Anthony Heath
Wouter Zwysen
Anthony Heath
Wouter Zwysen
This chapter asks whether and why disadvantage might become entrenched for some groups of natives with a migration background. Using the ad hoc module of the 2014 European Union Labour Force Survey, it compares the over- and underrepresentation in occupational levels of children of immigrants from different origins. In light of prior research, it goes on to pursue possible lines of enquiry to account for entrenchment of disadvantage, demonstrating that it cannot be explained solely by low socio-economic origins. Other potential factors such as differential minority/majority rates of intergenerational mobility, perverse fluidity and replenishment from the countries of origin, grandparental influences and discrimination are then considered. The discussion concludes with a description of the characteristics needed for a data set to eventually furnish conclusive answers.
Dr. Zwysen is funded through the European Union’s Horizon 2020 research project “Growth, Equal Opportunities, Migration and Markets (GEMM)” under grant agreement No 649255.
Entrenched disadvantage, i.e. one that is persistent across generations, remains a worrying possibility among native-born with immigrant parents in Europe, at least for some disfavoured groups such as those of North African and Middle Eastern origin. In Austria, Belgium, Greece, Italy and Sweden, one or the other of these groups is overrepresented in the most disadvantaged social class (unemployed and low skilled) and underrepresented in the most advantaged social class (the salariat).
It is doubtful that this entrenchment can be explained solely by the low initial starting points of the immigrant parents. It is instead likely that additional processes such as perverse fluidity, three-generational processes, replenishment and discrimination all play a part.
There are a number of possible lines of enquiry for exploring why disadvantage may become entrenched among particular groups, including low initial starting points, replenishment from the countries of origin, differential rates of relative intergenerational mobility, perverse fluidity, three-generational processes and discrimination. At present it is impossible to quantify reliably both the scale and incidence of the entrenched disadvantage, or the relative importance of the different explanations. Much better data than available at present are needed in order to chart dynamic processes over time and generations, as well as much better monitoring of the situation in order to understand the nature and scale of the problem.
There is accumulating evidence that in Europe, young natives with immigrant parents are overrepresented in disadvantaged positions in the labour market. This is by no means true of all groups with a migration background or in all countries, but it does raise worrying questions about the possible entrenchment of disadvantage in the labour market. That could potentially create a form of ethnic stratification alongside the more usual patterns of class stratification. “Ethnic stratification” is here used to describe a situation where some groups with a migration background are disproportionately concentrated in lower levels of the class structure as a direct or indirect consequence of their group membership (or of their assumed membership). In other words, these minorities may experience enduring “ethnic penalties” over and above any disadvantages deriving from their social class origins, lack of human capital, and the like.
This can be contrasted with a situation where ethnic distinctions lose their relevance for processes of socio-economic attainment, and where stratification processes operate in exactly the same way for minorities as they do for the majority group. Patterson has powerfully described the different modes of what he terms “ethno-somatic stratification” operating in Europe and the Americas, where he highlights the strong stratification that has persisted in some parts of the Americas long after the Emancipation Proclamation (Patterson, 2005).
A situation of enduring disadvantage from generation to generation for some groups of natives but not others raises questions about the extent to which there is a level playing field for natives with a migration background. If socio-economic processes operate differently for minorities and the majority group, then it is possible to talk of inequality of opportunity. This will potentially have major implications for social mobility, wastage of talent, social injustice and ultimately for social integration and social order.
However, it is important to recognise that overrepresentation of immigrants in disadvantaged positions is not in itself surprising, and indeed may not necessarily indicate inequality of opportunity. Many immigrants, especially migrants coming from less developed non-European countries, will have low levels of education and human capital, low familiarity with the language of the country of destination, and/or little knowledge of the workings of western labour markets. Their initial situation in the host society is therefore likely to be a disadvantaged one, relative to that of the better-educated majority group (although not necessarily disadvantaged relative to that of non-migrants who remained in the country of origin). One would expect the children of immigrants to have better outcomes than their parents, and indeed there is evidence of substantial upwards intergenerational mobility for these children from their parents’ low positions in the countries of destination (Li and Heath, 2016; Algan et al., 2010).
In the normal course of events, if descendants of immigrants experience the same rates of relative mobility as do natives without a migration background, one would also expect that the occupational positions of natives with a migration background would gradually come to approximate the distributions of natives without a migration background. That may take several generations, but in open societies with equality of opportunity one would not expect those with a migration background to remain grossly overrepresented in lower levels of the labour market, including for persons born in the country (Alba and Nee, 2003). Convergence is expected to be quicker in more socially mobile societies such as Canada, where relative rates of intergenerational mobility are fairly equal for people from different social class backgrounds, and it would be slower in less socially mobile societies such as Germany. But since there is a considerable degree of fluidity even in the least mobile European societies, one would nonetheless expect to see the extent of disadvantage of the descendants of immigrants in Europe gradually diminishing in all countries of destination. (For research on relative and absolute rates of social mobility in Europe, see Ganzeboom, Luijkx and Treiman, 1989; Breen; 2004; and Bukodi, Paskov and Nolan, 2017.)
The central question, then, is whether socio-economic disadvantage may become entrenched for minorities with a migration background, and if so, what processes might be responsible for any such entrenchment.
The evidence is not nearly as strong as one would think, because the kind of data needed is not routinely collected. However, as will be shown, the possibility must be taken seriously that in some countries, particular groups with a migration background may remain overrepresented at lower levels of the labour market many years after the initial migrations of these groups. This does not appear to hold equally true in all countries or for all groups. The picture is very much a differentiated one, with much more marked disadvantage experienced by some minority groups than by others.
The issue can be illustrated using the ad hoc module of the 2014 European Union Labour Force Survey (EU-LFS). Figure 6.1 compares the over- and underrepresentation in different occupational classes of adults born in the host country to at least one immigrant parent from different national origins. Note that these data are far from ideal because of small sample sizes. Ideally we would look at people both of whose parents were immigrants, as this might well show greater disadvantages. However, this would reduce sample sizes too much. Furthermore, the grandchildren of immigrants (the “third generation”) are invisible in the data, and so are classified as natives without a migration background in the EU-LFS. A definitive test of the hypothesis requires that grandchildren and great-grandchildren be distinguished in order to see whether convergence is occurring across generations. Since large numbers of labour migrants first began arriving in developed European countries in the 1950s and 1960s, substantial numbers of grandchildren and great-grandchildren will now be present in some countries. Nevertheless, the EU-LFS helps indicate which countries pose the highest risks of enduring disadvantage for which origin groups.
Because of the very small sample sizes, we are not able to take account of the fact that different waves of migrants arriving at different times may have different characteristics and experiences in the labour market. Small sample sizes also mean that we cannot analyse particular national origins but need to construct broad regional groupings (based on the country of birth of the mother or that of the father if the mother was born in the country of residence or her information was missing). The groupings are as follows:
EU-15
EU-13 (the Member States that joined in 2004 and after)
Other Europe (including Turkey)1
North America and Australia
North Africa and Middle East
Other Africa
East Asian and South Asian countries (including Pakistan)
South and Central America (including the Caribbean)
A condensed version of the European Socio-economic Classification (based on ISCO and employment status, and designed to harmonize different socio-economic indices) is used to define labour market positions. We exclude the economically inactive as the drivers of inactivity are rather different from those of occupational attainment. We also combine unemployment with low-skilled work (employees working in personal services or sales, agricultural work and elementary occupations) in the first, most disadvantaged category since both categories are highly disadvantaged in terms of income and security, and because there is considerable movement between the two categories. The second category is that of skilled industrial workers, consisting of employees in crafts and related trades (such as building, metal, electrical work or food processing) and employees working as plant and machine operators and assemblers. The third group is an intermediate group, comprising small entrepreneurs as well as clerks and employees carrying out skilled service work. The final category consists of those in technical, professional and managerial positions that tend to be the most advantaged in terms of pay, working conditions and security. The term used for this last category is the salariat.
Notes: Men and women combined, economically active respondents aged 15-59. Differences shown are statistically significant at p<0.05 and based on a cell size>=20. Non-significant differences are not shown.
Interpretation: Positive bars (rising above the horizontal) show overrepresentation whereas negative bars (dropping below the horizontal) show underrepresentation.
Source: European Union Labour Force Surveys (EU-LFS) and its Ad-Hoc Module (AHM), 2014.
Figure 6.1 shows the extent to which natives with immigrant parents from these eight broad regional groupings are over- and underrepresented at different occupational levels in the countries participating in the ad hoc module. The positive bars (above the horizontal) show overrepresentation whereas the negative bars (below the horizontal) show underrepresentation. The only bars shown are those where the degree of over- or underrepresentation is statistically significant at the 0.05 level.
The clearest picture is for natives with “Other European” and with “North African and Middle Eastern” backgrounds. Here it can be seen that in Austria, Belgium, Greece, Italy and Sweden, one or other of these groups is overrepresented in the most disadvantaged social class (unemployed and low skilled) and underrepresented in the most advantaged social class (the salariat).
Overrepresentation in disadvantaged positions and underrepresentation in advantaged ones are rare for the other main regional groupings. It appears, then, that disadvantage is not an invariable experience of all minority groups in all destination countries. It is a relatively common experience for some groups but quite rare for others. This warns us that combining all natives with a migration background into a single undifferentiated category may serve to obscure the extent to which specific groups experience entrenched disadvantage.
It should be remembered that the ability to find significant under- or overrepresentation will depend heavily on sample sizes. Many non-significant findings will be due to lack of statistical power, and thus could be regarded as “false negatives”. Even with the broad regional aggregations here, some origin groups are very small. On the other hand some, such as the two EU groups, are indeed large, with large enough sample sizes to show statistically significant under- or overrepresentation. What is striking however is that it is much rarer for these groups to experience significant disadvantage than it is for the “Other European” and the “North African/Middle Eastern” groups.
In the case of Austria, Belgium and Sweden, there is previous research showing that very similar patterns held true around twenty years ago (Heath and Cheung, 2007). While the data are far from ideal, the hypothesis that these patterns of disadvantage have become entrenched certainly cannot be rejected.
It has to be emphasised, however, that Figure 6.1 shows simple descriptive differences that do not take into account differences in the initial situations of these groups, such as their immigrant parents’ social class in the country of destination. These initial situations are the most obvious explanations of disadvantage. They and additional explanations are taken up in the remainder of this chapter. We should also note that, while the results in Figure 6.1 do not adjust for age, more detailed analysis indicates that the results after taking account of age differences between those with and without a migration background are closely in line with those shown here.
There are a number of possible lines of enquiry for exploring why disadvantage may become entrenched among particular groups:
low initial starting points
replenishment from the countries of origin (and differential fertility)
differential rates of relative intergenerational mobility
perverse fluidity
three-generational processes
discrimination.
The next section of this chapter pursues these different lines of enquiry.
Clearly, a basic factor in explaining the patterns shown in Figure 6.1 are the starting points of the immigrant parents when they arrive in the countries of destination. If a migrant group had been initially recruited, as many of the early labour migrant groups were, to fill low-skilled vacancies in European labour markets, then this could certainly explain why their adult children are disproportionately found in low-skilled positions. This could for example explain why the children of immigrants from Turkey are disproportionately disadvantaged in the destination country’s labour market: many of these Turkish labour migrants came from rural origins and had low levels of human capital. The Turkish immigrants were, in economists’ terminology, negatively selected2 and would have had little possibility of entering more advantaged positions in the salariat (Lee, 1966; Borjas, 1987; Chiswick, 1999).
However, disadvantaged starting points would not necessarily explain why disadvantage had become entrenched over time. The guest worker programmes were generally curtailed after the oil shock of the 1970s, and criteria for immigration to developed countries gradually became more selective. It might therefore be expected that later groups of labour immigrants would not have been quite so “negatively selected” (although migrants for family reunion might not have been affected in the same way). This is of course an empirical matter, which is relatively easy to establish – recent evidence suggests that many immigrant groups, particularly those from East and Southeast Asia and from sub-Saharan Africa, are “positively selected”. There are also are important differences between destination countries, with Canada, Sweden and the United Kingdom being particularly likely to have “positively selected” immigrants (Lessard-Phillips, Fleischmann and van Elsas, 2014).
Even from low initial starting points for the migrants, gradual convergence would be expected if the same patterns of intergenerational fluidity apply to immigrants’ descendants as to those without a migration background. The point is simply that the lower the initial starting point, the longer the process of convergence will take, other things being equal. Thus “positively selected” groups, such as the “Other African” and “South and East Asian” groups, may show little initial disadvantage and rapid convergence with native patterns.
A key question, therefore, is whether the patterns of minority disadvantage shown in Figure 6.1 can be explained by the negative selection and low starting points of the parents who migrated to these European countries. To investigate this, Figure 6.2 takes account of the level of education of the parents, distinguishing between parents with at most lower secondary qualifications and those where one or both parents had a higher, tertiary-level qualification. It then reports the extent of under- or overrepresentation at different levels of the labour market separately for those with lower- and higher-educated parents, respectively. In order to compare like with like, natives with a migration background are compared to natives without a migration background but with parents having the same level of education.
Notes: Men and women combined, economically active respondents aged 15-59, split by parental qualifications (at most lower secondary; upper secondary and above). Differences shown are statistically significant at p<0.05 and based on a cell size>=20. Non-significant differences are not shown. “L” stands for parents that had at most a lower secondary education; “H” stands for parents that had an upper secondary education or above. Positive bars (rising above the horizontal) show overrepresentation whereas negative bars (dropping below the horizontal) show underrepresentation.
Interpretation: In Portugal, natives whose parents were highly qualified and born in the EU-15 are significantly (p<0.05) more likely than their native counterparts whose parents were born in Portugal to be unemployed. They were also significantly less likely to work in the salariat. Those whose parents had lower qualifications were also significantly over-represented among the unemployed, but there was no significant difference in their chances of accessing the salariat.
Source: European Union Labour Force Surveys (EU-LFS) and its Ad-Hoc Module (AHM), 2014.
Figure 6.2 shows clearly that even when account is taken of the level of parental education, natives with immigrant parents from “Other European” or from “North African and Middle Eastern” origins continue to be overrepresented in more disadvantaged positions in the labour market, and underrepresented in more advantaged ones. The patterns of over- and underrepresentation shown in Figure 6.1 cannot therefore be explained solely in terms of the low initial starting positions of the immigrant parents.
Even more strikingly, it is clear that for some groups in some destination countries, the overrepresentation in disadvantaged positions is even greater for those with more highly educated parents. The discussion returns to this point later when it addresses what has been termed “perverse fluidity”.
These results may seem surprising, as it is often assumed that labour migrants in Europe are relatively poorly educated and that this will explain, at least in part, their children’s lack of success in European labour markets. However, while that may well have been true among the early labour migrants in the 1950s and 1960s, the data from the EU-LFS clearly shows that many immigrants today are highly educated. This will partly reflect the educational progress that many of the origin countries themselves have made over the past fifty years. But it will also reflect the changing labour market opportunities in the countries of destination, which have shifted to a considerable extent from an industrial to a post-industrial service economy.
Thus it is important to explore additional processes that might account for the entrenchment of disadvantage among some minority groups. The discussion turns next to the process of replenishment.
Replenishment refers to the continued immigration of new migrants from a given origin to join their co-ethnics in the country of destination (Jimenez, 2010; Waters, 2014). Labour migration from Turkey, North Africa, the Caribbean and South Asia started in earnest in the 1950s and 1960s, as did migration from southern European countries such as Greece, Italy, Spain and Portugal, to meet labour shortages in countries like Austria, Germany and the Netherlands. However, these migrations were not one-off phenomena; in many cases there have been continuing migrations from the same origin countries, involving family reunion or marriages between a native spouse and a migrant spouse. There has also been some return migration, which in principle should be taken into account – in migration/population data one should measure net replenishment, not simply continuing immigration from the same country of origin.
The extent to which net replenishment has occurred for different origin groups is in principle researchable, although data on return migration are not as plentiful as one would wish. Nevertheless, it seems likely that net replenishment has continued in many European countries at a high level among several groups, including notably from Turkey.
There are a number of ways in which replenishment could slow down processes of convergence. Perhaps most importantly, it may tend to inhibit the development of ethnic (that is, community) human capital, if newcomers arrive with relatively low levels of education. (One should not forget however that there has been rapid educational development in many of the origin countries, which have been far from static.) As Borjas (1995) has pointed out, collective human capital can play an important role in mobility chances, over and above individual characteristics. Thus grandchildren of immigrants in a replenished community may not make the same progress in terms of occupational mobility as those in a community whose members have made greater occupational strides. Replenishment may also hamper the development of bridging social ties with the wider native-born community and thus may further weaken inter-generational progress.
Enduring low levels of collective human capital as a result of replenishment are thus expected to slow down processes of convergence. However, this does assume that the replenishment takes the form of continued low-skilled migration. If replenishment brings in migrants who are rich in human capital (which may well be happening with some origin groups), it could have the reverse effect.
Differential fertility is a closely related issue. Large family size tends to be associated with lower educational attainment. While the evidence suggests that fertility patterns tend to converge with those of natives without a migration background, replenishment could potentially slow that convergence down.
A third important factor is a differential rate of relative mobility for natives with and without a migration background. That is, if there is a stronger effect of parental social class on adult children’s social class among those with a migration background than among those without a migration background, combined with low initial starting points, then this would tend to delay processes of convergence. (This would also apply in reverse to groups with higher-than-average initial starting points, who would therefore tend to remain advantaged.)
True, if this were found to be the case, then a key task would be to understand why fluidity is lower among natives with a migration background. One possibility is that groups with stronger family structures may show lower fluidity than groups with weaker family structures (for example with higher rates of marital dissolution). The role of family structure in social mobility is not well understood, but it is plausible that marital dissolution may weaken the intergenerational transmission of advantage if it means that the family of origin comes to play a weaker role in socialisation or intergenerational transmission of cognitive or material resources.
There is circumstantial evidence that family structures will vary between different origin groups. EU-15 origin groups may be expected to have similar structures to those in the countries of destination. Some origin groups from sub-Saharan Africa for example may have more extensive family commitments, and some South and East Asian traditions may encourage stronger family structures than those in more individualistic developed countries. Conversely, there are suggestions in the anthropological literature that the history of plantation economies in parts of the Caribbean and South America weakened family structures (although this has been controversial). Once again, then, merging all origin groups into a single category may hide crucial variations that could have important implications for relative mobility rates.
Some scholars such as Michael Hout (1984) have suggested that there may be greater fluidity among African Americans, but of a perverse kind with African-American parents with higher occupational positions being disproportionately unable to pass on their advantages to their children. The children therefore experience excess downwards mobility. This kind of argument can also be applied to the situation of natives with a migration background. Indeed the evidence of Figure 6.2, which suggested that overrepresentation in disadvantaged positions may be even greater among those with highly educated parents, is consistent with the concept of perverse fluidity. A phenomenon of this sort would therefore counteract the general expectation of convergence across generations, and could have important implications for the entrenchment of disadvantage among the groups affected.
A variety of social processes could be at the root of perverse fluidity. One important finding in mobility research is that parents in higher status positions seem to be able to protect their children from downward mobility, even if their children are less cognitively skilled and less educationally successful. The role of social ties and connections has been emphasised in explaining why elite groups are able to protect their less able members from downward mobility. Thus elite groups may be able to use their social ties and know‑how to secure positions for their children where soft skills as opposed to technical skills are valuable (Jackson, 2001). However, immigrant parents may not be able to protect their children from downward mobility to the same extent. One plausible mechanism is that minority parents who have secured higher status positions in the labour market may nonetheless find themselves socially excluded from the social networks of elite sections of European society. If families with a migration background lack access to these networks and/or soft skills, they may not be able to protect their less skilled children from downward mobility in the way that their peers without a migration background can.
Social mobility researchers have increasingly become interested in the possibility that grandparents’ resources might have consequences for grandchildren’s occupational positions, over and above parents’ positions (Chan and Boliver, 2013). In particular, grandparents from privileged positions may be able to help grandchildren’s education or occupational advancement, even if their own children (the parents) have been downwardly mobile. Mobility research therefore needs to move beyond the usual two-generational approach to a three-generational approach, which could be particularly relevant to the descendants of migrants, in a variety of ways. Firstly, it appears to be the case that some immigrants are downwardly mobile on arrival in the country of destination. It may be that the children of these immigrants will experience disproportionately greater upward mobility and will tend to return to the positions that their grandparents had.
Furthermore, grandparental influences may be stronger among communities that have a three-generational type of family structure where grandparents, parents and children co-reside (a stem family structure) – as is the case for example in some South and East Asian communities – than is typical of developed countries where two-generational nuclear families tend to predominate. There seems to be little evidence on the extent of grandparental influences on the mobility trajectories of people with a migration background; however, given the importance of three-generational families in some communities, it should not be assumed that grandparental influences will be the same as in high-income OECD countries with different intergenerational cultural traditions.
To be sure, it is also possible that grandparents who remain in their country of origin may have less impact on grandchildren’s outcomes in the country of destination than would be the case for natives without a migration background, where grandparents could be more likely to be on hand to help. That could explain why descendants of migrants make less progress in the labour market.
There is considerable evidence from field experiments that natives with a migration background experience discrimination in the labour market, and that this varies both between origin groups and between countries of destination. (For recent reviews of the literature see Pager and Shepherd, 2008; Heath, Liebig and Simon, 2013; Rich, 2014; and Schirnt and Ruedin, 2016) Meta-analyses of field experiments typically find that discrimination is significantly less against groups with a west European, Australian or North American background than against those whose origins were in Turkey, North Africa and the Middle East, or sub-Saharan Africa and the Caribbean. Discrimination therefore seems to be a plausible explanation for some of the disadvantage experienced by minorities with “Other European” and “North African and Middle Eastern” origins observed in Figure 6.1 and Figure 6.2. However, it seems less likely that discrimination will be important in explaining the disadvantages experienced by the “EU-15” and “EU‑13” groups in those two figures.
One particularly worrying finding is that discrimination appears to be equally prevalent against natives with an immigration background and the immigrants themselves (Midtbøen, 2016). On theoretical grounds one might have expected that natives with an immigration background, as a result of their greater familiarity with the language and culture of the destination country and their possession of host-country qualifications, would experience lesser discrimination than their immigrant parents did. However, this does not seem to be the case in practice. Statistical evidence on ethnic penalties has also suggested that natives with an immigration background appear to experience ethnic penalties in obtaining jobs of a magnitude similar to that experienced by their immigrant parents. At the same time, occupational attainments among those fortunate enough to obtain work are closer to those of the majority group than were the attainments of their immigrant parents (see for example Li and Heath, 2016). Given the serious consequences of unemployment, both for the careers of those affected and for their children’s outcomes, discrimination (or other forms of social exclusion) could be a major factor leading to entrenched disadvantage.
The provisional conclusion, then, is that there are serious risks of entrenched disadvantage among some, but not all minority groups. Entrenched disadvantage, of the kind that some African Americans experience in the United States after many generations, remains a worrying possibility, at least for some disfavoured groups such as those of North African and Middle Eastern origin. Moreover, it is doubtful that this entrenchment can be explained solely by the low initial starting points of the immigrant parents. It is likely that additional processes such as replenishment, perverse fluidity, three-generational processes and discrimination all play a part. Other social processes such as those involved in the educational system, housing market, and social relationships will also be important, although some of these will be endogenous, at least in part, and will be driven by occupational class and migration background.
At present it is impossible to quantify reliably either the scale and incidence of the entrenched disadvantage, or the relative importance of the different explanations. Much better data than available at present are needed in order to chart dynamic processes over time and generations. Also needed is much better monitoring of the situation in order to understand the nature and scale of the problem. Only then can it be ascertained whether the problem for particular origin groups has indeed remained entrenched both across historical time and across generations, or whether the patterns of over- and underrepresentation shown in Figure 6.1 can be understood simply as consequences of low initial starting points for the parents allied with the same rates of social fluidity as among natives without a migration background. Since full convergence is not expected to occur within two generations, it is important to identify the grandchildren’s situation and not solely that of the children of immigrants, which at the moment is simply not possible with sources such as the EU-LFS (although the ad hoc module used in this chapter is an important step forward). There is a therefore huge data gap. In the absence of better data, it must be assumed that there is a real problem that cannot simply be explained by low initial starting points.
Moreover, policy makers need a better diagnosis of the causes of the problem(s) in order to understand the nature of the challenge they face. What is the relative importance of replenishment and collective ethnic capital, differential fluidity and perverse fluidity, three-generational processes, and discrimination? Which groups are most adversely affected, and why?
The policy implications will differ considerably, depending on the nature of the diagnosis. Thus, if disadvantage among the children of immigrants is primarily due to low initial starting points, but with no differentials in rates of two- or three-generational mobility, then disadvantage will not become entrenched and the problem will largely resolve itself. In effect, this represents a benign scenario of equality of opportunity with high and similar rates of social fluidity for both those with and those without a migration background. The result will be a “shuffling of the cards” in each generation and relatively rapid convergence in occupational profiles (albeit slower convergence in less socially mobile countries). Under this benign scenario, major policy initiatives may not be needed.
In light of differential rates of relative social mobility, policy interventions may be needed, especially if they take the form of perverse fluidity, with higher rates of downwards mobility or simply greater “stickiness”, where people with a migration background have higher relative rates of remaining in lower-status positions. A deeper understanding of the processes involved is needed in order to suggest specific policy solutions. One might for example want to focus on the role of the educational system: if descendants of immigrants are disproportionately channelled into low-status streams in the educational system (perhaps because of their geographical concentration in deprived areas), then reforms will need to be targeted at equalising the playing field in education. On the other hand, if perverse fluidity occurs because of social exclusion by the advantaged social classes, strategies for increasing diversity and encouraging bridging social ties will be needed.
Finally, issues of discrimination suggest that policy interventions need to be focused on enforcement of existing legislation rather than radically new initiatives. Discrimination is against the law in all European and North American countries of destination, but the onus tends to be on the victims of discrimination to hold employers to account. External monitoring by equality commissions or similar public bodies to ensure fair employment is rare. One notable example of such monitoring is that in Northern Ireland, where an ambitious and successful monitoring and enforcement programme has been working for a number of years (Muttarak et al., 2013).
As a first step however, the priority must be to improve the availability of the data, so that the nature and scale of the problem can be identified. The ideal data set would be sufficiently large so as to furnish different kinds of information, and therefore insight. First, it would enable researchers to differentiate between the main ethnic minority groups. Since there is now powerful survey-based and field experimental research showing that disadvantage and discrimination vary among ethnic groups, it is no longer sufficient to combine all natives with a migration background into a single undifferentiated category. This undifferentiated approach runs the serious risk of underestimating the level of disadvantage experienced by some particularly disfavoured groups.
Secondly, the ideal data set would also enable one to distinguish between the grandchildren and even great-grandchildren of migrants, and not simply their children. Convergence with the majority-group’s occupational profile is not expected to occur within a single generation, but unfortunately present resources such as the EU-LFS ad hoc module only allows consideration of the experiences of the children of migrants. In other words, it provides only a partial view of the socio-economic outcomes of natives with a migration background – the grandchildren and later generations are invisible in the EU-LFS, and are merged with the natives without a migration background. It may possibly be the case that the grandchildren of migrants do experience equality of opportunity with the majority group, but there is no way of knowing at present whether this is the case or not. It will probably rarely be practicable in survey research to ascertain the countries of birth of all grandparents or great grandparents, and it is likely that other measures such as self-reported ancestry will be needed. (See for example the measure of ancestry that has been developed for the European Social Survey by Heath, Schneider and Butt, 2016.)
Social mobility is also a dynamic process that standard instruments do not capture especially well. The typical survey-based analysis of social mobility simply provides a snapshot of a single moment in time, whereas people’s occupations may, to greater or lesser extents, vary across the life cycle. One potential worry is that natives with a migration background may find themselves disproportionately in sectors of the labour market that do not provide much in the way of career progression (as indeed do many women who have interrupted careers and who take up part-time work). Thirdly then, the ideal data set would provide information on occupational trajectories, which could be collected either prospectively through (expensive) panel studies or retrospectively in cross-section surveys.
Fourth and most importantly, this data set would furnish necessary information about the socio-economic origins of both those with and without a migration background. Retrospective measures of mothers’ and fathers’ social class at the time when the respondents themselves were growing up are reasonably reliable and are often implemented in cross-section surveys, although they are not routinely collected in major official surveys. Furthermore, there are particular challenges if one wishes to measure intergenerational income mobility. Information about parental income cannot reliably be collected retrospectively, and standard practice is to draw on long-term panel studies where the same respondents are followed up from childhood to adulthood. (Linked tax records of adults and their parents are practicable in some countries, but data on migration background will not typically be available.)
Panel studies of this kind are very expensive, and inevitably new panels will take many years to generate results about intergenerational mobility. Moreover, panel studies themselves suffer from important issues of attrition and thus potentially of bias. Cross‑sectional studies investigating social class mobility are therefore likely to be the most practicable vehicles for this kind of research, without however overlooking the potential of the kind of register data available in the Nordic countries, and of linked census data that are available in the United Kingdom with the Longitudinal Survey. The highly regarded European Social Survey provides a model of what can be done, but the sample size is unfortunately too small for the kind of analysis needed.
In conclusion, the existing evidence suggests clearly enough that some minority groups in some countries are particularly at risk of falling into the trap of entrenched disadvantage, but we know far too little about the mechanisms involved or even whether the problems persist into later generations. The lack of current good data obscures many problems. But ‘absence of evidence’ is not the same as ‘evidence of absence’. If we do not collect the right sort of data, we may be lulled into a false sense of security and end up sleepwalking into problems of entrenched disadvantage.
Alba, Richard D., and Victor Nee (2003), Remaking the American Mainstream: Assimilation and Contemporary Immigration. Harvard University Press, Cambridge.
Algan, Y., C. Dustmann, A. Glitz and A. Manning (2010), “The economic situation of first and second-generation immigrants in France, Germany and the United Kingdom”, The Economic Journal, Vol. 120, p. 542: F4-F30.
Baert, S. (2017), ‘Hiring discrimination: an overview of (almost) all correspondence experiments since 2005’, Global Labor Organization Discussion Paper, No. 61.
Borjas, G.J. (1987), “Self-selection and the earnings of immigrants”, American Economic Review, Vol. 77, pp. 531-553.
Borjas, G. J. (1992), “Ethnic Capital and Intergenerational Mobility.” Quarterly Journal of Economics, Vol. 107, pp. 123–50.
Borjas, G. J. (1995), “Ethnicity, Neighbourhoods and Human Capital Externalities.” American Economic Review, Vol. 85, pp. 365–90.
Breen, R. (ed.) (2004), Social Mobility in Europe. Oxford University Press, Oxford.
Bukodi, E., Paskov, M. and Nolan, B (2017), “Intergenerational class mobility in Europe: a new account and an old story”, INET Oxford Working Paper no. 2017-03.
Chan, T.W. and V. Boliver (2013), “The Grandparents Effect in Social Mobility: Evidence from British Birth Cohort Studies”, American Sociological Review, Vol. 78, pp. 662-78.
Chiswick, B.R. (1999), “Are immigrants favourably self-selected?” American Economic Review, Vol. 89, pp. 181-5.
Feliciano, C. (2006), Unequal Origins: Immigrant Selection and the Education of the Second Generation. LFB Scholarly Publishing, New York.
Ganzeboom, H.B.G., R. Luijkx, and D. J. Treiman (1989), “Intergenerational Class Mobility in Comparative Perspective”, Research in Social Stratification and Mobility, Vol. 8, pp. 3-84.
Heath, A.F., and Y. Brinbaum (eds.) (2014), “Unequal Attainments: Ethnic Educational Inequalities in Ten Western Countries”. Proceedings of the British Academy, Vol. 196. Oxford University Press, Oxford.
Heath, A. and S.-Y. Cheung eds. (2007), “Unequal Chances: Ethnic Minorities in Western Labour Markets”, Proceedings of the British Academy, Vol. 137, Oxford University Press, Oxford.
Heath, A.F., T. Liebig and P. Simon (2013), “Discrimination against immigrants – measurement, incidence and policy instruments”, in OECD, International Migration Outlook 2013, OECD Publishing, Paris. http://dx.doi.org/10.1787/migr_outlook-2013-7-en.
Heath, A., S. Schneider and S. Butt (2016), “Developing a measure of socio-cultural origins for the European Social Survey”, GESIS papers 2016/16
Hout, M (1984), “Status, Autonomy, and Training in Occupational Mobility.” American Journal of Sociology, Vol. 89/6, pp. 1379–1409.
Jackson, M. (2001), “Non-meritocratic job requirements and the reproduction of class inequality”, Work, Employment and Society, Vol. 15, pp. 619-30.
Jimenez, T.R. (2010), Replenished Ethnicity: Mexican Americans, Immigration and IdentityUniversity of California Press, Berkeley.
Lee, E.S. (1966), “A theory of migration”, Demography Vol. 3, pp. 47-57.
Lessard-Phillips, L., Fleischmann, F. and van Elsas, E. (2014), “Ethnic minorities in ten western countries: migration flows, policies and institutional differences”, in Heath, A.F., and Y. Brinbaum, (eds.) “Unequal Attainments: Ethnic Educational Inequalities in Ten Western Countries”, Proceedings of the British Academy, Vol. 196. Oxford University Press, Oxford.
Li, Y. and A.F. Heath (2016), “Class matters: a study of minority and majority social mobility in Britain, 1982-2011”, American Journal of Sociology, Vol. 122, pp. 162-200.
Midtbøen, A. (2016), ‘Discrimination of the second generation: evidence from a field experiment in Norway’, Journal of International Migration and Integration, Vol. 17, pp. 253-72.
Muttarak, R., H. Hamill, A. Heath and C. McCrudden (2013), “Does Affirmative Action Work? Evidence from the Operation of Fair Employment Legislation in Northern Ireland”. Sociology, Vol. 47/3, pp. 560-79.
Pager, D. and H. Shepherd (2008) “The Sociology of Discrimination. Racial Discrimination in Employment, Housing, Credit and Consumer Markets”, Annual Review of Sociology, Vol. 34, pp. 181–209.
Patterson, O. (2005), “Four modes of ethno-somatic stratification: the experience of Blacks in Europe and the Americas”.in G.C. Loury, T. Modood and S.M. Teles (eds.), Ethnicity, Social Mobility and Public Policy: Comparing the US and the USA. Cambridge University Press, Cambridge.
Rich, J. (2014), “What do field experiments of discrimination in markets tell us? A meta analysis of studies conducted since 2000.” IZA discussion paper 8584.
Schirnt, E. and D. Ruedin (2016), “Ethnic discrimination in hiring decisions: A meta analysis of correspondence tests 1990-2015”, Journal of Ethnic and Migration Studies, Vol. 42, pp. 1115-1134.
Waters, M.C. (2014), “Defining difference: the role of immigrant generation and race in American and British immigration studies”, Ethnic and Racial Studies, Vol. 37, pp. 10-26.
Yaish, M., and R. Andersen. 2012. “Social Mobility in 20 Modern Societies: The Role of Economic and Political Context.” Social Science Research. Vol. 41, pp. 527–38.
Origin and destination countries |
Unemployed/ low skilled |
Skilled industrial |
Intermediate |
Salariat |
---|---|---|---|---|
Austria |
||||
Other Europe |
16 |
6 |
-5 |
-18 |
Belgium |
||||
EU-15 |
5 |
-4 |
||
EU-13 |
18 |
-17 |
||
Other Europe |
20 |
9 |
-25 |
|
North Africa/ Middle East |
18 |
-21 |
||
Spain |
||||
EU-15 |
15 |
-8 |
-7 |
|
North Africa/ Middle East |
18 |
-8 |
-11 |
|
Central/South America |
-7 |
|||
France |
||||
Other Europe |
||||
North Africa/ Middle East |
-4 |
-3 |
5 |
|
Other Africa |
-10 |
|||
South/East Asia |
-11 |
15 |
||
Greece |
||||
EU-15 |
-15 |
|||
Other Europe |
20 |
-12 |
-11 |
|
North Africa/ Middle East |
21 |
-22 |
||
Italy |
||||
EU-15 |
4 |
|||
Other Europe |
18 |
-12 |
||
North Africa/ Middle East |
12 |
-6 |
||
Central/ South America |
20 |
-12 |
||
Luxembourg |
||||
EU-15 |
4 |
|||
Norway |
||||
South/East Asia |
17 |
-9 |
||
Portugal |
||||
EU-15 |
16 |
|||
Other Africa |
7 |
-12 |
||
Sweden |
||||
EU-13 |
-6 |
|||
Other Europe |
7 |
|||
North America and Australia |
14 |
|||
North Africa/ Middle East |
16 |
|||
United Kingdom |
||||
EU-15 |
-5 |
7 |
||
EU-13 |
-5 |
|||
North America and Australia |
11 |
|||
North Africa/ Middle East |
11 |
|||
Other Africa |
-7 |
16 |
||
South/East Asia |
-6 |
5 |
||
Central/South America |
5 |
-3 |
Note: Men and women combined, economically active respondents aged 15-59. Differences shown are statistically significant at p<0.05 and based on a cell size>=20.
Source: European Union Labour Force Surveys (EU-LFS) and its Ad-Hoc Module (AHM), 2014.
Unemployed/ low skilled |
Skilled industrial |
Intermediate |
Salariat |
|||||
---|---|---|---|---|---|---|---|---|
Low |
High |
Low |
High |
Low |
High |
Low |
High |
|
Austria |
||||||||
Other Europe |
+14 |
+15 |
+9 |
-13 |
-11 |
-17 |
||
Belgium |
||||||||
EU-15 |
+5 |
|||||||
EU-13 |
+25 |
-22 |
||||||
Other Europe |
+14 |
+25 |
-18 |
|||||
North Africa/ Middle East |
+17 |
-7 |
-15 |
|||||
Spain |
||||||||
EU-15 |
+22 |
+11 |
-11 |
-11 |
-12 |
|||
North Africa/ Middle East |
+20 |
+21 |
||||||
Central/South America |
+13 |
|||||||
France |
||||||||
Other Europe |
||||||||
North Africa/ Middle East |
+5 |
-6 |
-6 |
+8 |
||||
Other Africa |
-13 |
-7 |
||||||
South/ East Asia |
||||||||
Greece |
38 |
39 |
9 |
5 |
36 |
20 |
18 |
36 |
EU-15 |
||||||||
Other Europe |
+18 |
+24 |
-18 |
-21 |
||||
North Africa/ Middle East |
||||||||
Italy |
28 |
21 |
17 |
6 |
31 |
28 |
24 |
46 |
EU-15 |
||||||||
Other Europe |
12 |
+27 |
+6 |
-12 |
-21 |
|||
North Africa/ Middle East |
+18 |
-16 |
||||||
Central/South America |
+18 |
+24 |
-15 |
|||||
Luxembourg |
16 |
12 |
10 |
8 |
23 |
17 |
51 |
62 |
EU-15 |
+5 |
-6 |
||||||
Norway |
16 |
14 |
18 |
13 |
25 |
18 |
41 |
56 |
South/East Asia |
+22 |
-13 |
||||||
Portugal |
34 |
23 |
19 |
4 |
21 |
13 |
25 |
60 |
EU-15 |
+17 |
+22 |
-19 |
|||||
Other Africa |
+17 |
+7 |
-11 |
+7 |
-12 |
|||
Sweden |
17 |
19 |
21 |
13 |
26 |
19 |
36 |
49 |
EU-13 |
||||||||
Other Europe |
+13 |
-12 |
||||||
North America and Australia |
+16 |
|||||||
North Africa/ Middle East |
||||||||
United Kingdom |
25 |
20 |
12 |
7 |
27 |
22 |
37 |
52 |
EU-15 |
-6 |
+9 |
||||||
EU-13 |
+13 |
|||||||
North America/ Australia |
+15 |
|||||||
North Africa/ Middle East |
||||||||
Other Africa |
-10 |
+18 |
||||||
South and East Asia |
-7 |
|||||||
Central/South America |
-7 |
Notes: Men and women combined, economically active respondents aged 15-59, split by parental qualifications (at most lower secondary; upper secondary and above). Differences shown are statistically significant at p<0.05 and based on a cell size>=20.
Source: European Union Labour Force Surveys (EU-LFS) and its Ad-Hoc Module (AHM), 2014.
AT |
BE |
ES |
FI |
FR |
GR |
IT |
LU |
NO |
PT |
SE |
UK |
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
age |
35 |
34 |
25 |
31 |
36 |
38 |
30 |
33 |
34 |
30 |
34 |
34 |
(14) |
(13) |
(11) |
(15) |
(13) |
(13) |
(12) |
(13) |
(13) |
(12) |
(13) |
(14) |
|
Highest qualifications |
||||||||||||
Lower secondary |
27 |
35 |
45 |
37 |
25 |
32 |
41 |
35 |
26 |
48 |
23 |
21 |
(44) |
(48) |
(50) |
(48) |
(44) |
(47) |
(49) |
(48) |
(44) |
(50) |
(42) |
(41) |
|
Upper secondary |
49 |
40 |
28 |
43 |
45 |
43 |
44 |
43 |
35 |
29 |
48 |
41 |
(50) |
(49) |
(45) |
(50) |
(50) |
(50) |
(50) |
(50) |
(48) |
(45) |
(50) |
(49) |
|
Tertiary |
24 |
25 |
27 |
20 |
30 |
25 |
16 |
22 |
39 |
23 |
29 |
38 |
(43) |
(43) |
(45) |
(40) |
(46) |
(43) |
(37) |
(42) |
(49) |
(42) |
(46) |
(49) |
|
Dummy: both parents migrants |
42 |
44 |
53 |
17 |
39 |
76 |
15 |
57 |
25 |
29 |
32 |
55 |
(49) |
(50) |
(50) |
(38) |
(49) |
(43) |
(36) |
(50) |
(43) |
(46) |
(46) |
(50) |
|
Parental qualifications |
||||||||||||
Lower secondary |
27 |
50 |
37 |
27 |
54 |
39 |
41 |
35 |
19 |
48 |
19 |
50 |
(45) |
(50) |
(48) |
(45) |
(50) |
(49) |
(49) |
(48) |
(39) |
(50) |
(39) |
(50) |
|
Upper secondary |
46 |
27 |
24 |
36 |
24 |
39 |
44 |
42 |
33 |
23 |
37 |
22 |
(50) |
(44) |
(43) |
(48) |
(43) |
(49) |
(50) |
(49) |
(47) |
(42) |
(48) |
(41) |
|
Tertiary |
27 |
23 |
40 |
37 |
22 |
22 |
15 |
23 |
48 |
29 |
45 |
28 |
(44) |
(42) |
(49) |
(48) |
(42) |
(42) |
(36) |
(42) |
(50) |
(46) |
(50) |
(45) |
|
Occupational status |
||||||||||||
Inactive |
25 |
38 |
63 |
36 |
29 |
34 |
51 |
35 |
25 |
40 |
22 |
35 |
(44) |
(49) |
(48) |
(48) |
(45) |
(48) |
(50) |
(48) |
(43) |
(49) |
(41) |
(48) |
|
Unemployed |
6 |
9 |
12 |
11 |
8 |
17 |
9 |
5 |
4 |
12 |
7 |
5 |
(24) |
(29) |
(33) |
(32) |
(28) |
(37) |
(29) |
(22) |
(20) |
(32) |
(26) |
(22) |
|
Low-skilled work |
15 |
8 |
8 |
11 |
12 |
8 |
7 |
7 |
12 |
11 |
10 |
9 |
(36) |
(27) |
(27) |
(32) |
(33) |
(27) |
(26) |
(26) |
(32) |
(31) |
(29) |
(29) |
|
Skilled Industrial |
12 |
10 |
2 |
7 |
8 |
5 |
5 |
7 |
8 |
5 |
9 |
4 |
(32) |
(30) |
(15) |
(25) |
(27) |
(22) |
(21) |
(26) |
(27) |
(22) |
(29) |
(20) |
|
Intermediate |
14 |
13 |
6 |
15 |
14 |
17 |
12 |
15 |
13 |
12 |
16 |
15 |
(35) |
(34) |
(23) |
(35) |
(34) |
(38) |
(33) |
(36) |
(34) |
(32) |
(36) |
(36) |
|
Salariat |
28 |
22 |
9 |
20 |
29 |
19 |
16 |
31 |
38 |
21 |
36 |
32 |
(45) |
(42) |
(29) |
(40) |
(46) |
(39) |
(36) |
(46) |
(49) |
(41) |
(48) |
(47) |
|
Region of origin |
||||||||||||
EU15 |
27 |
55 |
42 |
41 |
35 |
22 |
33 |
90 |
60 |
18 |
59 |
32 |
(44) |
(50) |
(49) |
(49) |
(48) |
(41) |
(47) |
(30) |
(49) |
(38) |
(49) |
(47) |
|
EU13 |
29 |
5 |
2 |
5 |
3 |
12 |
9 |
2 |
4 |
0 |
8 |
6 |
(45) |
(21) |
(14) |
(22) |
(16) |
(33) |
(29) |
(13) |
(19) |
(5) |
(27) |
(23) |
|
Other Europe |
36 |
11 |
2 |
23 |
4 |
41 |
17 |
2 |
7 |
1 |
12 |
2 |
(48) |
(32) |
(14) |
(42) |
(20) |
(49) |
(37) |
(14) |
(25) |
(9) |
(32) |
(14) |
|
North America and Australia |
1 |
1 |
2 |
6 |
0 |
6 |
6 |
0 |
8 |
0 |
3 |
5 |
(11) |
(10) |
(13) |
(25) |
(6) |
(25) |
(23) |
(7) |
(27) |
(7) |
(17) |
(22) |
|
North Africa and Middle East |
4 |
18 |
16 |
10 |
45 |
16 |
15 |
0 |
3 |
0 |
8 |
3 |
(19) |
(38) |
(36) |
(30) |
(50) |
(36) |
(36) |
(6) |
(18) |
(5) |
(26) |
(17) |
|
Other Africa |
1 |
7 |
4 |
5 |
8 |
1 |
6 |
3 |
1 |
66 |
2 |
10 |
(8) |
(26) |
(18) |
(21) |
(27) |
(8) |
(24) |
(18) |
(12) |
(47) |
(14) |
(30) |
|
South and East Asia |
2 |
3 |
2 |
8 |
4 |
2 |
4 |
1 |
15 |
3 |
4 |
31 |
(15) |
(16) |
(15) |
(27) |
(20) |
(13) |
(19) |
(11) |
(36) |
(16) |
(19) |
(46) |
|
Central and South America |
1 |
1 |
31 |
3 |
1 |
1 |
11 |
1 |
2 |
12 |
5 |
11 |
(9) |
(9) |
(46) |
(17) |
(10) |
(8) |
(31) |
(9) |
(14) |
(32) |
(21) |
(31) |
|
N |
1 959 |
1 816 |
2 102 |
607 |
3 333 |
2 220 |
2 127 |
1 681 |
578 |
1 095 |
1 947 |
5 601 |
Source: European Union Labour Force Surveys (EU-LFS) and its Ad-Hoc Module (AHM), 2014.
← 1. The “Other Europe” category includes EFTA countries (Switzerland, Iceland, Liechtenstein, Norway), candidate countries (Former Yugoslav Republic of Macedonia and Turkey), and Andorra, Albania, Bosnia Herzegovina, Belarus, Faroe Islands, Monaco, Republic of Moldova, Montenegro, Serbia, Russian Federation, San Marino, Ukraine, Vatican City and Kosovo* (*This designation is without prejudice to positions on status, and is in line with United Nations Security Council Resolution 1244/99 and the Advisory Opinion of the International Court of Justice on Kosovo’s declaration of independence.).
← 2. Selectivity refers to the degree to which migrants deviate from the general population in their country of origin. The term positive selection is used if migrants are disproportionately recruited from the upper part of the distribution of motivation or skills, and negatively selected if they are disproportionately recruited from the lower part of the distribution in the origin country.