The housing experiences of new Canadians: Insights from the Longitudinal Survey of Immigrants to Canada (LSIC)

3. Moving from renter to owner

3.1 Introduction

As mentioned earlier in this report, the Chicago School of Sociology introduced the theoretical concept of assimilation to immigration research roughly 80 years ago, and following the perceived inability of this theory to explain several outcomes, subsequent researchers introduced stratification theory to supplement or even replace the original formulation. Research on immigrants today typically weighs the comparative merits of these two traditions. In Canada, the word “integration” is typically used in place of “assimilation.”

Although easy to grasp theoretically, operational definitions for integration and stratification are more difficult to pin down. Is economic convergence with the host society alongside persistent social and cultural differences reflective of integration or stratification? What does it mean when a person is indistinguishable from a certain segment of the population (such as those with chronic low-income), but readily identifiable in the broader society? Is a group integrated if its members have high levels of economic success, but live in an ethnic enclave?

These are difficult questions to answer, and for this report Alba and Nee’s definition provides a useful working definition. They define assimilation (and, for this report, integration) as “minority participation in mainstream socio-economic institutions (e.g., the labour market) on the basis of parity with ethnic-majority individuals of similar socio-economic origins” (Alba and Nee 1997); stratification is defined as the opposite of this, and is usually based on physical characteristics like skin colour, or on other factors like place of origin.

Given that roughly two-thirds of households in Canada are owned (Haan 2005), one of the expectations for those that successfully integrate is that over time most immigrants will participate in the mainstream housing market by also buying a home of their own. That is not to say that everyone will buy a home among LSIC respondents, but that roughly 2/3 of them will want to and be able to. Home buying not only increases their resemblance with the host society in terms of type of residence, it also provides them with greater access to the amenities (parks, schools, community centres, etc.) that are more likely to be found in neighbourhoods with predominantly owner-occupied dwellings. As such, it is a good indicator of integration into Canadian society.

As shown earlier in this report, that rate at which LSIC respondents move into owner-occupied housing broadly suggests that integration as described above is proceeding apace. Further evidence of this comes from LSIC respondents themselves, who, when asked at six months if they planned to buy a home in the coming years, overwhelmingly responded positively to the question. Well over half of those that did not already own stated that they planned to make a purchase in the coming years. Once again, this is not meant to imply that the desire for home ownership is universal amongst immigrants, but that the desire is strong. In fact, only about 1 in 7 explicitly stated that they did not plan to buy a house, suggesting that the rate of home ownership among the LSIC sample may one day be as high or higher than it is for Canadians overall (68.4% according to the 2006 census).

As this report has shown, the broad trend of rapid attainment does not translate into identical ownership rates across immigrant groups. Using Alba and Nee’s definition presented earlier as a guide, one of the questions about these divergences is whether the differences between groups stem from a lack of parity, or if groups differ because they hold different amounts of home ownership relevant resources (or “housing capital”) or preferences.

The remainder of this part of the report proceeds as follows: first, overall home ownership trajectories are illustrated, followed by a break-down by visible minority status. Then, results from several logistic regression models are shown, and used to create probability plots of home ownership attainment in the first four years. These plots are useful in that they present ‘standardized’ trajectories, and allow for an assessment of what attainment would look like over time if the LSIC sample had identical access to credit, entry wealth, income, employment prospects, etc. Finally, the role of changing statuses (getting a better job, acquiring citizenship, etc.) is modeled as it pertains to housing.

3.2 Immigrant homeownership trajectories in the first four years

Although Section two of this report displayed differences across Census Metropolitan Area, class of entry, visible minority status, age and several other factors, this section focuses specifically on explaining the differences in home ownership rates across visible minority categories. The processes that differentiate immigrants in the Canadian housing market are many, but looking at visible minority status is especially important. First, as a country that prides itself on multiculturalism and tolerance for diversity, immigrants to Canada should be able to move through the housing market uninhibited by discrimination. That said, the theoretical discussion outlined in Section one points to unexplained gaps across visible minority groups as a central indicator of a stratified housing market, and it is an important human rights determination to see if the evidence of stratification presented earlier is robust to deeper analysis.

In an egalitarian housing market, the gaps that exist in the descriptive results in Table 2-8 should be the direct cause of home ownership relevant characteristics like family structure, labour market success, affordability, etc. Evidence to the contrary points to a stratified housing market, manifested by persistent gaps between groups.

In Figure 3-1 the rapid rate at which immigrants moved in to homeownership is illustrated.

Figure 3-1: Number of months for immigrants to acquire owner-occupied housing, LSIC

Number of months for immigrants to acquire owner-occupied housing, LSIC

Figure 3-1 shows that about 8% of the sample that bought a home during the observation period did so immediately. Following that, there is a fairly monotonic increase over the four year period, so that over half of all immigrants in the LSIC sample live in owner-occupied accommodations by wave 3.Footnote 18 As is often the case with aggregate trends like the one denoted in Figure 3-1, beneath the surface lies variations (Figures 3-2 and 3-3).Footnote 19 Visible minorities are divided into two groups so as to be able to see the differences between them better in the figures.

Figure 3-2: Home ownership attainment trajectories of Chinese, South Asian, Black, Filipino, and Latin American immigrants in the four years after arrival, LSIC

Home ownership attainment trajectories of Chinese, South Asian, Black, Filipino, and Latin American immigrants in the four years after arrival, LSIC

Filipinos and South Asians both have the highest home ownership rates in Figure 3-2 at 10% at entry, a lead that they are able to maintain and enhance throughout the four-year period. They are followed by Chinese, Latin American, and, finally, Black immigrants. The attainment rate (or slopes of these lines) appears to be fairly similar in Figure 3-2 for all groups except Blacks, who exhibit a flatter attainment trajectory than the other four groups.

Figure 3-3: Home ownership attainment trajectories of South East Asian, Arab, West Asian, White, and other visible minority immigrants in the four years after arrival, LSIC

Home ownership attainment trajectories of South East Asian, Arab, West Asian, White, and other visible minority immigrants in the four years after arrival, LSIC

In Figure 3-3, there is a similar pattern of differentiation between groups. Here, Koreans, White, and Other Visible Minorities are the high attainment groups, and Arabs and West Asians have the lowest rates throughout the observation period. These differences emerge immediately at entry, with the only exception being Koreans, who enter Canada with low home ownership rates but quickly catch and surpass all other groups in both Figures 3-2 and 3-3, and have homeownership rates of over 70% after roughly four years in Canada.

To summarize these figures, it is clear that some groups vault quickly into home ownership, and have rates that approximate or even exceed the national average after only four years. Although several groups exemplify this trend, Filipinos, Koreans, South Asians, whites, and Other Visible Minorities are Canada’s housing ‘high achievers’, each with home ownership rates that approach or exceed 60% by wave 3. At the other end of the spectrum are Arabs, Blacks, and West Asians, who have home ownership rates that are nearly half of their higher achieving counterparts.

The extent to which these differences reflect stratification (defined as a lack of participation in mainstream institutions based on external, physical characteristics) remains an open question. Although it is certainly possible that some groups face discrimination in the housing market due to their cultural and/or physical characteristics, this is not the only explanation. There could be differences in home ownership relevant resources, such as income, access to credit, and stable employment. Alternatively, visible minority groups might hold different belief systems, thereby positioning them differently for garnering the necessary resources for buying a home. Third, some groups may disproportionately choose to live in cities with different beliefs about the importance of ownership over tenancy (such as Montréal) (Choko 1987). It is also possible that discriminatory experiences, if they exist, originate in the labour market, and are manifested in the Figures 3-2 and 3-3 because of the close connection that exists between housing and labour market success. All of these possibilities, along with numerous others, would explain the differences seen above, but would not reflect stratification that originates in the housing market.

3.3 Explaining differences between visible minority groupsFootnote 20

Most early immigration researchers hypothesized a smooth and fairly linear convergence with the native-born populationFootnote 21, implying that there should be few, if any, differences, both between immigrants and the general population, and between immigrant sub-groups, that could not be explained by looking at home ownership relevant characteristics like income, employment status, entry wealth, etc.

Researchers today give more thought to the processes that lead immigrants to differentially integrate into their host society. Although there is a secular trend towards the home ownership rates of Canadians overall, there are also considerable differences between groups, with groups like Arabs, Blacks, and West Asians posting home ownership rates that are nearly half of what Chinese, Filipinos and Koreans are able to achieve after four years. Furthermore, earlier research (Haan 2007; Skaburskis 1996) demonstrates that differences in homeownership trajectories that emerge in an immigrant group’s early years persist to become longer-term gaps, so that groups that do not move quickly into home ownership never completely catch up with those that do.

The policy questions that arise from these disparities require more detailed research on the reasons for the disparities above. Are divergent trajectories the result of labour market misfortunes? Are groups that buy homes quickly living with extended family members? Do they have entry wealth? Do they simply not want to buy a house? It is only once questions like these are answered that clear policy directions can be taken.

3.4 Adjusting homeownership trajectories

The extent to which the differences in attainment profiles shown in Figures 3-2 and 3-3 above stem from differential treatment in the housing market can only be determined by adjusting for differences in the home ownership relevant characteristics that exist across LSIC respondents and between groups.

In this section, the trajectories presented above are adjusted for differences in home ownership relevant resources. This is done by estimating a series of logistic regressions, one for each wave (as well as at time of entry), and using these data to predict probabilities of home owning for each group, holding all other aspects constant. These regressions include a number of socio-demographic, household structure, human capital, geographic, class of entry, credit constraints, entry wealth, receipt of assistance, and labour market characteristics. Central to the analysis is how these factors affect the statistical significance of the vector of visible minority indicators, which capture the differences in home ownership propensities not explained by other variables in the model.

Coding information and sample means for the variables in the regression models is presented in Table 3-1 below.Footnote 22

Table 3-1: Variable coding and description

Variables Means
Dependent Variable
Owner D 0.51
Socio-Demographic Characteristics
Age C 35.01
Spouse arrived before respondent D 0.13
# of Children C 0.93
Household Characteristics
Multiple Family D 0.10
Person Living Alone D 0.06
Lives with Non-Family Persons D 0.01
One Family with non-family persons D 0.05
One Family without non-family persons RC 0.77
Human Capital Characteristics
Less than High School RC 0.02
High School D 0.12
Post-Secondary D 0.31
University D 0.55
CMA Indicators
Toronto D 0.42
Montreal D 0.14
Vancouver D 0.15
Lives Elsewhere in Canada RC 0.29
Visible Minority Indicators
Chinese D 0.21
South Asian D 0.26
Black D 0.05
Filipino D 0.07
Korean D 0.04
Latin Am. D 0.03
Arab D 0.06
West Asian D 0.05
White RC 0.21
Other Visible Minority D 0.02
Region of Origin
Africa and the Middle East D 0.13
Asia and the Pacific D 0.64
South and Central America D 0.03
Europe and the United Kingdom RC 0.15
Other parts of the World D 0.05
Credit Constraints
Bring Savings D 0.74
Entry Wealth (logged) C 10.92
Religion
Roman Catholic D 0.19
Protestant D 0.13
Orthodox D 0.07
Jewish D 0.01
Muslim D 0.19
Eastern Religions D 0.18
No Religious Affiliation RC 0.23
Receipt of Assistance
Housing Advice Has Been Received D 0.01
Labour Market Characteristics
# of Jobs C 1.18
Income (logged) C 10.20
Primarily Came to Canada for School D 0.10
Works Full-time C 0.56
Spouse works Full-time D 0.49
Credentials Have not been Recognized D 0.03
Admission Category
Refugee D 0.06
Economic Class – Prin. Applicant D 0.35
Economic Class – Spouse and Dep. D 0.25
Other Economic Class D 0.06
Family Class RC 0.28

Note: ‘C’ denotes continuous variables, ‘D’ = dichotomous, and ‘RC’ = Reference Category
Note: For variables that vary over time (such as age and income), results are shown for Wave 3.
Note: For variables that vary over time (such as age, owner and income), results are shown for Wave 3.
Each number refers to the proportion of the population that resides in each category. For example ‘0.51’ for Owner denotes that 51% of respondents were homeowners by Wave 3.

Socio-demographic characteristics include age, arrival of spouse before respondent, and the number of children. This information is taken from the wave of data that is used for estimating the regression results. As such, it is possible for a respondent to have different information across waves (as would be the case with the addition of another child).

Similarly, human capital characteristics, Census Metropolitan Area of residence, religion, receipt of assistance, and most labour market characteristics (all but primary reason for coming to Canada for schooling) can all change between waves. As such, each of these characteristics reflects an individual’s response in the wave for which regressions are being estimated. For visible minority status, class of entry, credit constraints, and primary reason for coming to Canada, individual’s responses are the same for each wave.

Each of these variables likely differs for members of different visible minority groups, and could therefore form part of the explanation for both differences at a single point in time and for attainment over time. The visible minority coefficients, which reflect the difference between a particular group and the reference group (whites), are expected to shrink in terms of strength and significance with the introduction of controls, levelling the gaps in attainment and explaining the unadjusted differences seen in Figures 3-2 and 3-3.

In Table 3-2, the regression results appear in four separate modelsFootnote 23, one for time of entry, then at six months, two years, and four years. Each model, estimated as logistic regressions, represents the predictors of homeownership at a single point in time. The first model looks models home ownership propensities in the first month after arrival, and enables a determination of who buys immediately, followed by an additional model for each of the other three waves. The regression results for time of entry homeownership propensities are essentially a ‘synthetic’ wave of the LSIC, generated from wave 1 retrospective information. As a result, it contains less complete information than the regressions for waves 1-3, but is useful in that it denotes homeownership propensities at time of arrival.Footnote 24

Table 3-2: Logistic regression results for home ownership among LSIC respondents

Variables Wave “0” Wave 1 Wave 2 Wave 3
Odds Sig. Odds Sig. Odds Sig. Odds Sig.
Chinese 0.669 ** 0.498 *** 0.644 *** 0.571 ***
South Asian 0.544 *** 0.378 *** 0.666 *** 0.796 *
Black 0.545 ** 0.347 *** 0.607 ** 0.466 ***
Filipino 0.43 *** 0.285 *** 0.385 *** 0.517 ***
Korean 0.271 *** 0.318 *** 0.623 ** 0.618 **
Latin Am. 0.539 * 0.432 *** 0.69 * 0.876  
Arab 0.574 * 0.374 *** 0.344 *** 0.499 ***
West Asian 0.756   0.604 * 0.775   1.052  
Other Immigrants 0.971   0.697   0.92   0.775  
Age 1.015 *** 1.013 *** 1.009 *** 1.006 *
High School 1.061   0.953   0.761 ** 0.923  
Post-Secondary 0.805   0.978   1.084   0.969  
University 0.715 ** 0.937   0.873   0.884  
# of Children     0.959   1.059   1.074 *
Spouse arrived before respondent 0.809 * 0.836   0.89   1.086  
Multiple Family     4.696 *** 3.788 *** 2.847 ***
Person Living Alone     0.593 ** 0.242 *** 0.2 ***
Lives with Non-Family Persons     1.112   0.314 *** 0.26 ***
One Family with non-family persons     3.604 *** 1.896 *** 1.805 ***
Roman Catholic 1.297   1.573 *** 1.752 *** 1.736 ***
Protestant 1.368 * 1.95 *** 2.072 *** 1.85 ***
Orthodox 0.489 *** 0.355 *** 0.685 * 0.739 *
Jewish 0.452   0.744   1.857 * 1.442  
Muslim 0.579 ** 0.653 * 0.931   0.802  
Eastern Religions 1.219   1.343   1.136   1.112  
# of Jobs     0.86 ** 1.005   0.953  
Income (logged)     1.013   1.002   1.039 **
Primarily Came to Canada for School 0.732 * 1.106   1.188   1.147  
Works Full-time     0.789 * 1.103   1.263 ***
Spouse works Full-time     0.977   1.182 ** 1.34 ***
Credentials Have not been Recognized     0.641 * 0.827   0.589 ***
Bring Savings 0.792   2.418 *** 2.91 *** 2.839 ***
Entry Wealth (logged) 1.03   1.232 *** 1.232 *** 1.202 ***
Refugee 0.046 *** 0.065 *** 0.082 *** 0.215 ***
Economic Class – Prin. Applicant 0.148 *** 0.204 *** 0.446 *** 0.76 **
Economic Class – Spouse and Dep. 0.149 *** 0.242 *** 0.507 *** 0.814 *
Other Economic Class 0.529 *** 0.935   1.606 *** 1.897 ***
Toronto 0.819 * 0.693 *** 0.64 *** 0.678 ***
Montreal 0.237 *** 0.223 *** 0.233 *** 0.236 ***
Vancouver 0.63 *** 0.602 *** 0.547 *** 0.551 ***
Received Housing Advice     0.802   1.341   0.97  
Bayesian Information Criteria 5356   6067   8538   9345  

Source: Longitudinal Survey of Immigrants to Canada

3.4.1 Home ownership at time of entry

At time of entry, age, a university degree, a spouse that was in Canada before the respondent, several religions, class of entry, and CMA of residence are the significant results. In addition to these results, the coefficients for visible minorities show that most groups have significantly lower homeownership propensities than reference group white immigrants. Only West Asians, and Other Immigrants can not be distinguished from Whites at time of entry.

3.4.2 After six months

By six months, with the addition of several variables, a more interesting story begins to emerge. First, those that live in non-nuclear family dwellings differ significantly from those that do. For people living alone, this means that home ownership rates are nearly half those of the reference group (a husband-wife single family household). For multiple family dwellings (which includes single family dwellings with one or more relatives) and single family dwellings with non-family members, homeownership propensities are much higher, suggesting that these forms of cohabitation may signify a strategy for early attainment. Several significant differences across religions are evident, with Roman Catholics and Protestants posting higher rates than those with no stated religion (the reference group), and Muslims and Orthodox Christians positioned well below the reference group.

In terms of employment, simultaneously holding more jobs hampers home ownership prospects, as does working full-time and not having recognized (self-declared) credentials. The findings for credential recognition seem to be straightforward, but the other two are not. Perhaps holding more jobs reflects the quality of the jobs held, and working full-time might compare negatively to not having to work full-time because of the attendant differences in resource levels. Although the result for credential recognition is intuitive, trends for the number of jobs is not. For the number of jobs, the only significant result occurs at six months, and there is no clearly ascending or descending pattern in the coefficients. For working full-time, there is a fairly clear trend towards increased ownership over time, speaking to the importance of stable employment for making a home purchase. An interesting result is that entry wealth has a sizeable effect on the ability to buy a home at wave 1, but that it did not at time of entry. This suggests that immigrants may wait to buy a home even when they have the resources to buy immediately.

Wide differences can also be seen by class of entry, with low propensities of owning for Refugees, and both Economic Class categories, after adjusting for other factors. The reference group (family class) has the highest propensities of any group, although this might be expected since these respondents are joining family members that may already be somewhat established. Home ownership propensities are lower in Montréal, Toronto, and Vancouver than they are in the rest of the country. The extent to which this reflects differences due to the nature of these housing markets versus unobserved factor that lead immigrants to settle outside of the ‘big three’ can not be determined here. Finally, judging from the coefficient for housing advice, it appears that the advice of others is too diverse to elicit a discernable impact on the propensity to buy.

Looking at the visible minority characteristics, most of the significant differences at time of entry (“wave 0”) still exist after six months. In fact, for all groups with significant differences but Koreans, the gap with reference group whites actually widens, suggesting that whites initially moved into home ownership faster than any other group.

3.4.3 After 2 years

Moving on to wave 2, many of the trends witnessed at wave 1 continue to be evident. Age continues to have a positive effect on ownership, education still has a minimal impact, and the disadvantage of low levels of education (defined as less than high school) become evident. The advantage of having a spouse arriving prior to the respondent continues to be negligible. All four household types differ from the reference group in interesting ways. For multiple family dwellings and one family households with unrelated persons, differences with the reference group shrink, suggesting that the tremendous advantage conferred by these cohabitation strategies early on lessens with time. At the same time, not being in a family has a growing negative effect on respondents, with persons living alone and with non-family persons showing home ownership propensities that are far below the reference group.

The results for employment variables at wave 2 change little from wave 1. The effect of the number of jobs an individual holds remains negative, but having a spouse that works full-time now positively predicts home ownership. It is interesting to note that although income continues to elicit no effect on home ownership propensities, entry wealth continues to be a strong and significant determinant.

After two years in Canada, all admission categories differ from the family class, with refugees, and both Economic Class categories tracking well behind the reference group, and, for the first time, Other Economic Class (which are largely composed of Provincial Nominees) surpass the Family Class with their homeownership propensities. Differences across CMAs are essentially unchanged from before, and now receiving housing advice has no effect on home ownership propensities.

At wave 0, whites had considerably higher homeownership rates than all groups but West Asians and Other Immigrants, and only Koreans narrow the gap slightly by wave 1. By wave 2, nearly all groups but Arabs converge slightly with whites relative to their wave 1 standing.

3.4.4 After 4 years

Most of the trends denoted for wave 2 continue on to wave 3. The effect of age attenuates slightly, though it is still significant. Education continues to have little effect, and the number of children now has a discernibly positive effect on propensities. The differences across household formation strategies noted above continue between years two and four. There are few changes to trends by religion, except that the effects continue to shrink in most cases. Jews are once again no longer significantly different from the reference group.

As with wave 2 regression results, having many jobs continues to have no effect on home ownership, and, for the first time, income is statistically significant and has the expected positive effect. Having a working spouse is a stronger predictor of home ownership than the respondent working him or herself.

Once again, refugees have much lower propensities than reference group family class participants, though the gap continues to shrink. Both Principal Applicant and Spouse and Dependent Economic Class respondents have lower propensities than those from the Family Class, but the gap is shrinking. For Other Economic Class, which largely consists of provincial nominees, there continues to be a strong and significant propensity for ownership.

The differences across geographic areas once again changes very little, which might be surprising given the price increases in Toronto and Vancouver over this time.

Looking at visible minority groups, in wave 3 there is only limited evidence of further convergence. Chinese, Black and Korean LSIC respondents lose ground on the reference group white immigrants, whereas South Asians, Filipinos, and Arabs post gains in home ownership. For Latin Americans, this gain is enough to erase the significance of the difference with whites. West Asian and Other Immigrants continue to resemble white immigrants.

Regarding the effect of visible minority status on the ability to own, there is support here for both integration and stratification theories. One group (Latin Americans) that was significantly different from reference group whites at “wave 0” is not by wave 3, bringing the number of indistinguishable groups to three. On the other hand, five groups continue to have home ownership rates that are lower than reference group whites, so the tally lends slightly more support for a stratified housing market, as argued by several other Canadian scholars (Murdie 1994; Henry 1989; Hulchanski 1997; Skaburskis 1996).

To further put this into perspective, however, it is important to note that Figures 3-1 and 3-2 show all groups moving into home ownership in the first four years, with several groups posting home ownership rates that are on par with that of white immigrants. What this suggests is that some groups are working against what might be deemed the forces of stratification. As one example of this, some groups that are known to form multiple family dwellings, such as Chinese and South Asians (Yu and Myers 2007; Haan 2007), may do so in part to secure an owned dwelling, so that even if there is reduced access compared to reference group whites, their strategies for overcoming hardships in the Canadian housing market may nonetheless allow them to buy a home and to integrate into the owner-occupied market.Footnote 25 The rate at which these hardships can be overcome seems to differ; within four years, most groups resemble the reference group more closely in terms of homeownership than they did at arrival, as judged by the visible minority regression coefficients. Finally, although Chinese, Blacks, and Arabs diverge from whites, in each instance the widening of the gap is slight, particularly when compared with the rate at which most groups converge with whites. Arabs actually narrow the gap between waves 2 and 3.

In the section below, adjustments are made for group differences by predicting home ownership propensities while holding all variables but visible minority group at mean values, providing an answer to the question: “what would home ownership rates be if visible minority status was the only difference between LSIC respondents?”

3.5 Plotting adjusted attainment probabilities

In Figures 3-4 and 3-5, homeownership probabilities are predicted at time of entry, and six months, two years, and four years after entry using the regression results presented in Table 3-2 above. The first prediction point denotes home ownership propensities after the first month in Canada, followed by expectations at each wave.

Figure 3-4: Home ownership attainment trajectories of Chinese, South Asian, Black, Filipino, and Latin American Immigrants in the four years after arrival, LSIC

Home ownership attainment trajectories of Chinese, South Asian, Black, Filipino, and Latin American Immigrants in the four years after arrival, LSIC

When holding all variables but visible minority group status at their mean values, predicted differences at time of entry are now largely negligible, with a roughly 2 percentage point gap separating the highest and lowest attainment groups. Furthermore, when confidence intervals (not shown) rather than means are considered, these groups are statistically indistinguishable, suggesting that the difference between groups at time of entry in Figure 3-2 almost entirely stem from compositional characteristics. Other than differences in resource levels, visible minority groups at time of entry more or less face the same frontier. This point is important, because it suggests that visible minority immigrants are relatively equal in terms of access as they approach Canada’s housing market.

What is important to note is that, over time, there is divergence, particularly after wave 1, when groups are still quite tightly clustered. By wave 2, the gap between high (Latin American) and low (Filipino) attainment groups expands to roughly 10 percentage points. This gap, though significant in itself, nearly doubles by wave 3, although it is now Blacks rather than Filipinos with the lowest expected homeownership probabilities. Latin Americans continue to have the highest expected homeownership rates of all the groups in Figure 3-4, at over 50%.

Figure 3-5: Home ownership attainment trajectories of Korean, Arab, West Asian, White, and other visible minority immigrants in the four years after arrival, LSIC

Home ownership attainment trajectories of Korean, Arab, West Asian, White, and other visible minority immigrants in the four years after arrival, LSIC

The groups in Figure 3-5 portray a similar story of growing divergence, which would be expected under stratification. Although the entry gap is larger at nearly 10 percentage points, by wave 3 the roughly 20 percentage point gap between high and low attainment groups (West Asians and Arabs, respectively) is about the same as the five groups in Figure 3-4. Whites, which represent the benchmark group for considering levels of integration, do have one of the highest expected rates at wave 3, but they are not that far ahead of other groups, as might be expected under stratification.

In both instances, these plots hold all other variables constant, so earnings, family structure, and every other home ownership relevant characteristic is identical across groups. These figures therefore represent what home ownership rates ‘should’ be in the face of equal opportunity and constraint.

3.6 Comparing predicted and actual homeownership rates

The probability plots above are useful in that they provide an indication of what homeownership propensities would be for immigrant visible minority groups with identical resource levels. Comparing the gaps in Figures 3-2 and 3-3 to those in 3-4 and 3-5 (Table 3-3), it appears that roughly half of the differences between groups in home ownership propensities stems from compositional characteristics, and that attainment trajectories would be much more similar if other group characteristics were the same. The other half of the difference remains unexplained, once again providing equal support for integration and stratification.

In many ways, however, this brings us back to the question posed at the beginning of this section regarding the difficulties around defining how immigrant groups integrate into Canadian society. Table 3-3 below, which was generated by subtracting predicted values from observed rates at each time point illustrates this more poignantly. This table shows that most groups actually post higher than expected homeownership rates, and that although there are differences between them, nearly all groups perform better than expected in terms of home ownership attainment. For some groups, like Chinese, South Asians, and Filipinos, this difference is substantial. Of all groups, whites approximate expectations most closely, with a less than one percentage point gap between actual and observed rates at wave 3.

Table 3-3: Percentage point difference between observed and expected homeownership rates

Visible Minority Status Time Point
0 1 2 3
Chinese 4.1 7.4 7.2 6.1
South Asian 11.3 13.6 11.8 8.3
Black 6.5 5.8 0.7 -1.4
Filipino 11.3 14.6 18.6 19.4
Latin American 6.5 6.9 6.8 2.1
Korean 1.0 4.2 15.8 8.7
Arab -2.4 -6.9 -7.3 -11.4
West Asian -0.6 -1.4 -7.0 -17.5
White 5.2 2.2 2.3 -0.7
Other Visible Minority 7.0 11.7 10.1 8.3

Source: Longitudinal Survey of Immigrants to Canada
Note: Numbers above represent the gap in percentage points between actual and expected homeownership rates.

These higher than expected rates suggest that attainment is proceeding more quickly than expected for most groups, and that the gaps between groups largely exist at higher than expected home ownership levels.Footnote 26 Further evidence of this comes from the fact that nearly three-quarters of all groups have home ownership rates that approach or surpass 50% after only four years in Canada. Considering that new immigrants are often treated like new labour market entrants (Picot and Sweetman 2005), this rate of attainment is especially impressive, and is unlikely to be matched by new Canadian-born labour market entrants.

At the same time, there are significant disparities for both Arab and West Asian immigrants, with homeownership rates that are 11.4% and 17.5% lower than what they ‘should’ be based on respondent characteristics at wave 3. Under stratification theory, these disparities suggest that these groups face a different reception in Canadian society than other immigrant groups.

An alternative explanation for the disparity, one that does not involve differential treatment by the host society, is that over 4/5 of Arab and West Asian LSIC respondents identified as Muslim in the survey, and that this in itself alters their access to homeownership due to beliefs concerning Riba.Footnote 27 Given that the primary means through which houses are purchased by immigrants in Canada (a mortgage) is not permissible for many Muslims, it is not surprising to see lower than expected home ownership rates for Arabs and West Asians.

Changes to the mortgage industry would likely have a profound effect on access to home ownership for Arabs and West Asians. A small but growing sector of the mortgage industry has begun to offer ‘Islamic Mortgages’, which adhere to the mandates of Sharia law. It will be interesting to see how this affects home ownership propensities for these groups in the future.

3.7 Changing characteristics over time

In Table 3-3 we saw considerable evidence of many groups moving in to homeownership at higher than expected rates, after entering Canada with more or less equal adjusted homeownership rates at time of entry. After spending some time in Canada, however, a nearly 20 percentage point gap emerges between visible minority groups. Arabs, West Asians, Blacks, and Whites to a much lesser extent, did not buy homes at rates that would be expected with their socioeconomic resources, whereas all other groups bought homes at higher than expected rates.

What is interesting about this finding is that the differences seem to emerge while immigrants are in Canada, and part of the LSIC observation period. It is therefore possible to measure whether the ability to buy a home stemmed from a change in state, such as the recognition of credentials, receipt of citizenship, the addition of another child, a change in household composition, or improvements in labour market performance.

In the final section of this report, we take a look at how changing characteristics in the lives of LSIC respondents alters their probability of ownership. The model below looks at the determinants of owning at wave 3, looking only at those that did not immediately buy upon entering Canada. The focus is on how the transition is accompanied by other changes in an LSIC respondent’s life circumstances.

The dependent variable for this logistic regression model is dichotomous, set to ‘1’ if a household buys during the observation period and ‘0’ if they did not. Most of the variables in Table 3-1 are used as predictors here, but in addition to these variables are several new pieces of information. First, rather than looking at household cohabitation characteristics as static characteristics, the model below looks at the possibility of a change between waves 1 and 3 (such as moving from a multiple family dwelling to a single family dwelling). In addition to this, acquiring citizenship, the addition of a child, and improvements in employment situation are also included as factors that possibly changed between waves 1 and 3. In all instances, the new variables are dichotomous and set to one if a change occurred. Zero denotes no change in status on the variable of interest.

Table 3-4: The changing lives of LSIC respondents and the relationship to home ownership

Variables Odds Sig.
Chinese 0.613 ***
South Asian 1.002  
Black 0.47 ***
Filipino 0.656 **
Korean 0.742  
Latin Am. 0.948  
Arab 0.561 ***
West Asian 1.153  
Other Immigrants 0.651  
Age 1.008 **
High School 0.912  
Post-Secondary 0.935  
University 0.915  
Spouse arrived before respondent 1.157  
Roman Catholic 1.829 ***
Protestant 1.753 ***
Orthodox 0.844  
Jewish 1.22  
Muslim 0.835  
Eastern Religions 1.263  
Primarily Came to Canada for School 1.144  
Bring Savings 3.222 ***
Entry Wealth (logged) 1.226 ***
Refugee 0.294 ***
Economic Class – Principal Applicant 0.878  
Economic Class – Spouses and Dependants 1.091  
Other Economic Class 2.051 ***
Toronto 0.709 ***
Montreal 0.241 ***
Vancouver 0.584 ***
Income (logged) 1.093 ***
Changing Characteristics between Waves 1-3
Addition of Another Child 1.143 *
Respondent gets Canadian Citizenship 0.757 ***
Moving from multiple dwelling to single 0.671 ***
Respondent Stays in Mult-family dwelling 2.429 ***
Spouse gets a full-time job 1.335 *
Respondent gets a full-time job 1.111  
Respondent’s credentials receive recognition 0.784 *
Respondent’s work experience receives recognition 1.249  

Source: Longitudinal Survey of Immigrants to Canada

When comparing the model above to those in Table 3-2, several interesting differences emerge. First, there are now only four significantly different visible minority groups, further suggesting that groups more or less proceed into home ownership at the same rate. Once again, however, there are some standouts. On the low end in terms of attainment, Arabs, Blacks, Filipinos, and the Chinese emerge. For South Asians and Koreans, the differences that exist between most groups appear to dissipate when transitions (Table 3-4), rather than absolute levels (Table 3-2), are studied.

Once again, only Roman Catholics and Protestants differ from those with no religion. In both instances, the rate of attainment is much higher than it is for those with no religion. It is difficult to know for certain why this is the case, although it may be because adherents to these two religions gain access to Canadian-born and more established immigrant communities, and can therefore rely on their fellow church members to help them navigate the real estate and mortgage market. Without further research, however, it is difficult to know for certain.

Furthermore, many of the variables designed to measure change are statistically significant. The addition of a child encourages households to buy a home, suggesting that children are an important part of a household’s dwelling tenure decision. For some reason, gaining citizenship puts negative pressure on ownership, possibly because the time and cost associated with gaining citizenship takes resources away from homebuying. Once again, strong evidence emerges about household formation being a strong determinant of tenure type. Respondents that are in multiple family dwellings at both points in time are more than twice as likely to make the transition to owner than are those in single family dwellings across the period (the reference group). Moving from a multiple family dwelling to a single family dwelling hurts home ownership propensities. Too few households went from a single family dwelling in wave 1 to a multiple family dwelling in wave 3 to enable an analysis of this transition, although it too would have been interesting to look at.

As with citizenship (and perhaps for the same reason), having credentials recognized detracts from homebuying, whereas work experience recognition elicits a positive effect. Spouses also play a role in home buying, and their transition into full-time employment increases the probability of home buying by roughly 34%.

This model supports earlier assertions that many groups proceed in to home ownership at more or less the same pace, relative to white immigrants, once resource levels are controlled for. Of the ten visible minority groups in the models, only four have distinct trajectories from whites. If any evidence for stratification exists in Table 3-4 (and Table 3-2), it comes from the trend among Arab, Black and Chinese immigrants. Since Arab immigrants are relatively recent arrivals, however, there is very little research among this group, so it is probably premature to interpret their housing trends as evidence of stratification, even though the evidence above supports such a claim. More work needs to be done to understand why this group diverges so significantly from both white immigrants and several other groups before such claims can be made more forcefully.

Both Haan (2007) and Painter, Gabriel and Myers (2000) highlight Chinese immigrants as having one of the highest rates of home ownership of all immigrant groups, at times exceeding even whites and the native-born population. Given this, it is interesting to see how Chinese in the LSIC sample do not fare as well as their predecessors, a finding that may be due in part to the fact that many came from mainland China instead of Hong Kong, and may therefore not have the same resource levels. In any event, it is hard to imagine that Chinese attainment levels stem from stratification, since it appears to be cohort-specific rather than systematic.

Regarding Black immigrants, low rates of attainment have been documented elsewhere in both Canadian (Haan 2007) and US research (Painter, Gabriel, and Myers 2000). The reasons behind low attainment rates are unknown. US research has suggested that Black immigrants are pushed into the racialized African-American underclass, which would explain low rates in the US. Comparative research on Black immigrants in Canada, however, finds similarly low attainment rates, without the racialized underclass. Even so, the only remaining group that consistently supports stratification theory is Blacks.

3.8 Conclusion

This report suggests that many newcomers in the LSIC sample face considerable difficulties with housing market affordability when they first arrive in Canada. They not only spend much more on housing than Canadians on average, but they also earn much less, contributing to a ‘double-burden’. The situation appears to improve quickly after arrival, although the initial transition could be extremely difficult for many households. These experiences differ by age, census metropolitan area of residence, class of entry, education, region of origin, and visible minority status.

Not surprisingly, there are also differences by admission category; economic class migrants do not face the hurdles that refugees do, for example, although even refugees move quickly into ownership. Those who come under the family class have fewer difficulties with affordability, moving into ownership, although this may stem in part from having family members already in the country, thereby giving them a ‘head-start’ compared to other LSIC respondents. Skilled workers and provincial nominees both enter Canada with low propensities, but they make considerable gains over the four-year period, suggesting that they too require some time to settle in to life in Canada before they begin to experience residential mobility.

Wide differences also exist across census metropolitan areas. Although this report could not determine the extent to which this stems from peculiarities inherent in each private housing market versus the availability of subsidized transition housing, both factors likely play a role. In Edmonton, for example, the waiting list for subsidized housing far exceeds the supply, so newcomers must often rely on charity or the private market for accommodations. Too often, this presents them with an affordability crunch or no choice but to live in sub-standard housing, or both.

The primary purpose of this report, particularly section 3, was to expand upon differences by visible minority, and to focus on the factors that lead to such rapid changes in housing tenure. Focusing on home ownership as an outcome, this report has sought to understand why Canada’s immigrant visible minority groups experience the housing market differently. Findings reveal substantial unexplained differences across groups, equating to a 20 percentage point gap, both between groups and between the observed and expected home ownership propensities of a group. This suggests that some groups face additional hurdles in Canada, and that this may be an area for policy intervention. Canada’s public and private housing market leads to different housing experiences among visible minority groups. This report flags Muslim immigrants and their financing considerations in particular. Others have noted the housing difficulties faced by blacks (Murdie 1994), which this report supports, although it would be difficult to prescribe an intervention without looking at the housing issues of this group more closely.

This report also highlights the considerable accomplishments of many Canadian newcomers. Home ownership rates on average approach 50% by the end of the observation period (for those that remain in the sample), which is impressive given that many newcomers struggle to navigate the Canadian labour market, Canadian society, and, despite this, were able to find a home and community to invest in.

The results of section 3 suggest that religion may be a more effective way to examine the early experience of immigrant groups rather than by visible minority status at least in terms of housing. Informed by the substantial body of (often US) literature that points to the importance of skin colour for understanding access to socio-economic resources, researchers have paid too little attention to religion, even though it is more directly reflective of a belief system than skin colour. As argued over 100 years ago by Max Weber, belief systems are at the core of so many human behaviours, yet most research tends to ignore this. A follow-up study to this report could look at the role of religion on housing outcomes more directly.

Another interesting finding in this report is that entry wealth does not have an appreciable effect on homeownership propensity until after respondents have spent some time in Canada. This is somewhat counterintuitive, since presumably LSIC respondents with entry wealth should buy homes immediately. Perhaps this is because individuals choose to wait until they know more about their new destinations before making an investment. An interesting follow-up study would be to investigate how newcomers use their wealth in the housing market, and whether this varies by other characteristics (like admission category or census metropolitan area of residence).

Although the intention of most studies is to be able to derive a series of generalizable statements about the early experiences of all Canadian immigrants, several caveats in this report complicate the potential for this to occur here. First, given the importance of the business cycle for determining an immigrant’s early success, arriving in 2000-2001 is likely to have implications for generalizability. Canada did not experience the early 2000s recession as deeply as the United States, but it was also not immune from it. Arriving during a recession tends to affect immigrants on average, but it may have elicited differential effects on immigrant sub-groups.

The broad storyline of this report, however, is that immigrants do quite well in Canada’s housing market, and that many are willing to take extraordinary steps to move in to owner-occupied housing. Differences across groups do exist, but for most the differences are between groups that have higher than expected home ownership propensities.

Page details

Date modified: