Toronto and Vancouver bound: the location choice of new Canadian immigrants.
Subject: Urban population (Research)
Ethnic groups (Behavior)
Immigrants (Behavior)
Immigrants (Forecasts and trends)
Author: McDonald, James Ted
Pub Date: 06/22/2004
Publication: Name: Canadian Journal of Urban Research Publisher: Institute of Urban Studies Audience: Academic Format: Magazine/Journal Subject: Social sciences Copyright: COPYRIGHT 2004 Institute of Urban Studies ISSN: 1188-3774
Issue: Date: Summer, 2004 Source Volume: 13 Source Issue: 1
Topic: Event Code: 310 Science & research; 010 Forecasts, trends, outlooks Computer Subject: Market trend/market analysis
Product: Product Code: E198450 Immigrants
Geographic: Geographic Scope: Canada Geographic Name: Canada; Canada Geographic Code: 1CANA Canada
Accession Number: 115498147
Full Text: Abstract

Ethnic clustering plays an important role in the location choice of new immigrants to Canada. A concentration of people in the same geographic area who are of similar ethnic background, culture and language can be an important source of financial and personal support, information and guidance, and social mores. Evidence is found that, after controlling for a range of other observable and unobservable factors important to the location decision, the concentration of people of a particular ethnic group in a particular area has a significant effect on the chance that new immigrants of the same ethnic group will choose to live there. The size of the effect, however, depends on the personal characteristics of immigrants. The attraction of ethnic concentrations of immigrants is significantly smaller for new immigrants with a university degree and for those who normally speak English at home. Characteristics of the local ethnic community related to education and language profiles are also significant determinants of location choice.

Keywords: Immigrants, Mobility, Neighborhood Effects

Resume

L'existence de grappes de groupes ethniques joue un role important dans le choix d'etablissement de nouveaux immigrants au Canada. La concentration en une region geographique de gens d'une ethnicite, d'une culture et d'une langue semblables a celles des nouveaux arrives peut s'averer pour eux une source importante d'appui financier et moral, d'informations, de conseils et de moeurs. Apres avoir neutralise certains facteurs observables et non observables jouant un role important dans la decision quant au lieu d'etablissement, nous avons trouve que la concentration de gens appartenant a un groupe ethnique donne dans un endroit donne augmente de facon significative la possibilite que des nouveaux immigrants du meme groupe ethnique choisiront de s'etablir dans la meme region. Toutefois, l'ampleur de cet effet depend des caracteristiques des immigrants. Les nouveaux immigrants qui detiennent un diplome universitaire ou qui parlent anglais a la maison sont moins attires par le rassemblement en grappes de groupes ethniques. Les caracteristiques de la communaute ethnique locale quant a la scolarite et le profil linguistique constituent egalement des facteurs determinants importants dans le choix de lieu d'etablissement.

Mots-cles: immigrants, mobilite, externalite de voisinage

Introduction

It is well known that many recent immigrants to Canada choose to live in Canada's largest cities, in particular, Toronto, Vancouver and Montreal. Over the period 1996-2001, Toronto gained 445,000 new immigrants (or 19 per 1000 of population). Vancouver gained 180,000 new immigrants and Montreal gained 126,000 new immigrants (CIC, 2000). Less understood is what drives these location decisions. One obvious reason is that immigrants, like native-born Canadians, are drawn to Canada's major cities where economic conditions are good and jobs are readily available. The presence of large concentrations of immigrants already resident in some areas also acts as a magnet for new immigrants, thereby adding to population growth rates in areas already experiencing significant population increases because of good economic conditions. Canada's largest cities already have large, established immigrant populations. In 1996, over 60% of Canada's total population of people born overseas was located in Toronto, Montreal and Vancouver, compared with approximately 27% of the Canadian-born population.As new immigrants to Canada continue to settle primarily in Toronto, Vancouver and Montreal, federal and provincial governments are becoming increasingly concerned with the problem of how to attract new immigrants to other areas. (1) In fact, there is a perception among many people in government and the media that some areas get "too many" immigrants while many other areas get "too few" immigrants. Thus, it is argued that interventionist policies are needed to induce immigrants to settle in other areas. (2) Clearly, location matters. The location decisions made by immigrants directly affect the economic welfare of the domestic population. (3) Positive effects come from stimuli to the local economy provided by new immigration, particularly if immigrants have investment capital or skills in short supply. Negative effects come from increased strain on urban infrastructure and increased use of health services, income support and other social programs. As well, location decisions by immigrants affect their own subsequent social and economic adjustment. This is due to employment opportunities, access to settlement and language programs, and support from the local ethnic community which, in turn, affects the contributions that immigrants can make to the local economy. This is an important research question. Recent work by Picot and Hou (2003) generally finds little significant evidence that labour market outcomes of immigrants are affected by exposure to own-ethnic group enclaves within Canada's largest cities.

The objective of this paper is to identify the extent to which geographic clustering of immigrants already resident in Canada affects the location decisions of new immigrants. A concentration of people in the same geographic area who are of similar ethnic background, culture and language might be a necessary source of financial or personal support, information and guidance, and social mores for new immigrants. More generally, the ethnic community can provide a sense of belonging and security to immigrants moving to a new and unfamiliar country, and large communities of immigrants may imply greater availability of services directly tailored to new immigrants. If geographic clustering does play a significant role in location choice, then improved economic conditions on their own might not be sufficient to attract immigrants to areas where there are not already established immigrant communities.

The plan of the paper is as follows. After a review of previous work, I outline the data sources, methodology, sample selection and creation of the measures of ethnic geographic clustering. The key empirical difficulty is in separating the role that ethnic concentrations play in attracting new immigrants from the myriad other factors that might influence the location choices of immigrants. I then illustrate and discuss the main results, with reference to predicted geographic distributions of recent immigrants under alternative assumptions about the geographic distribution of immigrants already in Canada, holding all other factors constant, I conclude with implications for economic policy.

Literature and Methodology

The importance of ethnic clustering to location choice links directly to the concept of ethnic networks that has been developed in the literature. Ethnic networks refer to the externalities, both positive and negative, that arise from living in proximity to people of similar culture, ethnicity or language. The role of ethnic enclaves in immigrant location decisions has been a feature of American research. (See, for example, Bartel, 1989; Zavodny, 1999; Funkhouser, 2000; and Chiswick and Miller, 2002b.) A typical result of these papers is that the presence of other foreign-born people is an important determinant of the location decision of immigrants. The links between ethnic enclaves and other dimensions of economic behavior of immigrants have been the subject of a growing body of research. This includes links with educational attainment (Borjas, 1995; Cardak and McDonald, 2004), welfare participation (Borjas, 1999; Bertrand et.al., 2000; Dodson, 2001), language and earnings (Chiswick and Miller, 2002a), and occupation, employment and earnings (Picot and Hou, 2003). There is also an important literature that has focused on the effects of immigrant location choices on the mobility patterns of the native-born population or immigrants already resident in the US. See, for example, Frey (1996), Borjas (1998), and Funkhouser (2000). An extension of this concept is that the demographic and economic dimensions of the local immigrant community are also important. It is not just the size of the community but its characteristics that affect location choice. For example, ethnic communities made up of younger people or more recent arrivals might imply stronger links with the home country language and customs. This may have a strong attraction for new arrivals. In the empirical analysis, I consider a range of measures that capture important demographic and labour market characteristics of particular ethnic groups already resident in particular Canadian regions.

Canadian economic literature specifically on the geographic mobility of immigrants is more limited. Lin (1998) uses the Labour Market Activity Survey of Statistics Canada to compare the inter-provincial mobility of foreign-born and native-born Canadians. He finds that, while immigrants are relatively less likely to move provinces, this difference is due to distributional and compositional differences between immigrants and native-born Canadians rather than to structural differences in the determinants of mobility. Using the 1986 Canadian Census to examine inter-provincial migration of immigrants to Canada, Newbold (1996) finds that, after controlling for personal and ecological effects, country-of-birth explains relatively little of the observed differences in mobility between immigrants and the Canadian born. However, neither paper considers the initial location choice of immigrants. (4)

In this paper, I model the actual choice of where an immigrant chooses to settle in Canada by using a conditional logit specification that includes both characteristics of the choice (in particular, measures of ethnic clustering) and characteristics of the chooser to affect location choice. It is possible to measure the influence of ethnic enclaves in a variety of ways. I follow Bertrand et.al. (2000) and compute the proportion of a region's population that is from a particular region of origin and scale it by dividing by that ethnic group's share of the total Canadian population. (5) It might also be expected that the attraction of local enclaves for new immigrants would be stronger for immigrants who are more likely to experience difficulties in terms of employment, skill recognition, or social adjustment, such as individuals from non-English speaking backgrounds or individuals without tertiary educational qualifications.

In order to identify the true effects of ethnic clustering, it is necessary to control for a range of observable and unobservable factors that might also affect an individual's choice about where to settle. These include but are not limited to earnings and employment opportunities, generosity of social assistance, cost of living, amenities, climate, and financial and non-financial moving costs. To reflect these factors, I include a set of individual-specific economic and demographic characteristics typically found to be significant determinants of mobility in the literature. These are age (which enters as a quadratic), education level, marital status, presence of children in the household, language fluency, and local economic conditions. Thus, factors such as language and education can have both an indirect effect on location choice by affecting the magnitude of any ethnic clustering effect, as well as a direct effect.

It is also important to control, as far as possible, for potentially important unobserved factors that may affect location choice and be correlated with measures of ethnic clustering. Failing to control for any unobserved factors that affect location choice may give rise to a finding of a significant impact of ethnic enclaves on location choice when none in fact exist. For example, Toronto may be an attractive destination choice for new immigrants from a certain region because of its amenities or economic opportunities. A relatively high concentration of immigrants from that region already resident in Toronto will reflect the same general factors. Failing to control for these factors will, however, give the impression that the high concentration of immigrants itself leads to more recent immigrants choosing to settle in Toronto. I include a set of region dummy variables to reflect time-invariant region-specific characteristics that are common to all immigrants (such as urban infrastructure and climate). I also include a time dummy variable interacted with each of the regions to reflect time-varying region-specific characteristics (such as economic conditions, population growth, changes in cost of living or changes in the availability of government settlement programs). (6) As well, I include a full set of ethnic group indicator variables that allow particular ethnic groups to have intrinsic preferences for particular locations. For example, proximity to the Pacific Rim might make Vancouver an attractive destination for immigrants from China and Southeast Asia. (7)

Data Sources

The data are drawn from the public use files of the Statistics Canada Census for 1986, 1991 and 1996. In order to focus on the initial location decisions of new immigrants, I assemble data on two groups of immigrants. The first group consists of those people in the 1991 Census born outside Canada who arrived in Canada within the five-year period 1986-1991. The second group consists of those people in the 1996 Census born outside Canada who arrived in Canada within the five-year period 1991-1996. In order to exclude those people whose initial location decision reflected choice of educational facility, I restrict the sample to adults aged to 25 and 64 years only. I also restrict the sample to men.

I define the following distinct regions in the analysis: Montreal, rest of Quebec, Toronto, rest of Ontario, Prairies (Manitoba and Saskatchewan), Alberta, Vancouver, rest of British Columbia. Region-of-birth categories for immigrants living in the Atlantic Provinces and Canadian Territories are too broadly defined to be useful. Thus, I omit individuals living in these areas of Canada from the sample. (8) I also define 16 distinct regions of birth outside of Canada, including seven for Europe, U.K. and Ireland, French-speaking Europe, other Western European countries, Mediterranean countries, Poland, Slavic countries, and other European countries. I define four regions for Asia, Middle Eastern and Western Asian countries, South Asian countries, China and Hong Kong, and South East Asian countries; and five for the rest of the world: African countries, the Caribbean, South and Central America, USA, and Other regions.

Measures of ethnic enclaves and their characteristics are computed using characteristics of the resident Canadian population inferred from the most recent Canadian Census prior to the five-year migration window of each arrival group. Thus, for recent arrivals in the 1991 Census, I compute ethnic group characteristics by region using statistics drawn from the 1986 Census. For recent arrivals in the 1996 Census, I use the 1991 Census. To measure the characteristics of particular ethnic clusters, I construct time-, region- and ethnic group-specific measures of the following characteristics: average age, unemployment rate, participation rate, average earnings, average years since migration, proportion of immigrants who arrived in Canada in the last 10 years, proportion of immigrants who normally speak another language at home, and proportion of immigrants with a university degree.

Empirical Results

In McDonald (2003), I provide a descriptive overview and discussion of the settlement choices and the socio-economic and demographic characteristics of recent immigrants to Canada, compared to both earlier immigrants and native-born Canadians. To summarize, recent immigrant men aged 21-65 are significantly more likely than either earlier immigrants or native-born Canadians reside in Canada's three largest cities. (There are over 40% of recent immigrants located in Toronto alone). They are also more likely to be younger and to have a university degree, compared with other immigrants and Canadian-born men, and substantially more likely to speak a language other than French or English at home.

Since coefficient estimates obtained from estimation of the conditional logit do not have a simple economic interpretation, I discuss estimation results in two parts. First, I present coefficient estimates that indicate direction of effect and statistical significance. For brevity, I only report results relevant to the effects of ethnic enclaves on location choice. (9) Secondly, I examine the magnitude of the effects of key variables by predicting individual location choice for recent immigrants across a range of personal characteristics and over alternative population structures in order to gauge economic significance.

In the first column of Table 1, it can be seen that neither the relative size of the enclave (relative concentration) nor two characteristics of the enclave (average age and other language normally spoken at home) has an effect on location decision, although the educational profile of the local ethnic group does seem to matter. The larger the proportion of that group with a university degree, the more likely it is that a new immigrant of the same group will choose to live in that area. However, as can be seen in column 2, allowing the effects to vary by education and language spoken at home, yields markedly different results. Focusing first on the relative concentration interactions, an immigrant who normally speaks another language at home is more likely to choose to settle in an area with a higher relative concentration of people from the same ethnic group. This is indicated by the positive and significant coefficient on the relative concentration measure interacted with an indicator variable for language spoken at home. (10)

In contrast, both the coefficient on the interaction of relative concentration and an indicator variable for university degree are negative and significant. This implies that, for immigrants with a degree that speak another language at home, the enclave effect is significantly reduced. For immigrants with degrees who speak French or English at home, the enclave effect is actually negative. This latter group of immigrants is less likely to choose to live in an area with a relatively high concentration of people from their ethnic group, ceteris paribus. For immigrants who speak English or French at home and who do not have a degree, the location decision is not significantly affected by the relative concentration of people from the same region of origin.

Turning next to the other characteristics of the immigrant enclaves, all three dimensions included (average age, proportion of people speaking another language at home, and proportion of people with a university degree) significantly affect a new immigrant's location choice. First, new immigrants are more likely to choose to settle in a region where current residents from the same region of origin are generally younger. Since immigrants themselves tend, on average, to be relatively younger than the general population, recent immigrants might share closer social and economic ties to a younger resident immigrant population in a particular region. The effect is, however, significantly reduced for new immigrants with a university degree. This is consistent with the results reported above on the relative size of the ethnic enclave that university education mitigates the size of the enclave effect. (11)

The second enclave characteristic included is the proportion of the local ethnic enclave that speaks a language other than English or French at home. This variable is negative and highly significant. Its interaction with individual language spoken at home is positive and also highly significant. The net additional effect of these terms is zero for most recent immigrants, so that the clustering effects discussed earlier are largely unaffected. However, for ethnic groups where the usual language at home is not French or English, new immigrants who do speak French or English at home are significantly less likely to settle in areas where more of their ethnic cohort speaks another language. Thus, it appears that new immigrants who perhaps are already more acculturated into Canada (using language as an indicator) are also more likely to settle where their ethnic group is more acculturated.

The third enclave characteristic included is the proportion of the local ethnic enclave that has a university degree. For all immigrants without a degree, this characteristic has no effect on their location choice. For immigrants with a degree, there is a positive correlation, implying relatively educated immigrants are more likely to settle where their ethnic community is also relatively educated. Thus, in addition to the direct effects (not reported here) of educational attainment on location choice, it also affects the magnitude and direction of effect of enclave characteristics on a new immigrant's location choice.

A potentially important extension to the basic specification is to incorporate a role for visa category to affect an immigrant's location decision. For example, immigrants arriving on "family" class visas will have residential ties to immigrants already resident in Canada. As well, they might be more likely than immigrants arriving in the "independent" or "business" class to expect to rely on support from concentrations of people of the same ethnic background. Although visa category is not available in the Census data, some insight into the importance of visa category can be gained by combining Citizenship and Information Canada's immigrant yearly flows data by visa category and country of origin with the Census data. Using these data, I am able to control for the average visa characteristics of each individual's particular ethnic group in the year that the individual arrived in Canada. Though not reported here, results indicate that the positive effect of relative concentration on the location choice of new immigrants is even larger for those immigrants who are more likely to have arrived on a family class visa. However, the coefficient estimate is significant only at the 10% level and its inclusion has little effect on the other results.

Predictions

In the first part of the econometric results, evidence was found that the relative concentration and characteristics of ethnic enclaves exert a statistically significant influence on the location decisions of recent immigrants. To investigate whether the results also have economic significance, I use them to predict the probability that a recent immigrant with a given set of characteristics will settle in each of the eight Canadian regions I have defined. I then simulate what the probabilities would have been under the alternative hypothesis that immigrants from that ethnic group already resident in Canada are instead distributed across Canadian regions identically to native-born Canadians, holding everything else (including unobservable effects) constant. Comparing predicted location choices across these alternative distributions of immigrants already resident in Canada gives some insight into the magnitude (and so economic importance) of the effect of enclaves on location choice. For exposition purposes, I use an individual with the following set of characteristics as my "base case": age 30, single, no children at home, and high school education only. The general conclusions of the simulations are not sensitive to the choice of "base case," except as noted.

In Table 2a, I present the predicted location choices of a base-case individual from China who speaks another language at home and does not have a university degree. In the first row of the Table, predictions are based on the actual relative concentrations of immigrants from China already resident in Canada at the time of 1986 Census. It is clearly evident that this base-case immigrant is far more likely to settle in Toronto (47.6% probability) and Vancouver (29.8% probability) than anywhere else in Canada. Row 2 contains the predicted location choices of the same immigrant but under the hypothetical assumption that immigrants from China already resident in Canada are dispersed according to the distributional pattern of native-born Canadians. Under this assumption, recent Chinese immigrants would be much less likely to settle in Vancouver (11.4% compared to 29.8%) and more likely to settle in most other regions of Canada. That is, new Chinese immigrants to Canada would be distributed significantly more widely across Canada if Chinese immigrants already resident in Canada were not so concentrated in Vancouver. Specifically, they would be more likely to settle in Ontario (outside Toronto), Montreal, the Prairies and Alberta. However, regardless of the distributional assumption, close to 50% of recent Chinese immigrants would still choose to reside in Toronto.

The next two rows of Table 2a repeat the simulation for the same hypothetical immigrant from China except that the immigrant is assumed to speak English at home. Compared to rows 1 and 2, this immigrant is more likely to reside in Toronto and less likely to reside in Vancouver. The magnitude of the clustering effect is uniformly smaller. In other words, the relative concentration of people of the same ethnic group has only a small effect on the location choice of immigrants who speak English at home. If the immigrant has a university degree (and speaks another language at home), the predicted location choices are similar to row 1 but changing the distribution of Chinese already in Canada has a smaller effect on these choices. For example, the proportion of university-educated immigrants who choose to live in Vancouver is predicted to fall from 27.7% to 15.6%, if the Chinese in Canada were distributed across regions similarly to native-born Canadians. These results illustrate the reduced attraction of ethnic concentrations for university-educated immigrants.

Given the importance of ethnic group to location choice, Tables 2b and 2c report results of simulations for immigrants from other regions of origin. Table 2b presents results for recent immigrants from Poland. It shows that new Polish immigrants are predicted to reside primarily in Toronto and in the rest of Ontario. However, if immigrants from Poland already resident in Canada were distributed across Canadian regions similarly to native-born Canadians, the probability that a recent Polish immigrant will choose to settle in Toronto falls from 40.7% to 31.7%, and the probability the immigrant will reside elsewhere in Ontario increases from 32.7% to 44.7%. For those immigrants who speak English at home, the magnitude of effect of ethnic clustering has little effect on the predicted location choice of new immigrants. Similarly, if the new immigrant has a university degree, the enclave effect is substantially smaller than for comparable immigrants who do not have a degree.

Table 2c reports results for immigrants from non-Arab Africa. Since a significant proportion of these immigrants speak French, I simulate settlement patterns for French-speaking immigrants from this region. The probability that a new immigrant from Africa who speaks another language at home settles in Montreal is relatively high (25.6 %), compared to other ethnic groups. As before, if existing immigrants from these areas were distributed across Canada similarly to native-born Canadians, new immigrants would be less likely to settle in either Montreal or Toronto, but language spoken at home is key. If the new immigrant speaks French at home, he is overwhelmingly likely to reside in Montreal (probability = 84%). This figure falls only marginally if the distribution of immigrants from Africa already resident in Canada resembles the Canadian-born distribution. Thus, once again, ethnic clustering effects are significant only for immigrants who normally speak a language other than French or English at home. (12)

Other simulations (available on request) indicate that the magnitude of effect of changes in visa category, age, education and language characteristics of the ethnic enclaves can also be economically significant. For example, if the proportion of Chinese in Vancouver with a university degree increased by 5%, new immigrants from China would be 3% more likely to choose to settle there. If the proportion of Chinese immigrants residing in Vancouver who speak another language at home rose by 5%, new immigrants from China who speak English at home would be 2.5% less likely to settle there. Finally, if the average age of Chinese immigrants residing in Vancouver fell by 2 years, new immigrants would be 2% more likely to settle there.

Discussion and Conclusions

The economic opportunities offered in Canada's largest cities are a magnet to people living in other areas, including immigrants to Canada. Thus, policymakers hoping to encourage new immigrants to settle in other areas of Canada face an intrinsically difficult task. The results of this paper provide additional reasons why dispersing immigrants across Canada will be difficult, but it also offers some insights into how policymakers might move forward.

The key result of this paper is that, even after controlling for a wide range of observable and unobservable determinants of location choice, many new immigrants choose to locate in particular regions because relatively high concentrations of immigrants from the same ethnic group already reside there. However, the magnitude of this effect varies markedly by personal demographic characteristics. In particular, the attraction of ethnic concentrations is strongest for immigrants who normally speak another language at home and who do not have a university degree. For example, simulations indicate that immigrants in this category who emigrated from China would be 18% less likely to settle in Vancouver, if Chinese immigrants already present in Canada were distributed across the Canadian regions similarly to the Canadian-born. The clustering effects are weakest for immigrants who normally speak English or French at home or who have a university education. There are no observable clustering effects for immigrants from English-speaking countries. Thus, given the high concentration of new immigrants that originate from non-English speaking regions, it is likely that the geographic distribution of future immigrants will continue to reflect a strong concentration in Canada's largest cities because of the current concentration of immigrants in those cities.

The results also suggest that immigrants with more human capital are more likely to base their location choice on economic opportunity. This means that the absence of a strong immigrant base in an area would not preclude increased international migration to the area as long as economic prospects were promising. If local ethnic communities can become established in other areas, then those areas will become more attractive destinations for new immigrants from the same ethnic background. Simulations reported in the paper suggest that the magnitude of this spillover effect, in terms of future immigration, could be substantial. Some indirect support for this conclusion is provided by anecdotal evidence from Manitoba's successful provincial nominees program in which the presence of (and involvement by) ethnic communities already resident in Manitoba is a key factor in attracting new immigrants to that province.

Finally, evidence indicates that the age and education profiles of the resident ethnic community also exert a significant influence on the location decision of new immigrants with similar age and education levels. This suggests that the attraction of an ethnic concentration is more complex than a simple size effect. While young and well-educated immigrants will still be drawn to the economic opportunities offered in Canada's largest cities, concentrations of young and well-educated immigrants in other areas will increase the attraction of these areas to new immigrants with similar characteristics. This would clearly be a positive outcome for those regions of Canada in need of the injection of skills and energy that immigration would bring.

Acknowledgment

I would like to thank Charles Beach, Jeff Borland, Barry Chiswick, Don Devoretz, an anonymous referee and the guest editors for helpful comments.

Notes

(1) The Canadian government continues to pursue a large-scale immigration program that targets new immigration inflows at approximately 1% of the Canadian population per year. Over 229,000 immigrants entered Canada in 2000 alone.

(2) For example, the issue of attracting immigrants to regions of Canada outside Toronto and Vancouver was a key topic of discussion at a conference of Federal and Provincial immigration ministers held in Winnipeg in October 2002.

(3) Related work by Myles and Hou (2003) takes a different approach and estimates 'locational attainment' models in which neighborhood characteristics (such as median income or percentage of the population that is white) are expressed as a function of family socio-economic and demographic characteristics. They find that initial settlement of visible minorities is in immigrant enclaves from which longer-term, more successful immigrants subsequently exit, but the results vary significantly by ethnicity and whether English is usually spoken at home.

(4) An alternative is to compute the share of each ethnic group that lives in each region of Canada (see Bartel, 1989; Borjas, 1999). This measures how dispersed a particular ethnic group is across Canada. The main results are not sensitive to the use of this alternative to measure ethnic clustering effects.

(5) In the empirical analysis I experiment with additional interaction terms to allow for changes over time in a region's characteristics to have differential effects on different groups of people in the sample. For example, including a time dummy variable interacted with region dummy variables and education levels allows for changed economic conditions across regions over time to affect people with a university degree differently from people with less than high school education.

(6) See McDonald (2003) for a detailed description of the econometric methodology employed, including discussion of the specific assumptions necessary to identify ethnic clustering effects in the presence of region-specific and ethnic group-specific characteristics that also affect the initial location choice of new immigrants.

(7) Also, only 1.4% of the sample of recent immigrants in each Census year reside in Atlantic Canada and the Canadian Territories.

(8) A longer version of the current paper (available from the author) discusses all of the regression results plus a number of extensions and sensitivity checks of the regressions reported here. As well, regression results for the variables not reported here are generally comparable to what is reported in McDonald (2003).

(9) Because the regressions include community-level variables in an individual-level regression, it is possible that the standard errors might be downward biased because of correlation in the error term across individuals within particular communities (see Moulton, 198e). However, bootstrapped standard errors are only trivially larger than the conventionally robust standard errors reported in the table, and inferences about the statistical significance of all of the estimated variables are identical to what is reported.

(10) Recall that these results are based on a specification that also allows for a recent immigrant's age, education, language spoken and other variables directly to affect that person's choice about where to settle in Canada. It is also important to emphasize that this estimated effects are not due to general differences in average age profiles either across Canadian regions or across ethnic groups, since both of these effects are captured by the ethnic group and region dummy variables.

(11) Though not reported here, the predicted location choice of new immigrants from the UK closely resembles the native-born English-speaking population distribution. Not surprisingly, changing the distribution of UK immigrants already resident in Canada has almost no effect on location choice of new immigrants from the UK.

References

Bartel, A. 1989. Where Do the New U.S. Immigrants Live? Journal of Labor Economics, 7(4): 371-391.

Bertrand, M., E. Luttmer and S. Mullainathan. 2000. Network Effects and Welfare Cultures. Quarterly Journal of Economics 115: 1019-1055.

Borjas, G. J. 1995. Ethnicity, Neighborhoods, and Human-Capital Externalities. American Economic Review 85: 365-390.

Borjas, G.J. 1999. Immigration and Welfare Magnets. Journal of LaborEconomics, 17(4): 607-637.

Cardak, B. and J. T. McDonald. 2004. Neighborhood Effects, Preference Heterogeneity and Immigrant Educational Attainment. Applied Economics, in-press.

Chiswick, B. and P. Miller. 2002a. Do Enclaves Matter in Immigrant Adjustment?" Unpublished mimeo.

Chiswick, B. and P. Miller. 2002b. Where Immigrants Settle in the United States. Unpublished mimeo.

Citizenship and Immigration Canada. 2000. Facts and Figures: Immigration Overview. Ottawa: Government of Canada: C+I-291-08-01E.

Dodson, M. 2001. Welfare generosity and location choices among new United States immigrants. International Review of Law and Economics 21:47-67.

Frey, W. H. 199e. Immigration, Domestic Migration and Demographic Balkanization in America: New Evidence for the 1990s. Population and Development Review 22(4): 741-763.

Funkhouser, E. 2000. Changes in the Geographic Concentration and Location of Residence of Immigrants. International Migration Review 34(2): 489-510.

Hou, F. and G. Picot. 2003. Visible Minority Neighbourhood Enclaves and Labour Market Outcomes of Immigrants. Statistics Canada Analytical Studies Branch Research Paper 204.

Lin, Z. 1998. Foreign-Born vs Native-Born Canadians: A Comparison of their Inter-Provincial Labour Mobility.

Statistics Canada Analytical Studies Branch Research Paper 114.

McDonald, J. T. 2003. The Location Choice of New Immigrants to Canada: the Role of Ethnic Networks. Canadian Immigration Policy for the 21st Century, edited by C. Beach, A. Green and J. Reitz. Kingston, Ontario: John Deutsch Institute for the Study of Economic Policy, Queen's University.

Moulton, B. 198e. Random group effects and the precision of regression estimates. Journal of Econometrics, 32(3): 385-397.

Myles, J. and F. Hou. 2003. Neighbourhood Attainment and Residential Segregation Among Toronto's Visible Minorities. Statistics Canada Analytical Studies Branch Research Paper 206.

Newbold, K. 1996. Internal Migration of the Foreign-Born in Canada. International Migration Review, 30(3): 728-747.

Zavodny, M. 1999. Determinants of Recent Immigrants' Locational Choices. International Migration Review, 33(4): 1014-1030.
Table 1: Selected Econometric Results--Ethnic Enclave Effects (a,b)

(conditional logit coefficient estimates)

Variable                                          (1)           (2)

Relative Concentration (RC)                     0.140          -0.069
                                               (0.162)         (0.179)

RC *(Other Home Language) (c)                                  0.281 **
                                                               (0.076)

RC *(University Degree)                                        -0.108 **
                                                               (0.041)

Average age of local ethnic enclave
(E.Age)                                        -0.038          -0.079 **
(0.26) (0.034)

E.Age *(Other Home Language)                                   0.045
                                                               (0.028)

E.Age *(University Degree)                                     0.037 *
                                                               (0.020)

Proportion of local ethnic enclave that
speaks other home language (E. Lang)            0.473          -2.488 **
                                               (1.020)         (1.237)

E.Lang *(Other Home Language)                                  4.345 **
                                                               (0.777)

E.Lang *(University Degree)                                    -0.458
                                                               (0.621)

Proportion of local ethnic enclave that has
university degree (E.Univ)                      1.667 **       0.191
                                               (0.774)         (0.961)

E.Univ *(Other Home Language)                                  0.326
                                                               (0.724)

E.Univ *(University Degree)                                    2.527 **
                                                               (0.523)

Log likelihood                                 -13533.6        -13434.7
Pseudo R2                                        0.298         0.303

(a): White's heteroskedastic-consistent standard errors in parentheses

(b): * denotes significance at the 10% level, ** at the 5% level

(c): language normally spoken at home is neither French nor English

Table 2a: Predicted distribution of location choice for recent
immigrants from East Asia (a)

                                  Montreal   Rest of
                                             Quebec     Toronto

Measure of ethnic enclave used

1986 distribution of immigrants
from East Asia                     0.065      0.002     0.476

1986 distribution of Canadian
Residents                          0.122      0.005     0.482

Speaks English at Home

1986 distribution of immigrants
from East Asia                     0.053      0.001     0.569

1986 distribution of Canadian
Residents                          0.061      0.001     0.574

Has a University Degree

1986 distribution of immigrants
from East Asia                     0.073      0.002     0.487

1986 distribution of Canadian
Residents                          0.110      0.004     0.501

                                  Rest of
                                  Ontario    Prairies   Alberta

Measure of ethnic enclave used

1986 distribution of immigrants
from East Asia                     0.049      0.025     0.079

1986 distribution of Canadian
Residents                          0.126      0.041     0.101

Speaks English at Home

1986 distribution of immigrants
from East Asia                     0.050      0.027     0.097

1986 distribution of Canadian
Residents                          0.061      0.030     0.103

Has a University Degree

1986 distribution of immigrants
from East Asia                     0.047      0.030     0.079

1986 distribution of Canadian
Residents                          0.086      0.042     0.094

                                              Rest of
                                  Vancouver   B.C.

Measure of ethnic enclave used

1986 distribution of immigrants
from East Asia                     0.298      0.006


1986 distribution of Canadian
Residents                          0.114      0.009

Speaks English at Home

1986 distribution of immigrants
from East Asia                     0.195      0.010

1986 distribution of Canadian
Residents                          0.160      0.011

Has a University Degree

1986 distribution of immigrants
from East Asia                     0.277      0.006

1986 distribution of Canadian
Residents                          0.156      0.008

(a) Predictions are generated for an individual with the following
characteristics unless otherwise noted: speaks a language other than
French or English at home, unmarried with no children, high school
education, 30 years of age, and gained Canadian residency between
1987 and 1991 inclusive. Measures of the characteristics of the
local immigrant population already resident in the region are
computed at the sample mean for the relevant ethnic group.

Table 2b: Predicted distribution of location choice for recent
immigrants from Poland (a)

                                              Rest of
                                   Montreal   Quebec     Toronto

Measure of ethnic enclave used

1986 distribution of immigrants
from Poland                         0.058      0.004      0.407

1986 distribution of Canadian
Residents                           0.059      0.006      0.317

Speaks English at Home

1986 distribution of immigrants
from Poland                         0.036      0.001      0.487

1986 distribution of Canadian
Residents                           0.036      0.001      0.469

Has a University Degree

1986 distribution of immigrants
from Poland                         0.059      0.004      0.431

1986 distribution of Canadian
Residents                           0.061      0.005      0.378

                                   Rest of
                                   Ontario    Prairies   Alberta

Measure of ethnic enclave used

1986 distribution of immigrants
from Poland                         0.327      0.047      0.098

1986 distribution of Canadian
Residents                           0.447      0.037      0.082

Speaks English at Home

1986 distribution of immigrants
from Poland                         0.247      0.034      0.105

1986 distribution of Canadian
Residents                           0.269      0.033      0.103

Has a University Degree

1986 distribution of immigrants
from Poland                         0.282      0.049      0.095

1986 distribution of Canadian
Residents                           0.350      0.043      0.087

                                              Rest of
                                   Vancouver   B.C.

Measure of ethnic enclave used

1986 distribution of immigrants
from Poland                         0.048      0.010

1986 distribution of Canadian
Residents                           0.043      0.009

Speaks English at Home

1986 distribution of immigrants
from Poland                         0.076      0.014

1986 distribution of Canadian
Residents                           0.076      0.014

Has a University Degree

1986 distribution of immigrants
from Poland                         0.071      0.010

1986 distribution of Canadian
Residents                           0.068      0.009

a Predictions are generated for an individual with the following
characteristics unless otherwise noted: speaks a language other than
French or English at home, unmarried with no children, high school
education, 30 years of age, and gained Canadian residency between
1987 and 1991 inclusive. Measures of the characteristics of the local
immigrant population already resident in the region are computed
at the sample mean for the relevant ethnic group.

Table 2c: Predicted distribution of location choice for recent
immigrants from Africa (non-Arab countries) (a)

                                               Rest of
                                  Montreal     Quebec     Toronto
Measure of ethnic enclave used

1986 distribution of immigrants
from Africa                         0.256       0.009      0.407

1986 distribution of Canadian
Residents                           0.222       0.013      0.322

Speaks English at Home

1986 distribution of immigrants
from Africa                         0.134       0.002      0.507

1986 distribution of Canadian
Residents                           0.129       0.002      0.494

Speaks French at Home

1986 distribution of immigrants
from Africa                         0.838       0.051      0.045

1986 distribution of Canadian
Residents                           0.824       0.057      0.045

Has a University Degree

1986 distribution of immigrants
from Africa                         0.231       0.008      0.424

1986 distribution of Canadian
Residents                           0.200       0.010      0.387

                                  Rest of
                                  Ontario      Prairies   Alberta

Measure of ethnic enclave used

1986 distribution of immigrants
from Africa                         0.169       0.029      0.071

1986 distribution of Canadian
Residents                           0.306       0.032      0.063

Speaks English at Home

1986 distribution of immigrants
from Africa                         0.160       0.029      0.085

1986 distribution of Canadian
Residents                           0.180       0.031      0.085

Speaks French at Home

1986 distribution of immigrants
from Africa                         0.048       0.002      0.010

1986 distribution of Canadian
Residents                           0.055       0.002      0.011

Has a University Degree

1986 distribution of immigrants
from Africa                         0.157       0.034      0.070

1986 distribution of Canadian
Residents                           0.228       0.039      0.069

                                               Rest of
                                  Vancouver    B. C.
Measure of ethnic enclave used

1986 distribution of immigrants
from Africa                         0.053      0.006

1986 distribution of Canadian
Residents                           0.036      0.007

Speaks English at Home

1986 distribution of immigrants
from Africa                         0.072      0.011

1986 distribution of Canadian
Residents                           0.069      0.011

Speaks French at Home

1986 distribution of immigrants
from Africa                         0.004      0.002

1986 distribution of Canadian
Residents                           0.004      0.002

Has a University Degree

1986 distribution of immigrants
from Africa                         0.070      0.007

1986 distribution of Canadian
Residents                           0.060      0.007

(a) Predictions are generated for an individual with the following
characteristics unless otherwise noted: speaks a language other
than French or English at home, unmarried with no children,
high school education, 30 years of age, and gained Canadian
residency between 1987 and 1991 inclusive. Measures of the
characteristics of the local immigrant population already resident
in the region are computed at the sample mean for the relevant
ethnic group.
Gale Copyright: Copyright 2004 Gale, Cengage Learning. All rights reserved.