Immigrant earnings performance in Canadian cities: 1981 through 2001.
Wages (Demographic aspects)
Warman, Casey R.
|Publication:||Name: Canadian Journal of Urban Research Publisher: Institute of Urban Studies Audience: Academic Format: Magazine/Journal Subject: Social sciences Copyright: COPYRIGHT 2004 Institute of Urban Studies ISSN: 1188-3774|
|Issue:||Date: Summer, 2004 Source Volume: 13 Source Issue: 1|
|Topic:||Event Code: 530 Labor force information; 250 Financial management Computer Subject: Company financing; Salary|
|Product:||Product Code: E198450 Immigrants|
|Geographic:||Geographic Scope: Canada Geographic Code: 1CANA Canada|
A comparison of immigrant and Canadian-born resident earnings in eight major Canadian urban centres, or Census Metropolitan Areas (CMAs), using the 1981 to 2001 Censuses is presented. Unlike most other studies, this one presents only averages for large groups and does not control for observable characteristics such as education, and it takes a largely graphical approach. Akin to the results found in earlier studies at the national level, the relative earnings of immigrant men and women are generally observed to decline for more recent arrival cohorts from 1966 to 1995 in each of the cities studied. Moreover, the relative performance of immigrant and Canadian-born workers living in CMAs shows a lower level of immigrant economic integration than the comparison between all immigrants and all Canadian-born persons. However, there is a small reversal for the 1996-2000 arrival cohort of men. Their earnings are generally higher than the earnings of men in the earlier 1991-95 arrival cohort in the first five years after arrival in Canada. This same improvement is much smaller for immigrant women.
Keywords: Immigrants, Earnings, Cities, Canada
Cet article puisi dans les recensements effectues entre 1981 et 2001 pour presenter une comparaison entre les salaires d'immigrants et ceux de Canadiens nes dans le pays vivant dans huit grandes agglomerations urbaines au Canada, ou Regions metropolitaines de Recensement (RMR). Cette etude se distingue de la plupart des autres en ne presentant que les moyennes pour les grands groupes, en ne tenant pas compte de l'effet de caracteristiques observables telles la scolarite et en adoptant une approche basee surtout sur les graphiques. Nos resultats refletent ceux de recherches anterieures effectuees sur le plan national: de facon generale, les salaires relatifs des immigrants et des immigrantes baissent chez les cohortes nouvellement arrivees entre 1966 et 1995 et ce, dans chacune des villes etudiees. De plus, dans les RMR, la performance relative des travailleurs immigrants et ceux nes dans le pays indique un niveau d'integration economique inferieur a celui qui ressort d'une comparaison entre tous les immigrants et tousles travailleurs nes au Canada. Toutefois, il y a un leger retournement de la situation chez la cohorte d'hommes ayant immigre entre 1996 et 2000 et dont le salaire, pendant les cinq premieres annees au Canada, est generalement plus eleve que celui des hommes etant arrives entre 1991 et 1995. Chez les immigrantes, cette augmentation est beaucoup moins importante.
Mots cles: immigrants, salaire et renumeration, villes, Canada
There is evidence of a decline in the labour market outcomes of recent immigrants in Canada. Findings at the national level indicate that more recent cohorts have had entry earnings lower than earlier landing cohorts and earnings integration has been modest at best. However, these studies have compared immigrants and the Canadian born at the aggregate level, ignoring potential differences across regions and cities. This study diverges from earlier ones by recognizing that immigrants have not settled randomly across the country. If immigrants tend to reside in regions that are economically different than the national average, a national-level analysis may obscure differences in earnings. (1) Our analysis focuses on large cities where most immigrants have settled.
Studies in Canada analyzing immigrant earnings have discovered that recent immigrant cohorts have not fared as well as earlier immigrant cohorts. Baker and Benjamin (1994) find lower entry earnings and enduring earning differences across male immigrant cohorts. Bloom, Grenier and Gunderson (1995) also find lower entry earnings and deteriorating rates of economic integration for recent male immigrant cohorts. They also observe that entry conditions for women have been deteriorating and that the rate of economic integration has been very slow. (See also Beach and Worswick, 1993.) As well, recent evidence also highlights the diminishing returns to foreign experience. (2)
Studies in the U.S. have also indicated declining outcomes for recent immigrants. Borjas (1995) finds that the entry wage of the 1970 and 1980 cohorts were lower than that of earlier cohorts and that the wage disadvantage between recent immigrant cohorts and the native population have persisted. While most of the U.S. literature has been focused at the national level, some studies have examined immigrant earnings at the city level. Ellis (2001) shows that the "spatial scale of analysis" influences the outcomes of landing cohorts. Examining the wage differentials across five major cities using the 1980 and 1990 U.S. Census, Ellis finds that there are substantial differences in the wage differential between immigrants and American-born persons at both the national and city levels. (3) His findings indicate that the wage differential is larger at the city level than at the national level. (4)
Immigrants to Canada have not dispersed randomly across the country but have instead concentrated in large metropolitan areas, with around 70 percent of immigrants residing in Toronto, Montreal and Vancouver. Comparing earnings between immigrants and the Canadian born at the national level may lead to incorrect conclusions about relative economic performance. (5) If immigrants have tended to settle disproportionately in areas that have not performed as well economically as the national average, it may appear that immigrants are doing worse than they actually are. Conversely, if immigrants have tended to settle in more prosperous regions, then national comparisons may understate the declining earnings of immigrants and overstate their economic integration.
There are several reasons why cities may be a more appropriate "spatial scale of analysis". First, urban centres tend to have above average earnings. Therefore, comparing immigrants with all Canadian-born persons means comparing immigrants with a comparison group that has lower average earnings than the comparison group of Canadian-born persons living in Census Metropolitan Areas (CMAs). Secondly, the cost of living varies greatly across the country. Within cities, there will be less variance in the cost of living and, therefore, comparisons at the city level may be more appropriate. Thirdly, macroeconomic cycles vary by region. While the country may be in a recession, a particular region or city may not be experiencing the same conditions. If areas where immigrants are concentrated experience different macro conditions than the rest of the country, conclusions regarding their earning differentials and economic integration will be misleading.
This article proceeds in the following manner. In Section 2, the data is described. In Section 3, we first examine the earnings of immigrants and the Canadian born at the national level and then look at the aggregate earnings differential between these two groups in urban areas. (6) We next compare earnings differentials at the city level. The CMAs with the highest portion of immigrants are examined. These prime immigrant destinations are Toronto, Vancouver, Montreal, Calgary, Ottawa, Edmonton, Hamilton, and Winnipeg. The final Section contains the conclusion.
Data and Methodology
The 1981, 1986, 1991, 1996 and 2001 one in five Canadian Census micro data files are used. The sample used in this study includes all Canadian-born and immigrant workers between the ages of 25 and 59. (7) Individuals younger than 25 are excluded since their labour market participation will vary more because of school/employment decisions. Individuals older than 59 are also excluded because of the higher variance in their participation rates relating to retirement/employment decisions. Both men and women are included in the study but they are examined separately. The landing cohorts examined are the 1966-70, 1971-75, 1976-1980, 1981-85, 1986-90, 1991-95 and 1996-2000 cohorts. Immigrants who arrived in Canada in the Census reference year are not included in that year, since the Census reports earnings for the year prior to the Census enumeration. Further, immigrants who landed the year prior to the Census enumeration year are also excluded from the calculations for the mean earnings of the cohort in the year of landing. Since an immigrant landing in the reference year may have landed partway through the year, there is a disparity in the opportunity for labour market activity. This potentially biases conclusions on economic integration. (8) The earnings are transformed by using the Canadian Consumer Price Index with 2000 as the base year. Individuals for whom the majority of earnings come from wages and salaries are included in the analysis. (9) Unlike almost all previous studies that present regression adjusted economic integration profiles, the profiles presented here are based on simple means for the relevant sex-immigrant status group. Thus, we do not control for such differences as age or education.
Table 1 contains the mean earnings of immigrants and the Canadian born by sex, CMA status and Census year. Earnings of Canadian-born men and women who reside in CMAs are higher and these differences grow over the 20-year period. Immigrant women living in a CMA have higher earnings than immigrant women living outside CMAs in all Census years. Interestingly, this pattern is reversed for immigrant men. Those residing in non-CMA areas have higher average earnings than immigrant men living in CMAs in all of the Census years. However, the magnitude of these differences is significantly smaller than the respective differences in earnings between immigrant women and men and Canadian-born women and Canadian-born men.
With immigrant men outperforming the Canadian born in non-CMAs, it may seem inconsistent to argue that including the earnings of non-CMA residents bias the conclusions about male earnings differentials. However, it is important to recognize that only a small percent of immigrants live in non-CMAs, relative to the Canadian born. Looking at Table 2, between 39 and 45 percent of Canadian-born men and women live in non-CMAs, while only between 9 and 14 percent of immigrants reside in non-CMAs. Therefore, comparing earnings at the national level underestimates the earning differentials between immigrant and Canadian-born men. This is because the larger proportion of Canadian-born men living in non-CMAs will dilute the overall mean earnings of Canadian-born men.
Earnings of the Canadian born By City
The two panels of Figure 1 contain the mean log earnings of Canadian-born men and women in the eight major immigrant-receiving cities for the Census years, 1980, 1985, 1990, 1995 and 2000. This first look at the data allows us to see the differences across these urban labour markets, in terms of the performance of the Canadian born as well as the role of changing macroeconomic conditions over time. Men's earnings were generally highest in Toronto, Ottawa, and Hamilton and lowest in Winnipeg and Montreal. The ordering for women is similar, although women in Hamilton have relatively lower earnings and women in Ottawa have a particularly large advantage, when compared with the earnings distribution for men across these cities. For men, the business cycle effects on earnings are pronounced. Mean earnings dip between the 1980 and 1985 Census years as well as 1990 and 1995. This likely results from the impacts of severe recessions, which began after the Census earning year in both the early 1980s and the early 1990s. For women, the general pattern is one of growth in earnings over the 20-year period, with positive trends relative to macroeconomic fluctuations. Given that annual earnings rise with both the hours of work in a typical week as well as the number of weeks worked in the year, this upward trend likely stems from the increasing attachment of women to the labour market across time.
[FIGURE 1 OMITTED]
The relationships found in Figure 1 inform our understanding of immigrant labour market performance. Given the large proportion of immigrants living in Toronto, Montreal and Vancouver, the trends and macroeconomic fluctuations experienced by workers in those cities are likely to be important factors in the labour market experiences of immigrants more generally. Consequently, the analysis of this paper delves deeply into these relationships. Immigrant performance, relative to the Canadian born, will be carried out at both the national and city levels. In the latter case, the Canadian born living in the city under study will be employed so city-specific factors, such as the overall wage level and the macroeconomic fluctuations, will be the same for both immigrants and the Canadian born.
Earnings of Immigrants
In the two panels of Figure 2, we present fairly standard differences in log earnings of immigrants relative to the Canadian born, by years since migration and by immigrant arrival cohort. (For example, c6670 represents immigrants who arrived between 1966 and 1970.) This is similar to previous studies by Baker and Benjamin (1994) and Bloom, Grenier and Gunderson (1995). However, as mentioned, the analysis here does not employ a regression framework to control for covariates, such as education, age and region of residence. The dark horizontal line at zero on the vertical axis relates to the situation where immigrants have the same mean log earnings as the Canadian born. Points above this line indicate that the mean log earnings are higher for immigrants; points below the line show that mean log earnings are lower for immigrants, relative to the Canadian born.
[FIGURE 2 OMITTED]
While immigrants who arrived in the late 1960s and early 1970s have mean log earnings that are comparable to, or above, those of the Canadian born, more recent immigrant arrival cohorts have not fared as well. Successive immigrant cohorts have had lower earnings than their predecessors, when the comparison is made at the same number of years since migration. This general result is well documented in the literature. (See, for example, Baker and Benjamin, 1994; Bloom, Grenier and Gunderson, 1995; Reitz, 2001; and Green and Worswick, 2003). The main exception is the 1996-2000 cohort. It is difficult to see this group of men in the diagram. Their mean log earnings are roughly equal to those of the 1981-85 and 1986-90 cohorts. They are, however, roughly 0.2 log points higher (or roughly 20 percent higher) than that of the 1991-95 cohort, when compared at one to five years in Canada over the period. This indicates that the pattern of declining earnings has been reversed for more recent cohorts or at least for the most recent cohort. In addition, the 1986-90 cohort had low earnings at arrival in Canada but they also experienced very high earnings growth with time in Canada, as indicated by the steepness of their profile for one-to-five and six-to-ten years since migration. The relationships are generally similar for women, with the exception of the 1991-95 cohort of women who fared relatively better than men.
Urban/Rural Residency and Immigrant Earnings
Immigrants are much more likely to reside in urban areas, especially the larger urban areas, than the Canadian born. Consequently, comparing all immigrants to all Canadian born without controlling for this difference in residence means that differences in earnings between immigrant and Canadian-born persons may partially reflect differences in earnings between urban and rural areas.
In order to explore this possibility, the next part of the analysis focuses on immigrants residing in CMAs. The two panels of Figure 3 contain comparisons for immigrants and the Canadian born residing in CMAs. The differences across immigrant arrival cohorts are very similar to those found for both groups across Canada in Figure 2. This is hardly surprising since most immigrants reside in CMAs; therefore, the sample exclusion does not have a significant impact on the immigrant sample. However, since relatively more Canadian-born persons than immigrants reside in rural areas, the main effect of the restriction to CMA residents is the change in the composition of the Canadian-born sample. Excluding individuals who do not reside in a CMA generally makes the immigrant performance look worse, relative to that of the Canadian born. This is true for both immigrant men and women.
[FIGURE 3 OMITTED]
In the two panels in Figure 4, differences in log earnings between immigrants and the Canadian born living outside CMAs are displayed. Immigrant men tend to fare better relative to the Canadian born in non-CMAs for most of the landing cohorts, while the outcomes for women in non-CMAs resemble more closely the outcomes in CMAs. As a higher proportion of the Canadian born reside in rural areas, where wages are lower, there is a dilution of earnings for the Canadian born. This could lead to the wrong conclusions about immigrant outcomes in national-level analysis and is made apparent in a comparison of Figures 2, 3 and 4. For example, immigrants could perform worse relative to the Canadian born in both CMAs and non-CMAs but still appear to perform relatively better at the national level. For example, the 1971-75 landing cohort has a negative earnings differential for women living in both CMAs (Figure 3) and non-CMAs (Figure 4) at six-to-ten years since migration. However, it appears that at the national level this cohort has a positive earnings differential at six-to-ten years since migration. Since rural earnings are lower and a much higher proportion of the Canadian born live in non-CMAs, the average earnings of the Canadian born is artificially pushed down, when immigrants are compared to the Canadian born at the national level.
[FIGURE 4 OMITTED]
Immigrant Earnings Across Cities
The eight panels of Figure 5 contain differences in log earnings between immigrant men and Canadian-born men for eight Canadian cities. These are Montreal, Ottawa, Toronto, Hamilton, Winnipeg, Calgary, Edmonton and Vancouver. The patterns are broadly similar across the cities, indicating that local labour market conditions do not have differential impacts on immigrants when compared with the Canadian born. As indicated in Figure 1, large differences in earnings exist across cities in Canada and there have been changes in earnings, which are city-specific, across time. However, the similarities across the city-specific panels of Figure 5 indicate that these city labour market effects seem to be similar for immigrants and Canadian-born men.
[FIGURE 5 OMITTED]
However, some differences in earnings performance across the different cities are worth highlighting. For example, the performance of the earlier cohorts was especially high in Ottawa, where many cohorts appear to have experienced rapid economic integration. For the most recent cohorts, the patterns across cities are more complicated. For the 1991-95 cohort, performance was relatively poor in some cities, especially Montreal, Winnipeg, and Edmonton. In general, the 1996-2000 cohort performed better in terms of earnings shortly after arrival than the 1991-95 cohort did. This difference or "bounce-back" is especially pronounced in Ottawa, Toronto, Winnipeg and Edmonton but it is apparent in all cities. Hamilton is the sole exception, where the performance of this cohort is comparable to that of the 1991-95 cohort.
The eight panels of Figure 6 contain equivalent log earnings differences for immigrant and Canadian-born women in the eight major cities. The patterns are similar to those found for men in the eight panels of Figure 5. The general pattern of a cross-cohort decline in earnings, evaluated at the same number of years since migration, is apparent for women in each of the eight cities. The strong bounce-back in the entry earnings of the 1996-2000 cohort of immigrant men found in most cities is not nearly as pronounced for women. One possibility is the increased emphasis placed on education for economic class immigrants. Since the education criteria only applies to the principal applicant and the principal applicant is often a man in a married household, it may be that the impact of the increased emphasis on education has primarily affected the composition of immigrant men.
[FIGURE 6 OMITTED]
The earnings of immigrant men and women have been analyzed using Census data over the period 1981 through 200l. The role of urban status and, in particular, city of residence has been investigated. Immigrants are much more likely to live in CMA than in non-CMA areas and are more likely live in CMAs than the Canadian born. In addition, when the earnings of immigrants living in CMAs are compared with the earnings of Canadian-born residents of CMAs, the performance of immigrants falls, relative to a national comparison. This indicates that the treatment of "urban effects" is important in earnings comparisons between immigrants and the Canadian born.
We also investigate differences in immigrant performance across eight major immigrant-receiving cities in Canada. The results indicate similar patterns of cross-cohort decline over the arrival period 1966-1995 in the earnings of immigrant men and women in each of the cities studied. This suggests that the experience of immigrants in the Canadian labour market does not differ greatly across cities, when the comparison group is the Canadian born living in the same city.
The 1996-2000 arrival cohort of men generally fared better than did the men in the 1991-95 arrival cohort in the first five years after arrival in Canada. This same improvement is much smaller for immigrant women. The size of the effect for immigrant men varies across Canadian cities with the effect being especially pronounced in Ottawa, Toronto, Winnipeg and Edmonton. It is apparent in all cities, except Hamilton, where the performance of this cohort is comparable to that of the 1991-95 cohort.
Overall, the results indicate that the CMA/non-CMA distinction is an important determinant of differences in the earnings of immigrants and the Canadian born. The patterns of earnings differences across arrival cohorts of immigrants in Canadian major cities are found to be similar. However, the differences underline the importance of investigating the role of local labour market conditions in studies of immigrant labour market performance.
This work was carried out in the Family and Labour Studies Division of Statistics Canada. We would like to thank Miles Corak, Ted McDonald and Garnett Picot for their comments. Financial support was provided by a SSHRC Standard Research Grant 410-2003-0409. All errors are attributable to the authors.
(1) Ellis (2001) notes that multivariate analysis with control variables for region or city will not cure the problem if wage structures vary between regions.
(2) Aside from finding a further drop in the entry earnings of immigrants in the 1990s relative to the 1980s, Green and Worswick (2003) find a fall in returns to foreign experience. See also Schaafsma and Sweetman (2001) and Hum and Simpson (1999).
(3) The cities are New York, Chicago, San Francisco, Los Angeles and Miami.
(4) Using regression analysis to determine the causes of the decline in relative wages between immigrants and the American born, Ellis (2001) finds that analysis at the national level indicates that immigrant characteristics are the major cause whereas city level analysis highlights a change in wage structures.
(5) While most Canadian studies on discrimination have focused on national-level analysis, Pendakur and Pendakur (2002) did examine the earnings gaps of Canadian-born ethnic minorities at the city level.
(6) For the urban examination, all Census Metropolitan Areas, based on the 1981 definition, are included.
(7) All permanent residents who are not immigrants are included in the Canadian-born group.
(8) Immigrants who landed in the Census year appear in the next Census calculations of the cohort mean log earnings. Immigrants who landed in the reference year are also included in future calculations, since there should be no correlation between the time of the year in which an individual immigrates and future earnings.
(9) If the absolute value of self-employment earnings made up more than 20 percent of earnings (wages and salaries + the absolute value of self-employment earnings), then the individual is not included in the analysis.
Baker, M. and D. Benjamin. 1994. The performance of immigrants in the Canadian labour Market. Journal of Labor Economics 12: 369-405.
Beach, C. and C. Worswick. 1993. Is there a Double-Negative Effect on the Earnings of Immigrant Women? Canadian Public Policy 19(1): 36-53.
Bloom, D.E., G. Grenier and M. Gunderson. 1995. The changing labour market position of Canadian immigrants. Canadian Journal of Economics 28:987-1005.
Borjas, G. J. 1995. Assimilation and Changes in Cohort Quality Revisited: What Happened to Immigrant Earnings in the 1980s? Journal of Labor Economics 13 (2): 201-239.
Ellis, M. 2001. A Tale of Five Cities? Trends in Immigrant and Native-Born Wages. In Strangers at the Gates: New Immigrants in Urban America, ed. Roger Waldinger. Berkeley: University of California Press.
Green, D.A. and C. Worswick. 2003. Immigrant Earnings Profiles in the Presence of Human Capital Investment: Measuring Cohort and Macro Effects. Society of Labor Economists Conference, Toronto, Canada, September.
Hum, D. and W. Simpson. 1999. Wage Opportunities for Visible Minorities in Canada. Canadian Public Policy 25 (3): 379-394.
Pendakur, K. and R. Pendakur. 2002. Colour My World: Have Earnings Gaps for Canadian-Born Ethnic Minorities Changed Over Time? Canadian Public Policy 28 (4): 489-511.
Reitz, J.G. 2001. Immigrant Success in the Knowledge Economy: Institutional Change and the Immigrant Experience in Canada, 1970-1995. Journal of Social Issues 57(3): 579-613.
Schaafsma, J. and A. Sweetman 2001. Immigrant earnings: age at immigration matters. Canadian Journal of Economics 34 (4): 1066-1099.
TABLE 1. EARNINGS BY CMA/NON-CMA LOCATION Year Sex Status Location 1980 1985 1990 CMA 24,333 25,134 27,613 Canadian Born Non-CMA 19,281 19,310 20,783 Women CMA 21,851 22,470 25,179 Immigrant Non-CMA 19,372 19,172 21,112 CMA 46,805 44,620 45.133 Canadian Born Non-CMA 39,393 36,525 37,191 Men CMA 41,449 39,772 40,425 Immigrant Non-CMA 45,325 42,434 42,292 Year Sex Status Location 1995 2000 CMA 28,792 32,246 Canadian Born Non-CMA 21,846 24,042 Women CMA 24,653 27,504 Immigrant Non-CMA 21,730 24,432 CMA 44,246 49,760 Canadian Born Non-CMA 36,744 39,215 Men CMA 36,859 41,735 Immigrant Non-CMA 40,324 43,414 TABLE 2. PROPORTION IN NON-CMA LOCATIONS Year Sex Status 1981 1986 1991 1996 2001 Canadian Born 41.92 39.49 40.51 40.70 40.13 Women Immigrant 12.51 11.29 11.48 11.08 9.81 Canadian Born 45.42 42.22 42.38 42.42 41.29 Men Immigrant 14.13 12.20 11.62 10.97 9.60
|Gale Copyright:||Copyright 2004 Gale, Cengage Learning. All rights reserved.|