Canada’s immigration policy includes a system of points assigned according to various criteria for selection of economic migrants. Using data from a 2002 retrospective survey of women and men admitted to live in Quebec on economic grounds, Julie Lacroix, Alain Gagnon, and Vincent Lortie analyse these immigrants’ access to a first job, and to a first job corresponding to their qualifications, over the four years following their arrival. Taking an intersectional approach, they analyse differences in labour market outcomes of immigrants from different countries, distinguishing between men and women. They show that disparities between geographical origin groups are more marked for men, and that knowledge of the French language makes a greater difference for women.
1Women immigrating to Canada have traditionally been admitted either for reasons of family reunification or as dependents of another migrant, generally a man. This administrative category has long relegated women to the status of economic dependents, obscuring both their involvement in the decision process behind migration and their participation in the economic sphere. Consequently, researchers have focused on the situation of these women within the family unit, and in certain sectors of precarious economic activity where they are particularly present (Catarino and Morokvasic, 2005). They tend to concentrate on women’s work in socially undervalued, unskilled sectors (domestic work, care work), or on the problems of exploitation and violence that these women may face. While these studies are based on a statistical representation, they do not capture the full diversity of migration trajectories, and notably the growing numbers of highly skilled immigrant women.
2To be admitted as permanent residents of Quebec, immigrants must be recognized as refugees, accepted through family reunification, or selected as economic immigrants (either as principal applicants or as dependents thereof). Although the majority of women are admitted as dependents, increasing numbers come as principal applicants; their proportion rose from 23% to 39% between 1980 and 2010 (CIC, 2013). Skilled workers [1] are selected on the basis of a selection grid that is specific to Quebec, [2] consisting of a set of weighted socio-occupational criteria, mainly linked to human capital and employment. More than 60% of individuals admitted as permanent residents between 2009 and 2013 were selected in this way (MIDI, 2014).
3Despite the attractive profile of skilled female immigrants to Quebec, their labour market trajectories may differ from those of men due to their family trajectories, their concentration in certain sectors of employment and training, or different forms of discrimination and exclusion (Boyd, 1996). National statistics on employment rates and salaries show that despite high levels of education, female immigrants are disadvantaged (Conseil du statut de la femme, 2010). The difficulties faced by women of certain origins in host country labour markets are often attributed to “culture” (see Frank and Hou, 2013, for Canada). It is argued that gender roles specific to each region make some women reluctant to take on employment, regardless of the labour market structure or their capacity to find a job. Unequal treatment of immigrants (both men and women) is also explained in terms of discrimination on the basis of national and ethnic origin, both in Quebec (Piché et al., 2002) and elsewhere in Canada (Pendakur and Pendakur, 2010; Reitz, 2005). According to these studies, access to employment varies by region of origin, with an advantage for traditional countries of immigration, such as Europe and the United States (Adsera and Chiswick, 2007; Boudarbat and Gontero, 2008; Wanner and Ambrose, 2003).
4The present study aims to shed light on the role of gender and national origin (country of birth) in the employment trajectories of immigrant women and men chosen by the government of Quebec for their productive potential in the economic sector. First, we seek to identify effects linked to gender and national origin, and then to determine whether the effect of origin differs for men and women. Second, we investigate whether other individual characteristics, identified by the criteria in the selection grid, have differential effects on women’s and men’s ease of access to employment. The use of separate models by gender highlights determinants that are gender-specific, first in access to any job, and second in access to a job corresponding to the individual’s pre-migration level of education. The interaction terms show whether the effect of gender is merely additive, or combines differently with regions of origin through multiplier effects. Finally, to limit confusion between factors that play a role in the decision to join the labour market and other factors attributable to obstacles to employment, such as unfavourable individual characteristics and hiring discrimination, we restrict our analyses to principal applicants in the category of skilled workers. Factors linked to “cultural barriers”, believed to limit women’s access to employment, are thus neutralized in our comparison of different groups seeking to enter the labour market.
I. Literature review
Access to employment and (over-)education
5Although the proportion of selected immigrants has increased over the last 30 years, the situation of recent immigrants on the labour market has deteriorated relative to both previous immigrant cohorts and the native-born population (Boudarbat and Boulet, 2007; Cousineau and Boudarbat, 2009; Girard et al., 2008). Occupational over-qualification, or “holding a job that requires lesser qualifications or that pays less than would theoretically be available to people having the same level of education” (OECD, 2007), is relatively widespread in the Canadian population, but particularly among persons born outside the country, and even more so among immigrant women (Chicha and Charest, 2008).
6One of the explanations generally offered for this difference in economic integration is the difficulty of transferring qualifications, skills, and experiences obtained abroad (Chiswick and Miller, 2009). Factors cited include, among others, the quality of educational qualifications, the recognition of training and experience, and the protection of certain occupations, such as engineering and medicine, by their professional orders. On this topic, Reitz (2001) argues that the growing importance of education in a knowledge economy has increased the educational level of Canadian natives, and consequently, due to employers’ preference for locally obtained qualifications, decreased the relative value of immigrants’ formal skills. Other studies have highlighted the importance of language skills in access to high-quality employment (Grondin, 2007; Helgertz, 2010; Thompson, 2000). In a thesis on access to employment among selected workers in Quebec, Bégin (2010) identified a positive effect of knowledge of French and English on individuals’ chances of obtaining a job corresponding to their pre-migration level of education.
7Many studies have examined measures of over-education and over-qualification. They have indicated a variable tendency towards over-qualification among immigrants, depending on their region of origin (Chicha and Charest, 2008). Using the median level of education in different occupations in Great Britain, Battu and Sloane (2004) showed that ethnic minorities are more often over-qualified for their positions, particularly when their qualifications were obtained abroad. In an analysis of speed of access to employment by socioeconomic status, Frank (2013) arrived at similar conclusions. According to Frank, visible minorities’ slower access to jobs with high socioeconomic status reflects a form of “social closure”. Other authors have argued that it is more appropriate to examine workers’ skills, rather than their level of education (Allen and Van Der Velden, 2001; Green and Zhu, 2010). They argue that individuals are not over-qualified if they make full use of their knowledge and skills in the tasks that they perform in their job, even if their level of education exceeds that normally expected for the job. Other studies posit that the observable economic differences between groups of immigrants and natives are driven by the considerable variability in the quality of educational systems and of skills acquired in different origin countries (Bonikowska et al., 2010; Ferrer et al., 2006; Sweetman, 2004). For equivalent results in standardized tests (such as reading or problem-solving) the authors find that immigrants and natives receive similar wages, a result which, they argue, rules out the hypothesis of discrimination.
8It has also been suggested that the low return on experience acquired abroad, principally among the most educated individuals (Helgertz, 2010), explains a large proportion of the wage differences between immigrants and natives (Ferrer et al., 2006). Under this hypothesis, it is crucial for immigrants to acquire local experience as soon as they arrive, which they can only do by entering the labour market through jobs requiring a skill level lower than their own. These first experiences on the labour market are transitory, and are followed by upward socio-occupational mobility (Chiswick, 1978). However, using data from the Longitudinal Survey of Immigrants to Canada (LSIC), Sharaf (2013) observed that the frequency of over-education decreases only slightly after four years of residence. Moreover, Renaud and Cayn (2006) highlight that the selected immigrants may be less tolerant of poor working conditions than immigrants from other categories of admission, in keeping with the expectations raised by a targeted and restrictive selection process.
National origin and gender
9Inequalities linked to gender, immigrant status, and ethnicity or race (according to U.S. terminology) or visible minority status (according to Canadian terminology) are sometimes presented as cumulative disadvantages. This has been qualified as the “double” or “triple disadvantage”, by some authors and was first illustrated by Boyd (1984). In her study, Boyd noted an increased concentration of immigrant women in certain sectors of precarious activity, and, by extension, lower wages, particularly for women from developing countries. An alternative to the double disadvantage hypothesis is the hypothesis of intersectionality. This concept refers to combined forms of discrimination and domination wherein factors of inequality (gender, ethnicity, class) have both additive and multiplicative effects (Bürkner, 2012).
10A few statistical measures have been proposed to dissociate these two potential aspects of disadvantage, either additive (double effect) or multiplicative (intersectionality). One commonly used technique is that of Oaxaca (1973). It consists of decomposing wage differences into a portion that is explained in terms of returns on human capital and an unexplained portion, interpreted as a factor of discrimination. For example, in Canada, Shamsuddin (1998) concluded that the size of the double-negative effect on immigrant women, in comparison to native-born men, is 71.9%. This effect is then decomposed into a portion attributable to the dominant effect of gender (70.4%), corresponding to the difference between native-born individuals, and another portion due to the effect of origin (1.5%), i.e. the difference between women. However, Lopez (2012) emphasizes that this approach to measuring separate effects of gender and origin is not necessarily equivalent to considering the hypothesis of intersectionality, which instead should be examined through the interconnection of these different components. In the United States, for example, Kim (2009) concluded that black women face not only two distinct penalties linked to gender and ethnicity, but a third disadvantage due to the intersection of gender and ethnicity. Lopez (2012) observes that immigrant status explains a larger proportion of the disadvantage in the United States, while in Canada, the effect of gender apparently dominates that of country of origin (Shamsuddin, 1998; Tastsoglou and Preston, 2012).
11Another approach consists of including interaction terms between categories of interest in multivariate models, as in the study of Greenman and Xie (2008), which examined variation in the ethnicity effect in wage differences between men and women in the United States. Interactions between ethnic minority status and gender indicated a smaller penalty for ethnicity among women.
12A number of longitudinal studies on differences in immigrants’ access to employment by origin have been conducted in Quebec. Some, using data from the Enquête sur l’établissement des nouveaux immigrants (Survey on the settlement of new immigrants, ENI), have examined different dimensions of economic integration, such as access to a first job and number of weeks of full-time work (Piché et al., 2002); socioeconomic status of jobs and long-term income (Renaud et al., 2003); and fields of employment before and immediately after arrival (Girard et al., 2008). These studies found an advantage for immigrants from Western Europe and the United States. Renaud et al. (2003), however, reported evidence of a long-term adjustment process that varies with region of origin. Evidence on the effect of gender, on the other hand, is mixed. Although many studies from Quebec have reported differences between the genders – always to the advantage of men – in ease of access to a first job (Bélanger et al., 2010; Lacroix, 2013), number of weeks of full-time work (Piché et al., 2002), and similarity between fields of pre- and post-migration employment (Girard et al., 2008), other studies have failed to find such differences (Bégin, 2010; Martin, 2007; Renaud and Cayn, 2006; Slaoui, 2008). In these studies, gender is treated only as a control variable, bringing out only its net or “main” effect, thus masking interactions between gender and national origin.
13Finally, we have found only a single study on the career trajectories of immigrant women in Quebec (Fihri et al., 2004). It examines access to and continuation in employment, but does not look at potential differences linked to the nature of the job. According to this study, there are similarities between the genders, principally in employment sequences, but also a number of differences, notably in ease of access to employment (to women’s advantage) and in the proportion of individuals entering the labour market (to men’s advantage). In the case of France, Houseaux and Tavan (2005) highlight the finding that it is not so much employment status (stable employment, unemployment, or instability) that distinguishes immigrant women from immigrant men, but whether or not they participate in the labour market at all. This dynamic is not applicable to our population of workers selected to enter the labour market.
II. Data and methods
Data on selected workers
14The present study is based on results from the Enquête sur les travailleurs sélectionnés (Survey of selected workers, or ETS) carried out in 2002 [3] in Quebec by the Ministère des Relations avec les citoyens et de l’immigration. These retrospective data can be used to study the labour market integration of principal applicants from the category of selected workers. The principal applicant is the person formally evaluated in an immigration application (Godin et al., 2004). Only one individual in a couple is assessed, often the one with the best chances of acceptance (the highest points total on the selection grid). [4] This individual is considered as the principal applicant, and the accompanying person as their spouse.
15From the initial sample of 1,875 individuals representative of arrivals in Quebec between January 1997 and June 2000, we removed 296 spouses of principal applicants as they were only partially assessed using the selection grid. Likewise, 38 workers (including 37 women) admitted through the Live-in Caregiver Program, were also excluded from our analyses as they were accepted to work in a particular employment sector. Finally, 153 individuals with missing values for certain variables of interest were removed from the sample for that reason. Our analyses were thus performed on a sample of 1,388 individuals, including 1,014 men and 374 women. Their median length of residence in Quebec at the time of the interview was nearly 3 years, with a minimum of 21 months and a maximum of 63 months.
16Three sources were combined to form the database: administrative files; information from the selection grid; and the survey data. The first dataset is derived from federal entry permits and the Quebec selection certificate (Certificat de sélection du Québec, or CSQ). These provide information notably on gender, age, selection programme, country of birth, [5] level of education, and knowledge of French and English (Renaud and Cayn, 2006). The second component contains information on the points obtained for each of the criteria in the selection grid. Finally, the answers to the retrospective questionnaire contain a week-by-week record of training undertaken in Quebec and career trajectory, and of the characteristics of jobs occupied before and after migration.
17The analyses presented here include both fixed and time-varying variables. The fixed variables mainly consist of the migrants’ individual characteristics on arrival, as well as the points allocated for each of the criteria in the selection grid. They include age, education, work experience, admissions programme, [6] whether or not the individual has a qualification preferred by the government, [7] family unit, [8] and knowledge of French and English. Language skills are validated during an interview with a government agent, and are expressed in the form of percentage points. Activities during the first years in Quebec (additional training, language courses) are also included in the analyses, given their potential influence on the ease of entry into employment and on the quality of the job obtained.
18Births and union formation may also affect the career trajectories of women in particular, notably access to employment and working time. Unfortunately, the information on the family unit only describes their situation at the time of migration. However, among principal applicants in a selection programme aimed at labour market entry, access to a first job occurs within the first weeks after arrival (Figure 1, p. 429). In most cases, then, any changes in family composition will have occurred after this event. Thus, while these variables correspond to a fixed moment in time, they reliably reflect the family circumstances influencing access to employment.
Labour market integration
19Here, we measure labour market integration through individuals’ rate of access to employment. First, we look at determinants of access to a first job of any kind. Second, as immigrants’ access to skilled jobs is both a priority for the government and a motivation for these migrants, we then analyse access to the first job corresponding to the individual’s level of education. We interpret this as an indicator of over-education, as measured by the difference between the individual’s pre-migration level of education and the educational requirements of their job in Quebec. [9] This double measure enables us to dissociate immigrants who succeed in maximizing their human capital from those who remain unemployed or who have to accept jobs requiring qualifications below the level of education they had when they migrated. A code from the National Occupational Classification (HRDC, 2001) was assigned to each of the jobs reported by the respondents (before and after migration). This code is used to describe the occupations of Canadians, and refers to the skills (level of education and training) normally required to perform a particular job. It is divided into four levels of education: university education; college education or apprenticeship training; secondary school and/or occupation-specific training; and on-the-job training. Note that no specific code is assigned to management jobs (upper and middle management), as these positions are not necessarily distributed based on the individual’s educational qualifications (Bégin, 2010). Here, these occupations were treated in the same way as those requiring an undergraduate university degree. Finally, the university education level includes all levels (bachelor’s, master’s, and PhD). Given this grouping, over-education is likely underestimated for individuals with a university education.
20To describe access to employment, we use Kaplan-Meier curves representing the proportion of individuals who have not experienced the event over time. The observation period begins at the time of arrival in Quebec and continues up to the date of first employment or the survey date (right-censored). We then present semi-parametric Cox models in order to measure the influence of the covariates on the occurrence of these events and the role of multivariate relationships (Blossfeld et al., 2009). These are based on the proportional hazards condition, and thus assume that the effect of each category of a categorical variable (with respect to the reference category) is constant over time. Where this condition is not satisfied, a simple interaction with time is added.
21Interaction terms between the categories of each variable bring out additional information that cannot be observed through the effects of gender and origin taken separately. This “intercategorical” approach (McCall, 2005) makes it possible to more directly disentangle multiplicative from additive effects; if there was no interaction effect between gender and national origin, then the situation of immigrants could be described as an additive function of the separate disadvantages of gender and origin (Greenman and Xie, 2008). The opposite finding would show that gender has a differential impact on national minorities’ access to employment, supporting the hypothesis of intersectionality.
III. Results
22Although one might expect Quebec’s selection grid to standardize the characteristics of the admitted population, Table 1 shows that these characteristics are unequally distributed among men and women at arrival. The chi-square tests and Student’s t-tests show significant gender differences.
Table 1. Distribution of individual characteristics at arrival and of training activities during the first four years of residence, by gender
Table 1. Distribution of individual characteristics at arrival and of training activities during the first four years of residence, by gender
23First of all, the proportion of individuals who migrated from Western Europe and the United States is higher among women than men (46% versus 36%). The second-largest group is from Africa (principally the Maghreb), with a larger proportion of men (33%) than women (21%). A low proportion of women have a qualification preferred by the government of Quebec (18%); the proportion among men is twice as high (41%). Additionally, while a majority of these skilled workers were admitted through the employability and occupational mobility programme, the proportion is higher for women (71% versus 58%), while their male counterparts are more strongly represented in Quebec’s occupations in demand programme (32% versus 14%). In most cases, the workers migrated with neither children nor spouse, particularly women (79%); men were more likely to be accompanied (42% had a spouse and 30% at least one child on arrival). Another criterion differentiating men and women is whether the individual has previously spent time in Quebec. This is the case of a majority of women (54%), versus 45% of men. Finally, the men are slightly older (32.3 years) than women (30.8 years) on average. Other characteristics, such as level of education, language skills, or having taken a course in Quebec, do not seem to significantly differ between men and women.
24Certain occupations attract or recruit more men than women. Table 2 gives the distribution of immigrants’ first jobs in Quebec and first jobs corresponding to educational level in Quebec across economic sectors. While men are over-represented in the natural and applied sciences, women are more likely to work in business, finance, and administration, as well as in the social sciences, public administration, education, and health. The qualifications (educational qualifications, language skills) sought by employers differ between fields, and may influence the effects of individual characteristics on men’s and women’s access to employment, depending on the sectors they are recruited into.
Table 2. Occupational field of first job and of first job corresponding to pre-migration level of education, by gender
Table 2. Occupational field of first job and of first job corresponding to pre-migration level of education, by gender
Speed of access to employment
25In keeping with the results of previous studies on skilled workers in Quebec (Renaud and Cayn, 2006), there does not seem to be a significant difference between men’s and women’s speed of access to a first job.
26According to the log-rank test, there is no difference between the two Kaplan-Meier curves illustrating the proportion of individuals who have not yet experienced the event of interest (chi²(1) = 0.07; p = 0.790). The median time spent in Quebec before obtaining a first job was 14 weeks for men and 15 weeks for women. At the end of the four-year observation period, around 10% of men and women still did not have a job. (Figure 1)
Figure 1. Proportion (%) of immigrants without a first job, or a first job matching their pre-migration skill level, by time since arrival and gender
Figure 1. Proportion (%) of immigrants without a first job, or a first job matching their pre-migration skill level, by time since arrival and gender
27However, in an analysis of the time taken to find a job corresponding to the pre-migration level of education – through the comparison of job characteristics and the National Occupational Classification – differences between men and women become visible. As could be expected, the time taken to obtain a job matching one’s skill level is considerably longer than the time taken to obtain any job, and the waiting time is still longer for women. Here the median time increases to 43 weeks for men and 64 weeks for women. A log-rank test showed that these curves differ (chi²(1) = 5.24; p = 0.022), with the proportion of men having experienced the event of interest being higher throughout the observation period. After four years of residence in Quebec, 30% of the men and 36% of the women had still not obtained a first job corresponding to their qualifications.
28Region of origin is often identified as a determinant of labour market integration. Men and women from Western Europe and the United States show an advantage (Figure 2). East Asia and Oceania is the only group of origin in which more than half of men do not find a first job matching their qualifications in the four years following their arrival. For women on the other hand, more than half of the immigrants from Eastern Europe and the former Soviet Union, West Asia and the Middle East, and East Asia and Oceania find themselves in this situation. Overall, a certain convergence between regional groups can be seen in the speed of access to qualified work for men, whereas groups of women seem to be more heterogeneous in this respect.
Figure 2. Proportion (%) of immigrants without a first job matching their pre-migration skill level, by region of origin and gender
Figure 2. Proportion (%) of immigrants without a first job matching their pre-migration skill level, by region of origin and gender
Determinants of access to a first job in Quebec
29To assess the role of the determinants of access to a first job in Quebec, we first use hierarchical models, specifically semi-parametric Cox models, by gender (Table 3A). We then use a model including both genders as well as an interaction term between gender and each individual characteristic (Table 3B). This last model allows us to determine whether the influence of these factors differs statistically between men and women.
Table 3. Relative risks of access to a first job, by gender
Table 3. Relative risks of access to a first job, by gender
30Looking first at the isolated effect of origin, it can be seen that all regional groups (with the exception of women from Latin America) took longer to find their first job than the group from Western Europe and the United States (Table 3A, Model 1). These results are largely consistent with those of Piché et al. (2002). At this step, the interaction term between gender and region of origin is not significant (Appendix Table A.1). Each regional group’s disadvantage relative to the group from Western Europe and the United States is statistically equivalent for men and women. The effect of gender alone is significant, however, indicating that women from Western Europe and the United States find a job less quickly than men from the same region (relative risk, RR = 0.82; p < 0.05).
31When other individual characteristics and training in Quebec are added (Table 3A, Model 2), the effect of origin considerably decreases for women. Only women from Africa remain disadvantaged, with chances of obtaining a first job one third lower than those of the reference category (RR = 0.68; p < 0.05). For men, on the other hand, very few differences emerge when the other variables are introduced. Several regional groups continue to show a substantial disadvantage, with a likelihood divided by two or even three with respect to the reference group (RR between 0.39 and 0.59; p < 0.001).
32Adding interaction terms between gender and each of the individual characteristics (Table 3B) allows us to formally test the hypothesis of intersectionality. The results of the model with interactions reproduce those of the separate models by gender, with an additional significant gender difference in certain cases: the differences between men and women are smaller for the groups from Eastern Europe and the former Soviet Union (p < 0.05) and West Asia and the Middle East (p < 0.01), relative to migrants from Western Europe or the United States. Marginally significant differences (at the 10% level) are also found for the groups from Latin America and Africa. In summary, women’s likelihood of entering employment is roughly equal to men’s, but the difference due to region of origin is smaller for these four groups of women than for men. It would be an error, however, to interpret the substantial negative effect of gender in this model (RR = 0.18; p < 0.01) as a general disadvantage for women, as it only applies in interaction with all other individual characteristics. The model takes other interactions into account, notably the interaction between gender and knowledge of French (RR = 3.11; p < 0.05). The survival curves also reflect these differences by sex (Figure 3).
Figure 3. Proportion (%) of persons without a first job, by time since arrival, gender and level of knowledge of French at arrival
Figure 3. Proportion (%) of persons without a first job, by time since arrival, gender and level of knowledge of French at arrival
33For women who arrive with limited knowledge of the French language, access to employment is much slower, whereas men are unaffected. It is thus the intersection of gender and knowledge of French, rather than the effect of region of origin, which marks the differences between women and men in access to a first job. This may be partly explained by differences in the distribution of sectors of activity where men and women work (Table 2). Women’s first job is mainly in sectors where language skills are more important (social sciences, public administration, teaching, business management). Men, on the other hand, are present mainly in the natural and applied sciences, where knowledge of French is not always indispensable.
34Another objective of these analyses is to determine whether human capital, as measured by the selection grid, has differential effects on the employability of immigrant men and women. As expected, having previously spent time in Quebec has a positive effect for men and women. But the effect attributable to a number of other factors in the grid differs by gender. Men admitted in the category of assured and certified jobs and who took an English course appear to have found a job more easily, whereas this was not the case for women (Table 3A). Surprisingly, and contrary to men, women admitted through the “occupations in demand” category were less likely to find a job than those admitted through the “employability and occupational mobility” programme. The latter category of admissions applies to individuals with a high potential for adaptability, defined overall, without aiming to direct them into a particular field of employment.
Determinants of access to employment that matches the pre-migration level of education
35This section examines the correspondence between individuals’ pre-migration level of education and the formal demands of the jobs they obtain. As in the case of access to a first job, the effects of region of origin differ in size between women and men when no other variable is included in the model (Table 4A). Women from all regions other than Western Europe and the United States take longer than this reference group to find a first job that matches their qualifications. However, contrary to the findings on entry into a first job (qualified or not), adding individual characteristics makes virtually no difference: for women of all origins, with the exception of those from West Asia and the Middle East, there is still a significant gap with respect to the reference group. Likewise, among men, each origin group takes longer to enter qualified employment than the reference group. The size of the effects varies little between models, with the exception of men from East Asia and Oceania, whose disadvantage with respect to the reference decreases after individual characteristics are included.
Table 4. Relative risk of finding a first job corresponding to the pre-migration level of education in Quebec, by gender
Table 4. Relative risk of finding a first job corresponding to the pre-migration level of education in Quebec, by gender
36The results obtained after adding interaction terms between gender and national origin (table 4B) do not show differences between the genders in access to employment that matches educational level. [10] There is thus inter-regional variability in men’s and women’s access to skilled employment, but the differences for particular regional groups do not appear to depend on gender. [11] Similarly, the effects of the characteristics in the selection grid and those acquired after arrival in Quebec do not statistically differ by gender (Table 4B). However, a limited number of individuals obtained a first job that matched their skills (as opposed to a first job of any kind), and certain categories are represented by few individuals (for example, few respondents both took a French course and obtained a first job that matched their skills). This clearly represents a limitation for a model involving multiple interaction terms.
37In the separate models, the sizes of the effects of characteristics from the selection grid and of training activities vary by gender. But the effects generally run in the same direction: that which is advantageous for men is also advantageous for women. Generally speaking, the less educated the individual, the better their chances of obtaining a job corresponding to their level of education. These results are intuitive insofar as the possibilities for over-education are more limited at the bottom of the educational scale (Helgertz, 2010; Renaud and Cayn, 2006). Having trained in an area preferred by the government, as well as having previously stayed in Quebec to work or study, have a significant positive effect on the likelihood of finding a first job that matches one’s skill level, for both men and women. These are qualities often associated with the development of a network and the acquisition of experience in Quebec, and the recognition of experience, skills, and training.
38Here, contrary to the results for women’s access to a first job (matching their qualifications or not), it is not language skills on arrival that are the important factor, but training activities undertaken in Quebec. For men, taking an English course multiplies the chances of obtaining a job that matches their skills by a factor of more than three (RR = 3.19; p < 001). This effect decreases with each week spent in Quebec, as shown by the interaction with time (RR = 0.985; p < 0.01). For women, on the other hand, access to a first job that matches their skills is positively affected by taking French language courses or other academic (technical or university) courses. In addition to this effect, there is a marginally significant interaction between sex and taking a French course (at the 10% level: RR = 1.95; p < 0.1) (Model B), highlighting the particular importance of French language skills for women, as in the previous analysis. The influence here of these courses taken in Quebec, as opposed to previous knowledge of French and English, can be explained by the additional time needed to enter a first job corresponding to one’s skill level. The first weeks after arrival may be occupied with training courses to fill possible knowledge gaps and to adapt one’s human capital to the demands of the Quebec labour market.
IV. Discussion and conclusion
39This study demonstrates that the interrelationships between gender, national origin, and individual characteristics differ between access to a first job of any kind and access to the first job corresponding to the pre-migration level of education. Note, however, that our data are representative of arrivals to Quebec between 1997 and 2000. Changes made to the selection grid since 1996, notably the withdrawal of the employability and occupational mobility programme, may have influenced these dynamics. For women’s access to a first job, the results indicate that differences by origin are mainly due to individual characteristics on arrival. Some variables that have a substantial effect on labour market integration, such as knowledge of French, are strongly correlated to national origin. Once these characteristics have been taken into account in analyses, the speed of access to employment of different regional groups is very similar, with the exception of female immigrants from Africa (principally the Maghreb), who are disadvantaged. For men, however, the advantage of men from Western Europe and the United States persists even after individual characteristics have been taken into account. This result is consistent with a survey of Quebec employers which found a hiring preference for North American and Western European immigrants, with those from Eastern Europe and the Middle East ranking last in the list of preferences (Antonius and Tadlaoui, 2003). Another study found employers’ fear of religious and cultural differences to be the second-largest barrier to migrant integration, after language skills (FCEI, 2007). The present results indicate that along with visible minority status, often cited as the principal factor in discrimination, region of origin is a marker of differences in labour market integration.
40Our results contrast with those of studies which generally focus on migrant women within the family unit, often admitted as dependents, who are found to be disadvantaged. Note, however, that our population is a highly selected one, composed of individuals who chose to (and succeeded in) immigrating to Quebec as principal applicants in the category of skilled workers. The homogeneous characteristics and results found for women from different regions of origin can be explained by two selection processes: success in overcoming the administrative hurdles set in place by the government of Quebec, and differential self-selection by region of origin upstream of migration. All the female immigrants in the population studied here were admitted as principal applicants, and the majority arrived without a spouse or children. The “cultural barriers” to employment that are linked to gender roles in the family sphere must therefore have operated upstream of migration, when deciding to apply as a principal applicant, rather than after arrival in Quebec, contrary to the situations examined in studies on married female immigrants (Boudarbat and Gontero, 2008; Reimers, 1985).
41Our analysis of immigrants’ access to a first job corresponding to their pre-migration level of education yielded somewhat different results. Migrants’ speed of access to employment is still strongly affected by their region of origin even after controlling for their individual characteristics. However, no effect was found either for the interaction between gender and region of origin, or for the effect of gender alone. The disadvantage in access to a first skilled job linked to region of origin thus seems to affect men and women equally.
42The present study has certain limitations. For example, the interactions tested between gender and region of origin may be non-significant due to the small size of the sample of women, a finding which would undermine our conclusions on intersectionality. But although the small sample sizes make it difficult to interpret the interaction terms, sensitivity testing using larger groups of regions of origin did not change the results. In addition, the same sample size yields significant effects in the analysis of access to a first job, suggesting that the statistical power is sufficient for the purposes of the study. Other mechanisms may explain the absence of significant results on intersectionality. In this respect, the selection that applies to principal applicants, especially women, and particularly those from non-Western regions, may be an important factor. The interaction coefficients would doubtless have been very different if the analysis had covered all immigrant women, regardless of category of admission. Our results must thus be understood within the limits of our sample, that was small in size, but more importantly, selected.
43All in all, immigrant women from non-Western regions remain doubly disadvantaged in access to employment in Quebec, due both to their origin and to their gender, although these effects are not multiplicative. Men, on the other hand, seem to be affected mainly by their region of origin, independently of the type of job. In other words, our results suggest that employers evaluate the qualifications of Western men most favourably, at the expense of all other groups. The effect of language skills seems to be important, but is divergent between women and men. In access to a first job, knowledge of French and English is very important for women, whereas it has little effect for men. For employment that matches the level of education, however, it is language courses that matter: while women benefit from taking French courses, men seem instead to draw an advantage from English courses. These results shed new light on the linguistic dynamics applying to men and women in the context of Quebec. Several previous studies have found modest effects of knowledge of French and English on immigrants’ access to employment. However, separating these effects by gender reinforces the relationship. Targeted analyses using a larger database could inform discussions around the mode of selection of immigrants and efforts to encourage French speaking, which are a focus of constant political debate.
Table A.1. Relative risks of access to a first job and a first job corresponding to pre-migration level of education
Table A.1. Relative risks of access to a first job and a first job corresponding to pre-migration level of education
Notes
-
[*]
University of Geneva.
Correspondence: Julie Lacroix, Université de Genève, Institut de démographie et de socioéconomie, Geneva, Switzerland, email: Julie.Lacroix@unige.ch -
[**]
Université de Montréal
-
[1]
The skilled workers programme is the largest component of economic immigration.
-
[2]
Quebec’s selection grid, established in 1996, differs from its Canadian counterpart mainly in the number of points allocated to knowledge of French. The Quebecois grid also targets employability based on general skills rather than a direct match with hiring needs.
-
[3]
A survey similar to the 2002 ETS was conducted in 2011 to evaluate the changes made to the selection grid in 2006. The sample of this second survey was considerably smaller (135 women), so could not be used in a multivariate analysis.
-
[4]
To gain admission, individuals must obtain a score of at least 65 points for a single migrant and 70 points when accompanied by a spouse. The evaluated factors are broken down as follows: education (23 points), employment (15 points), work experience (10 points), age (10 points), language skills (23 points), spouse’s characteristics (17 points), age of children (8 points), financial autonomy (1 point) and adaptability (31 points).
-
[5]
The grouping by region of origin is detailed in Appendix 1 of the report of Godin et al. (2004).
-
[6]
“Admissions programme” refers to the category of employment from the selection grid: 1) employability and occupational mobility programme; 2) occupations in demand in Quebec; and 3) assured and certified jobs. This last category also includes individuals admitted by exemption.
-
[7]
This is a list of educational qualifications, created by the government of Quebec, which offer good employment prospects.
-
[8]
That is, having one or more children and a spouse on arrival in Quebec.
-
[9]
The measure of the match between level of education and job held is described in detail in the dissertation of Bégin (2010, pp. 111-114).
-
[10]
As in the previous analysis, the main effect of gender is significant when it is used in a model including an interaction term with region of origin (Appendix Table A.1), whereas this main effect is not significant in Table 4, which includes all individual characteristics.
-
[11]
In order to increase statistical power, we grouped together all regions other than the one for Western Europe and the United States. The interaction effect between these two groups remains non-significant (RR = 0.999; p > 0.1).
