A socio-demographic profile of migrants in Canada according to the 2006 Census

View the most recent version.

The previous section described internal migration flows in Canada between 2001 and 2006 and how the populations of Canada’s various geographic regions were affected. The wealth of data from the 2006 Census supplements this analysis by providing a socio-demographic profile of recent migrants.

Canadian censuses contain a mass of information on respondents’ demographic and social characteristics: in addition to the age, the sex and details about their recent migrations, there is information on the marital status, the education level, the immigrant status, the Aboriginal identity, the visible minority status as well as details on the family structure and the place of residence. This section of the document specifically examines all these characteristics in order to flesh out the profile of Canadians who recently migrated.

Moreover, because migrants’ characteristics can vary depending on the destination chosen (for example, youth and people who are more educated choose more often to settle in large urban centres than do older or less educated people), this analysis looks at the individual characteristics of migrants by type of destination.

Methods and concepts

A multivariate analysis, based on a statistical model, was preferred to a descriptive analysis, which is quite often used to establish the socio-demographic profiles of population groups (see text box). The statistical model was chosen in order to isolate the net effect of the characteristics studied. It is thus possible to measure the association between a socio-demographic characteristic of an individual and his or her migration path while neutralizing the effect of all other characteristics. The results of the model can be used to estimate the probability of having migrated in the year preceding the 2006 Census according to various characteristics considered and by type of destination.

Six types of destination are considered in the model: the central municipalities of Toronto, Montreal and Vancouver census metropolitan areas; the peripheral municipalities of Toronto, Montreal and Vancouver census metropolitan areas; other census metropolitan areas; mid-size urban centres; rural areas located near an urban centre; and remote rural areas or territories (see appended definitions). This urban-rural gradient has the advantage of covering Canada’s entire geographic space and creating groups that exhibit some homogeneity as to their characteristics (presence of universities, types of industries, distance from cities, residential characteristics, etc.).

In this section, migrants are defined as those persons who changed municipality (census subdivision) in the twelve months preceding the 2006 Census. Since only the socio-demographic characteristics at the end of the period, that is, at the time of the 2006 Census, are known, it was preferable to restrict the study period to the migration observed during the year preceding the census, thereby limiting the chances that the characteristics studied changed after migration.¹

It should be noted that according to the definition of migrants that is used in this section, an individual could migrate while remaining in a same type of area in the urban-rural gradient. For example, the characteristics of a person who moved from Calgary to Edmonton—two census metropolitan areas included in the “other census metropolitan areas” category—would be accounted for in the model by recording the destination in the “other census metropolitan areas” category. Thus, the model takes account of all migrants, not only those who migrated from one type of area to another.²

The text box provides more details on the sample and the model used.

Results

The results of the multivariate statistical analysis are presented in the form of estimated probabilities of migrating. For each characteristic presented, the model brings out the net effect of this characteristic, that is, the effect with all things being otherwise equal.

A comparison of the estimated probabilities sheds light on the strength of the association between the various socio-demographic characteristics and migration, and it brings out nuances that would not emerge from a simple descriptive analysis.

Table 2.1 shows the estimated probabilities of migrating according to different types of destinations. Table 2.2 provides a percentage distribution of the estimated probabilities of migrating to the different types of destinations. It shows the geographic distribution of migrants with an equal probability of migrating and facilitates comparisons of probabilities between places of destination.

Young people aged 20 to 29 are more likely to migrate

Age is often seen as reflecting the position of individuals in the life cycle. For example, while young persons are generally more mobile, this is due in part to the great number of transitions experienced during one’s youth, such as the beginning of post-secondary education, changes in marital status or entry into the labour market.

The results of the model show that indeed, all things being otherwise equal, migration is strongly associated with age (table 2.1). The probability of migrating is relatively high for youths aged 15 to 19 (6.77%); it peaks between ages 20 and 29 (11.08%), then gradually but substantially declines in the older age groups. The stronger propensity of young people aged 20 to 29 to migrate is observed for all types of destinations.

The fact remains that the various types of destinations do not attract persons in the different age groups equally. Table 2.2 shows the estimated percentage distribution of probabilities of migrating for each characteristic and by type of destination. It appears that all things being otherwise equal, the proportion of migrants choosing a remote rural area or a territory as a type of destination is higher among persons aged 45 and over than in the other age groups. The probabilities show that out of 100 migrants aged 45 and over, at least 20 choose to settle in a remote rural area or a territory, compared to 17 or fewer in the other age categories. Probably some events related to the life cycle of these persons—the departure of children from the parental home, retirement—generate, in part at least, migrations to less urbanized areas.

Also, relatively speaking, migrants aged 30 and over exhibit a greater preference for the peripheral municipalities of Montreal, Toronto or Vancouver than persons under 30 years of age (table 2.2). For their part, young people under 30 years of age seem, more than persons in other age groups, to prefer the central municipalities of the Montreal, Toronto and Vancouver census metropolitan areas as well as other census metropolitan areas.

Single persons migrate less

In general, the results of the model show that all things being otherwise equal, single persons have a lower probability of migrating (3.81%) than persons who are married or in a common-law relationship (5.49%) or persons who are divorced or separated (7.72%) or widowed (6.37%). These findings go against the common-sense notion that single persons are more mobile. Part of the explanation lies in the fact that most single persons are young and have no children, two characteristics associated with high mobility.

In addition, living with one’s parents is another characteristic often observed among single persons. More than half (56.1%) of single persons included in the study population were living with their parents at the time of the 2006 Census. Living in the parental home might well be an impediment to mobility, given the sometimes-prohibitive cost associated with a first migration. Moreover, this effect probably varies according to the place of residence. Persons living with their parents in small towns and rural areas are more likely to leave the parental home earlier than persons living in a city with a population of one million or more.⁵ One possible reason for this phenomenon is that whereas persons living in the parental home in a large city have access to post-secondary educational institutions, those living in small towns and rural areas must often leave the family home to continue their education.⁶

Another factor to consider is that unlike single status, the other marital statuses—married or common-law, separated or divorced, or widowed—may be the consequence of a transition that occurred shortly before the date of the census. Such transitions, e.g., getting married or losing a loved one, are often causes of migration.⁷

Divorced or separated persons have the highest probabilities of migrating, all destinations combined (7.72%). For many of them, migration is probably a consequence of the breakdown of their conjugal relationship.

The distribution of migrants by type of destination (table 2.2) also shows that given the same probability of migrating, single persons have the highest propensity to choose the central municipalities of Montreal, Toronto or Vancouver as their destination type. Divorced or separated persons and widowed persons are also more inclined to choose this type of destination than those living in a couple relationship. This is probably explained in part by the opportunities that these places offer in terms of education, employment and leisure facilities.⁸

The results in table 2.2 also show that the propensity to choose to settle in the peripheral municipalities of Montreal, Toronto or Vancouver is greater among migrants who are married or living in a common-law relationship and widowed persons than it is among single persons or divorced or separated persons. Finally, widowed persons are generally less likely to choose mid-size urban centres or rural areas when they migrate.

Having children reduces the probability of migrating

The results of the model show that all things being otherwise equal, the probability of migrating for persons who have children at home is lower than for persons who do not. Various reasons may explain this phenomenon. For example, the economic costs associated with migration often increase with the number of children and may sometimes become prohibitive. Also, the number of ties that must be broken when migrating is greater in larger families and may be an impediment to migration.⁹

The greater propensity to migrate among childless persons is mainly evident where the destination is either one of the central municipalities of Montreal, Toronto or Vancouver or another census metropolitan area. As regards the central municipalities of Montreal, Toronto and Vancouver, the probability of migrating for childless persons (0.57%) is more than double that of persons who have children but none born recently (0.18%), or that of persons who have recently had a first child (0.25%), a second child (0.17%) or a third or higher child (0.15%).

The greater mobility of childless persons was an expected result, but the results concerning the recent arrival of a couple’s first child are instructive. The results show that the recent birth of a first child has a major impact on the probability of migrating.

The fact is that parents of a first child have a slightly higher probability of migrating to a peripheral municipality of Montreal, Toronto or Vancouver (1.35%) than childless persons (1.29%). Furthermore, data from the 2006 Census showed that in the Montreal, Toronto and Vancouver census metropolitan areas, the proportions of households comprised of couples with children was higher in the peripheral municipalities than in the central municipalities.¹⁰

The probabilities of migrating to a mid-size urban centre or a rural area for persons with a first child born shortly before the date of the Census are comparable to those for childless persons.

These phenomena could be linked to the desire to change one’s place of residence to better meet the new needs created by the arrival of a first child, or more specifically to purchase a home more easily or at less cost.¹¹ These findings agree with those of earlier studies conducted elsewhere, notably in France. According to a study on spatial mobility in that country, the probability of migrating from one area to another is relatively high in the year following a first birth.¹² Another study showed that the probability of migrating to cities declines with each birth and the probability of migrating to a rural area increases with family size.¹³

More educated persons migrate to large urban centres

The link between migration and education is fairly well known: in general, mobility increases with educational attainment. This phenomenon might be explained in part by the fact that individuals with high education levels tend to have job opportunities over a very wide geographic area, which would lead to greater mobility.¹⁴

Data from the 2006 Census confirm this general finding. For example, in 2006, persons aged 25 to 64 with a university diploma accounted for 23% of the population, whereas they represented 33% of persons who were not living in the same province five years earlier.¹⁵

The results of the model also confirm the association between education and migration: the propensity to migrate gradually increases with education level, going from 4.26% for persons with less than a high school diploma to 5.66% for those with a bachelor’s degree or a university diploma higher than the bachelor’s.

Also, the probabilities of migrating vary according to the type of destination. Thus, persons with a bachelor’s or a higher degree have three times greater a propensity to migrate to large urban centres such as the central municipalities of Montreal, Toronto and Vancouver than persons who have not completed high school and twice the propensity of persons with a high school or trade school diploma.

Conversely, persons with a university degree have lower probabilities of migrating to a rural area. This result might be linked with the nature of the jobs usually found in these areas.

Immigrants and visible minority persons tend more to migrate to the Montreal, Toronto and Vancouver census metropolitan areas

The variables that reflect immigrant status and belonging to a visible minority group were cross-tabulated in the model to take account of the correlation between the two: in 2006, two-thirds (66.3%) of the population belonging to a visible minority group were immigrants to Canada.¹⁶

All things being otherwise equal, recent immigrants—namely, those who came to Canada between 1996 and 2005—generally have a greater propensity to migrate than immigrants who came prior to that period and persons born in Canada, except for those who both are Canadian-born and do not belong to a visible minority group.

At the same time, the results indicate that persons belonging to a visible minority group are overall less mobile than persons not belonging to such a group, regardless of immigrant status.

A few interesting nuances may be added to this picture of migrations on the basis of immigrant status and visible minority status. Compared to the other groups studied, recent immigrants to Canada, especially those belonging to a visible minority group, have a strong propensity to migrate to Canada’s three large metropolises, namely Montreal, Toronto and Vancouver, either to the peripheral municipalities or to the central municipality. These results agree with other results from the 2006 Census that showed that immigrants and persons belonging to a visible minority group were more concentrated in these three large census metropolitan areas. According to data from the 2006 Census, the three census metropolitan areas of Montreal, Toronto and Vancouver together accounted for nearly 72% of visible minority persons in 2006 (namely 11.6%,  42.9% and 17.3% respectively)¹⁷ and nearly two-thirds of persons born outside Canada.¹⁸

At the same time, members of a visible minority group and, to a lesser extent, immigrants have relatively low probabilities of migrating to rural areas or mid-size urban centres. This suggests that internal migrations in Canada have only a limited effect on how the ethnocultural diversity of the population is distributed throughout the country.

A number of reasons may be cited to explain immigrants’ migratory movements. According to studies, two major factors to consider in this regard are the economic opportunities offered by potential destinations¹⁹ and the draw of ethnic communities already established in some large cities.²⁰

Aboriginals migrate more

With a median age of 27 years, the Aboriginal population is on average younger than the rest of Canada’s population (median age of approximately 40 years).²¹ It is also more concentrated in rural areas often remote from large urban centres. Because of these two characteristics of the Aboriginal population, this population is more likely to migrate than others.

With a probability of migrating of 5.82%, the results of the model show that Aboriginal persons are indeed more mobile than non Aboriginal persons (4.90%). This result is even more striking since the model controls for the effect of age and rural/urban living environment.

However, the Aboriginal population’s greater propensity to migrate is observed in only three types of destinations: remote rural areas and the territories, mid-size urban centres, and to a lesser extent, other census metropolitan areas. In particular, the Aboriginal population’s probability of migrating to a remote rural area or a territory (1.52%) is nearly double that of non-Aboriginals (0.81%).

Of course, migration patterns vary among Aboriginal groups, and the results yield only a general picture for the Aboriginal population as a whole. For example, between 1991 and 1996, urban-to-urban migration movements accounted for 37% of migrations for Registered Indians, 59% for Non-Status Indians, 53% for Métis, but only 24% for Inuit. In the latter group, the dominant type of migration consisted of rural-to-rural movements (38%).²²

Persons living in rural areas are more mobile than those living in urban areas

Persons living in a rural area in 2005 had a greater probability of migrating (5.67%) than those living in an urban area (4.68%).

The propensity to migrate to the central or the peripheral municipalities of Montreal, Toronto or Vancouver is higher for persons living in an urban area. Elsewhere, the probabilities of migrating are higher for persons living in a rural area. Considering mid-size urban centres in particular as a destination, the probability of persons living in rural areas migrating to them is nearly double that of persons living in urban areas.

This finding may perhaps be explained in part by sequential migration, in which persons from rural areas go to mid-size urban centres before possibly going to larger urban centres. Internal migrations would thus contribute to the larger phenomenon of the urbanization of Canada’s population.

---

Notes

1. Similarly, the selection of variables that could be included in the model was restricted to those applicable to the beginning of the studied period. Therefore, variables such as employment status and type of occupation have been excluded from the analysis.

2. The probabilities of changing category of place of residence were calculated for comparison purposes (see tables A-2.2 and A-2.3, appended). It was found that while the distribution of probabilities among the different types of destination differs slightly from the original model, differences between the probabilities associated with the independent variables are, for their part, of similar orders to magnitude.

3. King, Gary and Langche Zeng. 2001. “Logistic Regression in rare Events Data”. Political Analysis. Volume 9. Number 2. pp. 137 to 163.

4. Hosmer, David W. and Stanley Lemeshow. 2000. Applied Logistic Regression, Second Edition. John-Wiley & sons.

5. Beaupré, Pascale, Pierre Turcotte and Anne Milan. 2006. “When is junior moving out? Transitions from the parental home to independence”. Canadian Social Trends. Statistics Canada Catalogue number 11-008. Winter. pp. 8 to 15. Ottawa.

6. Turcotte, Martin. 2006. “Parents with adult children living at home”. Canadian Social Trends. Statistics Canada Catalogue number 11-008. Spring. Volume 80. pp. 2 to 12.

7. Sandefur, Gary D. and Wilbur J. Scott. 1981. “A dynamic analysis of migration: an assessment of the Effects of Age, Family and Career Variables”. Demography. Volume 18. Number 3. pp. 355 to 368.

8. Feijten, Peteke and Maarten Van Ham. 2007. “Residential mobility and migration of the divorced and separated”. Demographic research. Volume 17. Article 21. pp. 623 to 654. December 20.

9. Sandefur, Gary D. and Wilbur J. Scott. 1981. “A dynamic analysis of migration: an assessment of the Effects of Age, Family and Career Variables”. Demography. Volume 18. Number 3. pp. 355 to 368.

10. Milan, Anne, Mireille Vézina and Carrie Wells. 2007. Family portrait: Continuity and change in Canadian families and households in 2006: 2006 Census. Statistics Canada Catalogue number 97-553-X.

11. Détang-Dessendre, Cécile, Florence Goffette-Nagot and Virginie Piguet. 2004. “Life-cycle position and migration to urban and rural areas: estimations of a mixed logit model on French data”. Groupe d’analyse et de théorie économique, W.P. 04-03. April.

12. Detang-Dessendre, Cécile, Virginie Piguet and Bertrand Schmitt. 2002. “Les déterminants micro-économiques des migrations urbain-rural : leur variabilité en fonction de la position dans le cycle de vie”. Population. 57th year. Number 1. January-February. pp. 35 to 62.

13. Courgeau, Daniel. 1989. “Family formation and urbanization”. Population : An english selection 44(1): pp. 123 to 146; cited in:  Hulu, Hill and Nadja Milewski. 2007. “Family change and migration in the life course: An introduction”. Demographic Research. Volume 17. Article 19. pp. 567 to 590. Published December 20, 2007.

14. Courgeau, Daniel. 1984. “Relations entre cycle de vie et migrations”. Population. 39th year. Number 3. May-June. pp. 483 to 513.

15. Statistics Canada. 2008. Educational Portrait of Canada, 2006 Census. Statistics Canada Catalogue number 97-560-X.

16. Chui, Tina, Kelly Tran and Hélène Maheux. 2008. Canada’s Ethnocultural Mosaic, 2006 Census. Statistics Canada Catalogue number 97-562-X.

17. Ibid.

18. Chui, Tina, Kelly Tran and Hélène Maheux. 2007. Immigration in Canada: A Portrait of the Foreign-born Population, 2006 Census. Statistics Canada Catalogue number 97-557-X.

19. Newbold, K. 1996. “Internal Migration of the Foreign-Born in Canada”. International Migration Review. 30(3). pp. 728 to 747.

20. Moore, Eric G. and Mark W. Rosenberg. 1991. Factors influencing the redistribution of immigrant groups in Canada. Queen’s University.

21. Statistics Canada. 2008. Aboriginal Peoples in Canada in 2006: Inuit, Métis and First Nations, 2006 Census. Statistics Canada Catalogue number 97-558-X.

22. Norris, Mary Jane and Stewart Clatworthy. 2003. “Aboriginal Mobility and Migration within Urban Canada: Outcomes, Factors and Implications”. In the publication: Not Strangers in these Parts: Urban Aboriginal Peoples. Edited by David Newhouse and Evelyn Peters, Policy Research Initiative