The sometimes passionate discussion surrounding Canada’s always controversial immigration policy is often shaped by opinion rather than research. How many and what kind of immigrants we should welcome speaks to the type of society we wish to become. Will immi- grants integrate smoothly into mainstream Canadian socie- ty, or will they tend to form enclaves? How long will it take them to achieve the same economic status as native-born Canadians, particularly if they are visible minorities from countries whose culture is vastly different from Canada’s? Two-thirds of Canada’s current immigrants are members of visible minorities. Do they face different opportunities than native-born whites? If so, what accounts for these differ- ences? Are they due to differences in education and train- ing? In skills and work experience? In cultural or language factors? Or just plain discrimination? Is it simply that assim- ilation takes time?
There are any number of reasons to expect a ”œwage gap” between newly landed immigrants and native-born Canadians, but over time this gap should narrow as immigrants become accustomed to Canadian labour markets. Accordingly, one indicator of economic assimilation of immigrants is the rate of growth of their wages compared to those of other Canadians. If they are in fact closing the gap, their wages should grow more quickly than those of other Canadians. Assimilation can be considered complete only when immigrants’ wage levels are indistinguishable from those of comparably qualified native-born Canadians. Wages are a more appropriate measure of labour market assimilation than income or earnings because immigrants may simply adjust to low wage offers by working longer hours to obtain higher earnings.
In the Dec. 2000 issue of Policy Options we examined how being a member of a visible minority affects the wage gap, taking into account factors such as age, gender, educa- tion, language knowledge, work experience, regional loca- tion, the number of years since migrating to Canada, age at migration and the like. We found that, after adjustment for these ”œproductivity relevant” differences, being a visible minority was a significant disadvantage only for immigrant men. For Canadian-born men, visible minority membership was generally insignificant. The wage gap was mainly a problem for immigrants.
Conventional wisdom holds that immigrants face initial disadvantages when first arriving in Canada, a phenomenon often referred to as ”œa negative entry effect.” This disadvantage erodes over time, and eventually immigrants match or even surpass the economic performance of natives (thus producing ”œa positive assimilation rate”). Over the last decade, different researchers have employed different measures of economic assimilation, including earnings, hours worked and wages, using a variety of data sources, but all have used cross-sectional methods, that is, they have looked across individuals at a given point in time.
Despite its small size, the research literature on the assimilation of Canadian immigrants is quite diverse. Even so, a few patterns emerge. Upon entry, male immigrants appear to work fewer hours in less secure jobs and for lower wages than do comparable native-born males. This is especially true of later immigrant cohorts, most likely because they contain fewer immi- grants from Britain, Europe and the U.S. and therefore more visible minorities. On balance, the literature does indicate some assimilation along most dimensions, although its pace is unclear. For the most part, female immigrants follow roughly the same pattern, though women in two-immigrant marriages appear to work high initial hours in dead-end jobs, possibly to finance their husband’s education or training.
But the fact that evidence about the labour market performance of Canadian immigrants is based exclusively on cross-sectional studies is potentially a problem. If there are systematic dif- ferences between immigrants and native-born Canadians that aren’t captured in the variables controlled for (which typically include educa- tion, labour market experience, time in Canada and so on), then any remaining wage gap between immigrants and the native-born may be incorrectly attributed to place of birth, and that may lead to error in trying to estimate how immi- grants will assimilate into the labour market. From a policy standpoint, as well, this can be a problem if it focuses attention on factors that aren’t really at the heart of the wage gap that society would like to reduce.
In order to address this problem with the cross-sectional data, we look instead at panel data””that is, data gathered from the same indi- viduals over a number of years rather than just a single period””from Statistics Canada’s survey of Labour and Income Dynamics (SLID), covering the period from 1993 to 1997. In particular, we investigate how the growth of individual work- ers’ wages between 1993 and 1997 was related to their personal characteristics at the beginning of the period as well as to any changes in their observed characteristics over the same period. This allows us to determine if there are any sys- tematic differences in the evolution of the wage patterns of visible minorities or immigrants. As a simple matter of logic, if immigrants are ever to catch up they must at some time experience faster wage growth than native-born Canadians.
We employ a model that uses three sets of variables to explain wage growth over time: (1) whether workers are immigrants or members of a visible minority (2) how much training and edu- cation (or ”œinvestment in human capital”) they received over the five years and (3) how much edu- cation they had at the beginning of the period.
This framework has the potential to resolve a number of unanswered questions in the litera- ture. The first set of variables allows us to exam- ine whether wages grow faster for foreign-born than Canadian-born workers, and if so, whether wage assimilation differs for visible minorities, as was found in the cross-sectional studies. The sec- ond and third sets allow us to examine the role of human capital in the growth of wages for foreign- born and native-born Canadians. Specifically, the second set explains how wage growth is affected by the accumulation of human capital during the panel period and lets us answer such questions as, for example, how much does an additional year of schooling or work experience increase wage growth? The third set explains how schooling and work experience accumulated before the panel period affect subsequent wage growth. For instance, do highly trained immigrants experi- ence faster wage growth than Canadians in gen- eral? And what is the story for low-skilled immi- grants? We also include regional and community size measures as control variables to capture dif- ferences in the rate of wage growth in different locations.
What results do we get? To simplify our presentation, we report the change in wages or percentage wage growth over the entire panel period, 1993-1997, for those individuals reporting a wage in both 1993 and 1997. That is, we calculate the difference between the wage received in 1997 and the wage received in 1993, expressed as a percentage of the 1993 wage. Although foreign-born men earned more than Canadian-born men in both 1993 and 1997, this difference is entirely explained by other charac- teristics. Indeed, when these other relevant observable characteristics are taken into account foreign-born men were at a wage disadvantage. A similar story exists for women. In particular, for- eign-born men and women are older, have more labour market experience, are more likely to live in Ontario or the West, and are more likely to live in large cities than their native-born counter- parts””all of which is associated with higher than average wages. After adjusting for these factors, however, comparably qualified foreign-born men and women actually earn lower wages than Canadians.
As for whether the foreign-born ”œcatch up” over time as they assimilate into the Canadian labour market, our findings provide only modest evidence of assimilation for men. The average wage growth between 1993 and 1997 for all for- eign-born men was $2.49, or 13.2 per cent, com- pared to $2.30, or 12.9 per cent, for the native- born. For women, wages of the native-born increased by $1.56 or 10.8 per cent over the peri- od 1993-97, which was clearly faster than the $1.10 or 7.5 per cent catch-up experienced by foreign-born women. This is consistent with a commonly held view that immigrant women often take jobs to maintain family consumption standards while immigrant men invest in human capital for better paying jobs later on.
But, again, these raw numbers are not of great interest. What happens when differences in education and experience both before and during the panel period, as well as differences in regional and urban growth patterns are con- trolled for? Our results are that visible minority membership and immigration status, both sepa- rately and taken together, have no effect on wage growth. For the years 1993-97, we find no evidence that the wages of foreign-born men grow faster than those of their native-born coun- terparts. Indeed, if anything, our estimates indi- cate somewhat slower wage growth for immi- grants and visible minorities””which obviously poses a problem for the conventional wisdom on economic assimilation.
The inner workings of government
Keep track of who’s doing what to get federal policy made. In The Functionary.
The Functionary
Our newsletter about the public service.
Nominated for a Digital Publishing Award.
An even more interesting result, in our judg- ment, involves the effect of human capital accu- mulation, particularly work experience, on wage growth. Looking first at both foreign- and native- born men taken together, an additional year of full-time work experience acquired during the period 1993-97 raises wage growth by eight per cent (not eight percentage points) more than would occur with one year less experience. Workers with more prior experience have signifi- cantly slower wage growth. In fact, having an additional year of experience prior to 1993 reduces subsequent wage growth by 1.7 per cent, although this effect declines as additional years of experience are accumulated. (Note that the hourly wage rate in 1993 for all men was $17.95, and $20.27 in 1997.)
As for investments in new human capital, an additional year of schooling obtained during the panel period raises wage growth by four per cent, although this effect is not quite statistically sig- nificant at the five per cent level of significance. It appears that additional work experience is more relevant than additional schooling in rais- ing wages.
Among our other findings are that people who live in larger cities and residents of British Columbia experience significantly faster wage growth, while wages grew more slowly in Atlantic Canada compared to Ontario. Whether a person’s mother tongue is English or French has no sig- nificant effect on wage growth. Unfortunately, we were not able to include any measure for lan- guage proficiency, which can obviously vary over time, particularly for the foreign born.
We also naturally investigated whether these results differ between foreign-born and native-born men. Membership in a visible minority remains insignificant in explaining the rate of wage growth for the foreign-born, but it is significant and negative for the native-born: wage growth is slower for Canadian-born””but not foreign-born””visible minorities, a result that runs counter to some cross-sectional research suggesting that only foreign-born workers who were members of a visible minority group experi- enced problems in wage assimilation. Our results therefore confirm that evidence from cross-sec- tional data may be prone to bias arising from unobserved differences in workers and should be interpreted cautiously.
However, human capital accumulation again plays an important role in explaining wage growth. An additional year of experience is statistically significant for the native born, raising wage growth by almost nine per cent, compared to less than five per cent (a number that is not statistically significant) for foreign- born men. The effects of schooling are general- ly insignificant although they are noticeably larger for the foreign born. Residence in larger cities or in British Columbia resulted in signifi- cantly higher wage growth for native-born men only. Foreign-born men in larger cities did not experience significantly higher wage growth than their counterparts in smaller cities, other factors considered.
In general, our research results provide only weak evidence of differences in the wage growth experienced by foreign-born and native born men. In fact, the hypothesis that the for- eign-born and native-born samples are statistical- ly different is rejected by formal criteria. Hence, our conclusion is that there is no hard evidence that foreign-born male workers were assimilating economically during the 1990s in Canada.
We obtain similar results for women. Wage growth is similar (about seven per cent over the five-year period) across all groups, since there are no significant differences by immigration and vis- ible minority status. The only factors that are sta- tistically significant in explaining the wage growth of women are prior work experience, which is only significant for native-born women, and region. The wage growth of women in Atlantic Canada and the Prairie provinces was significantly slower than for native-born women in Ontario. We also cannot reject the hypothesis that wage growth in the foreign-born and native-born samples is iden- tical. Thus there is no evidence of economic assimilation for foreign-born female workers.
Where does this time-series analysis leave us? The standard story of immigrant assimilation is appealing. Immigrants face tough challenges upon arrival, but with hard work they eventually become full members of Canadian society. And our examination is admittedly exploratory and incomplete. Other measures of economic assimi- lation are available, and different interpretations of our results are also possible. So it would be rash to interpret findings from this first investigation using panel data as ”œconclusive evidence,” or to suggest that firm policy recommendations should flow from its results. For this reason, we advise readers to treat our initial efforts at under- standing wage growth differences between immi- grants and non-immigrants with caution. Yet our results do suggest that economic assimilation is neither immediate nor automatic. The lack of strong evidence of faster wage growth by immi- grants means either that assimilation during the mid-1990s was largely absent or that immigrants need more than five years to realize the acceler- ated wage gains assimilation requires. Adjusting to a new labour market isn’t easy and may take time. But either way, the speed and means of eco- nomic adjustment by immigrants to Canada remains an important policy topic.
Â
This paper follows on their recent work in “Closing the Wage Gap: Economic Assimilation of Canadian Immigrants Reconsidered,” Journal of International Migration and Integration (I: 4, 2000).
