over time also allows us to assess whether – and in which policy areas – party preferences on immigration have converged or become more polarized. We also investigate the importance of international diﬀusion dynamics and policy trade-oﬀs in immigration policymaking. Disaggregating immigration policy reform Figure 1 displays changes in migration policy restrictiveness, as well as the prevalence of left- and right-wing parties in governments over time. It shows that right-wing parties have dominated governments between the mid-1980s and the mid-2000s, while left-wing dominated governments have prevailed across the 21 countries understudy in the 1970s and since the mid-2000s. The patterns for parliaments reﬂect similar patterns (see Figure A1 in the Appendix). Crucial for this paper is that in both instances, the linear trendline capturing changes in migration policy restrictiveness seems not related to the inverted-U trend characterizing shifts in partisan compositions of governments or parliaments over time. In fact, the ﬁgures show that the average annual change in immigration policy restrictiveness has POLITICAL RESEARCH EXCHANGE 9 Figure 1. Trends in government composition and changes of migration policy restrictiveness (1970– 2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The govern- ment composition index gives an insight into the overall ideological positioning of the government. (Far) left-wing domi- nated governments have a (strongly) negative index score, whereas (far) right-wing governments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and government composition. remained below zero, and that therefore, since the 1970s, immigration policies have con- sistently been liberalized. The observable upward trend does therefore not represent an increase in restrictiveness, but the fact that over recent decades restrictive and liberal policy reforms have increasingly balanced each other out, indicating a decelerated liberal- ization of immigration policy change (de Haas, Natter, and Vezzoli 2018). This contradicts common perceptions that migration policies have become more restrictive, indicating a considerable gap between migration discourse and migration policy practice, and/or the tendency of restrictive policies to receive more media attention. It also underlines that the deceleration of liberalization has occurred regardless of the ideological compo- sition of parliaments or governments. What this ﬁgure does not show is that although policies have not become more restric- tive, immigration policymaking has gone through a process of considerable intensiﬁcation over the past decades. Both left- and right-wing parliaments having nearly tripled the average number of immigration-related laws they enact annually between the late 1980s and early 2000s. This seems to reﬂect the growing political salience of migration since the end of the Cold War, as well as the increasingly complex and selective nature of immigration policies. The intensity of immigration policymaking has stabilized since the early 2000s, suggesting either a decreasing political salience of migration because of the rise of other issues (such as ‘terrorism’), or a certain level of saturation after a major wave of policymaking on new border control technologies and skill-speciﬁc immi- gration policies. The heatmap (Figure 2) allows to look simultaneously at changes in immigration policy restrictiveness and government composition over time. In general, no clear pattern emerges: Apart from the early 1970s, during the ‘guest-worker’ era, when left-wing gov- ernments enacted particularly liberal policy changes, and the late 1990s and early 10 K. NATTER ET AL. Figure 2. Changes in migration policy restrictiveness according to the (right-left) political ideology of governments (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The govern- ment composition index gives an insight into the overall ideological positioning of the government. Areas coloured light or dark blue show governments dominated by a right-wing political ideology, areas coloured in green, yellow or red show governments dominated by a left-wing political ideology. 2000s, when right-wing governments enacted more restrictive policies, governments across the entire political spectrum seem to enact liberal and restrictive policies to a largely similar extent. A comparable pattern is also visible when looking at migration policy changes according to parliament composition over time (see Figure A2 in the Appendix). The fact that trends in immigration policy restrictiveness are largely detached from government composition might suggest that the restrictiveness of immigration pol- icies is primarily determined by broader economic, geopolitical and social factors rather than by ideological preferences and party politics per se. It is however also possible that these graphs conceal diﬀerences between immigration policies targetting particular policy areas or migrant groups. Figure 3 shows that restrictive policy reforms mainly focus on border control and exit policies (which include deportation) – and this regardless of the political party in power. Integration policies have moved into a more liberal direction over the entire 1970–2012 period under both right- and left-wing dominated governments, but the tendency towards liberalization has been signiﬁcantly stronger under left-wing governments. Entry policies – the very core of immigration policymaking – have shown the most consist- ent liberalizing tendency, with right-wing parties tending towards slightly more liberal positions in this policy area. Again, a comparable pattern is visible when looking at the restrictiveness of particular immigration policy areas according to the right-left compo- sition of parliaments, although right-wing dominated parliaments seem to be on average associated with slightly more restrictive border control policies and less liberal entry policies when compared to trends under right-wing governments (see Figure A3 in Appendix). When disaggregating immigration policy reform according to migrant categories (Figure 4), we see that there is no left/right diﬀerentiation with regards to policies POLITICAL RESEARCH EXCHANGE 11 Figure 3. Average change in migration policy restrictiveness according to the (right-left) political ideol- ogy of governments, disaggregated by policy area (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated governments have a (strongly) negative index score, whereas (far) right-wing governments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and government compo- sition by migration policy area. targetting family migrants and that both right- and left-wing dominated governments tend to back policies that expand the rights of (high- and low-skilled) labour migrants and restrict rights for undocumented migrants. However, in general, right-wing domi- nated governments seem to enact comparatively more restrictive policy changes towards most migrant categories, particularly undocumented migrants and asylum seekers. The major exception is high-skilled migration policies, which have become Figure 4. Average change in migration policy restrictiveness according to the (right-left) political ideol- ogy of governments, disaggregated by migrant group targeted (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated governments have a (strongly) negative index score, whereas (far) right-wing governments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and government compo- sition by migrant target group. 12 K. NATTER ET AL. more liberal across the political spectrum but seem particularly popular among right-wing dominated governments. Comparable patterns are visible when looking at trends in policy restrictiveness towards certain migrant groups according to the ideological composition of parliaments (see Figure A4 in Appendix). To verify whether these ﬁndings hold when controlling for the eﬀects of other migration policy drivers, the following sections move to multivariate regression analysis. We ﬁrst concentrate on the importance of parliamentary and governmental party compo- sition and discuss the role of structural factors captured by economic, political, welfare and demographic control variables. We then disaggregate the analysis by policy areas and migrant groups and discuss changes over time. We also zoom into the role of international policy diﬀusion and the question whether there are trade-oﬀs between diﬀerent policy issues in terms of restrictiveness. Putting partisan eﬀects in context Table 2 presents the results of the logistic regressions (models 1–2, 5–6), as well as all ordinary least square regression (models 3–4, 7–8) pooled over the entire 1970–2012 period. Overall, no clear pattern along a left–right continuum emerges and the association between the political ideology of parties in power and the restrictiveness of immigration policy change is rather weak and inconclusive. Interestingly, the analysis seems to suggest that – if at all – then the partisan composition of parliaments matters more for immigration policy restrictiveness than the partisan composition of governments, with right-wing dominated parliaments enacting on average more restrictive migration policies. This could point to the role of right-wing opposition parties in pushing the entire political spec- trum towards adopting more restrictive positions on immigration. However, this eﬀect (in model speciﬁcations 6 and 8) almost disappears once we control for the eﬀect of other migration policy drivers. A more ﬁne-grained analysis is, however, necessary to identify whether the eﬀects (or the lack thereof) hold when we disentangle the analysis of partisan eﬀects along policy areas and migrant groups targetted (see Section ‘Disentangling parti- san eﬀects’). As this analysis does not show a clear relation between the political ideology of parties in power and the overall restrictiveness of immigration policy change, it is important to assess the role of other plausible drivers of immigration policy reform. The analysis shows that economic growth is strongly and signiﬁcantly associated to the adoption of liberal policy changes. This is consistent with the idea that business cycles directly aﬀect the political willingness for allowing immigration. Irrespective of the political orientation of governments, economic lobbies pressuring governments toward more liberal policies gain power during the time of economic growth and high labour demand (Massey 1999; Meyers 1995). The pro-immigration tendency of a certain governement may thus be reinforced under circumstances of high economic growth and maybe counterbalanced in times of economic adversity when public pressure to reduce migration may be higher. Surprisingly, we do not ﬁnd any signiﬁcant eﬀect of unemployment on overall changes in immigration policy restrictiveness. The weak relation between unemployment and immigra- tion policy reform might be explained by the fact that the GDP growth variable absorbs part of the expected eﬀects. Alternatively, although unemployment is correlated with GDP growth, the specialized and segmented nature of labour markets can explain why structural POLITICAL RESEARCH EXCHANGE 13 Table 2. Determinants of immigration policy reform in 21 countries, 1970–2012. DV: Immigration (1) (2) (3) (4) (5) (6) (7) (8) policy change 0/1 0/1 −4/+4 −4/+4 0/1 0/1 −4/+4 −4/+4 Cabinet composition 0.00231 0.00200 0.00515 0.00381 (0.00333) (0.00429) (0.00320) (0.00358) Parliament 0.0419* 0.0351 0.0679** 0.0675** composition (0.0241) (0.0307) (0.0283) (0.0266) Policy diﬀusion 0.224*** 0.232*** 0.257*** 0.263*** (0.0582) (0.0584) (0.0775) (0.0774) Leg. 0.0753 0.0691 0.116* 0.112* fractionalisation (0.0474) (0.0474) (0.0637) (0.0638) Presidential system −0.346* −0.379** −0.510** −0.540** (0.178) (0.180) (0.238) (0.240) Federalism (ref: no) Weak −0.247 −0.324* −0.293 −0.359 (0.191) (0.196) (0.250) (0.260) Strong −0.341** −0.329** −0.421** −0.415** (0.135) (0.135) (0.181) (0.181) Trade openness −0.00131 −0.000848 −0.00198 −0.00159 (0.00129) (0.00133) (0.00182) (0.00186) Union density −0.0121*** −0.0115*** −0.0152*** −0.0148*** (0.00385) (0.00385) (0.00506) (0.00506) GDP growth (lag) −0.0755*** −0.0765*** −0.0924*** −0.0931*** (0.0269) (0.0269) (0.0341) (0.0341) Unemployment (lag) 0.00864 0.00748 5.87e−05 −0.000672 (0.0172) (0.0172) (0.0229) (0.0229) Immigration rate −0.00162 −0.00189 −0.00259 −0.00282 (lag) (0.00320) (0.00316) (0.00354) (0.00354) Old age share (65+) −0.0401 −0.0403 −0.0755 −0.0758 (0.0373) (0.0365) (0.0499) (0.0490) Employment −0.430*** −0.376*** −0.458*** −0.416*** protection (0.115) (0.119) (0.150) (0.153) Social security (% 0.0692*** 0.0680*** 0.0754** 0.0746** GDP) (0.0251) (0.0251) (0.0333) (0.0332) Post-1989 (ref: pre- 0.369* 0.332 0.328 0.300 1989) (0.202) (0.202) (0.244) (0.244) Post-2001 (ref: pre- 0.442* 0.424* 0.528* 0.513* 1989) (0.230) (0.227) (0.277) (0.274) 0.368 0.265 0.921 0.839 −1.035** −1.175** −0.145 −0.311 Constant (0.457) (0.461) (0.612) (0.616) (0.449) (0.458) (0.235) (0.249) Observations 1690 1690 1757 1757 2332 2332 2621 2621 R-squared 0.0312 0.0323 0.034 0.034 0.0704 0.0716 0.069 0.071 Estimator Logit Logit OLS OLS Logit Logit OLS OLS Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. unemployment in particular sectors of the labour market may coincide with high GDP growth and labour shortages in other sectors (Piore 1979). Indeed, unemployment partly reﬂects structural mismatches between the supply and demand for labour. GDP growth, therefore, seems a better indicator of year-to-year variations in labour demand. The level of trade openness, which could perhaps also be seen as a proxy of the power of business lobbies, seems associated to liberal reforms, but the eﬀect is not signiﬁcant. Interestingly, union density is associated with liberal changes towards immigration. This may seem surprising because of the historical opposition – or at least scepticism – of labour unions towards labour immigration. The fact that union density is associated to more liberal entry policies might be explained by a shift in labour unions’ approach to migrant workers over time. Partly as a result of the increasing share of migrant workers among union members, unions conceive migrants not anymore as competition for native workers but rather as potential clientele. This also seems to conﬁrm that unions 14 K. NATTER ET AL. are more interesting in safeguarding the rights of migrants who are already in the country rather than opening the door to more immigration in terms of entry policies (Haus 1999). So, this eﬀect might particularly reﬂect eﬀorts by unions to achieve more favourable inte- gration policies so as to put migrant workers on an equal legal footing as native workers, and by so doing remove ‘unfair competition’ by rightless migrant workers. The ﬁndings do not conﬁrm the idea that high levels of past immigration cause a public backlash against immigration and calls for more restrictions (Massey 1999; Meyers 1995). The absence of any signiﬁcant eﬀect of recent immigration levels on overall changes in immigration policy restrictiveness might be linked to the so-called ‘opinion-policy gap’, which is the time between a change in circumstances and the moment public opinion shifts in reaction to it (Morales, Pilet, and Ruedin 2015). The absent eﬀect of past immigra- tion may also indicate that public opinion may only turn against immigration after sus- tained periods of high immigration and the build-up of large immigrant population, which is not captured by this variable. The number of over 65 old persons as a share of national populations has no clear eﬀects on immigration policy restrictiveness. This casts doubts on the idea that in ageing societies there would be more support for liberal immigration policies as a way to address labour shortages or to attract migrant care workers. More in general, it questions the popular assumption that demographic factors would directly impact migration trends (de Haas 2011). If such an eﬀect exists at all, it can also be counterbalanced by the fact that older persons tend to adopt more con- servative political positions. Looking at welfare systems, high social security spendings are associated to more restrictive immigration policy changes. This is consistent with the idea that immigration policies are a means to regulate the pathways of access into social welfare systems and herewith to limit access and safeguard national privileges. Generous welfare rights, there- fore, seem to create an incentive to restrict entry and to increase selection. This might conﬁrm that there is some trade-oﬀ in migration policymaking between the ‘numbers’ of migrants allowed in and the ‘rights’ granted to them (Ruhs 2013). In contrast, high employment protection is associated with liberal immigration reforms, which suggests that migrant labour is not necessarily an immediate threat for native workers in dual labour markets (see Piore 1979). In fact, high labour market protection allows governments to be more lenient towards immigration, as both business and trade unions may have an interest in favouring the immigration of migrants workers who take up jobs that native workers tend to shun. Vice-versa, immigration restrictions may be a policy tool for govern- ments to secure jobs for native workers in ﬂexible labour markets that oﬀer only low employment protection. Employment protection and immigration policies may thus be two alternative ways to deal with the consequences of labour migration for native workers. Depending on whether governments react to business lobbies or to electorate pressure, they might priviledge one policy instrument over the other. Political system variables also play a role in explaining the restrictiveness of migration policy reform, as federal and presidential systems tend to enact more liberal changes. This is consistent with the idea that the more actors are involved in policymaking, the higher the need for negotiation and political compromise. Federalism introduces the need to negotiate fundamental policy reforms not only with nationally elected representatives but also with regional representatives, which might ultimately water down initially restric- tive policy proposals. Similarly, presidentialism introduces a potential additional POLITICAL RESEARCH EXCHANGE 15 negotiation level between the executive and legislative and increases the necessity for political compromise and political bartering, particularly when the president and prime minister are not from the same political party. These dynamics, in turn, create a sustained gap between policy rhetoric and policy practice. In contrast, the tendency for fragmented party systems (proxied by legislative fractionalization) to enact more restrictive immigra- tion policies is not strongly signiﬁcant. This oﬀers grounds to question the widespread idea that in highly competitive party landscapes, parties in power are urged to ‘show per- formance’ in terms of immigration policy toughness. Finally, the two time dummies reveal only weak period eﬀects: the fall of the Berlin Wall in 1989 and the ensuing growth of (asylum) migration does not seem to have con- tributed to party-speciﬁc changes in immigration policy restrictiveness. This seems to counter the idea that the increasing political salience of immigration in the public sphere and the disappearance of the ‘Communist threat’ has led right-wing politicians to enact overall more restrictive immigration policies. In contrast, the 9/11 terrorist attacks had a small, but signiﬁcant eﬀect in driving restrictive immigration policy reforms regardless of the party in power, reinforcing the overall slow-down of the post-Second World War liberalization of immigration policies. At the same time, as our analysis in Section ‘Disentangling partisan eﬀects’ shows, these period-speciﬁc eﬀects are driven by policies targeting particular migrant groups and policy ﬁelds and do not aﬀect immigration regimes as a whole. In general, the analysis suggests that the political ideology of parties in power plays only a very limited role in explaining the restrictiveness of immigration policy reform in Western liberal democracies over the 1970–2012 period. Right- and left-wing dominated governments do not seem to enact opposite reforms on immigration, as they are faced with the same interest groups and international context in migration policymaking and therefore policy trends have been overall coherent across the left/right spectrum. As the analysis of control variables has shown, the ideological orientation of parties in power is but one of the many factors shaping migration policies, with economic cycles, employment protection, the power of trade unions as well as international policy diﬀusion playing a more important role in shaping the restrictiveness of immigration policy reforms. The following section zooms into the diﬀerent dimensions of immigration policy reform in order to test whether the absence of a clear eﬀect of political ideology holds if we disaggregate the analysis for particular policy areas or policies targetting par- ticular migrant groups. Disentangling partisan eﬀects The disaggregated analyses (Table 3) show that the partisan composition of governments and parliaments aﬀects ﬁrst and foremost integration policies, with right-wing dominated governments and parliaments enacting on average more restrictive integration policies. This clear partisan eﬀect, however, should not obscure the fact that integration policies have been liberalized in the past decades by parties from across the political spectrum. Other migration policy areas such as border control or exit policies, but most strikingly also entry policies – the core of immigration regimes – seem only weakly or not aﬀected by government or parliament composition. In particular, the absent eﬀect of par- tisan ideology on border controls seems to conﬁrm that there is a cross-partisan 16 K. NATTER ET AL. Table 3. Partisan determinants of immigration policy reform by policy area and by target group in 21 countries, 1970–2012. DV: Immigration policy (1) (2) (3) (4) (5) (6) (7) (8) change +4/−4 0/1 +4/−4 0/1 +4/−4 0/1 +4/−4 0/1 Policy area Cabinet composition Parliament interacted with policy composition area… interacted with policy area… Border −0.00637 0.00197 0.0839 0.102** (0.0105) (0.00732) (0.0614) (0.0453) Entry −0.00257 −0.00173 0.0304 0.0333 (0.00520) (0.00410) (0.0393) (0.0321) Integration 0.0290*** 0.0232*** 0.158*** 0.120*** (0.00677) (0.00548) (0.0557) (0.0447) Exit 0.00282 0.000968 0.0976 0.0952 (0.0120) (0.00860) (0.0808) (0.0609) Target group Cabinet composition Parliament interacted with target composition group… interacted with target group… All (unspeciﬁed) migration 0.00928 0.00715 0.0870* 0.0793** (0.00747) (0.00543) (0.0498) (0.0385) Labour migration 0.0217** 0.0168** 0.137* 0.118* (0.00916) (0.00753) (0.0803) (0.0627) Family migration 0.00900 0.00988 0.0618 0.0899 (0.00994) (0.00852) (0.0673) (0.0597) Skilled migration −0.0177** −0.0154** −0.0390 −0.0506 (0.00806) (0.00706) (0.0521) (0.0465) Undocumented migration 0.0176* 0.0135** 0.152** 0.130*** (0.00922) (0.00679) (0.0616) (0.0465) Asylum migration 0.0145* 0.0141** 0.0881* 0.0749* (0.00825) (0.00636) (0.0518) (0.0405) Constant −1.393** −1.090** −1.481** −1.168** −1.121* −0.854* −1.343** −1.046** (0.566) (0.451) (0.589) (0.457) (0.630) (0.467) (0.643) (0.476) Observations 2332 2332 2332 2332 2240 2240 2240 2240 (Pseudo)-R 0.089 0.075 0.085 0.073 0.087 0.074 0.085 0.074 Estimator OLS Logit OLS Logit OLS Logit OLS Logit Country FE Yes Yes Yes Yes Yes Yes Yes Yes Time FE Yes Yes Yes Yes Yes Yes Yes Yes Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. agreement that border controls are the central tool of immigration restrictions. This is in line with Massey et al. (1998, 288) who argued that ‘elected leaders and bureaucrats increasingly have turned to symbolic policy instruments to create an appearance of control’. These results suggest that – in contrast to entry policies where pro- and anti-immigra- tion positions cut across the political spectrum – policies targetting migrants socio-econ- omic and political post-entry rights tend to split the political spectrum more neatly along the right-left cleavage. Our results thus lend support to the idea put forward by Givens and Luedtke (2005) and Money (1999) that integration policies signiﬁcantly diﬀer under left- and right-wing governments, while politicians will open entry policies if economic and other contextual factors require it, regardless of their party aﬃliation (see Table 1). The most striking ﬁnding, however, is the absence of signiﬁcant partisan eﬀects on entry pol- icies, the core of a country’s immigration regime. If we look at policies targeting speciﬁc migrant groups, the most consistent and robust result is that left-wing dominated governments and parliaments seem to be more liberal in POLITICAL RESEARCH EXCHANGE 17 granting (post-entry) rights to labour migrants, asylum seekers and undocumented migrants, whereas right-wing dominated governments seem to be slightly more favour- able towards skilled migrants. For high-skilled migration, the analysis suggests that right-wing governments are more likely to enact liberal changes while the partisan com- position of parliament does not seem to play a signiﬁcant role. This might point at the more direct inﬂuence of business lobbies in governmental priorities as opposed to the leg- islative process. We do not see clear diﬀerences in partisan eﬀects with regards to family migration. Importantly, all these group-speciﬁc results seem to pertain to integration pol- icies, as we failed to ﬁnd a signiﬁcant eﬀect of partisan ideology on entry, exit and border control policies. Developments over time (Tables A2 and A3 in Appendix) provide additional insights: In fact, the results just discussed can be partly explained by general trends in particular periods. Most strikingly, the analysis suggests that the party landscape polarized on inte- gration policies particularly in the pre-1989 and post-2001 period. We could not detect consistent period eﬀects for other policy areas. When looking at migrant groups, it seems that the positive eﬀect of left-wing governments and parliaments on labour migration and asylum policy is particularly explained by a policy polarization on the issue after 2001. In contrast, the positive eﬀect of right-wing ideology for high-skilled migration policies seems to be relevant only for the pre-2001 period, after which there seems to be a cross-partisan agreement on attracting the highly skilled. Overall, the ideo- logical composition of parliaments seems to be slightly more relevant for explaining diﬀer- ences in policy outcomes in speciﬁc periods – for instance, the drive towards restrictive border control policies in the 1990s – compared to the ideological composition of governments. Table 4. Policy diﬀusion eﬀects and policy trade-oﬀs in immigration policy reforms of 21 countries, 1970–2012. (1) (2) (3) (4) (5) (6) (7) (8) Variables Border Admission Integration Exit Border Admission Integration Exit Cabinet −0.00126 −0.00176 0.00559*** 0.00145 composition (0.0013) (0.0025) (0.0018) (0.0012) Parliament 0.0107 0.0331* 0.0167 0.00633 composition (0.00879) (0.0190) (0.0141) (0.00794) Policy diﬀusion 0.468** 0.0889*** 0.429** −0.0474 0.467** 0.532** 0.451** −0.0372 (0.225) (0.0157) (0.213) (0.241) (0.225) (0.218) (0.213) (0.239) Border 0.0889*** 0.0494*** −0.0325*** 0.0880*** 0.0473*** −0.0332*** (0.0157) (0.0129) (0.00833) (0.0157) (0.0128) (0.00832) Admission 0.0249*** −0.0426*** 0.0123*** 0.0247*** −0.0439*** 0.0119*** (0.00466) (0.00657) (0.00424) (0.00464) (0.00658) (0.00426) Integration 0.0212*** −0.0715*** 0.0133** 0.0202*** −0.0734*** 0.0140** (0.00609) (0.0108) (0.00589) (0.00600) (0.0107) (0.00584) Exit −0.0347*** 0.0453*** 0.0312** −0.0353*** 0.0439*** 0.0327** (0.00892) (0.0160) (0.0135) (0.00891) (0.0161) (0.0134) Constant −0.0598 −0.181 0.166 −0.0639 −0.105 −0.300* 0.173 −0.0690 (0.0772) (0.172) (0.125) (0.0625) (0.0817) (0.181) (0.123) (0.0650) Observations 2621 2621 2621 2621 2621 2621 2621 2621 R-squared 0.056 0.045 0.066 0.053 0.056 0.046 0.063 0.053 Country FE Yes Yes Yes Yes Yes Yes Yes Yes Time FE Yes Yes Yes Yes Yes Yes Yes Yes Estimator OLS OLS OLS OLS OLS OLS OLS OLS Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. 18 K. NATTER ET AL. Lastly, Table 4 explores diﬀusion eﬀects and policy trade-oﬀs between diﬀerent policy areas. The ﬁndings clearly show that policy reform in other relevant countries leads to adaptations in national policy in a similar direction: Liberal reforms elsewhere lead to immigration liberalization, while restrictive reforms in reference countries lead to a a further closure of national immigration policies. International policy emulation seems thus a robust feature of national migration policymaking, supporting the ‘convergence’ hypothesis in immigration policy (Cornelius, Martin, and Holliﬁeld 1994). The fact that many countries in our sample are members of the European Union (EU) might have driven this result, as policy diﬀusion does not only happen ad hoc but is also institutiona- lized through the Europeanization of immigration and integration policies since the 1990s. Interestingly, international policy diﬀusion dynamics seem predominantly at play in the area of border, entry and integration policies, but not in the area of exit policies where gov- ernments seem to act rather independently from international trends. This might be explained by the fact that the attractiveness of a country is determined more strongly by the facilities it oﬀers to enter and stay in the country, as well as to access socio-econ- omic rights, rather than by regulations of exit and return. Table 4 also provides insights into the interdependency between diﬀerent policies issues. Most strikingly, when entry policies are liberalized, this often goes together with a restriction of integration measures and vice-versa. This mechanism might point towards the existence of an empirical trade-oﬀ between the numbers of migrants allowed into a country and the post-entry rights granted to them (Ruhs and Martin 2008). A similar dynamic is at play between border control and exit policies. Further research should investigate whether and to what extent showing ‘toughness’ on certain policy issues might indeed be used as a way to gain support for more liberal policies in other areas. Overall, however, the role of the political ideology of parties in power on the overall restrictiveness of immigration policy change remains limited, as restrictive and liberal policy changes tend to balance each other out and there is often a signiﬁcant gap between rhetoric and reality in migration policymaking. Party ideology seems to translate into more or less restrictive immigration policies only in the area of integration, as there are no signiﬁcant partisan diﬀerences with regards to the core of immigration regimes – entry policies. Lastly, the results also point towards high inter-party consensus on restric- tive border control policies, which are at the heart of symbolic migration politics but have limited eﬀects on overall immigration volumes. Conclusion This paper has oﬀered a comprehensive analysis of the partisan drivers of immigration policy reform in 21 Western countries between 1970 and 2012. The analysis showed that the ideological orientation of parties in power plays a relatively marginal role in deter- mining immigration reform. In fact, our results are very robust in not ﬁnding evidence of a clear left/right gradient on immigration policy restrictiveness, particularly when it comes to the ideological orientation of governmental parties. Because party preferences on immi- gration cut across the left–right spectrum, there is no strong eﬀect of the ideology of gov- erning parties on overall immigration policy restrictiveness. Political party ideologies only aﬀect certain migration policy areas, in particular integration policies. This supports POLITICAL RESEARCH EXCHANGE 19 previous research arguing that integration splits political parties more neatly between the left and the right (Duncan and Van Hecke 2008; Money 1999). Also, political party ideology only aﬀects immigration policy restrictiveness towards certain migrant groups, particularly those that stand at the centre of public debates, such as asylum seekers and undocumen- ted migrants. In contrast, political party ideology does not fundamentally shape decisions on the core of immigration regimes, particularly entry policies on labour and family migration. Although immigration is subject to heated debates in the public sphere and extensive political bargaining, the actual policies enacted seem to be driven by factors other than political ideology. Overall, the restrictiveness of immigration policy reform seems deter- mined by general structural factors such as economic cycles, political system features and the strength of welfare systems. If economies grow fast, governments will be more lenient towards immigration, irrespective of their political orientation. In contrast, high social security spendings and low employment protection regulations are associated to more restrictive changes. This suggests a high interconnection between diﬀerent policy ﬁelds, with employment protection policies and immigration policies seemingly oﬀering policymakers two alternative tools to regulate the consequences of labour immigration for native workers. Lastly, structural political system variables also play a role: Federal and presidential systems are associated to more liberal policy outcomes, potentially as a result of the higher number of negotiation levels between national and regional levels, as well as between executive and legislative actors that require political compromising and ‘bartering’. The need to form coalitions, therefore, results in policies that are watered-down versions of initial political positions. In sum, both right- and left-wing gov- ernments largely deal with the same interest groups and have to respond to the same international contexts in migration policymaking. Next to disentangling partisan from more structural drivers, one of the main inno- vations of this paper has been to disaggregate immigration policy and to explore interde- pendencies between policy areas. Given the diﬀerentiated preferences of political parties regarding certain migrant groups and policy areas, policymaking is likely to involve trade- oﬀs and compromises, leading to immigration policy reforms that contain ‘mixed bags’ of measures that target diﬀerent migrant groups and policy issues in often incoherent and contradictory ways. In particular, our results show that the liberalization of entry policies often goes together with a restriction of integration measures and vice-versa, pointing towards a trade-oﬀ between the numbers of migrants allowed into a country and the post-entry rights granted to them (Ruhs and Martin 2008). Immigration policy is also strongly aﬀected by the actions of neighbouring countries or alternative migration desti- nations. International policy diﬀusion accounts for the high degree of policy convergence across countries and can partly explain why similar policies are enacted regardless of which party is in power. Whether international policy diﬀusion drive policy restrictions or liberalizations depends on the issue at stake. In sum, our results point to the complex mechanisms through which party ideology shapes immigration policy reform, as well as the necessity to focus on speciﬁc migrant cat- egories and policy areas separately to understand such general trends. Ultimately, this analysis shows that – if at all – the partisan composition of governments and parliaments has only limited power in explaining immigration policy outcomes. 20 K. NATTER ET AL. Notes 1. Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States. 2. The DEMIG POLICY database, as well as all supporting documentation are publicly available online, see: https://www.imi-n.org/data. DEMIG POLICY tracks over 6500 changes in migration policies in 45 countries over the 1945–2014 period. Although it also includes countries in Latin America, Asia and Africa, we limit our analysis to countries that have been democratically ruled over this entire period to assess the eﬀect of party politics. The limitation to the post-1970 period is because of availability of partisan and structural factors. 3. The weighted policy change variable also includes zeros. However, these do not represent ‘no policy change’ but policy changes without a clear direction in terms of changes in restrictive- ness. The total absence of policy changes in a given year is not coded by this dataset, which implies that country-years without any policy activity are missing. 4. The binary policy change variable is used for logistic regressions whereas weighted policy changes are used in linear least square regressions. We also performed analyses both with one-year lagged and non-lagged policy changes to account for potential delays in policy changes introduced by new governments as well as lengthy legislative processes. However, given that results were not aﬀected by the lag (the potential delay only aﬀects years in which a new government or parliament is elected), we decided to run the ﬁnal analyses without lagging policy changes. 5. We have also tested geographical distance and contiguity as alternative connectivity variables. Disclosure statement No potential conﬂict of interest was reported by the author(s). Funding The research leading to these results is part of the DEMIG (Determinants of International Migration) project and has received funding from the European Community’s Seventh Framework Programme (FP7/2007-2013)/ERC Grant Agreement 240940. The paper has been ﬁnalized thanks to support by the MADE (Migration as Development) project funded by the European Community’s Horizon 2020 Programme (H2020/2015-2020)/ERC Consolidator Grant Agreement 648496. ORCID Katharina Natter http://orcid.org/0000-0001-9683-0567 References Abou-Chadi, T. 2016. “Political and Institutional Determinants of Immigration Policies.” Journal of Ethnic and Migration Studies 42 (13): 2087–2110. doi:10.1080/1369183X.2016.1166938. Alonso, S., and S. C. da Fonseca. 2011. “Immigration, Left and Right.” Party Politics 18 (6): 865–884. doi:10.1177/1354068810393265. Armingeon, K., V. Wenger, F. Wiedemeier, C. Isler, L. Knöpfel, D. Weisstanner, and S. Engler. 2017. “Comparative Political Data Set 1960–2015.” Bale, T. 2008. “Politics Matters: A Conclusion.” Journal of European Public Policy 15 (3): 453–464. doi:10. 1080/13501760701847721. POLITICAL RESEARCH EXCHANGE 21 Block, L., and S. Bonjour. 2013. “Fortress Europe or Europe of Rights? The Europeanisation of Family Migration Policies in France, Germany and the Netherlands.” European Journal of Migration and Law 15: 203–224. doi:10.1163/15718166-12342031. Bonjour, S. 2011. “The Power and Morals of Policy Makers: Reassessing the Control Gap Debate.” International Migration Review 45 (1): 89–122. doi:10.1111/j.1747-7379.2010.00840.x. Cornelius, W. A., P. L. Martin, and J. F. Holliﬁeld. 1994. Controlling Immigration: A Global Perspective. Stanford, CA: Stanford University Press. Cornelius, W. A., T. Tsuda, P. L. Martin, and J. F. Holliﬁeld. 2004. Controlling Immigration: A Global Perspective. 2nd ed. Stanford, CA: Stanford University Press. Czaika, M. 2018. High-Skilled Migration. Drivers and Policies. Oxford: Oxford University Press. Czaika, M., and H. de Haas. 2013. “The Eﬀectiveness of Immigration Policies.” Population and Development Review 39 (3): 487–508. doi:10.1111/j.1728-4457.2013.00613.x. Davis, A. 2012. “The Impact of Anti-Immigration Parties on Mainstream Parties’ Immigration Positions in the Netherlands, Flanders and the UK 1987–2010: Divided Electorates, Left-Right Politics and the Pull Towards Restrictionism.” PhD thesis, EUI, Florence. de Haas, H. 2011. “The Determinants of International Migration.” IMI Working Paper 32. de Haas, H., K. Natter, and S. Vezzoli. 2015. “Conceptualizing and Measuring Migration Policy Change.” Comparative Migration Studies 3: 15. doi:10.1186/s40878-015-0016-5. de Haas, H., K. Natter, and S. Vezzoli. 2018. “Growing Restrictiveness or Changing Selection? The Nature and Evolution of Migration Policies.” International Migration Review 52 (2): 314–367. doi:10.1111/imre.12288. DEMIG. 2015. “DEMIG TOTAL.” Version 1.5. Duncan, F., and S. Van Hecke. 2008. “Immigration and the Transnational European Centre-Right: A Common Programmatic Response?” Journal of European Public Policy 15 (3): 432–452. doi:10. 1080/13501760701847705. Eule, T. 2014. Inside Immigration Law: Migration Management and Policy Application in Germany. Farnham: Ashgate. Freeman, G. P. 1995. “Modes of Immigration Politics in Liberal-Democratic States.” International Migration Review 29 (4): 881–902. doi:10.1177/019791839502900401. Freeman, G. P. 2013. Immigration and Public Opinion in Liberal Democracies. New York: Routledge. Geddes, A. 2003. The Politics of Migration and Immigration in Europe. London: Sage. Givens, T., and A. Luedtke. 2005. “European Immigration Policies in Comparative Perspective: Issue Salience, Partisanship and Immigrant Rights.” Comparative European Politics 3: 1–22. doi:10. 1057/palgrave.cep.6110051. Gutekunst, M. 2015. “Language as a New Instrument of Border Control: The Regulation of Marriage Migration from Morocco to Germany.” Journal of North African Studies 20 (4): 540–552. doi:10. 1080/13629387.2015.1065038. Hatton, T. J., and J. G. Williamson. 2004. “Refugees, Asylum Seekers and Policy in Europe.” IZA Discussion Paper No. 1230. Haus, L. 1999. “Labor Unions and Immigration Policy in France.” International Migration Review 33 (3): 683–716. doi:10.1177/019791839903300306. Infantino, F. 2010. “La frontière au guichet. Politiques et pratiques des visas Schengen aux Consulat et àl’Ambassade d’Italie au Maroc.” Champ pénal/Penal ﬁeld, nouvelle revue internationale de crimin- ologie VII. Laakso, M., and R. Taagepera. 1979. “’Eﬀective’ Number of Parties: A Measure with Application to West Europe.” Comparative Political Studies 12 (1): 3–27. doi:10.1177/001041407901200101. Lavenex, S. 2001. The Europeanisation of Refugee Policies: Between Human Rights and Internal Security. Aldershot: Ashgate. Massey, D. S. 1999. “International Migration at the Dawn of the 21st Century: The Role of the State.” Population and Development Review 25 (2): 303–322. doi:10.1111/j.1728-4457.1999.00303.x. Massey, D. S., J. Arango, G. Hugo, A. Kouaouci, A. Pellegrino, and J. E. Taylor. 1998. Worlds in Motion: Understanding International Migration at the End of the Millennium. Oxford: Clarendon Press. Meyers, E. 1995. “A Theory of Immigration Policy.” In International Immigration Policy – A Theoretical and Comparative Analysis,5–26. London: Palgrave Macmillan. 22 K. NATTER ET AL. Meyers, E. 2000. “Theories of International Immigration Policy – A Comparative Analysis.” International Migration Review 34 (4): 1245–1282. doi:10.1177/019791830003400407. Meyers, E. 2002. “The Causes of Convergence in Western Immigration Control.” Review of International Studies 28 (1): 123–141. doi:10.1017/S0260210502001237. Money, J. 1999. “Deﬁning Immigration Policy: Inventory, Quantitative Referents, and Empirical Regularities.” Paper presented at the Annual Meetings of the American Political Science Association, Atlanta, GA. Morales, L., J.-B. Pilet, and D. Ruedin. 2015. “The Gap Between Public Preferences and Policies on Immigration: A Comparative Examination of the Eﬀect of Politicisation on Policy Congruence.” Journal of Ethnic and Migration Studies 41 (9): 1495–1516. doi:10.1080/1369183X.2015.1021598. Neumayer, E., and T. Plümper. 2010. “Spatial Eﬀects in Dyadic Data.” International Organization 64 (1): 145–166. doi:10.1017/S0020818309990191. Odmalm, P. 2011. “Political Parties and ‘the Immigration Issue’: Issue Ownership in Swedish Parliamentary Elections 1991–2010.” West European Politics 34 (5): 1070–1091. doi:10.1080/ 01402382.2011.591098. Perlmutter, T. 1996. “Bringing Parties Back In: Comments on ‘Modes of Immigration Politics in Liberal Democratic Societies’.” International Migration Review 30 (1): 375–388. doi:10.1177/ Piore, M. J. 1979. Birds of Passage. Migrant Labor and Industrial Societies. Cambridge: Cambridge University Press. Ruhs, M. 2013. “The Rights of Migrant Workers Reframing the Debate.” In The Price of Rights – Regulating International Labor Migration,1–12. Princeton, NJ: Princeton University Press. Ruhs, M., and P. L. Martin. 2008. “Numbers vs. Rights: Trade-Oﬀs and Guest Worker Programs.” International Migration Review 42 (1): 249–265. doi:10.1111/j.1747-7379.2007.00120.x. Schain, M. A. 2008. “Commentary: Why Political Parties Matter.” Journal of European Public Policy 15 (3): 465–470. doi:10.1080/13501760701847739. Schmidt, M. G. 1996. “When Parties Matter: A Review of the Possibilities and Limits of Partisan Inﬂuence on Public Policy.” European Journal of Political Research 30: 155–183. doi:10.1111/j. 1475-6765.1996.tb00673.x. Sciortino, G. 2000. “Toward a Political Sociology of Entry Policies: Conceptual Problems and Theoretical Proposals.” Journal of Ethnic and Migration Studies 26 (2): 213–228. doi:10.1080/ Sciortino, G., and A. Colombo. 2004. “The Flows and the Flood: The Public Discourse on Immigration in Italy, 1969–2001.” Journal of Modern Italian Studies 9(1):94–113. doi:10.1080/1354571042000179209. Shipan, C. R., and C. Volden. 2008. “The Mechanisms of Policy Diﬀusion.” American Journal of Political Science 52 (4): 840–857. doi:10.1111/j.1540-5907.2008.00346.x. Tichenor, D. J. 2002. Dividing Lines: The Politics of Immigration Control in America. Princeton, NJ: Princeton University Press. Tichenor, D. J. 2008. “Strange Bedfellows: The Politics and Pathologies of Immigration Reform.” Labor: Studies in Working-Class History of the Americas 5 (2): 39–60. doi:10.1215/15476715-2007-077. Triadaﬁlopoulos, T., and A. Zaslove. 2006. “Inﬂuencing Migration Policy from Inside: Political Parties.” In Dialogues on Migration Policy, edited by M. Giugni and F. Passy, 171–191. Lanham: Lexington Books. Van Der Valk, I. 2003. “Political Discourse on Ethnic Minority Issues. A Comparison of the Right and the Extreme Right in the Netherlands and France (1990–97).” Ethnicities 3 (2): 183–213. doi:10.1177/ Volkens, A., P. Lehmann, T. Matthieß, N. Merz, S. Regel, and B. Weßels. 2017. “The Manifesto Data Collection. Manifesto Project (MRG/CMP/MARPOR).” Version 2017b. Wutts, J. R. 1998. “Strange Bedfellows: How Spanish Labor Union Leaders and Employers Find Common Ground on Immigration.” Policy Studies Journal 26 (4): 657–675. doi:10.1111/j.1541- 0072.1998.tb01938.x. Zolberg, A. R. 2006. A Nation by Design – Immigration Policy in the Fashioning of America. Cambridge, MA: Harvard University Press. POLITICAL RESEARCH EXCHANGE 23 Appendix Table A1. Descriptive statistics of variables. Std. Variable Obs. Mean dev. Min Max Source Policy change (dummy) 2332 0.450 0.498 0 1 DEMIG Policy Policy change (weighted) 2621 −0.299 2.742 −4 4 DEMIG Policy Policy changes by policy area Border policy 2621 0.245 0.981 −4 4 DEMIG Policy Entry policy 2621 −0.366 1.831 −4 4 DEMIG Policy Integration policy 2621 −0.287 1.447 −4 4 DEMIG Policy Exit policy 2621 0.109 0.955 −4 4 DEMIG Policy Policy changes by targeted migrant group All migrants 2621 0.213 0.409 0 1 DEMIG Policy Labour migrants 2621 0.129 0.335 0 1 DEMIG Policy Family migrants 2621 0.073 0.260 0 1 DEMIG Policy High-skilled migrants 2621 0.115 0.319 0 1 DEMIG Policy Undocumented migrants 2621 0.154 0.361 0 1 DEMIG Policy Forced migrants 2621 0.171 0.377 0 1 DEMIG Policy Parliament composition 2621 0.003 2.738 −10.307 9.859 Comparative Manifesto Project Government composition 2621 −0.790 17.495 −58.0 48.458 CPDS/ Comparative Manifesto Project Electoral fractionalisation of party 2605 0.736 0.089 0.500 0.903 CPDS system Presidential system 2602 0.182 0.386 0 1 CPDS Federalism No 2602 0.597 0.491 0 1 CPDS Weak 2602 0.096 0.295 0 1 CPDS Strong 2602 0.307 0.461 0 1 CPDS Trade openness 2609 73.715 49.133 10.730 348.43 CPDS Union density (% of employees) 2533 35.187 18.609 7.548 87.427 CPDS Real GDP growth (in %) 2600 2.381 2.399 −8.539 11.306 CPDS Unemployment rate (% of labour 2600 6.947 3.753 0.002 24.171 CPDS force) Immigration rate (per ‘000 2229 10.436 18.074 .541 299.3 DEMIG Flow population) Population over 65 (in %) 2609 14.597 2.503 8.028 21.101 CPDS Employment protection regulation 2083 2.117 0.943 0.257 5 CPDS Social security transfers (% of GDP) 2603 13.620 3.534 3.473 23.894 CPDS 24 K. NATTER ET AL. Table A2. Period eﬀects in the restrictivenes of policy changes by policy area according to the political ideology of parties in power (21 countries). DV: Immigration policy change (1) (2) (3) (4) (5) (6) Pre-1989 1990–2001 Post-2001 Pre-1989 1990–2001 Post-2001 Cabinet composition interacted with Parliament composition interacted with policy area … policy area … Border 0.0153 0.0174 −0.0191* −0.0206 0.180*** 0.154 (0.0263) (0.0107) (0.0114) (0.225) (0.0679) (0.0982) Entry −0.00562 −0.0146** 0.0119* −0.0709 −0.0185 0.258*** (0.00965) (0.00703) (0.00674) (0.0791) (0.0472) (0.0829) Integration 0.0280*** 0.0186 0.0306*** 0.243*** 0.0257 0.339*** (0.0100) (0.0119) (0.00916) (0.0939) (0.0657) (0.108) Exit 0.0279 0.0187 −0.00536 0.511** 0.171** 0.0824 (0.0217) (0.0147) (0.0115) (0.254) (0.0788) (0.127) Constant −1.631*** 0.403 −1.170*** −1.926*** 0.208 −1.796*** (0.509) (0.281) (0.387) (0.584) (0.349) (0.491) Observations 439 782 1099 439 782 1099 R-squared 0.0547 0.0491 0.0548 0.0634 0.0478 0.0526 Country FE Yes Yes Yes Yes Yes Yes Time FE No No No No No No Estimator Logit Logit Logit Logit Logit Logit Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < 0.1. Table A3. Period eﬀects in the restrictiveness of policy changes by migrant group according to the political ideology of parties in power (21 countries). DV: Immigration policy change (1) (2) (3) (4) (5) (6) Pre-1989 1990–2001 Post-2001 Pre-1989 1990–2001 Post-2001 Cabinet composition interacted with Parliament composition interacted with target group … target group … All (unspeciﬁed) migration 0.0198 0.00354 0.00742 0.0677 0.0664 0.268*** (0.0120) (0.0105) (0.00795) (0.101) (0.0634) (0.0895) Labour migration −0.0175 −0.00549 0.0401*** −0.178 0.149 0.344*** (0.0155) (0.0159) (0.0118) (0.134) (0.106) (0.112) Family migration 0.00894 0.00434 0.0185 0.0229 0.0610 0.182 (0.0163) (0.0161) (0.0151) (0.158) (0.0745) (0.174) Skilled migration −0.0332 −0.0458*** −0.00109 −0.654*** −0.179** 0.229** (0.0219) (0.0135) (0.00933) (0.236) (0.0833) (0.0977) Undocumented migration 0.0285* 0.0299*** −0.00610 0.280* 0.238*** 0.0598 (0.0157) (0.0110) (0.0110) (0.145) (0.0673) (0.116) Asylum migration 0.0128 0.0106 0.0217* 0.106 0.0209 0.278*** (0.0133) (0.0110) (0.0115) (0.0857) (0.0688) (0.0956) Constant −1.944*** 0.476 −1.059** −2.136*** 0.188 −1.808*** (0.571) (0.303) (0.417) (0.627) (0.364) (0.507) Observations 426 744 1064 426 744 1064 R-squared 0.058 0.064 0.051 0.060 0.068 0.051 Country FE Yes Yes Yes Yes Yes Yes Time FE No No No No No No Estimator Logit Logit Logit Logit Logit Logit Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. POLITICAL RESEARCH EXCHANGE 25 Figure A1. Trends in parliament composition and changes in migration policy restrictiveness (1970– 2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The parlia- ment composition index gives an insight into the overall ideological positioning of the parliament. (Far) left-wing domi- nated parliaments have a (strongly) negative index score, whereas (far) right-wing parliaments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and parliament composition. The upward trend of the policy change line thus represents a decelerated liberalization and not a trend towards more restric- tiveness per se. Figure A2. Changes in migration policy restrictiveness according to the (right-left) political ideology of parliaments (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The parlia- ment composition index gives an insight into the overall ideological positioning of the parliament. Areas coloured blue or turquoise show parliaments dominated by a right-wing political ideology, areas coloured in green or red show parliaments dominated by a left-wing political ideology. 26 K. NATTER ET AL. Figure A3. Average change in migration policy restrictiveness according to the (right-left) political ideology of parliaments, disaggregated by policy area (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated parliaments have a (strongly) negative index score, whereas (far) right-wing parliaments have a (strongly) posi- tive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and parliament composition by migration policy area. Figure A4. Average change in migration policy restrictiveness according to the (right-left) political ideology of parliaments, disaggregated by type of migrant group (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated parliaments have a (strongly) negative index score, whereas (far) right-wing parliaments have a (strongly) posi- tive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and parliament composition by migrant target group. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Political Research Exchange Taylor & Francis http://www.deepdyve.com/lp/taylor-francis/political-party-ideology-and-immigration-policy-reform-an-empirical-VrtpwqLOyy

Loading next page...

References (58)

Manfred Schmidt (1996)
When parties matter: A review of the possibilities and limits of partisan influence on public policy
European Journal of Political Research, 30
R. Cholewinski (2007)
The Rights of Migrant Workers
E. Meyers (2002)
The causes of convergence in Western immigration control
Review of International Studies, 28
Sonia Alonso, Sara Fonseca (2012)
Immigration, left and right
Party Politics, 18
T. Perlmutter (1996)
Bringing Parties Back In: Comments on ‘Modes of Immigration Politics in Liberal Democratic Societies’
International Migration Review, 30
Tarik Abou-Chadi (2016)
Political and institutional determinants of immigration policies
Journal of Ethnic and Migration Studies, 42
S. Lavenex (2002)
The Europeanisation of Refugee Policies: Between Human Rights and Internal Security
Jerry Mitchell (2006)
The Politics of Migration and Immigration in Europe
Acta Politica
I. Valk (2003)
Political Discourse on Ethnic Minority Issues: A Comparison of the Right and the Extreme Right in the Netherlands and France (1990-97)
Ethnicities, 3
Kathleen Garces-Foley (2007)
The Dividing Lines
A. Zolberg, Michael Piore (1981)
Birds of Passage: Migrant Labor and Industrial Societies.
International Migration Review, 15
Tim Bale (2008)
Politics matters: a conclusion
Journal of European Public Policy, 15
(2012)
Average change in migration policy restrictiveness according to the (right-left) political ideology of parliaments, disaggregated by policy area (1970–2012
Pontus Odmalm (2011)
Political Parties and ‘the Immigration Issue’: Issue Ownership in Swedish Parliamentary Elections 1991–2010
West European Politics, 34
Simone Wannemaker (2016)
Worlds In Motion Understanding International Migration At The End Of The Millennium
A. Zolberg (2006)
A Nation by Design: Immigration Policy in the Fashioning of America
G. Sciortino, Asher Colombo (2004)
The flows and the flood: the public discourse on immigration in Italy, 1969–2001
Journal of Modern Italian Studies, 9
(2017)
The Manifesto Data Collection
G. Sciortino (2000)
Toward a political sociology of entry policies: Conceptual problems and theoretical proposals
Journal of Ethnic and Migration Studies, 26
(2017)
The Manifesto Data Collection. Manifesto Project (MRG/CMP/MARPOR)
H. Haas, Katharina Natter, S. Vezzoli (2014)
Growing Restrictiveness or Changing Selection? The Nature and Evolution of Migration Policies1
International Migration Review, 52
Terri Givens, A. Luedtke (2005)
European Immigration Policies in Comparative Perspective: Issue Salience, Partisanship and Immigrant Rights
Comparative European Politics, 3
D. Tichenor (2008)
Strange Bedfellows: The Politics and Pathologies of Immigration Reform
Labour, 5
Miriam Gutekunst (2015)
Language as a new instrument of border control: the regulation of marriage migration from Morocco to Germany
The Journal of North African Studies, 20
E. Meyers (2004)
A Theory of Immigration Policy
E. Neumayer, Thomas Plümper (2009)
Spatial Effects in Dyadic Data
International Organization, 64
D. Massey (1999)
International Migration at the Dawn of the Twenty-First Century: The Role of the State
Population and Development Review, 25
M. Schain (2008)
Commentary: Why political parties matter
Journal of European Public Policy, 15
Ineke Valk (2003)
Political Discourse on Ethnic Minority Issues
Ethnicities, 3
(2012)
Average change in migration policy restrictiveness according to the (right-left) political ideology of parliaments, disaggregated by type of migrant group (1970–2012
T. Hatton, J. Williamson (2004)
Refugees, Asylum Seekers and Policy in Europe
Development Economics
T. Eule (2014)
Inside Immigration Law: Migration Management and Policy Application in Germany
Laura Block, S. Bonjour (2013)
Fortress Europe or Europe of rights? : the Europeanisation of family migration policies in France, Germany and the Netherlands
European Journal of Migration and Law, 15
Leah Haus, W. Cornelius, Philip Martin, J. Hollifield (1995)
Controlling Immigration: A Global Perspective.
International Migration Review, 30
B. Bowdish
Birds of Passage
Scientific American
Hein Haas, Katharina Natter, S. Vezzoli (2015)
Conceptualizing and measuring migration policy change
Comparative Migration Studies, 3
(2006)
Influencing Migration Policy from Inside: Political Parties.
Charles Shipan, C. Volden (2008)
The mechanisms of policy diffusion
American Journal of Political Science, 52
(2017)
Comparative Political Data Set 1960–2015.
D. Tichenor (2002)
Dividing Lines: The Politics of Immigration Control in America
E. Meyers (2000)
Theories of International Immigration Policy — A Comparative Analysis 1
International Migration Review, 34
L. Morales, J. Pilet, D. Ruedin (2015)
The Gap between Public Preferences and Policies on Immigration: A Comparative Examination of the Effect of Politicisation on Policy Congruence
Journal of Ethnic and Migration Studies, 41
G. Freeman, Randall Hansen, David Leal (2013)
Immigration and public opinion in liberal democracies
(2015)
DEMIG TOTAL
(1999)
Defining Immigration Policy: Inventory, Quantitative Referents, and Empirical Regularities.
Michael Piore (1979)
Birds of passage: Index
M. Ruhs (2013)
Chapter 1. The Rights of Migrant Workers Reframing the Debate
Telésforo García (2019)
Reseña: High-skilled Migration. Drivers and Policies
Revista Latinoamericana de Población, 13
Federica Infantino (2010)
La frontière au guichet. Politiques et pratiques des visas Schengen à l’Ambassade et au Consulat d’Italie au Maroc
, 7
M. Laakso, R. Taagepera (1979)
“Effective” Number of Parties
Comparative Political Studies, 12
Leah Haus (1999)
Labor Unions and Immigration Policy in France 1
International Migration Review, 33
Julie Watts (1998)
Strange Bedfellows: How Spanish Labor Union Leaders and Employers Find Common Ground on Immigration
Policy Studies Journal, 26
Fraser Duncan, S. Hecke (2008)
Immigration and the transnational European centre-right: a common programmatic response?
Journal of European Public Policy, 15
S. Bonjour (2011)
The Power and Morals of Policy Makers: Reassessing the Control Gap Debate 1
International Migration Review, 45
M. Czaika, H. Haas (2013)
The Effectiveness of Immigration Policies
Population and Development Review, 39
M. Ruhs, Philip Martin (2008)
Numbers vs. Rights: Trade-Offs and Guest Worker Programs 1
International Migration Review, 42
G. Freeman (1995)
Modes of Immigration Politics in Liberal Democratic States 1
International Migration Review, 29
Amber Davis (2012)
The Impact of Anti-Immigration Parties on Mainstream Parties' Immigration Positions in the Netherlands, Flanders and the UK 1987-2010: Divided electorates, left-right politics and the pull towards restrictionism

Publisher: Taylor & Francis
Copyright: © 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
ISSN: 2474-736X
DOI: 10.1080/2474736X.2020.1735255
Publisher site: See Article on Publisher Site

Abstract

POLITICAL RESEARCH EXCHANGE 2020, VOL. 2, 1735255 https://doi.org/10.1080/2474736X.2020.1735255 RESEARCH ARTICLE Political party ideology and immigration policy reform: an empirical enquiry a b c Katharina Natter , Mathias Czaika and Hein de Haas a b Institute of Political Science, University of Leiden, Leiden, Netherlands; Department for Migration and Globalization, Danube University Krems, Krems, Austria; Department of Sociology, University of Amsterdam, Amsterdam, Netherlands ABSTRACT ARTICLE HISTORY Received 29 November 2018 What drives the restrictiveness of immigration reforms? To what Accepted 8 January 2020 extent does the political ideology of parties in government and parliament matter? Drawing on immigration policy data oﬀering KEYWORDS unprecedented historical and geographical coverage, we analyse Immigration policy; political the drivers of immigration reforms in 21 Western immigration parties; migration; policy countries between 1970 and 2012. Our results show that there is determinants; policy no robust eﬀect of the political ideology of governments and diﬀusion parliaments on the overall restrictiveness of immigration reforms. Partisan eﬀects are limited to certain migration policy areas, primarily to integration policies, and to certain migrant groups, particularly asylum seekers and undocumented migrants. In contrast, political party ideology does not fundamentally shape decisions on the core of immigration regimes, such as entry policies or policies towards labour and family migrants. Our ﬁndings also showcase the importance of international policy diﬀusion and of trade-oﬀs between reforms in diﬀerent policy areas. Overall, the analysis highlights that although immigration is subject to heated debates in the public sphere and extensive political bargaining, the actual policies enacted seem primarily driven by factors such as economic growth, social welfare protection and the structure of political systems that are largely independent of the political ideology of parties in power. Introduction It is widely assumed that left-wing governments tend to be pro-immigration and that right-wing governments are generally in favour of restrictive immigration policies. This assumption can be questioned on two theoretical grounds: First, left- and right-wing parties tend to be internally divided on immigration. Typically, right-wing economic market liberalists and left-wing cosmopolitan-humanitarian streams are thought to be pro-immigration, while right-wing cultural conservatives and left-wing economic protec- tionists are expected to adopt more restrictive policy positions (Perlmutter 1996). As a con- sequence, we can expect party positions to vary according to migrant categories and CONTACT Katharina Natter [email protected] © 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (http:// creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 K. NATTER ET AL. policy issues at stake. For instance, right-wing parties may favour labour immigration, while left-wing parties may see immigration of labour migrant as harming the interests of native workers. Second, the need to compromise and bargain at diﬀerent levels of policymaking – between the executive and the legislative, between parties of a coalition government, or between governments and important lobbies – may drive governments to water down initial ideological positions. This can go both ways: The inﬂuence of business lobbies may lead parties with a relatively ‘tough’ immigration rhetoric to ultimately adopt less stringent policies, while parties with liberal positions on immigration might decide to abandon some of their policy proposals in an electoral strategy to not lose voters worried about high immigration. For ideological and strategic reasons, we can thus expect party positions on immigra- tion to be systemically incoherent, varying according to the policy issue at stake and leading to the adoption of ‘mixed’ policy packages that are characterized by trade-oﬀs and compromises reﬂecting the diversity of interests within and between parties. This begs the question whether, ultimately, there is a diﬀerence in immigration policy reforms enacted under left- or right-wing dominated governments or parliaments. It also shows the need to assess the overall importance of partisan eﬀects compared to other immigration policy determinants, such as the inﬂuence of economic factors. In other words: To what extent does the political orientation of parties in power aﬀect changes in immigration policy restrictiveness? A number of prior studies have provided initial insights into the inﬂuence of party orien- tations on speciﬁc immigration policies (for a review, see: Bale 2008; Meyers 2000, 1257– 1261). Such studies generally yielded mixed results and were often limited to one country or a relatively short time period. There are two main exceptions: Givens and Luedtke (2005) studied immigration laws in France, Germany, and the UK over the 1990–2002 period and showed that partisanship does not signiﬁcantly aﬀect the restrictiveness of immigration laws, but is a signiﬁcant predictor for the restrictiveness of integration laws. More recently, Abou-Chadi (2016) analysed immigration policies of nine European countries, Australia and Canada between 1980 and 2006, showing that left-wing governments are more likely to pass liberal immigration reforms, but only if they are in control of both executive and legislative powers. However, the overall lack of pertinent data on immigration policy has limited the ability of researchers to conduct systematic, quantitative comparisons of the eﬀects of party politics on immigration policy restrictiveness over longer time periods and a large number of countries. In particular, it has obstructed an analysis of partisan eﬀects on diﬀerent areas of migration policy or across diﬀerent migrant groups. Given the selective nature of immigration policies (de Haas, Natter, and Vezzoli 2018), such diﬀerentiation is however crucial. In addition, a one-sided focus on the ideological orien- tation of political parties may easily overestimate their importance in shaping immigra- tion policies. A proper assessment of partisan eﬀects can only be achieved if we do so in conjunction with analysing other plausible drivers of immigration reform, such as econ- omic growth, unemployment levels, the structure of political system or the international diﬀusion of policy practices. This paper seeks to address these limitations and expands the empirical and theoretical scope of research on this issue (1) by investigating how partisan eﬀects shape diﬀerent types POLITICAL RESEARCH EXCHANGE 3 of immigration policies in distinct and potentially heterogeneous ways and (2) by embedding the analysis of partisan eﬀects within a broader assessment of structural determinants of immigration policy reform. To do so, this paper draws on the DEMIG POLICY database and investigates the drivers of immigration policy reform in 21 Western immigration countries over the 1970–2012 period. Party politics and immigration policy reform Since the 1990s, immigration has become an increasingly important topic into most Euro- pean and North American elections. Confronted with the rise of far-right populist parties, parties both at the left and the right of the political spectrum have sought to win over voters – or at least not to alienate their traditional constituencies – through promising tougher border controls, as well as stricter access for migrants to labour markets or welfare provisions (Alonso and da Fonseca 2011; Davis 2012). Restrictive discourses on immigration are thus not a prerogative of the right. Nonetheless, political parties continue to diﬀer signiﬁcantly in their discourses on migration, leaving the impression that right- wing parties are more restrictive towards immigration than left-wing parties. For instance, while the conservative Christian Democratic Union (CDU) and far-right Alternative for Germany (AfD) in Germany opposed dual nationality and the granting of local political rights to foreigners in the run-up to the 2017 parliamentary election, the Social Democratic Party (SPD), the Green party, the left Die Linke and also the liberal Free Democratic Party (FDP) supported such measures. Campaigning for the 2015 UK parliamentary elections, most parties promised to cap legal immigration, albeit to diﬀerent degrees: Labour suggested to only cap the number of non-EU workers, the Conservatives announced it would keep annual net immigration within the ‘tens of thousands’, while the far-right UK Independency Party (UKIP) suggested a cap of 50,000 highly skilled workers per year together with a complete ban on low-skilled workers for a 5-year period. A cursory look at the history of immigration policymaking, however, gives reason to question the idea that right-wing parties are more ‘anti-immigration’ than left-wing parties when it comes to the actually enacted immigration policy reforms. For instance, under the inﬂuence of industry lobbies, right-wing and centrist parties in the 1950s and 1960s often favoured (‘guest-worker’) labour immigration from Mediterranean countries to Western Europe. Right-wing parties were also more lenient to enshrine family reuniﬁca- tion rights for ideological reasons related to conservative family values (Bonjour 2011). Left-wing parties, on the other hand, were often more wary towards immigration, because this was seen as undermining the position of native workers and trade unions. At the same time, they advocated socio-economic and citizenship rights of already- entered migrants to defend their rights as workers and to prevent that their presence would cause a downward pressure on wages and employment conditions of native workers (Haus 1999). We, therefore, cannot simply classify the left as ‘pro’ and the right as ‘anti’ immigration. In fact, the migration issue does not neatly cut across the left–right spectrum (Massey 1999, 313; Odmalm 2011, 1076–1077; Schain 2008, 468). Sciortino (2000, 225) argued that migration divides each party internally between those close to the party’s economic tradition and those close to its socio-cultural tradition. Perlmutter (1996, 378) emphasized that left-wing parties have to accommodate conﬂicts between unions who traditionally 4 K. NATTER ET AL. Table 1. Internal party divisions on migration. Left Right Restrictive migration policy reform Dimension Economic tradition Socio-cultural tradition Ideology Market protectionism Value conservatism Actor Labour unions Cultural conservatives Liberal migration policy reform Dimension Socio-cultural tradition Economic tradition Ideology International solidarity Market liberalism Actor Liberal and ethnic groups Employer lobbies favour restrictive policies and human rights advocates or ethnic groups lobbying for more open policies; while right-wing parties tend to be divided between employers favouring immigration and cultural conservatives asking for immigration restrictions. Analyses of Spain (Wutts 1998) or the United States (Tichenor 2002, 2008; Zolberg 2006) have shown the prevalence of such ‘strange bedfellow’ coalitions between businesses and human rights advocates or between labour unions and cultural conservatives that cut right across the typical partisan cleavages. Table 1 summarizes these internal divisions. Several scholars have also argued that because of such potential internal conﬂicts and the electoral risk of their public exposure, mainstream parties have long tried to avoid public debate on migration (Perlmutter 1996; Triadaﬁlopoulos and Zaslove 2006). This ties in with the classic argument by Freeman (1995) that migration policies are often made through client politics in closed political arenas, which can only be inﬂuenced by highly organized groups. Particularly since the end of the Cold War in 1989 and the demise of the perceived ‘communist threat’, however, the political salience of immigration in the public sphere has increased (Freeman 2013; Geddes 2003; Sciortino and Colombo 2004; Van Der Valk 2003). Politicians and the media have frequently presented immigra- tion as a threat to employment and the welfare state, to social cohesion and cultural integ- rity. Since the attacks on the Twin Towers and the Pentagon in the United States in 2001, politicians have also linked migration to terrorism and have cast immigrants as an overall security threat. The concomitant rise of far-right anti-immigration parties has further shifted immigration out of client politics and into the public domain (Davis 2012; Givens and Luedtke 2005, 7; Perlmutter 1996, 377–378; Triadaﬁlopoulos and Zaslove 2006, 172). The increasing politicization of migration may thus have diminished the inﬂuence of business lobbies and civil society groups while increasing the role of party politics on immigration policymaking. Yet, immigration policymaking is not an exclusively domestic aﬀair. National govern- ments closely monitor policy developments in neighbouring or partner countries, facilitat- ing the spread of policy practices regardless of the party in power (Cornelius et al. 2004; Meyers 2002). For instance, if several countries introduce policies to attract the high- skilled, this may motivate other governments to introduce similar policies as part of an international competition for talent (Czaika 2018). The other way around, the introduction of more restrictive asylum policies by some countries may motivate other governments to also restrict asylum out of fear of a large-scale increase in asylum applications (Hatton and Williamson 2004). It seems plausible that globalization and the formation of regional gov- erning systems such as the European Union have increased the importance of inter- national policy diﬀusion (Block and Bonjour 2013; Lavenex 2001). The politicization of immigration and the growing importance of international policy diﬀusion suggest that parties’ impact on policy may have changed over time, but the POLITICAL RESEARCH EXCHANGE 5 two trends seem to work in diﬀerent directions. On the one hand, the growing political salience of migration and the need of mainstream parties to position themselves more clearly on the political spectrum suggests an increasing polarization of party ideologies over time. On the other hand, the growing importance of supra-national policymaking and the concomitant convergence of policy trends can be expected to decrease party inﬂuences on migration policymaking. To further complicate the picture, immigration policies are not homogenous but typi- cally consist of ‘mixed bags’ of policy measures targeting diﬀerent migrant groups and policy areas in distinct ways. Large-scale immigration reforms are rarely purely restrictive or liberalizing, but generally comprise measures that grant certain entry and post-entry (integration) rights to certain groups while closing opportunities for others (de Haas, Natter, and Vezzoli 2018). We know for instance that entry and post-entry rights of most labour and family migrants have generally increased since 1945 partly under the inﬂuence of human rights law and shifting ideologies (Bonjour 2011), while border con- trols and policies targeting unauthorized migrants and asylum seekers have generally become more restrictive (de Haas, Natter, and Vezzoli 2018). In this vein, Money (1999, 37) found that immigration control (entry) and immigrant integration (post-entry) policies follow diﬀerent political logics: While integration splits political parties neatly between the left and the right, pro- and anti-migration control forces exist both within the left and the right. This shows that migration policies can be contradictory and incoherent ‘by design’, because they are inﬂuenced by diﬀerent ideologies and interests, and because they target various immigrant groups through selective immigration policies. Immigration policymak- ing may, therefore, become subject of bartering, in which political parties compromise to guarantee the immigration of their ‘favourite’ immigrant group. This suggests that the restrictiveness of immigration policy reform enacted by a speciﬁc political party is likely to vary according to the type of policy (border controls, entry or inte- gration) and in terms of the migrant categories targeted (such as high- and low-skilled labour migrants, family members, or asylum seekers). We could, for instance, expect that left-wing parties favour more liberal policies towards undocumented migrants and refugees because values such as international solidarity and universal human rights are central to their party ideology, while right-wing parties favour immigration of both lower- and higher-skilled workers because of the inﬂuence of business lobbies. Similarly, left-wing parties are likely to favour integration policies that grant labour, welfare and citi- zenship rights to settled migrants, while right-wing parties may oppose such policies to maintain a more malleable, ﬂexible and ‘returnable’ immigrant workforce. Data and methodology To test these hypotheses and gain more systematic and comprehensive insights into the role of political party ideology in shaping the restrictiveness of immigration policy reform, this paper investigates 21 Western immigration countries over the 1970–2012 period, drawing on the DEMIG POLICY database. The database deﬁnes migration policies as the ‘rules (i.e. laws, regulations and measures) that national states deﬁne and [enact] with the objective of aﬀecting the volume, origin, direction, and internal composition of […] migration’ (Czaika and de Haas 2013, 489). For the 21 countries included in this paper, DEMIG POLICY recorded more than 2600 migration policy changes over the 6 K. NATTER ET AL. 1970–2012 period. The database records changes in ‘policies on paper’ and does therefore not capture the leverage bureaucrats (Eule 2014; Infantino 2010) and private actors (Gute- kunst 2015) have in the implementation of policy measures. Taking this into account would be a future valuable contribution to the literature. Because migration policy reforms generally consist of diﬀerent (and potentially con- tradictory) measures, DEMIG POLICY disaggregates policy packages by breaking them down into individual policy measures. Each policy measure was coded to identify the relevant policy area (border and land control; legal entry and stay; integration and post-entry rights; and exit and return regulations) and the migrant category targeted (e.g. high-skilled workers, low-skilled workers, family members, undocumented migrants, asylum seekers). This enables the analysis of policymaking trade-oﬀs between diﬀerent policy areas or migrant groups. For more information on the concep- tual and methodological underpinnings of DEMIG POLICY, see de Haas, Natter, and Vezzoli (2015). The dependent variable is the change in restrictiveness introduced by an immigration policy measure enacted in a speciﬁc country in a given year. Each policy measure was coded as a change towards more restriction (+1) or less restriction (−1), according to whether it made the existing legal framework more or less restrictive, that is, whether it increased or decreased the rights granted to the migrant group targeted by that policy change. This focus on changes in restrictiveness allows to explore the drivers of immigra- tion policy reform regardless of the level of absolute openness or restrictiveness of a coun- try’s immigration regime. To add nuance, we also weighted changes in restrictiveness according to their magnitude as ‘ﬁne-tuning’, ‘minor’, ‘mid-level’ or ‘major change’, leading to a 9-point change in restrictiveness scale from −4to 4. To provide a robustness check, all analyses in this paper were performed using both the weighted and unweighted variables. With regards to independent variables, our main variable of interest is the political orientation of parties in power. We have consistently run all model speciﬁcations using two alternative measures of parties in power: government composition and parliament composition. First, we used a variable from the Comparative Political Data Set (CPDS) that measures the composition of governments since 1960 through the percentage of cabinet posts held by each party in government (Armingeon et al. 2017). The percentage of each party in government is weighted by information on its right-left ideological pos- ition as provided by the Comparative Manifestos Project (CMP) dataset, which codes the party programmes of all electorally relevant parties in 54 countries since 1945 (Volkens et al. 2017). High (low) scores of this government composition variable represent govern- ments dominated by right-(left-)wing political ideologies. Second, we constructed a parlia- ment composition variable in a similar way, weighting the right-left position of parties represented in Parliament (provided by the CMP dataset) by the number of parliamentary seats they hold, with positive (negative) numbers indicating a dominance of right-(left- )wing parties in parliament (see Appendix Table A1 for descriptive statistics). The two vari- ables are signiﬁcantly correlated (0.622) as the government composition partly reﬂects the composition of the parliament, depending on the electoral system. There are exceptions, such as in the case of Austria over the 1990s and 2000s, where the government was a coalition between social-democrats and conservatives, but the far-right had a strong voice in the parliament. POLITICAL RESEARCH EXCHANGE 7 We also explored the role of international policy emulation and diﬀusion in explaining national immigration policy reform (Shipan and Volden 2008) by constructing a so-called spatial dependence variable (Neumayer and Plümper 2010). This variable captures immi- gration policy changes in all other countries k in the dataset and tests the inﬂuence of (weighted) trends in general immigration policy trends on national immigration policy- making y of country i at time t according to: it y = m v y + bX + 1 . (1) it ik kt it it The weighting matrix v proxies the relative connectivity between N (=45) countries i ik and N countries k using linguistic distance as connectivity variable (Neumayer and Plümper 2010). The spatial autoregression parameter m identiﬁes the eﬀect of the spatially lagged dependent variable on policy changes y . International policy diﬀusion it can impact national policymaking both towards more and less restriction: One the one hand, the increased role of supra-national institutions may limit the inﬂuence of national governments and party preferences on immigration restrictions. On the other hand, inter- national policy diﬀusion can also trigger a ‘race to the bottom’ towards progressively dis- mantling migrant rights, particularly for lower-skilled migrants or, alternatively, a race to attract the ‘best and brightest’, particularly with regards to higher-skilled migrants. To account for the structural context within which political parties operate and migration policies are made, we introduced four sets set of control variables measuring (1) demographic, (2) economic, (3) political and (4) welfare factors that may aﬀect the restrictiveness of immigration policies. First, demographic indicators include population ageing (measured through the share of the population over 65 years) and recent immigra- tion levels (measures through immigration rates as a percentage of total population drawn from the DEMIG TOTAL database (DEMIG 2015)) to capture the demand for migrant labour as well as perceptions of immigration. While it has been regularly argued that ageing societies create a higher demand for migrant labour, high levels of recent immigration may create political pressures to tighten entry policies. Second, we included two economic cycle variables (one-year lagged GDP per capita growth and one-year lagged change in unemployment rates) and two variables measuring economic structure (trade openness measured by the annual trade volume as percentage of GDP, and the strength of trade unions measured by net union membership as a pro- portion of employed wage and salary earners (Armingeon et al. 2017)). The economic cycle control variables were included based on the expectation that growing economies with low unemployment create more demand for migrant labour and that this may increase the power of economic lobbies favouring liberal immigration policies. With regards to economic structure variables, we can expect open economies to have more liberal immigration policies while trade union power may push policies in a more restric- tive direction. Third, we also expect political system variables to shape the inﬂuence of political party ideology on immigration policy reform. As Schmidt (1996) has shown, party preferences tend to be watered down in political systems that have a lot of inbuilt room for nego- tiation. We would, therefore, expect that both presidential systems (measured through a binary variable capturing presidentialism) and federal systems (measured through the absent, weak or strong levels of federalism) enact overall more liberal changes because 8 K. NATTER ET AL. of the need to negotiate laws over multiple decision-making structures. On the other hand, the more fragmented the party landscape, the more restrictive we would expect immigration reforms to be, given the better representation of fringe parties – both on the left and the right side of the political spectrum – with more radical views on immigra- tion. We expect that institutional fragmentation (measured through the fractionalization of 2 2 the party system, index by EF = 1/ v , where v captures the squared share of votes i=1 it it for party i in an election at time t (Laakso and Taagepera 1979)) increases the tendency of parties in power to introduce restrictive policies. Lastly, we included two variables to capture the structure of a country’s welfare system (measured by social security spending in the percentage of GDP and the strictness of employment protection) in the expectation that immigration policies will be more restric- tive in countries where migrants have access to strong welfare state arrangements. Next to these four sets of control variables, we also account for possible structural changes in general immigration policy trends (i) after the fall of the Berlin Wall in 1989 and (ii) after the 9/11 attacks in the United States in 2001. By introducing two time dummies, we test to what extent immigration policies have been aﬀected by the end of communism and the start of the so-called ‘war on terror’ which is often associated to the increasing ‘securitization’ of migration. An overview of all variables used can be found in Table A1 (Appendix). The empirical analysis is divided in a descriptive and multivariate part. The descriptive analysis in Section ‘Disaggregating immigration policy reform’ compares trends in migration policy restrictiveness under left- and right-wing governments. To account for the increasingly targeted and selective nature of immigration policies (de Haas, Natter, and Vezzoli 2018), we zoom in on immigration policy changes targetting diﬀerent policy areas and migrant groups. Sections ‘Putting partisan eﬀects in context’ and ‘Disen- tangling partisan eﬀects’ use multivariate regression analyses to investigate the extent to which the restrictiveness of immigration policy reform is driven by the political orientation of parties in power. The analyses are performed separately by policy areas and migrant groups targeted, as well as for the three diﬀerent time periods (1970–1988, 1989–2000, 2001–2012). Looking at changes over time also allows us to assess whether – and in which policy areas – party preferences on immigration have converged or become more polarized. We also investigate the importance of international diﬀusion dynamics and policy trade-oﬀs in immigration policymaking. Disaggregating immigration policy reform Figure 1 displays changes in migration policy restrictiveness, as well as the prevalence of left- and right-wing parties in governments over time. It shows that right-wing parties have dominated governments between the mid-1980s and the mid-2000s, while left-wing dominated governments have prevailed across the 21 countries understudy in the 1970s and since the mid-2000s. The patterns for parliaments reﬂect similar patterns (see Figure A1 in the Appendix). Crucial for this paper is that in both instances, the linear trendline capturing changes in migration policy restrictiveness seems not related to the inverted-U trend characterizing shifts in partisan compositions of governments or parliaments over time. In fact, the ﬁgures show that the average annual change in immigration policy restrictiveness has POLITICAL RESEARCH EXCHANGE 9 Figure 1. Trends in government composition and changes of migration policy restrictiveness (1970– 2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The govern- ment composition index gives an insight into the overall ideological positioning of the government. (Far) left-wing domi- nated governments have a (strongly) negative index score, whereas (far) right-wing governments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and government composition. remained below zero, and that therefore, since the 1970s, immigration policies have con- sistently been liberalized. The observable upward trend does therefore not represent an increase in restrictiveness, but the fact that over recent decades restrictive and liberal policy reforms have increasingly balanced each other out, indicating a decelerated liberal- ization of immigration policy change (de Haas, Natter, and Vezzoli 2018). This contradicts common perceptions that migration policies have become more restrictive, indicating a considerable gap between migration discourse and migration policy practice, and/or the tendency of restrictive policies to receive more media attention. It also underlines that the deceleration of liberalization has occurred regardless of the ideological compo- sition of parliaments or governments. What this ﬁgure does not show is that although policies have not become more restric- tive, immigration policymaking has gone through a process of considerable intensiﬁcation over the past decades. Both left- and right-wing parliaments having nearly tripled the average number of immigration-related laws they enact annually between the late 1980s and early 2000s. This seems to reﬂect the growing political salience of migration since the end of the Cold War, as well as the increasingly complex and selective nature of immigration policies. The intensity of immigration policymaking has stabilized since the early 2000s, suggesting either a decreasing political salience of migration because of the rise of other issues (such as ‘terrorism’), or a certain level of saturation after a major wave of policymaking on new border control technologies and skill-speciﬁc immi- gration policies. The heatmap (Figure 2) allows to look simultaneously at changes in immigration policy restrictiveness and government composition over time. In general, no clear pattern emerges: Apart from the early 1970s, during the ‘guest-worker’ era, when left-wing gov- ernments enacted particularly liberal policy changes, and the late 1990s and early 10 K. NATTER ET AL. Figure 2. Changes in migration policy restrictiveness according to the (right-left) political ideology of governments (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The govern- ment composition index gives an insight into the overall ideological positioning of the government. Areas coloured light or dark blue show governments dominated by a right-wing political ideology, areas coloured in green, yellow or red show governments dominated by a left-wing political ideology. 2000s, when right-wing governments enacted more restrictive policies, governments across the entire political spectrum seem to enact liberal and restrictive policies to a largely similar extent. A comparable pattern is also visible when looking at migration policy changes according to parliament composition over time (see Figure A2 in the Appendix). The fact that trends in immigration policy restrictiveness are largely detached from government composition might suggest that the restrictiveness of immigration pol- icies is primarily determined by broader economic, geopolitical and social factors rather than by ideological preferences and party politics per se. It is however also possible that these graphs conceal diﬀerences between immigration policies targetting particular policy areas or migrant groups. Figure 3 shows that restrictive policy reforms mainly focus on border control and exit policies (which include deportation) – and this regardless of the political party in power. Integration policies have moved into a more liberal direction over the entire 1970–2012 period under both right- and left-wing dominated governments, but the tendency towards liberalization has been signiﬁcantly stronger under left-wing governments. Entry policies – the very core of immigration policymaking – have shown the most consist- ent liberalizing tendency, with right-wing parties tending towards slightly more liberal positions in this policy area. Again, a comparable pattern is visible when looking at the restrictiveness of particular immigration policy areas according to the right-left compo- sition of parliaments, although right-wing dominated parliaments seem to be on average associated with slightly more restrictive border control policies and less liberal entry policies when compared to trends under right-wing governments (see Figure A3 in Appendix). When disaggregating immigration policy reform according to migrant categories (Figure 4), we see that there is no left/right diﬀerentiation with regards to policies POLITICAL RESEARCH EXCHANGE 11 Figure 3. Average change in migration policy restrictiveness according to the (right-left) political ideol- ogy of governments, disaggregated by policy area (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated governments have a (strongly) negative index score, whereas (far) right-wing governments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and government compo- sition by migration policy area. targetting family migrants and that both right- and left-wing dominated governments tend to back policies that expand the rights of (high- and low-skilled) labour migrants and restrict rights for undocumented migrants. However, in general, right-wing domi- nated governments seem to enact comparatively more restrictive policy changes towards most migrant categories, particularly undocumented migrants and asylum seekers. The major exception is high-skilled migration policies, which have become Figure 4. Average change in migration policy restrictiveness according to the (right-left) political ideol- ogy of governments, disaggregated by migrant group targeted (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated governments have a (strongly) negative index score, whereas (far) right-wing governments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and government compo- sition by migrant target group. 12 K. NATTER ET AL. more liberal across the political spectrum but seem particularly popular among right-wing dominated governments. Comparable patterns are visible when looking at trends in policy restrictiveness towards certain migrant groups according to the ideological composition of parliaments (see Figure A4 in Appendix). To verify whether these ﬁndings hold when controlling for the eﬀects of other migration policy drivers, the following sections move to multivariate regression analysis. We ﬁrst concentrate on the importance of parliamentary and governmental party compo- sition and discuss the role of structural factors captured by economic, political, welfare and demographic control variables. We then disaggregate the analysis by policy areas and migrant groups and discuss changes over time. We also zoom into the role of international policy diﬀusion and the question whether there are trade-oﬀs between diﬀerent policy issues in terms of restrictiveness. Putting partisan eﬀects in context Table 2 presents the results of the logistic regressions (models 1–2, 5–6), as well as all ordinary least square regression (models 3–4, 7–8) pooled over the entire 1970–2012 period. Overall, no clear pattern along a left–right continuum emerges and the association between the political ideology of parties in power and the restrictiveness of immigration policy change is rather weak and inconclusive. Interestingly, the analysis seems to suggest that – if at all – then the partisan composition of parliaments matters more for immigration policy restrictiveness than the partisan composition of governments, with right-wing dominated parliaments enacting on average more restrictive migration policies. This could point to the role of right-wing opposition parties in pushing the entire political spec- trum towards adopting more restrictive positions on immigration. However, this eﬀect (in model speciﬁcations 6 and 8) almost disappears once we control for the eﬀect of other migration policy drivers. A more ﬁne-grained analysis is, however, necessary to identify whether the eﬀects (or the lack thereof) hold when we disentangle the analysis of partisan eﬀects along policy areas and migrant groups targetted (see Section ‘Disentangling parti- san eﬀects’). As this analysis does not show a clear relation between the political ideology of parties in power and the overall restrictiveness of immigration policy change, it is important to assess the role of other plausible drivers of immigration policy reform. The analysis shows that economic growth is strongly and signiﬁcantly associated to the adoption of liberal policy changes. This is consistent with the idea that business cycles directly aﬀect the political willingness for allowing immigration. Irrespective of the political orientation of governments, economic lobbies pressuring governments toward more liberal policies gain power during the time of economic growth and high labour demand (Massey 1999; Meyers 1995). The pro-immigration tendency of a certain governement may thus be reinforced under circumstances of high economic growth and maybe counterbalanced in times of economic adversity when public pressure to reduce migration may be higher. Surprisingly, we do not ﬁnd any signiﬁcant eﬀect of unemployment on overall changes in immigration policy restrictiveness. The weak relation between unemployment and immigra- tion policy reform might be explained by the fact that the GDP growth variable absorbs part of the expected eﬀects. Alternatively, although unemployment is correlated with GDP growth, the specialized and segmented nature of labour markets can explain why structural POLITICAL RESEARCH EXCHANGE 13 Table 2. Determinants of immigration policy reform in 21 countries, 1970–2012. DV: Immigration (1) (2) (3) (4) (5) (6) (7) (8) policy change 0/1 0/1 −4/+4 −4/+4 0/1 0/1 −4/+4 −4/+4 Cabinet composition 0.00231 0.00200 0.00515 0.00381 (0.00333) (0.00429) (0.00320) (0.00358) Parliament 0.0419* 0.0351 0.0679** 0.0675** composition (0.0241) (0.0307) (0.0283) (0.0266) Policy diﬀusion 0.224*** 0.232*** 0.257*** 0.263*** (0.0582) (0.0584) (0.0775) (0.0774) Leg. 0.0753 0.0691 0.116* 0.112* fractionalisation (0.0474) (0.0474) (0.0637) (0.0638) Presidential system −0.346* −0.379** −0.510** −0.540** (0.178) (0.180) (0.238) (0.240) Federalism (ref: no) Weak −0.247 −0.324* −0.293 −0.359 (0.191) (0.196) (0.250) (0.260) Strong −0.341** −0.329** −0.421** −0.415** (0.135) (0.135) (0.181) (0.181) Trade openness −0.00131 −0.000848 −0.00198 −0.00159 (0.00129) (0.00133) (0.00182) (0.00186) Union density −0.0121*** −0.0115*** −0.0152*** −0.0148*** (0.00385) (0.00385) (0.00506) (0.00506) GDP growth (lag) −0.0755*** −0.0765*** −0.0924*** −0.0931*** (0.0269) (0.0269) (0.0341) (0.0341) Unemployment (lag) 0.00864 0.00748 5.87e−05 −0.000672 (0.0172) (0.0172) (0.0229) (0.0229) Immigration rate −0.00162 −0.00189 −0.00259 −0.00282 (lag) (0.00320) (0.00316) (0.00354) (0.00354) Old age share (65+) −0.0401 −0.0403 −0.0755 −0.0758 (0.0373) (0.0365) (0.0499) (0.0490) Employment −0.430*** −0.376*** −0.458*** −0.416*** protection (0.115) (0.119) (0.150) (0.153) Social security (% 0.0692*** 0.0680*** 0.0754** 0.0746** GDP) (0.0251) (0.0251) (0.0333) (0.0332) Post-1989 (ref: pre- 0.369* 0.332 0.328 0.300 1989) (0.202) (0.202) (0.244) (0.244) Post-2001 (ref: pre- 0.442* 0.424* 0.528* 0.513* 1989) (0.230) (0.227) (0.277) (0.274) 0.368 0.265 0.921 0.839 −1.035** −1.175** −0.145 −0.311 Constant (0.457) (0.461) (0.612) (0.616) (0.449) (0.458) (0.235) (0.249) Observations 1690 1690 1757 1757 2332 2332 2621 2621 R-squared 0.0312 0.0323 0.034 0.034 0.0704 0.0716 0.069 0.071 Estimator Logit Logit OLS OLS Logit Logit OLS OLS Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. unemployment in particular sectors of the labour market may coincide with high GDP growth and labour shortages in other sectors (Piore 1979). Indeed, unemployment partly reﬂects structural mismatches between the supply and demand for labour. GDP growth, therefore, seems a better indicator of year-to-year variations in labour demand. The level of trade openness, which could perhaps also be seen as a proxy of the power of business lobbies, seems associated to liberal reforms, but the eﬀect is not signiﬁcant. Interestingly, union density is associated with liberal changes towards immigration. This may seem surprising because of the historical opposition – or at least scepticism – of labour unions towards labour immigration. The fact that union density is associated to more liberal entry policies might be explained by a shift in labour unions’ approach to migrant workers over time. Partly as a result of the increasing share of migrant workers among union members, unions conceive migrants not anymore as competition for native workers but rather as potential clientele. This also seems to conﬁrm that unions 14 K. NATTER ET AL. are more interesting in safeguarding the rights of migrants who are already in the country rather than opening the door to more immigration in terms of entry policies (Haus 1999). So, this eﬀect might particularly reﬂect eﬀorts by unions to achieve more favourable inte- gration policies so as to put migrant workers on an equal legal footing as native workers, and by so doing remove ‘unfair competition’ by rightless migrant workers. The ﬁndings do not conﬁrm the idea that high levels of past immigration cause a public backlash against immigration and calls for more restrictions (Massey 1999; Meyers 1995). The absence of any signiﬁcant eﬀect of recent immigration levels on overall changes in immigration policy restrictiveness might be linked to the so-called ‘opinion-policy gap’, which is the time between a change in circumstances and the moment public opinion shifts in reaction to it (Morales, Pilet, and Ruedin 2015). The absent eﬀect of past immigra- tion may also indicate that public opinion may only turn against immigration after sus- tained periods of high immigration and the build-up of large immigrant population, which is not captured by this variable. The number of over 65 old persons as a share of national populations has no clear eﬀects on immigration policy restrictiveness. This casts doubts on the idea that in ageing societies there would be more support for liberal immigration policies as a way to address labour shortages or to attract migrant care workers. More in general, it questions the popular assumption that demographic factors would directly impact migration trends (de Haas 2011). If such an eﬀect exists at all, it can also be counterbalanced by the fact that older persons tend to adopt more con- servative political positions. Looking at welfare systems, high social security spendings are associated to more restrictive immigration policy changes. This is consistent with the idea that immigration policies are a means to regulate the pathways of access into social welfare systems and herewith to limit access and safeguard national privileges. Generous welfare rights, there- fore, seem to create an incentive to restrict entry and to increase selection. This might conﬁrm that there is some trade-oﬀ in migration policymaking between the ‘numbers’ of migrants allowed in and the ‘rights’ granted to them (Ruhs 2013). In contrast, high employment protection is associated with liberal immigration reforms, which suggests that migrant labour is not necessarily an immediate threat for native workers in dual labour markets (see Piore 1979). In fact, high labour market protection allows governments to be more lenient towards immigration, as both business and trade unions may have an interest in favouring the immigration of migrants workers who take up jobs that native workers tend to shun. Vice-versa, immigration restrictions may be a policy tool for govern- ments to secure jobs for native workers in ﬂexible labour markets that oﬀer only low employment protection. Employment protection and immigration policies may thus be two alternative ways to deal with the consequences of labour migration for native workers. Depending on whether governments react to business lobbies or to electorate pressure, they might priviledge one policy instrument over the other. Political system variables also play a role in explaining the restrictiveness of migration policy reform, as federal and presidential systems tend to enact more liberal changes. This is consistent with the idea that the more actors are involved in policymaking, the higher the need for negotiation and political compromise. Federalism introduces the need to negotiate fundamental policy reforms not only with nationally elected representatives but also with regional representatives, which might ultimately water down initially restric- tive policy proposals. Similarly, presidentialism introduces a potential additional POLITICAL RESEARCH EXCHANGE 15 negotiation level between the executive and legislative and increases the necessity for political compromise and political bartering, particularly when the president and prime minister are not from the same political party. These dynamics, in turn, create a sustained gap between policy rhetoric and policy practice. In contrast, the tendency for fragmented party systems (proxied by legislative fractionalization) to enact more restrictive immigra- tion policies is not strongly signiﬁcant. This oﬀers grounds to question the widespread idea that in highly competitive party landscapes, parties in power are urged to ‘show per- formance’ in terms of immigration policy toughness. Finally, the two time dummies reveal only weak period eﬀects: the fall of the Berlin Wall in 1989 and the ensuing growth of (asylum) migration does not seem to have con- tributed to party-speciﬁc changes in immigration policy restrictiveness. This seems to counter the idea that the increasing political salience of immigration in the public sphere and the disappearance of the ‘Communist threat’ has led right-wing politicians to enact overall more restrictive immigration policies. In contrast, the 9/11 terrorist attacks had a small, but signiﬁcant eﬀect in driving restrictive immigration policy reforms regardless of the party in power, reinforcing the overall slow-down of the post-Second World War liberalization of immigration policies. At the same time, as our analysis in Section ‘Disentangling partisan eﬀects’ shows, these period-speciﬁc eﬀects are driven by policies targeting particular migrant groups and policy ﬁelds and do not aﬀect immigration regimes as a whole. In general, the analysis suggests that the political ideology of parties in power plays only a very limited role in explaining the restrictiveness of immigration policy reform in Western liberal democracies over the 1970–2012 period. Right- and left-wing dominated governments do not seem to enact opposite reforms on immigration, as they are faced with the same interest groups and international context in migration policymaking and therefore policy trends have been overall coherent across the left/right spectrum. As the analysis of control variables has shown, the ideological orientation of parties in power is but one of the many factors shaping migration policies, with economic cycles, employment protection, the power of trade unions as well as international policy diﬀusion playing a more important role in shaping the restrictiveness of immigration policy reforms. The following section zooms into the diﬀerent dimensions of immigration policy reform in order to test whether the absence of a clear eﬀect of political ideology holds if we disaggregate the analysis for particular policy areas or policies targetting par- ticular migrant groups. Disentangling partisan eﬀects The disaggregated analyses (Table 3) show that the partisan composition of governments and parliaments aﬀects ﬁrst and foremost integration policies, with right-wing dominated governments and parliaments enacting on average more restrictive integration policies. This clear partisan eﬀect, however, should not obscure the fact that integration policies have been liberalized in the past decades by parties from across the political spectrum. Other migration policy areas such as border control or exit policies, but most strikingly also entry policies – the core of immigration regimes – seem only weakly or not aﬀected by government or parliament composition. In particular, the absent eﬀect of par- tisan ideology on border controls seems to conﬁrm that there is a cross-partisan 16 K. NATTER ET AL. Table 3. Partisan determinants of immigration policy reform by policy area and by target group in 21 countries, 1970–2012. DV: Immigration policy (1) (2) (3) (4) (5) (6) (7) (8) change +4/−4 0/1 +4/−4 0/1 +4/−4 0/1 +4/−4 0/1 Policy area Cabinet composition Parliament interacted with policy composition area… interacted with policy area… Border −0.00637 0.00197 0.0839 0.102** (0.0105) (0.00732) (0.0614) (0.0453) Entry −0.00257 −0.00173 0.0304 0.0333 (0.00520) (0.00410) (0.0393) (0.0321) Integration 0.0290*** 0.0232*** 0.158*** 0.120*** (0.00677) (0.00548) (0.0557) (0.0447) Exit 0.00282 0.000968 0.0976 0.0952 (0.0120) (0.00860) (0.0808) (0.0609) Target group Cabinet composition Parliament interacted with target composition group… interacted with target group… All (unspeciﬁed) migration 0.00928 0.00715 0.0870* 0.0793** (0.00747) (0.00543) (0.0498) (0.0385) Labour migration 0.0217** 0.0168** 0.137* 0.118* (0.00916) (0.00753) (0.0803) (0.0627) Family migration 0.00900 0.00988 0.0618 0.0899 (0.00994) (0.00852) (0.0673) (0.0597) Skilled migration −0.0177** −0.0154** −0.0390 −0.0506 (0.00806) (0.00706) (0.0521) (0.0465) Undocumented migration 0.0176* 0.0135** 0.152** 0.130*** (0.00922) (0.00679) (0.0616) (0.0465) Asylum migration 0.0145* 0.0141** 0.0881* 0.0749* (0.00825) (0.00636) (0.0518) (0.0405) Constant −1.393** −1.090** −1.481** −1.168** −1.121* −0.854* −1.343** −1.046** (0.566) (0.451) (0.589) (0.457) (0.630) (0.467) (0.643) (0.476) Observations 2332 2332 2332 2332 2240 2240 2240 2240 (Pseudo)-R 0.089 0.075 0.085 0.073 0.087 0.074 0.085 0.074 Estimator OLS Logit OLS Logit OLS Logit OLS Logit Country FE Yes Yes Yes Yes Yes Yes Yes Yes Time FE Yes Yes Yes Yes Yes Yes Yes Yes Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. agreement that border controls are the central tool of immigration restrictions. This is in line with Massey et al. (1998, 288) who argued that ‘elected leaders and bureaucrats increasingly have turned to symbolic policy instruments to create an appearance of control’. These results suggest that – in contrast to entry policies where pro- and anti-immigra- tion positions cut across the political spectrum – policies targetting migrants socio-econ- omic and political post-entry rights tend to split the political spectrum more neatly along the right-left cleavage. Our results thus lend support to the idea put forward by Givens and Luedtke (2005) and Money (1999) that integration policies signiﬁcantly diﬀer under left- and right-wing governments, while politicians will open entry policies if economic and other contextual factors require it, regardless of their party aﬃliation (see Table 1). The most striking ﬁnding, however, is the absence of signiﬁcant partisan eﬀects on entry pol- icies, the core of a country’s immigration regime. If we look at policies targeting speciﬁc migrant groups, the most consistent and robust result is that left-wing dominated governments and parliaments seem to be more liberal in POLITICAL RESEARCH EXCHANGE 17 granting (post-entry) rights to labour migrants, asylum seekers and undocumented migrants, whereas right-wing dominated governments seem to be slightly more favour- able towards skilled migrants. For high-skilled migration, the analysis suggests that right-wing governments are more likely to enact liberal changes while the partisan com- position of parliament does not seem to play a signiﬁcant role. This might point at the more direct inﬂuence of business lobbies in governmental priorities as opposed to the leg- islative process. We do not see clear diﬀerences in partisan eﬀects with regards to family migration. Importantly, all these group-speciﬁc results seem to pertain to integration pol- icies, as we failed to ﬁnd a signiﬁcant eﬀect of partisan ideology on entry, exit and border control policies. Developments over time (Tables A2 and A3 in Appendix) provide additional insights: In fact, the results just discussed can be partly explained by general trends in particular periods. Most strikingly, the analysis suggests that the party landscape polarized on inte- gration policies particularly in the pre-1989 and post-2001 period. We could not detect consistent period eﬀects for other policy areas. When looking at migrant groups, it seems that the positive eﬀect of left-wing governments and parliaments on labour migration and asylum policy is particularly explained by a policy polarization on the issue after 2001. In contrast, the positive eﬀect of right-wing ideology for high-skilled migration policies seems to be relevant only for the pre-2001 period, after which there seems to be a cross-partisan agreement on attracting the highly skilled. Overall, the ideo- logical composition of parliaments seems to be slightly more relevant for explaining diﬀer- ences in policy outcomes in speciﬁc periods – for instance, the drive towards restrictive border control policies in the 1990s – compared to the ideological composition of governments. Table 4. Policy diﬀusion eﬀects and policy trade-oﬀs in immigration policy reforms of 21 countries, 1970–2012. (1) (2) (3) (4) (5) (6) (7) (8) Variables Border Admission Integration Exit Border Admission Integration Exit Cabinet −0.00126 −0.00176 0.00559*** 0.00145 composition (0.0013) (0.0025) (0.0018) (0.0012) Parliament 0.0107 0.0331* 0.0167 0.00633 composition (0.00879) (0.0190) (0.0141) (0.00794) Policy diﬀusion 0.468** 0.0889*** 0.429** −0.0474 0.467** 0.532** 0.451** −0.0372 (0.225) (0.0157) (0.213) (0.241) (0.225) (0.218) (0.213) (0.239) Border 0.0889*** 0.0494*** −0.0325*** 0.0880*** 0.0473*** −0.0332*** (0.0157) (0.0129) (0.00833) (0.0157) (0.0128) (0.00832) Admission 0.0249*** −0.0426*** 0.0123*** 0.0247*** −0.0439*** 0.0119*** (0.00466) (0.00657) (0.00424) (0.00464) (0.00658) (0.00426) Integration 0.0212*** −0.0715*** 0.0133** 0.0202*** −0.0734*** 0.0140** (0.00609) (0.0108) (0.00589) (0.00600) (0.0107) (0.00584) Exit −0.0347*** 0.0453*** 0.0312** −0.0353*** 0.0439*** 0.0327** (0.00892) (0.0160) (0.0135) (0.00891) (0.0161) (0.0134) Constant −0.0598 −0.181 0.166 −0.0639 −0.105 −0.300* 0.173 −0.0690 (0.0772) (0.172) (0.125) (0.0625) (0.0817) (0.181) (0.123) (0.0650) Observations 2621 2621 2621 2621 2621 2621 2621 2621 R-squared 0.056 0.045 0.066 0.053 0.056 0.046 0.063 0.053 Country FE Yes Yes Yes Yes Yes Yes Yes Yes Time FE Yes Yes Yes Yes Yes Yes Yes Yes Estimator OLS OLS OLS OLS OLS OLS OLS OLS Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. 18 K. NATTER ET AL. Lastly, Table 4 explores diﬀusion eﬀects and policy trade-oﬀs between diﬀerent policy areas. The ﬁndings clearly show that policy reform in other relevant countries leads to adaptations in national policy in a similar direction: Liberal reforms elsewhere lead to immigration liberalization, while restrictive reforms in reference countries lead to a a further closure of national immigration policies. International policy emulation seems thus a robust feature of national migration policymaking, supporting the ‘convergence’ hypothesis in immigration policy (Cornelius, Martin, and Holliﬁeld 1994). The fact that many countries in our sample are members of the European Union (EU) might have driven this result, as policy diﬀusion does not only happen ad hoc but is also institutiona- lized through the Europeanization of immigration and integration policies since the 1990s. Interestingly, international policy diﬀusion dynamics seem predominantly at play in the area of border, entry and integration policies, but not in the area of exit policies where gov- ernments seem to act rather independently from international trends. This might be explained by the fact that the attractiveness of a country is determined more strongly by the facilities it oﬀers to enter and stay in the country, as well as to access socio-econ- omic rights, rather than by regulations of exit and return. Table 4 also provides insights into the interdependency between diﬀerent policies issues. Most strikingly, when entry policies are liberalized, this often goes together with a restriction of integration measures and vice-versa. This mechanism might point towards the existence of an empirical trade-oﬀ between the numbers of migrants allowed into a country and the post-entry rights granted to them (Ruhs and Martin 2008). A similar dynamic is at play between border control and exit policies. Further research should investigate whether and to what extent showing ‘toughness’ on certain policy issues might indeed be used as a way to gain support for more liberal policies in other areas. Overall, however, the role of the political ideology of parties in power on the overall restrictiveness of immigration policy change remains limited, as restrictive and liberal policy changes tend to balance each other out and there is often a signiﬁcant gap between rhetoric and reality in migration policymaking. Party ideology seems to translate into more or less restrictive immigration policies only in the area of integration, as there are no signiﬁcant partisan diﬀerences with regards to the core of immigration regimes – entry policies. Lastly, the results also point towards high inter-party consensus on restric- tive border control policies, which are at the heart of symbolic migration politics but have limited eﬀects on overall immigration volumes. Conclusion This paper has oﬀered a comprehensive analysis of the partisan drivers of immigration policy reform in 21 Western countries between 1970 and 2012. The analysis showed that the ideological orientation of parties in power plays a relatively marginal role in deter- mining immigration reform. In fact, our results are very robust in not ﬁnding evidence of a clear left/right gradient on immigration policy restrictiveness, particularly when it comes to the ideological orientation of governmental parties. Because party preferences on immi- gration cut across the left–right spectrum, there is no strong eﬀect of the ideology of gov- erning parties on overall immigration policy restrictiveness. Political party ideologies only aﬀect certain migration policy areas, in particular integration policies. This supports POLITICAL RESEARCH EXCHANGE 19 previous research arguing that integration splits political parties more neatly between the left and the right (Duncan and Van Hecke 2008; Money 1999). Also, political party ideology only aﬀects immigration policy restrictiveness towards certain migrant groups, particularly those that stand at the centre of public debates, such as asylum seekers and undocumen- ted migrants. In contrast, political party ideology does not fundamentally shape decisions on the core of immigration regimes, particularly entry policies on labour and family migration. Although immigration is subject to heated debates in the public sphere and extensive political bargaining, the actual policies enacted seem to be driven by factors other than political ideology. Overall, the restrictiveness of immigration policy reform seems deter- mined by general structural factors such as economic cycles, political system features and the strength of welfare systems. If economies grow fast, governments will be more lenient towards immigration, irrespective of their political orientation. In contrast, high social security spendings and low employment protection regulations are associated to more restrictive changes. This suggests a high interconnection between diﬀerent policy ﬁelds, with employment protection policies and immigration policies seemingly oﬀering policymakers two alternative tools to regulate the consequences of labour immigration for native workers. Lastly, structural political system variables also play a role: Federal and presidential systems are associated to more liberal policy outcomes, potentially as a result of the higher number of negotiation levels between national and regional levels, as well as between executive and legislative actors that require political compromising and ‘bartering’. The need to form coalitions, therefore, results in policies that are watered-down versions of initial political positions. In sum, both right- and left-wing gov- ernments largely deal with the same interest groups and have to respond to the same international contexts in migration policymaking. Next to disentangling partisan from more structural drivers, one of the main inno- vations of this paper has been to disaggregate immigration policy and to explore interde- pendencies between policy areas. Given the diﬀerentiated preferences of political parties regarding certain migrant groups and policy areas, policymaking is likely to involve trade- oﬀs and compromises, leading to immigration policy reforms that contain ‘mixed bags’ of measures that target diﬀerent migrant groups and policy issues in often incoherent and contradictory ways. In particular, our results show that the liberalization of entry policies often goes together with a restriction of integration measures and vice-versa, pointing towards a trade-oﬀ between the numbers of migrants allowed into a country and the post-entry rights granted to them (Ruhs and Martin 2008). Immigration policy is also strongly aﬀected by the actions of neighbouring countries or alternative migration desti- nations. International policy diﬀusion accounts for the high degree of policy convergence across countries and can partly explain why similar policies are enacted regardless of which party is in power. Whether international policy diﬀusion drive policy restrictions or liberalizations depends on the issue at stake. In sum, our results point to the complex mechanisms through which party ideology shapes immigration policy reform, as well as the necessity to focus on speciﬁc migrant cat- egories and policy areas separately to understand such general trends. Ultimately, this analysis shows that – if at all – the partisan composition of governments and parliaments has only limited power in explaining immigration policy outcomes. 20 K. NATTER ET AL. Notes 1. Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States. 2. The DEMIG POLICY database, as well as all supporting documentation are publicly available online, see: https://www.imi-n.org/data. DEMIG POLICY tracks over 6500 changes in migration policies in 45 countries over the 1945–2014 period. Although it also includes countries in Latin America, Asia and Africa, we limit our analysis to countries that have been democratically ruled over this entire period to assess the eﬀect of party politics. The limitation to the post-1970 period is because of availability of partisan and structural factors. 3. The weighted policy change variable also includes zeros. However, these do not represent ‘no policy change’ but policy changes without a clear direction in terms of changes in restrictive- ness. The total absence of policy changes in a given year is not coded by this dataset, which implies that country-years without any policy activity are missing. 4. The binary policy change variable is used for logistic regressions whereas weighted policy changes are used in linear least square regressions. We also performed analyses both with one-year lagged and non-lagged policy changes to account for potential delays in policy changes introduced by new governments as well as lengthy legislative processes. However, given that results were not aﬀected by the lag (the potential delay only aﬀects years in which a new government or parliament is elected), we decided to run the ﬁnal analyses without lagging policy changes. 5. We have also tested geographical distance and contiguity as alternative connectivity variables. Disclosure statement No potential conﬂict of interest was reported by the author(s). Funding The research leading to these results is part of the DEMIG (Determinants of International Migration) project and has received funding from the European Community’s Seventh Framework Programme (FP7/2007-2013)/ERC Grant Agreement 240940. The paper has been ﬁnalized thanks to support by the MADE (Migration as Development) project funded by the European Community’s Horizon 2020 Programme (H2020/2015-2020)/ERC Consolidator Grant Agreement 648496. ORCID Katharina Natter http://orcid.org/0000-0001-9683-0567 References Abou-Chadi, T. 2016. “Political and Institutional Determinants of Immigration Policies.” Journal of Ethnic and Migration Studies 42 (13): 2087–2110. doi:10.1080/1369183X.2016.1166938. Alonso, S., and S. C. da Fonseca. 2011. “Immigration, Left and Right.” Party Politics 18 (6): 865–884. doi:10.1177/1354068810393265. Armingeon, K., V. Wenger, F. Wiedemeier, C. Isler, L. Knöpfel, D. Weisstanner, and S. Engler. 2017. “Comparative Political Data Set 1960–2015.” Bale, T. 2008. “Politics Matters: A Conclusion.” Journal of European Public Policy 15 (3): 453–464. doi:10. 1080/13501760701847721. POLITICAL RESEARCH EXCHANGE 21 Block, L., and S. Bonjour. 2013. “Fortress Europe or Europe of Rights? The Europeanisation of Family Migration Policies in France, Germany and the Netherlands.” European Journal of Migration and Law 15: 203–224. doi:10.1163/15718166-12342031. Bonjour, S. 2011. “The Power and Morals of Policy Makers: Reassessing the Control Gap Debate.” International Migration Review 45 (1): 89–122. doi:10.1111/j.1747-7379.2010.00840.x. Cornelius, W. A., P. L. Martin, and J. F. Holliﬁeld. 1994. Controlling Immigration: A Global Perspective. Stanford, CA: Stanford University Press. Cornelius, W. A., T. Tsuda, P. L. Martin, and J. F. Holliﬁeld. 2004. Controlling Immigration: A Global Perspective. 2nd ed. Stanford, CA: Stanford University Press. Czaika, M. 2018. High-Skilled Migration. Drivers and Policies. Oxford: Oxford University Press. Czaika, M., and H. de Haas. 2013. “The Eﬀectiveness of Immigration Policies.” Population and Development Review 39 (3): 487–508. doi:10.1111/j.1728-4457.2013.00613.x. Davis, A. 2012. “The Impact of Anti-Immigration Parties on Mainstream Parties’ Immigration Positions in the Netherlands, Flanders and the UK 1987–2010: Divided Electorates, Left-Right Politics and the Pull Towards Restrictionism.” PhD thesis, EUI, Florence. de Haas, H. 2011. “The Determinants of International Migration.” IMI Working Paper 32. de Haas, H., K. Natter, and S. Vezzoli. 2015. “Conceptualizing and Measuring Migration Policy Change.” Comparative Migration Studies 3: 15. doi:10.1186/s40878-015-0016-5. de Haas, H., K. Natter, and S. Vezzoli. 2018. “Growing Restrictiveness or Changing Selection? The Nature and Evolution of Migration Policies.” International Migration Review 52 (2): 314–367. doi:10.1111/imre.12288. DEMIG. 2015. “DEMIG TOTAL.” Version 1.5. Duncan, F., and S. Van Hecke. 2008. “Immigration and the Transnational European Centre-Right: A Common Programmatic Response?” Journal of European Public Policy 15 (3): 432–452. doi:10. 1080/13501760701847705. Eule, T. 2014. Inside Immigration Law: Migration Management and Policy Application in Germany. Farnham: Ashgate. Freeman, G. P. 1995. “Modes of Immigration Politics in Liberal-Democratic States.” International Migration Review 29 (4): 881–902. doi:10.1177/019791839502900401. Freeman, G. P. 2013. Immigration and Public Opinion in Liberal Democracies. New York: Routledge. Geddes, A. 2003. The Politics of Migration and Immigration in Europe. London: Sage. Givens, T., and A. Luedtke. 2005. “European Immigration Policies in Comparative Perspective: Issue Salience, Partisanship and Immigrant Rights.” Comparative European Politics 3: 1–22. doi:10. 1057/palgrave.cep.6110051. Gutekunst, M. 2015. “Language as a New Instrument of Border Control: The Regulation of Marriage Migration from Morocco to Germany.” Journal of North African Studies 20 (4): 540–552. doi:10. 1080/13629387.2015.1065038. Hatton, T. J., and J. G. Williamson. 2004. “Refugees, Asylum Seekers and Policy in Europe.” IZA Discussion Paper No. 1230. Haus, L. 1999. “Labor Unions and Immigration Policy in France.” International Migration Review 33 (3): 683–716. doi:10.1177/019791839903300306. Infantino, F. 2010. “La frontière au guichet. Politiques et pratiques des visas Schengen aux Consulat et àl’Ambassade d’Italie au Maroc.” Champ pénal/Penal ﬁeld, nouvelle revue internationale de crimin- ologie VII. Laakso, M., and R. Taagepera. 1979. “’Eﬀective’ Number of Parties: A Measure with Application to West Europe.” Comparative Political Studies 12 (1): 3–27. doi:10.1177/001041407901200101. Lavenex, S. 2001. The Europeanisation of Refugee Policies: Between Human Rights and Internal Security. Aldershot: Ashgate. Massey, D. S. 1999. “International Migration at the Dawn of the 21st Century: The Role of the State.” Population and Development Review 25 (2): 303–322. doi:10.1111/j.1728-4457.1999.00303.x. Massey, D. S., J. Arango, G. Hugo, A. Kouaouci, A. Pellegrino, and J. E. Taylor. 1998. Worlds in Motion: Understanding International Migration at the End of the Millennium. Oxford: Clarendon Press. Meyers, E. 1995. “A Theory of Immigration Policy.” In International Immigration Policy – A Theoretical and Comparative Analysis,5–26. London: Palgrave Macmillan. 22 K. NATTER ET AL. Meyers, E. 2000. “Theories of International Immigration Policy – A Comparative Analysis.” International Migration Review 34 (4): 1245–1282. doi:10.1177/019791830003400407. Meyers, E. 2002. “The Causes of Convergence in Western Immigration Control.” Review of International Studies 28 (1): 123–141. doi:10.1017/S0260210502001237. Money, J. 1999. “Deﬁning Immigration Policy: Inventory, Quantitative Referents, and Empirical Regularities.” Paper presented at the Annual Meetings of the American Political Science Association, Atlanta, GA. Morales, L., J.-B. Pilet, and D. Ruedin. 2015. “The Gap Between Public Preferences and Policies on Immigration: A Comparative Examination of the Eﬀect of Politicisation on Policy Congruence.” Journal of Ethnic and Migration Studies 41 (9): 1495–1516. doi:10.1080/1369183X.2015.1021598. Neumayer, E., and T. Plümper. 2010. “Spatial Eﬀects in Dyadic Data.” International Organization 64 (1): 145–166. doi:10.1017/S0020818309990191. Odmalm, P. 2011. “Political Parties and ‘the Immigration Issue’: Issue Ownership in Swedish Parliamentary Elections 1991–2010.” West European Politics 34 (5): 1070–1091. doi:10.1080/ 01402382.2011.591098. Perlmutter, T. 1996. “Bringing Parties Back In: Comments on ‘Modes of Immigration Politics in Liberal Democratic Societies’.” International Migration Review 30 (1): 375–388. doi:10.1177/ Piore, M. J. 1979. Birds of Passage. Migrant Labor and Industrial Societies. Cambridge: Cambridge University Press. Ruhs, M. 2013. “The Rights of Migrant Workers Reframing the Debate.” In The Price of Rights – Regulating International Labor Migration,1–12. Princeton, NJ: Princeton University Press. Ruhs, M., and P. L. Martin. 2008. “Numbers vs. Rights: Trade-Oﬀs and Guest Worker Programs.” International Migration Review 42 (1): 249–265. doi:10.1111/j.1747-7379.2007.00120.x. Schain, M. A. 2008. “Commentary: Why Political Parties Matter.” Journal of European Public Policy 15 (3): 465–470. doi:10.1080/13501760701847739. Schmidt, M. G. 1996. “When Parties Matter: A Review of the Possibilities and Limits of Partisan Inﬂuence on Public Policy.” European Journal of Political Research 30: 155–183. doi:10.1111/j. 1475-6765.1996.tb00673.x. Sciortino, G. 2000. “Toward a Political Sociology of Entry Policies: Conceptual Problems and Theoretical Proposals.” Journal of Ethnic and Migration Studies 26 (2): 213–228. doi:10.1080/ Sciortino, G., and A. Colombo. 2004. “The Flows and the Flood: The Public Discourse on Immigration in Italy, 1969–2001.” Journal of Modern Italian Studies 9(1):94–113. doi:10.1080/1354571042000179209. Shipan, C. R., and C. Volden. 2008. “The Mechanisms of Policy Diﬀusion.” American Journal of Political Science 52 (4): 840–857. doi:10.1111/j.1540-5907.2008.00346.x. Tichenor, D. J. 2002. Dividing Lines: The Politics of Immigration Control in America. Princeton, NJ: Princeton University Press. Tichenor, D. J. 2008. “Strange Bedfellows: The Politics and Pathologies of Immigration Reform.” Labor: Studies in Working-Class History of the Americas 5 (2): 39–60. doi:10.1215/15476715-2007-077. Triadaﬁlopoulos, T., and A. Zaslove. 2006. “Inﬂuencing Migration Policy from Inside: Political Parties.” In Dialogues on Migration Policy, edited by M. Giugni and F. Passy, 171–191. Lanham: Lexington Books. Van Der Valk, I. 2003. “Political Discourse on Ethnic Minority Issues. A Comparison of the Right and the Extreme Right in the Netherlands and France (1990–97).” Ethnicities 3 (2): 183–213. doi:10.1177/ Volkens, A., P. Lehmann, T. Matthieß, N. Merz, S. Regel, and B. Weßels. 2017. “The Manifesto Data Collection. Manifesto Project (MRG/CMP/MARPOR).” Version 2017b. Wutts, J. R. 1998. “Strange Bedfellows: How Spanish Labor Union Leaders and Employers Find Common Ground on Immigration.” Policy Studies Journal 26 (4): 657–675. doi:10.1111/j.1541- 0072.1998.tb01938.x. Zolberg, A. R. 2006. A Nation by Design – Immigration Policy in the Fashioning of America. Cambridge, MA: Harvard University Press. POLITICAL RESEARCH EXCHANGE 23 Appendix Table A1. Descriptive statistics of variables. Std. Variable Obs. Mean dev. Min Max Source Policy change (dummy) 2332 0.450 0.498 0 1 DEMIG Policy Policy change (weighted) 2621 −0.299 2.742 −4 4 DEMIG Policy Policy changes by policy area Border policy 2621 0.245 0.981 −4 4 DEMIG Policy Entry policy 2621 −0.366 1.831 −4 4 DEMIG Policy Integration policy 2621 −0.287 1.447 −4 4 DEMIG Policy Exit policy 2621 0.109 0.955 −4 4 DEMIG Policy Policy changes by targeted migrant group All migrants 2621 0.213 0.409 0 1 DEMIG Policy Labour migrants 2621 0.129 0.335 0 1 DEMIG Policy Family migrants 2621 0.073 0.260 0 1 DEMIG Policy High-skilled migrants 2621 0.115 0.319 0 1 DEMIG Policy Undocumented migrants 2621 0.154 0.361 0 1 DEMIG Policy Forced migrants 2621 0.171 0.377 0 1 DEMIG Policy Parliament composition 2621 0.003 2.738 −10.307 9.859 Comparative Manifesto Project Government composition 2621 −0.790 17.495 −58.0 48.458 CPDS/ Comparative Manifesto Project Electoral fractionalisation of party 2605 0.736 0.089 0.500 0.903 CPDS system Presidential system 2602 0.182 0.386 0 1 CPDS Federalism No 2602 0.597 0.491 0 1 CPDS Weak 2602 0.096 0.295 0 1 CPDS Strong 2602 0.307 0.461 0 1 CPDS Trade openness 2609 73.715 49.133 10.730 348.43 CPDS Union density (% of employees) 2533 35.187 18.609 7.548 87.427 CPDS Real GDP growth (in %) 2600 2.381 2.399 −8.539 11.306 CPDS Unemployment rate (% of labour 2600 6.947 3.753 0.002 24.171 CPDS force) Immigration rate (per ‘000 2229 10.436 18.074 .541 299.3 DEMIG Flow population) Population over 65 (in %) 2609 14.597 2.503 8.028 21.101 CPDS Employment protection regulation 2083 2.117 0.943 0.257 5 CPDS Social security transfers (% of GDP) 2603 13.620 3.534 3.473 23.894 CPDS 24 K. NATTER ET AL. Table A2. Period eﬀects in the restrictivenes of policy changes by policy area according to the political ideology of parties in power (21 countries). DV: Immigration policy change (1) (2) (3) (4) (5) (6) Pre-1989 1990–2001 Post-2001 Pre-1989 1990–2001 Post-2001 Cabinet composition interacted with Parliament composition interacted with policy area … policy area … Border 0.0153 0.0174 −0.0191* −0.0206 0.180*** 0.154 (0.0263) (0.0107) (0.0114) (0.225) (0.0679) (0.0982) Entry −0.00562 −0.0146** 0.0119* −0.0709 −0.0185 0.258*** (0.00965) (0.00703) (0.00674) (0.0791) (0.0472) (0.0829) Integration 0.0280*** 0.0186 0.0306*** 0.243*** 0.0257 0.339*** (0.0100) (0.0119) (0.00916) (0.0939) (0.0657) (0.108) Exit 0.0279 0.0187 −0.00536 0.511** 0.171** 0.0824 (0.0217) (0.0147) (0.0115) (0.254) (0.0788) (0.127) Constant −1.631*** 0.403 −1.170*** −1.926*** 0.208 −1.796*** (0.509) (0.281) (0.387) (0.584) (0.349) (0.491) Observations 439 782 1099 439 782 1099 R-squared 0.0547 0.0491 0.0548 0.0634 0.0478 0.0526 Country FE Yes Yes Yes Yes Yes Yes Time FE No No No No No No Estimator Logit Logit Logit Logit Logit Logit Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < 0.1. Table A3. Period eﬀects in the restrictiveness of policy changes by migrant group according to the political ideology of parties in power (21 countries). DV: Immigration policy change (1) (2) (3) (4) (5) (6) Pre-1989 1990–2001 Post-2001 Pre-1989 1990–2001 Post-2001 Cabinet composition interacted with Parliament composition interacted with target group … target group … All (unspeciﬁed) migration 0.0198 0.00354 0.00742 0.0677 0.0664 0.268*** (0.0120) (0.0105) (0.00795) (0.101) (0.0634) (0.0895) Labour migration −0.0175 −0.00549 0.0401*** −0.178 0.149 0.344*** (0.0155) (0.0159) (0.0118) (0.134) (0.106) (0.112) Family migration 0.00894 0.00434 0.0185 0.0229 0.0610 0.182 (0.0163) (0.0161) (0.0151) (0.158) (0.0745) (0.174) Skilled migration −0.0332 −0.0458*** −0.00109 −0.654*** −0.179** 0.229** (0.0219) (0.0135) (0.00933) (0.236) (0.0833) (0.0977) Undocumented migration 0.0285* 0.0299*** −0.00610 0.280* 0.238*** 0.0598 (0.0157) (0.0110) (0.0110) (0.145) (0.0673) (0.116) Asylum migration 0.0128 0.0106 0.0217* 0.106 0.0209 0.278*** (0.0133) (0.0110) (0.0115) (0.0857) (0.0688) (0.0956) Constant −1.944*** 0.476 −1.059** −2.136*** 0.188 −1.808*** (0.571) (0.303) (0.417) (0.627) (0.364) (0.507) Observations 426 744 1064 426 744 1064 R-squared 0.058 0.064 0.051 0.060 0.068 0.051 Country FE Yes Yes Yes Yes Yes Yes Time FE No No No No No No Estimator Logit Logit Logit Logit Logit Logit Notes: Robust standard errors in parentheses. ***p < .01, **p < .05, *p < .1. POLITICAL RESEARCH EXCHANGE 25 Figure A1. Trends in parliament composition and changes in migration policy restrictiveness (1970– 2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The parlia- ment composition index gives an insight into the overall ideological positioning of the parliament. (Far) left-wing domi- nated parliaments have a (strongly) negative index score, whereas (far) right-wing parliaments have a (strongly) positive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and parliament composition. The upward trend of the policy change line thus represents a decelerated liberalization and not a trend towards more restric- tiveness per se. Figure A2. Changes in migration policy restrictiveness according to the (right-left) political ideology of parliaments (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). The parlia- ment composition index gives an insight into the overall ideological positioning of the parliament. Areas coloured blue or turquoise show parliaments dominated by a right-wing political ideology, areas coloured in green or red show parliaments dominated by a left-wing political ideology. 26 K. NATTER ET AL. Figure A3. Average change in migration policy restrictiveness according to the (right-left) political ideology of parliaments, disaggregated by policy area (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated parliaments have a (strongly) negative index score, whereas (far) right-wing parliaments have a (strongly) posi- tive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and parliament composition by migration policy area. Figure A4. Average change in migration policy restrictiveness according to the (right-left) political ideology of parliaments, disaggregated by type of migrant group (1970–2012, 21 countries). Notes: Changes in migration policy restrictiveness can range from −4 (least restrictive) to 4 (most restrictive). (Far) left-wing dominated parliaments have a (strongly) negative index score, whereas (far) right-wing parliaments have a (strongly) posi- tive score. The lines show best curvi-linear ﬁt predictions of changes in policy restrictiveness and parliament composition by migrant target group.

Journal

Political Research Exchange – Taylor & Francis

Published: Jan 1, 2020

Keywords: Immigration policy; political parties; migration; policy determinants; policy diffusion

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Political party ideology and immigration policy reform: an empirical enquiry

Political party ideology and immigration policy reform: an empirical enquiry

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Political party ideology and immigration policy reform: an empirical enquiry

Political party ideology and immigration policy reform: an empirical enquiry

References (58)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies