Trade unions and migrant workers in Western Europe

Abstract

This article examines the gap between the unionization rate of local and migrant workers in 14 Western European countries. The analysis reveals that the lower unionization rate of migrant workers can be attributed only in part to the impact of labour market segregation. Moreover, the gap between the unionization rate of local and migrant workers varies substantially across countries. We find that this gap is larger in those countries in which unions enjoy organizational security either in the form of state financing or a single dominant confederation.

Keywords

Europe migrant workers organizational security trade unions unionization

Introduction

Migrant workers have always presented trade union movements with a series of challenges and dilemmas. Above all, unions have been wary of their potential threat as a source of cheap labour, able to undercut existing pay levels and working conditions (Bonacich, 2001 [1972]). The waves of immigration into the countries of Western Europe over the last 25 years have served to accentuate these longstanding dilemmas. With migrants now forming increasing proportions of national workforces, ‘trade unions have a direct stake in the foreign labour issue’ (Avci and McDonald, 2000: 197). This has led to growing scholarly focus on the question of how unions relate to immigration policy in general and migrant workers in particular. Yet, as McGovern (2007) points out, systematic cross-national research on unions and migrant workers is relatively rare. While many valuable studies have recently emerged of unions’ changing stances on immigration policy, and of the problems unions face in engaging with migrant workers, these have tended to be either single-country case studies or small-n comparisons (for example, Hassel, 2007; Haus, 2002; Holgate, 2005; Kahmann, 2006; Marino and Roosblad, 2008; Penninx and Roosblad, 2000; Richards, 2008, 2011; Watts, 2002). In this study we use a systematic cross-national comparison of 14 Western European countries to explain why migrant workers are consistently less unionized than native workers.

Any definition of ‘migrant workers’ is problematic. We use here a straightforward distinction between citizens and non-citizens, as we consider this the simplest and least controversial definition. As such, the terms ‘migrant’ and ‘non-citizen’ are used interchangeably throughout the text. This definition is based on the idea that the distinction between citizens and non-citizens is best able to capture the incorporation (or not) of migrant workers into host societies on the basis of equal rights. At the same time, we acknowledge that it is an imperfect definition, given that citizenship rules vary across countries.

It is also important to note that we focus on the relations between unions and migrants who have already entered the host country. Following the theory of ethnic antagonism (Bonacich, 2001 [1972]), two aspects of union strategies may be distinguished. The first is whether they support either exclusionary or inclusionary immigration policies. The second is whether they attempt to incorporate migrant workers into their own ranks or exclude them. This study focuses exclusively on the second issue.

Research on Western European countries has long suggested that migrant workers are less unionized than local workers. However, in the last two decades there has been no systematic cross-national examination of the current rate of migrant worker unionization, even though less systematic evidence suggests that this varies across countries. This neglect is surprising, since migrant workers’ unionization has become even more important, yet at the same time even more complicated, given overall declining unionization in Western economies. The present study thus aims to bridge this gap in the literature. First, we examine the rate of migrant unionization compared to that of local workers in 14 Western European countries, using data from four rounds of the European Social Survey (ESS). Second, we assess the extent to which the relatively low unionization rate of foreign workers can be attributed to labour market segregation between foreigners and locals, as previous research has suggested. Third, we explore the degree to which the institutional position of trade unions in a given society can explain cross-national differences in the gap between unionization of migrants and local workers. By doing so, we aim to contribute not only to debates on the contemporary trajectory of trade unionism but also to the literature on migrants’ incorporation into host societies. Since union membership is, potentially, a critical means for migrant workers to exercise an important range of social rights, we argue that it should be viewed as one of the most important indicators of migrant workers’ incorporation into host societies.

Theoretical considerations and previous research

In theory, unions should have every incentive to recruit migrant workers. Union density and the absolute number of union members have fallen throughout the advanced capitalist democracies during the last 25 years, and this decline forms the focus for a huge body of literature. At the same time, migrant workers represent an increasing proportion of national workforces; in most countries, immigration has contributed significantly to the expansion of the working population and to increasing rates of labour market participation. Yet migrant workers tend invariably to occupy the most precarious, unstable and marginalized sectors of the labour market (Holgate, 2005; King, 2002; Richards, 2011; Stalker, 1994; TUC, 2003; Wrench et al., 1999). Though migrant workers in Western Europe form a diverse population, with different countries of origin within and beyond the European Union, non-citizens are widely perceived as an out-group and as such they often become a target for discriminatory attitudes and practices (Pettigrew, 1998; Semyonov et al., 2006). Many lack the same employment rights as native workers. According to the Migrants’ Integration Policy Index (MIPEX) (Niessen et al., 2007), in all the countries considered here, with the exception of Sweden, migrants from outside the EU are not eligible for equal access to employment as EU nationals. In several countries (for example, Germany and the Netherlands), migrants face limitations on their right to change jobs. Thus they are in particular need of union representation and protection while forming a potentially critical and expanding constituency for union movements eager for new recruits. We should therefore expect unionization rates amongst migrant workers to be at least the same as for native workers.

Yet unionization rates for migrant workers lag consistently behind those for local workers, though the extent to which they do so varies considerably cross-nationally. These two phenomena merit explanation.

Trends in overall unionization rates

According to the ICTWSS database (Visser, 2011), there is a general downward trend in unionization rates for the 14 countries we analyze, though with cross-national variations. Density fell in all countries between 1980 and 2008, and in all countries except Belgium in the most recent period, 2000–2008.

In accounting for this downward trend, much research has emphasized the adverse effects of structural changes in the economy which have undermined the traditional foundations of union strength while promoting the expansion of sectors where the presence of unions has traditionally been weak or non-existent. Thus the historic constituency of trade unionism, skilled and semi-skilled workers employed on a permanent or relatively secure basis in traditional heavy industries and manufacturing, is shrinking. In contrast, the traditionally less-unionized private service sector has expanded significantly, as in many countries has the largely unregulated informal economy (Richards, 2001; Visser, 2002). Such structural economic shifts have been accompanied by changes in the profile of employment, with the expansion of temporary and part-time contracts which are correlated strongly with lower unionization rates. Moreover, unskilled, precarious or under-employed workers in the informal economy are difficult to organize and consequently remain largely beyond the ambit of trade unionism (Charlwood, 2002; Francesconi and García-Serrano, 2004; Heery, 1998; Sverke and Goslinga, 2003; Waddington and Whitson, 1997).

Nonetheless, while global economic pressures and structural economic change may have been broadly similar cross-nationally, their effects on national union movements have been mediated by the institutional and political contexts in which they are located. Such contexts affect the unions’ ability and willingness to organize new members; some unions may have stronger incentives than others to recruit the non-unionized as a means of halting their own decline.

The institutional and political factors which encourage or inhibit unionization are, of course, multiple. At country level, involvement in the Ghent system, whereby unions are involved in the administration of important social benefits, has been shown to be a powerful explanation for consistently higher rates of unionization (Van Rie et al., 2011; Western, 1997). Additionally, some union movements enjoy institutional supports, such as works councils or union elections, which help them retain influence and membership in the workplace, while others do not. In some countries, the legal obstacles to establishing and joining a union have traditionally been severe or have become increasingly so, thereby increasing the costs for unions of attempting to expand membership through campaigns to organize the non-unionized. Some union movements have also had to endure politically hostile governments and anti-union legislation (above all, Britain in the 1980s), thereby reducing their mobilization capacity and, consequently, their ability to retain existing members and gain new ones.

Nonetheless, while such factors may help explain cross-national variation and trends in unionization rates in general, they tell us little about why unionization rates for migrant workers should consistently lag behind those for local workers. Although reliable data on national (let alone cross-national) unionization rates for migrants are scarce, there is a strong consensus that across the advanced capitalist economies these are disproportionately low. In the following sections, we discuss two sets of explanations for the relatively low rate of unionization amongst migrant workers and its variation across countries.

The labour market position of migrant workers

There is general consensus that a high degree of segregation exists in European labour markets between local and migrant workers. The latter are disproportionately concentrated in the least favourable sectors, either in the unregulated informal economy or in those occupations characterized by longer working hours, lower pay, less employment stability, higher accident rates and fewer (if any) opportunities for acquiring skills and gaining promotion (Castles and Kosack, 1985; Holgate, 2005; Kogan, 2006; Stalker, 1994). These are precisely the labour market sectors where unions are least likely to be present. Thus while numerous studies have demonstrated the positive impact of a union presence on the recruitment of new union members, migrants, in comparison to other workers, are much more likely to be working in sectors where union presence is either weak or non-existent (Badigannavar and Kelly, 2005; Charlwood, 2002, 2004; D’Art and Turner, 2002; Richards, 2011; Stirling, 2005; Waddington and Whitson, 1997). Yet ironically, migrants would presumably gain the most from union protection and representation. What Waddington and Whitson (1997: 518) conclude about precariously situated workers in general may be applied in particular to migrants:

a key explanation of non-membership appears to be the inability of unions to make contact with, or provide sufficient support to, potential members, rather than a principled opposition to unionism. Workers in expanding areas of employment are not individually more predisposed to non-unionism, but they are less likely to have a union available.

The institutional position of unions

While unionization rates for migrants lag consistently behind those for native workers, there is nonetheless considerable cross-national variation in the extent to which they do so. This could be due partly to cross-national variation in the distribution of industrial sectors: for example, the construction industry everywhere is notable for its extensive use of migrant labour, yet the industry itself is particularly prominent in Spain. Yet we contend that the institutional position of unions also shapes both their willingness and ability to gain a greater presence amongst migrant workers, thereby affecting their unionization rate.

In our study, we attempt to discern the effect of what we term ‘organizational security’ as one of the aspects of the institutional position of trade unions. Organizational security may, of course, be attained in several ways. Visser’s database incorporates a series of factors, such as the right to strike, the structure and scope of wage bargaining, or the degree of involvement in government decisions on social and economic policy, all of which shape the degree of security enjoyed by a given union movement. Here, we present two alternative factors which we claim provide it and which, for the purposes of our core argument, affect the incentives that union movements will have to recruit migrant workers into their ranks. First, state financing of unions, either directly or indirectly, may constitute a crucial source of organizational security, as is the case in Greece and Spain. In these countries, membership itself may not be the most important basis of union security, and therefore not the principal concern of union movements. Thus we would expect that state financing, by guaranteeing organizational survival, would diminish unions’ interest in recruiting non-unionized migrant workers, especially given that the latter represent the least accessible and hence the most costly potential constituency. Second, we argue that the organizational configuration of unions is also an important component of their security. A single dominant union confederation (as, for example, in Austria, Ireland and Britain), holding a monopoly position, enjoys greater organizational security and therefore may have fewer incentives to organize migrant workers. In contrast, where there is no single dominant organization, competing union confederations may be more willing to recruit migrant workers as a means of gaining organizational strength at their rivals’ expense. To reiterate, our objective is not to provide a comprehensive explanation for the gap between local workers’ unionization rates and those of migrants, but merely the role played by organizational security.

For a systematic large-scale comparison, we focus on the union confederation, rather than individual affiliated unions, in order to facilitate and simplify the analysis. In doing so, we acknowledge that we lose information regarding individual unions’ propensity to organize migrant workers, especially since there is considerable cross-national variation in the degree of authority that union confederations exert over their affiliates (ranging, for example, from very high in the case of Austria to low or non-existent in the case of Britain).¹ Certainly, future research undertaken at the level of individual affiliates would be valuable. Nonetheless, two important qualifications are needed here. First, we make no assumption regarding any possible correlation between the number of confederations and the degree of confederal authority over affiliated unions. Thus Austria and Britain both have single dominant confederations, yet with sharply contrasting degrees of authority over their affiliates. Switzerland, meanwhile, has competing union confederations with low authority over their affiliates, yet unions have less organizational security than in the British case of a single dominant confederation with no authority over its affiliates. Second, if we draw again on the ICTWSS database and compare and contrast the values of the ‘external demarcation between union confederations’ (DEME) with those of the ‘internal demarcations within union confederations’ (DEMI) for each of the 14 countries under consideration, there is no significant difference between the degree of competition at the confederal and individual union levels. In general, the DEME value is somewhat higher than the DEMI value.²

Finally, in analytical terms there is an important distinction to be made between the organizational security of a union movement, and its strength (or weakness) measured by its influence in the workplace and in the legislative and political arenas (and captured in the ICTWSS database with respect, for example, to the status of works councils or social pacts, or the degree of union involvement in government decisions on economic and social policy). In general, across Western Europe, unions have lost such influence over the last 25 years, or at least seen it severely weakened. Krings (2009), for example, links the declining power and influence of the British TUC to its recent and growing support for more inclusionary policies of immigration and its commitment to actively organizing migrant workers, given the latter’s huge potential as a growing constituency for union membership. His account is plausible; but why did British unions take so long to renew their commitment to recruiting new members, given the long-term process of their own decline? We argue that while the TUC is indeed weak in industrial and political terms, it is nonetheless more organizationally secure than confederations elsewhere that have to compete with rivals for membership and workers’ loyalties. This diminished its incentives to organize migrant workers or, at a minimum, explains its extreme tardiness in developing and adopting measures to do so as a means of responding to long-term membership decline.

In short, our basic argument is that a union confederation enjoying organizational security will have fewer incentives to organize a costly constituency such as migrant workers. In contrast, unions which are organizationally less secure will have greater incentives to attempt to recruit migrant workers as a means of compensating for membership loss. In the following analysis, we examine systematically: 1) the unionization rate of migrant workers compared to that of local workers across 14 European countries; 2) the extent to which the gap between the unionization rate of locals and migrants can be explained by socio-demographic characteristics and labour market segregation; and 3) to what extent the organizational security of trade unions explains the variation across countries in the gap between local and migrant workers’ unionization rates.

Analysis and findings

Data and variables

Our data were obtained from the four rounds (2002, 2004, 2006 and 2008) of the ESS. In each country that participated in the survey, information was gathered from a random probability national sample of the eligible resident populations aged 15 and over. The analysis reported here was restricted to the currently employed working population of 14 Western European countries which participated in at least three rounds of the ESS (with the majority in all four rounds). Hence we exclude Italy because it features in only the first two rounds of the ESS. France is also excluded for two principal reasons. First, as Visser (2000: 253) notes, ‘membership data of French unions are hard to come by and contrasting claims and uncertainty surround published statistics. Data are not regularly published if at all.’ Possibly for this reason, we note a significant discrepancy in relative terms between the unionization rate for France generated by the ESS data and that presented by the OECD. Second, at the time of the data analysis, certain key variables for France were omitted from the international file of the ESS, indicating possible deviations (from other countries’ data) or problems during the data collection process.

We are well aware that problems may arise in the estimation of the unionization rate of non-citizens from their relatively small sample size in our research and from under-representation of non-citizens in the surveys. However, in addition to the fact that the ESS is a highly reliable data source, we find two justifications for using this sample. First, the overall rates of unionization obtained from other sources (such as the ICTWSS database) are in line with those obtained from the ESS pooled data. Second, we argue that the unionization rate of non-citizens is lower than citizens and try to explain this difference. The survey data sample under-represents in particular undocumented immigrants and non-citizens who do not speak the host country language; in general, these groups have an even lower chance of becoming a union member. Thus it is reasonable to suggest that the real differences between the unionization rate of citizens and non-citizens will actually be higher than those which we find using the ESS data.

The 14 countries include 12 members of the EU and two non-members (Norway and Switzerland). Again, we define migrant workers as non-citizens. In order to increase the number of cases, especially the number of non-citizens, to reach more reliable statistical estimates we pool all four rounds into one sample, controlling for year of survey in the analysis. ESS data provide us with information with which to construct a series of key individual-level variables of the study. They are: migrant worker (non citizen = 1), being a union member (union member = 1), type of work contract (unlimited contract = 1), part-time/full-time job (less than 30 weekly work hours = 1), plus a series of dummy variables representing classification of the economic activities of respondents. The series of dummy variables were constructed by merging two-digit items of NACE 1.1 (Statistical Classification of Economic Activities in the European Community) into six categories: manufacturing and agriculture (hereafter MA); retail, hotels, restaurants and domestic services (RHRD); transport and utilities (TU); financial and business services (FB); public administration, education, health (PAEH) and construction (C). In addition, we use several socio-demographic control variables: age (in years), gender (male = 1), education (in terms of formal years of schooling), rural versus urban residence (rural = 1) and living in a country for only one year (since short-term migrants – who are particularly likely to come from other EU countries under the provisions for free movement of labour – may be less likely to seek union membership in the host country).

We also incorporate in our analysis two country-level dummy variables: the presence or absence of one of the two factors (either state financing of trade unions or a single dominant union confederation)³ that provides organizational security to trade unions (organizational security = 1) and belonging or not to the Ghent system (Ghent = 1). The Ghent system is introduced into the models mostly as a control variable, since this may substantially increase the incentives to join a union.

Descriptive overview

In Table 1 we present the unionization rates among locals and migrants (in other words, among citizens and non-citizens), migrant unionization rates as a proportion of those of locals, plus two country-level characteristics: belonging or not to the Ghent system and the presence or absence of factors that provide organizational security to trade unions. The average unionization rate of locals in Europe (29.5%) is considerably higher than that of migrants (16.1%). Although both rates vary considerably across countries, the unionization rate of migrants is strongly correlated with that of locals. The higher the unionization rate of local workers in a country, the higher is that of migrant workers.

Table 1.

Rate of unionization among locals and migrants (%), and country-level variables.

	Locals	Migrants	Migrants/locals ratio	Ghent system	State financing	Dominant confederation
AT	34	13	0.38	No	No	Yes
BE	51	33	0.65	Yes	No	No
CH	19	13	0.68	No	No	No
DE	20	18	0.90	No	No	No
DK	86	79	0.92	Yes	No	No
ES	16	6	0.38	No	Yes	No
FI	78	72	0.92	Yes	No	No
GR	19	8	0.42	No	Yes	No
IE	41	18	0.44	No	No	Yes
NL	29	13	0.45	No	No	No
NO	61	39	0.64	No	No	No
PT	15	6	0.40	No	No	No
SE	78	70	0.90	Yes	No	No
UK	29	16	0.55	No	No	Yes
Total	29.5	16.1	0.55

Source: ESS, authors’ own elaboration; TUW (2005).

The data show that the unionization rate of migrants is lower than that of locals in each of the countries presented in the study, but that the differences vary significantly across countries. In the highly unionized Scandinavian countries, union density for migrant workers almost reaches that of the local population: about 79 percent in Denmark, 72 percent in Finland and 70 percent in Sweden. In contrast, in Spain and Portugal the unionization rate of migrants, at 6 percent, is less than half that of local workers. In addition, a very low relative unionization rate of migrants is found in Austria, where one third of the local population but only 13 percent of the migrant population are union members. Conversely, while the unionization rate of migrant workers in Germany is also quite low (18%), in relative terms it is very close to that of local workers (20%).

As mentioned above, differences between migrant and local workers within the industrially divided labour market (in other words, labour market sector segregation) are often viewed as a key explanation for the differences in their unionization rates. Table 2 displays the distributions of two sub-population groups, local workers and migrants, across six industries and the overall rate of unionization in these sectors.

Table 2.

Occupational distributions (economic activities) among migrant and local workers, and overall rate of unionization by economic activity (%).

	TU	PAEH	MA	C	FB	RHRD
Locals	7.2	35.8	19.7	6.7	14.2	16.5
Migrants	4.4	22.2	22.7	12.4	12.7	25.6
Unionization rate	39.9	38.8	29.8	21.1	20.0	14.2

Source: ESS, authors’ own elaboration; TUW (2005).

Migrant workers tend to be less concentrated than locals in the public, education and health sectors (PEAH) and in the transport and utilities industries (TU). In contrast, they are overrepresented in retail, hotel, restaurants and domestic services (RHRD) and in construction (C). There are no considerable differences between the proportions of locals and migrants working in manufacturing and agriculture (MA), and even fewer differences between the two groups in financial and business services (FB). The data also reveal considerable variation in unionization rates across different industries. The TU and PEAH sectors have the highest rates (39.9 and 38.8%, respectively) while C, FB, and RHRD have the lowest (21.1, 20 and 14.2%, respectively). Thus migrant workers are heavily concentrated in poorly unionized sectors while locals are overrepresented in more unionized sectors. These findings are in line with previous research suggesting that migrant workers are often overrepresented in poorly unionized industrial sectors (Vranken, 1990). However, examining the unionization rates in each sector separately for migrants and locals (results are not presented here) also reveals that despite some differences between the unionization rates of migrants and locals in each sector, the TU and PEAH sectors still have the highest unionization rates for both migrants and locals, while the RHRD sector has the lowest unionization rates for both.

As for the country-level variables, the data reveal that unions in five of the 14 countries enjoy organizational security (see Table 1). In Spain and Greece, they obtain state financing, and in Austria, Ireland and Britain a single union confederation predominates. The descriptive findings also reveal that all these five cases display below average relative rates of unionization for migrant workers. (Although the relative rates of migrant unionization in Portugal and Netherlands, where union movements do not enjoy organizational security, are also below average.)

Multivariate analysis

Although the data provided in the descriptive overview are interesting and meaningful, they do not provide accurate estimates of the extent to which migrant status affects the relative probability of an individual becoming a union member and how far this effect can be attributed to the differences in the distributions of migrant and local workers across sectors. Since the sub-groups (migrants and locals) possess different socio-demographic characteristics and a different distribution across labour market sectors, it is not clear from the descriptive data whether the unionization rate of migrants fits that to be expected on the basis of their characteristics. Thus in the analyses that follow we estimate, first, the effects of migrant status on the odds of being a union member (versus not being a union member) net of the individual level attributes. We then examine the effect that country-level characteristics, namely the Ghent system and the organizational security of unions, exert on the slope of migrant status.

Since in our sample individuals are nested in 14 countries and since we assume that the effect of migrant status on the relative odds of being a union member may vary across countries and can be explained by country-level attributes, we use hierarchical models using HLM software. This is a statistical procedure that allows the estimation of country-level effects while variations in individual-level characteristics are controlled, and vice versa (Bryk and Raudenbush, 1992).

To examine the relative odds of migrants as compared to local workers being union members, we estimate four bi-level logistic regression equations. In all equations, the intercept and slope of migrant status are allowed to vary across countries. The first equation predicts the odds of being a union member (versus not being a union member) as a function of socio-demographic characteristics and migrant status. The values of the estimated coefficients for migrant status serve as indicators of the net effect of migrant status on the odds of being a union member in comparison to similar odds for local workers. To the second equation we add a series of dummy variables representing the sectors in which individuals work (manufacturing and agriculture are the omitted category). To the third equation we add two variables capturing job stability, namely, having an unlimited contract and working in a part-time job. The fourth equation includes, in addition to the individual-level attributes, country-level variables (the Ghent system and organizational security) and interaction terms between organizational security and migrant status. In other words, in the fourth equation, the slope of migrant status also becomes a dependent variable that is predicted by country-level characteristic. The results of the analysis are presented in Table 3.

Table 3.

Logistic regression coefficients (standard errors) from predicting odds of being union member (versus not being union member) for employees currently in active labour force^a.

	1	2	3	4
Intercept	-1.12**	-1.06**	-1.53**	-1.89**
Individual-level variable and interaction terms
Being migrant	-0.51**	-0.42**	-0.34**	-0.17
	(0.15)	(0.10)	(0.09)	(0.11)
Being migrantorganizational security*				-0.37*
				(0.17)
Men	0.19	0.30**	0.20**	0.20**
	(0.18)	(0.02)	(0.03)	(0.03)
Age	0.03**	0.03**	0.03**	0.02**
	(0.003)	(0.001)	(0.001)	(0.001)
Rural	-0.03	-0.06*	-0.05*	-0.05*
	(0.05)	(0.02)	(0.02)	(0.02)
Education	0.04*	0.03**	0.02**	0.02**
	(0.019)	(0.003)	(0.003)	(0.003)
Only one year in a country	-2.43**	-2.41**	-2.32**	-2.28**
	(0.89)	(0.87)	(0.87)	(0.87)
Labour market position (AM)
PAEH	—	0.32**	0.38**	0.38**
		(0.03)	(0.03)	(0.03)
TU	—	0.37**	0.38**	0.38**
		(0.05)	(0.05)	(0.05)
C	—	-0.50**	-0.46**	-0.46**
		(0.06)	(0.06)	(0.06)
FB	—	-0.72**	-0.71**	-0.71**
		(0.04)	(0.04)	(0.04)
RHRD	—	-0.91**	-0.85**	-0.85**
		(0.04)	(0.04)	(0.04)
Unlimited contract	—	—	0.65**	0.65**
			(0.04)	(0.04)
Under 30 weekly hours	—	—	-0.36**	-0.36**
			(0.03)	(0.03)
Country-level variables
Organizational security	—	—	—	0.17
				(0.30)
Ghent system	—	—	—	2.18**
				(0.44)
Country-level variance component for intercept	0.80**	0.83**	0.84**	0.27**
Country-level variance component for being migrant	0.09**	0.08**	0.06**	0.04

Intercept and slope of being migrant vary across countries. The slopes of all other variables are constrained to be identical across countries. Continuous variable have been centred around their grand means. Dummy variables are uncentred. The models also include survey year as control variables (the coefficients are not presented)

p < 0.01; *p < 0.05.

The data presented in equation 1 reveal that the odds of being a union member tend to increase with age and education. Net of socio-demographic characteristics (included in the model) the relative odds of being a union member tend to be lower among migrant workers than among locals. More specifically, the odds of being a union member for migrant workers are 0.60 times the odds of those for local workers, as demonstrated by the negative and significant coefficient: b = -0.51 (exp (-0.51) = 0.60).

The effect of labour market sectors on the odds of being a union member is examined in equation 2. The results show clearly that the industry in which individuals are employed exerts a considerable effect on the odds of being a union member. All else being equal, people working in retail, hotels, restaurants and domestic services have the lowest odds of being a union member, while those in the public sector, education and health sectors, as well as in the transportation and utilities industries, have the highest odds. At the same time, the inclusion of six sectors in equation 2 has changed only slightly the effect of migrant status on the relative odds of becoming a union member. The absolute value of the migrant status coefficient has decreased from b = 0.51 in equation 1 to b = 0.42 in equation 2, but has remained highly significant. This finding implies that the gap between migrants’ odds of being a union member and citizens’ odds of being a union member can be attributed only in part to the differences in the sector distributions between these two groups.

In equation 3 we include two additional variables (having an unlimited contract and working in a part-time job) that may explain not only the odds of being a union member but also the differences in such odds between migrants and locals. The results presented in equation 3 show that the relative odds of being a union member tend to be higher among workers with an unlimited labour contract and a full-time job. The inclusion of the last two variables in the equation has slightly reduced the value of the coefficient of migrant status; however, its effect still remains significant, negative and substantial (b = -0.34). Thus controlling for socio-demographic characteristics, education, sector, work contract and part-time employment, the relative odds of migrants being union members are still lower than those of locals: 0.71 times (exp (−0.34) = 0.71) the odds for local workers.

The findings presented until now show us that individual-level characteristics (included in the study) cannot explain fully the differences in the rate of unionization between migrant and local workers. Moreover, the effect of migrant status on the odds of being a union member (in comparison to local workers) varies across countries in all three models discussed, as implied by the statistically significant variance of the slope.

In order to examine whether the effect of migrant status varies by the presence or absence of organizational security of trade unions, we estimate equation 4. This incorporates (in addition to individual-level variables) two country variables, the Ghent system and the organizational security of trade unions, and interaction terms between the organizational security of trade unions and migrant status.

In line with theoretical expectations, the results reveal that, in general, the odds of being a union member are much higher in Ghent countries, as implied by the statistically significant coefficient (b = 2.18). Organizational security by itself (that is, without the interaction term with migrant status) also exerts a statistically significant negative effect on the general rate of unionization – but only if introduced into the model separately without the Ghent system (this model is not presented). However, our main theoretical interest is the interaction term between organizational security and migrant status. The coefficient of interaction terms between organizational security and migrant status is statistically significant and negative (b = -0.37). It implies that the net relative odds of migrants being union members in countries in which trade unions enjoy organizational security are lower than those in countries in which unions do not enjoy such security. These findings support our theoretical expectations that the organizational security of unions tends to reduce incentives to unionize migrant workers. Indeed, we have shown that the presence of factors providing organizational security to trade unions (either state financing or a single dominant confederation) is associated with a bigger gap in the unionization rate between migrants and local workers than in countries where such factors are absent.

Conclusions

This article represents a first attempt to analyze in a systematic comparative and cross-national manner the gap between the unionization rate of local and foreign workers in 14 Western European countries. We have examined the role of labour market segregation and the organizational security of trade unions in accounting for this gap. Our results show that the unionization rate of migrant workers in Europe is lower than that of locals, and this can be attributed only in part to the impact of labour market segregation. Even when migrants work in the same industrial sectors as locals, they fail to reach the same level of unionization. Thus migrant workers who may be in greatest need of union representation because of their vulnerable status lag behind local workers in their rate of unionization. The gap between the unionization rates of locals and migrants is one reflection of the extent to which the incorporation of migrant workers into the labour market, and their acquisition and exercising of certain economic and social rights, remain problematic. The research presented here is in keeping with the general finding that migrant workers are far less likely to acquire, and be able to exercise, such rights in the labour market (Castles and Davidson, 2000).

Nonetheless, while the findings here are in line with the well-established argument that migrant workers are less likely to come into contact with unions in the workplace, our research also suggests strongly that the institutional position of unions affects the rates of migrant worker unionization relative to those of local workers. We have shown that organizational security, either in the form of state financing or a single dominant union confederation, is associated with lower relative rates of migrant worker unionization (in comparison to locals); conversely, the absence of state funding and competition between union confederations are associated with higher relative rates. We argue that the absence (presence) of organizational security increases (diminishes) union incentives to organize migrant workers, affecting the latter’s unionization rates accordingly.

Future research should include the use of more comprehensive data on migrant unionization rates across countries. Ideally, this would generate more precise information, for example on migrant workers’ positions in the labour market, or their previous experience of, and contact with, trade unions in their countries of origin. We have presented here a first attempt to examine existing comparable cross-national data on this issue. Likewise, the analysis of the institutional position of unions should be expanded to incorporate additional aspects which may also affect the unionization rate of migrant workers.

Nonetheless, our findings still have important implications for the relationship between union movements and migrant workers. The latter represent, across Western Europe, a growing section of the national workforce and, potentially at least, an increasingly important constituency for unions. However, as we have shown, the extent to which migrant workers are unionized varies substantially across countries and this variation cannot be fully attributed to labour market structure. Instead, the institutional context in which unions operate is likely to affect their recruitment of migrant workers. As such, we find a role for union agency, not just in terms of determining their own strategies for renewal and revival but also in terms of fulfilling their potential as critical actors in the integration of migrant workers into host societies.

Footnotes

Acknowledgements

The authors thank Richard Hyman, Guglielmo Meardi and three anonymous reviewers for comments on earlier drafts.

Funding

The research was partially funded by the People Programme (Marie Curie Actions) of the EU Seventh Framework Programme (FP7/2007-2013) under REA grant agreement PCIG11-GA-2012-321730.

Notes

Author biographies

Anastasia Gorodzeisky is a Lecturer at Ben-Gurion University of the Negev, Israel.

Andrew Richards is Senior Research Fellow at the Carlos III-Juan March Institute of Social Sciences, Carlos III University, Madrid, Spain.

References

Avci

McDonald

(2000) Chipping away at the fortress: Unions, immigration and the transnational labour market. International Migration 38(2): 191–213.

Badigannavar

Kelly

(2005) Why are some union organizing campaigns more successful than others? British Journal of Industrial Relations 43(3): 515–535.

Bonacich

(2001 [1972]) A theory of ethnic antagonism: The split labor market. In: Grusky

(ed.) Social Stratification. Class, Race, and Gender in Sociological Perspective, 2nd edn. Boulder, CO: Westview Press, pp. 555–568.

Bryk

Raudenbush

(1992) Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, CA: SAGE.

Castles

Davidson

(2000) Citizenship and Migration: Globalization and the Politics of Belonging. New York: Routledge.

Castles

Kosack

(1985) Immigrant Workers and Class Structure in Western Europe. London: Oxford University Press.

Charlwood

(2002) Why do non-union employees want to unionize? Evidence from Britain. British Journal of Industrial Relations 40(3): 463–491.

Charlwood

(2004) Influences on trade union organizing effectiveness in Britain. British Journal of Industrial Relations 42(1): 69–93.

D’Art

Turner

(2002) The decline of worker solidarity and the end of collectivism? Economic & Industrial Democracy 23(1): 7–34.

10.

Francesconi

García-Serrano

(2004) Unions and flexible employment in Britain and Spain: A descriptive note. Industrial Relations 43(4): 674–682.

11.

Hassel

(2007) The curse of institutional security: The erosion of German trade unionism. Industrielle Beziehungen 14(2): 176–191.

12.

Haus

(2002) Unions, Immigration, and Internationalization: New Challenges and Changing Conditions in the United States and France. New York: Palgrave Macmillan.

13.

Heery

(1998) Campaigning for part-time workers. Work, Employment and Society 12(2): 351–366.

14.

Heery

Simms

Delbridge

Salmon

Simpson

(2000) The TUC’s Organising Academy: An assessment. Industrial Relations Journal 31(5): 400–415.

15.

Holgate

(2005) Organizing migrant workers: A case study of working conditions and unionization in a London sandwich factory. Work, Employment and Society 19(3): 463–480.

16.

Kahmann

(2006) The posting of workers in the German construction industry: Responses and problems of trade union action. Transfer 12(2): 183–196.

17.

King

(2002) Towards a new map of European integration. International Journal of Population Geography 8: 89–106.

18.

Kogan

(2006) Labour markets and economic incorporation among recent immigrant in Europe. Social Forces 85(2): 687–721.

19.

Krings

(2009) A race to the bottom? Trade unions, EU enlargement and the free movement of labour. European Journal of Industrial Relations 15(1): 49–69.

20.

Marino

Roosblad

(2008) Migration and trade unions. A comparison between Dutch and Italian trade union actions and strategies. Transfer 14(4): 625–638.

21.

McGovern

(2007) Immigration, labour markets and employment relations: Problems and prospects. British Journal of Industrial Relations 45(2): 217–235.

22.

Niessen

Huddleston

Citron

(2007) Migrant Integration Policy Index. Brussels: British Council and Migration Policy Group.

23.

Penninx

Roosblad

(eds) (2000) Trade Unions, Immigration and Immigrants in Europe, 1960–1993: A Comparative Study of the Attitudes and Actions of the Trade Unions in Seven West European Countries. Oxford: Berghahn.

24.

Pettigrew

(1998) Reaction toward the new minorities of Western Europe. Annual Review of Sociology 24: 77–103.

25.

Richards

(2001) The crisis of union representation. In: Van Gyes

de Witte

Pasture

(eds) Can Class Still Unite? The Differentiated Work Force, Class Solidarity and Trade Unions. Aldershot: Ashgate, pp. 13–36.

26.

Richards

(2008) El sindicalismo en España. Documento de trabajo 135/2008. Madrid: Fundación Alternativas.

27.

Richards

(2011) Progressive failure: Government, unions and the continuing marginalisation of immigrants in Spain, 2004–8. In: Field

(ed.) Spain’s ‘Second Transition’? The Socialist Government of José Luis Rodríguez Zapatero. London and New York: Routledge, pp. 91–107.

28.

Semyonov

Raijman

Gorodzeisky

(2006) The rise of anti-foreigner sentiment in European societies, 1988–2000. American Sociological Review 71(3): 426–449.

29.

Stalker

(1994) The Work of Strangers: A Survey of International Labour Migration. Geneva: ILO.

30.

Stirling

(2005) There’s a new world somewhere: The rediscovery of trade unionism. Capital and Class 87: 43–63.

31.

Sverke

Goslinga

(2003) The consequences of job insecurity for employers and unions: Exit, voice and loyalty. Economic & Industrial Democracy 24(2): 241–270.

32.

TUC (2003) Overworked, underpaid and over here. Migrant workers in Britain. European Union and International Relations Department, July. London. TUC.

33.

TUW (2005) Trade Unions of the World, 6th edn. London: Harper.

34.

Van Rie

Marx

Horemans

(2011) Ghent revisited: Unemployment insurance and union membership in Belgium and the Nordic countries. European Journal of Industrial Relations 17(2): 125–139.

35.

Visser

(2000) France. In: Ebbinghaus

Visser

(eds) Trade Unions in Western Europe since 1945. Basingstoke and Oxford: Macmillan, pp. 237–277.

36.

Visser

(2002) Why fewer workers join unions in Europe: A social custom explanation of membership trends. British Journal of Industrial Relations 40(3): 403–430.

37.

Visser

(2011) ICTWSS: Database on institutional characteristics of trade unions, wage setting, state intervention and social pacts in 34 countries between 1960 and 2007. Available at: http://www.uva-aias.net/208.

38.

Vranken

(1990) Industrial rights. In: Layton-Henry

(ed.) The Political Rights of Migrant Workers in Western Europe. London: SAGE, pp. 47–73.

39.

Waddington

Whitson

(1997) Why do people join unions in a period of membership decline? British Journal of Industrial Relations 35(4): 515–516.

40.

Watts

(2002) Immigration Policy and the Challenge of Globalization. Unions and Employers in Unlikely Alliance. Ithaca, NY and London: ILR Press.

41.

Western

(1997) Between Class and Market. Postwar Unionization in the Capitalist Democracies. Princeton, NJ: Princeton University Press.

42.

Wrench

Rea

Ouali

(eds) (1999) Migrants, Ethnic Minorities and the Labour Market. Basingstoke: Macmillan.