![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
Although rates of residential crowding in Canada declined between 1971 and 2001, the drop for immigrants was not nearly as pronounced as it was for the Canadian-born. This paper determines the extent to which the differential trends in residential crowding can be attributed to changes in educational attainment, household composition, economic characteristics (including housing values), Census Metropolitan Area (CMA) of residence, and skin colour (visible minority status). The prospect of changing effects over time for several factors is also assessed. There are two main findings. The first is that basic household characteristics explain a significant portion of the 30-year immigrant/Canadian-born divergence. Secondly, including time interactions reveals that differences in propensities across CMAs emerged post-1971, and that there have been almost no significant changes in the propensity to crowd among visible minority groups since then.
Notes
1. A room is defined as it was by Statistics Canada for the 2001 census. Partially divided L-shaped rooms are treated as separate rooms if they are considered as such by the respondent.
2. Where a household contains one adult and one or more children, or more than one adult with or without children.
3. The list of negative outcomes is lengthy, but a short list could include poor mental health and lower life satisfaction (Gove and Hughes 1983), labour productivity (Hacker Citation1999), child academic performance (Evans et al. Citation1998), plus increased risk of tuberculosis, shigellosis and pneumococcal infections (www.phac-aspc.gc.ca/publicat/ccdr-rmtc/03vol29/dr2905ea.html).
4. Exceptions include Gove and Hughes (Citation1983), Gillis et al. (Citation1986) and Myers et al. (1996).
5. ‘Redlining’ is the practice of denying, or increasing the cost of, services such as banking and insurance, and access to jobs, healthcare or even supermarkets to residents in certain, often racially determined, areas.
6. Other standard household characteristics are age, education, years since migration and knowledge of English or French, but descriptive results for these factors are not shown here because the changes over time suggest that there would be even larger differential gaps between groups. Furthermore, in the regression models presented later, there is a count of children and adults rather than the combinations presented above for illustration purposes.
7. These rates pertain only to the highest household earner, who is more likely to be employed than the labour force population overall.
8. Comparing CMAs in Canada over time is complicated by changes in boundaries over time. I chose to use current boundaries, bearing in mind that most metropolitan areas have grown considerably across the study period.
9. Interestingly, however, the Canada Mortgage and Housing Corporation recently reversed its trend of extending amortisation periods in July 2008, by reducing the maximum period from 40 to 35 years.
10. A more common definition used in Canada is that of the National Occupancy Standard, which also looks at the age and distribution of individuals across rooms. Since many datasets do not contain this information, comparability is compromised to some extent, so this study instead uses total number of persons/number of rooms to identify crowding.
11. ‘Visible minority status’ is a distinctly Canadian term to denote differences in skin colour, and is measured in accordance with the Employment Equity Act. It is not intended to denote actual differences beyond that of skin colour. As with any other racial marker, these are socially constructed signifiers of difference.
12. My use of visible minority indicators is not intended to denote within-group homogeneity. Since race is socially constructed and imbued with meaning, it is my intention to capture the social process of collapsing heterogeneity, labelling groups ‘Black’ or ‘White’, and using these markers to establish treatment.
13. Ideally, all characteristics would be tested as a time interaction but, as discussed later in the paper, several could not be due to problems with multicollinearity.
14. There is actually some discussion around this issue. Some would say that the interpretation actually does change, since the reference group becomes more specific with the inclusion of controls.
15. It is important to note that the order in which variables are entered is not meant to imply causal prioritisation, but merely intended to reflect the theoretical order of importance often implied or stated in the literature.
16. Researchers typically assess the significance of reductions in BIC by using critical chi-square values with k degrees of freedom. These reductions are all well above the threshold value, so little attention is paid to chi-square assessments. Also, the values themselves do not have a straightforward interpretation, but BIC values can be used to compare the relative explanatory value of each set of variables.
17. The decision to exclude these variables was made by running tests for collinearity (using a linear probability model and variance inflation factors), and determining that these variables excessively inflate variance estimates. Furthermore, they had little impact on the coefficients of interest, further prompting their exclusion.