Moving Forward with Ethnoburb and Income Inequality - A Preliminary Analysis Rafael Harun, PhD Candidate, School of Planning, University of Waterloo

Abstract Ethnoburb has gained wide acceptance as a new model to describe immigrant settlement patterns in the last twenty years. Many studies have adopted the model and applied in cities in the USA, Canada, Australia, and New Zealand. However, the studies have embedded inconsistencies resulting from the lack of operationalization of the concept. One of the apparent inconsistencies is the use of visible minority or ethnicity variables from the census data to characterise ethnoburbs empirically. Ethnicity or visible minority variable includes information on immigrants and the subsequent generations. Thus, using the variable in the ethnoburb delineation process is daunting. Also, research to date has been limited to explore the ethnic concentrations and to analyse the socioeconomic configurations of ethnoburbs. A lack of effort is noted to explore socioeconomic issues pertaining inequalities, health, and education challenges, that have clear planning implications. Based on the identified limitations this study investigated – i) the efficacy of using visible minority as a variable to capture immigrant settlement dynamics, and ii) the prevalence of income inequality issues in ethnoburbs. The study was conducted on four major Canadian metropolia, Greater Area, , Montreal, and Calgary by using the Canadian census data from 1996 to 2011. The results indicate that visible minority variable can not capture the immigrant settlement dynamics efficiently and thus the validity of the use of the variable in ethnoburb delineation should be revisited. Also, the prevalence of income inequality concerns was valid in some metropolitan regions that vary based on ethnicity. The study strongly encourages to use individual level data to track the movement of the immigrant population and recommends to further investigate the dynamics of income inequality for better understanding and planning. Keywords: Ethnoburb; Income Inequality; Canada; Immigrant

1.0 Introduction The global economic restructuring promoted transnational economic linkages and established the need for highly skilled workers in many developed countries. Countries such as the USA, Canada, and Australia experienced a high influx of immigrant population predominantly from the Asian countries over the last two decades. Of all, Canada has experienced the highest growth in their foreign-born population among the G7 countries, and it constitutes about 20.6% of the total population. Researchers have used traditional models such as place stratification, spatial assimilation, and enclaves to study immigrant settlement patterns. A new addition to the steam of models explaining immigrant settlement pattern is ethnoburb. Ethnoburb is widely accepted by the research community and is applied to study multiple ethnic minority groups in many developed countries. However, due to a lack of operationalization of the concept, inconsistencies in the characterization of ethnoburb is observed in the existing literature. The objective of this paper to identify one of the inconsistencies regarding the selection of variable to study ethnoburbs in Canada and to explore the existence of the critical issue of income inequality in the Canadian ethnoburbs. Ethnoburbs are multiethnic communities that have a high concentration of non-White ethnic minorities but does not constitute the majority, and their economic tie extends from local to global outposts (Wei Li 1998a). In an ethnoburb, the highly skilled and affluent non-White immigrants who migrate to the developed countries avoid the torn down city centres and settle down directly into that offer better housing quality, school services, and ambience. The people living in the dispersed suburban locations soon create their own communities, preserve ethnic identity through social and religious gatherings, and establish networks from local to transnational scales (Li, Skop, & Yu, 2016). Also, they materialise their ethnic businesses in the ethnoburbs that are integrated into the mainstream economy and get actively involved in mainstream politics, making their place in the society ( Li et al., 2016). There is an increasing interest in ethnoburb research, but it is mostly descriptive by nature. Many of the research without making an effort to verify if a location fits the profile of an ethnoburb, sporadically uses the term to describe a that has high ethnic concentration. Wei Li has clearly identified multiple aspects to characterise ethnoburbs including ethnic density, socioeconomic composition, economic activities, cultural , and political involvement. But a lack of operationalization of the concept has resulted in inconsistencies in the ethnoburb delineation process. Studies attempting to find ethnoburbs are few, and they don’t always address all aspects as shown in Table 1. The existing literature has focused mainly on ethnic densities in the ethnoburb delineation process. A range of ethnic density threshold values are applied in studies to locate ethnoburbs. To differentiate ethnoburbs from enclaves and ghettos, Wei Li has emphasised that the ethnic density in ethnoburbs can be as low as 10% to 15% in many locations (Li, 1998b). Similar threshold values ranging between 10% to 35% has been applied in many studies (Hoalst-Pullen et al. 2013; Phillips 2016; Johnston, Poulsen, and Forrest 2011). A few studies have also applied higher thresholds of ethnic densities (Chan 2012; Wang and Zhong 2013). There is a lack of ethnoburb research that investigates critical social issues. Since ethnoburbs create a distinct footprint in urban and social landscapes and have the potential of creating unique social, health, and educational challenges with succinct urban planning implications. A study of Richmond, BC revealed that the plan to create designated sites for religious buildings and combining with farming gained high appreciation in planning discourse, but did not attract the community due to a lack of understanding of priorities of the community (Dwyer, Tse, and Ley 2016). Also, a study of the Asians in ethnoburb identified that the workers are constrained by the availability of health insurance, and accessibility to health care access, a lack of information on available health services that indicates the prevalence of inequality issues (Pih, Hirose, and Mao 2012). The transnationalism has also taken a new turn in the offspring settlement patterns that gave rise to ‘satellite kids’, which has created unique education challenges. In a case study of Vancouver Canada, Waters (2003) introduced a phenomenon occurring in ethnoburbs, where the affluent parents settle down their kids in Canada and move back to their country of origin to handle their businesses. The phenomenon creates challenges for education provider as they can hardly get access to the parents to discuss issues and experienced problems with fake relatives. Recently, the decision made by the Peel Region District School Board in Canada to accommodate for the Muslim Friday prayer created a significant backlash in the entire community. Since Peel region hosts many ethnoburbs, the changing demography must have triggered the proposed changes. As ethnoburbs increases with more immigrations in place, more policy and planning challenges will be posed making it essential to learn about the dynamics of ethnoburbs and keep track of the changes. The efforts of identifying ethnoburbs in Canada are very limited. Wang & Zhong (2013) is probably one of the major studies that attempted to identify ethnoburbs empirically. The authors consider ethnic density and socioeconomic configuration and make an additional effort to evaluate ethnic businesses in ethnoburbs in their study. However, the author used visible minority variable to assess ethnic concentrations. The use of the visible minority population is fundamentally flawed as the variable includes the immigrant population as well as a population that are born in Canada but belong to a particular ethnicity. The immigrants and their following generations born in Canada may have substantial differences in their preferences. Ethnoburb is an immigrant phenomenon, and the use of the variable may fail to capture the immigrant dynamics. A similar problem is witnessed in many other studies ( for example Ishizawa & Arunachalam, 2014; Johnston et al., 2011; Wen, Lauderdale, & Kandula, 2009) where the focus of the study was on ethnicity, not immigrants. Based on the identified limitations, this study attempted to empirically determine if the visible minority is a good choice of a variable to capture the immigrant dynamics. Also, the study explored the prevalence of a critical social issue of income inequalities in the ethnoburb. The objective was attained by answering three questions with a focus on two ethnic minority groups: 1) Are there differences in the characteristics of the two ethnic minority groups?; 2) Where are the ethnoburbs of the two ethnic minority groups located?; and 3) Are there issues with income inequality in the ethnoburbs? The Canada Population Census data from 1996 to 2011 was used to address the questions.

2.0 Research Design 2.1 Data Sources This analysis is based on the Canadian Census of Population micro-data from the 20% long-form sample for the years 1996 to 2011. Each census year consists of a broad cross-section of observations, which permits analysis at the census tract levels. The four census years were selected based on the similarity in the variables among the census years that are used in this study. However, the variables were redefined to ensure coherence. The study had three major focus: foreign-born immigrant population, visible minority, and income, and the ethnic groups of interests were East and Southeast Asians and Southern Asians. The two ethnic groups were selected because the majority of the population who are immigrating to Canada since the 1990s are from Asia, predominantly from Eastern Asian, Southeast Asian, and Southern Asian countries (Statistics Canada 2013). The Canadian census collects and reports immigrants’ information by their places of birth. The information provided under the category refers to the foreign-born population who are, or who have been landed immigrants in Canada, which corresponds to the first-generation. Using the variable, the immigrants who were born in East and Southeast Asian and Southern Asian countries were grouped into two following the Canadian Census 2006 as a standard and referring to the United geoscheme (UNSD, 2017) that divides countries into regional and sub- regional groups. The population from East and Southeast Asian countries were grouped as Asian immigrants, and the population belonging to Sothern Asia countries were grouped as South Asian immigrants in this study. The second key variable used in this study was the visible minority population. Visible minority refers to any person who is non-Caucasian or non-white in colour and does not belong to Aboriginal population groups (Statistics Canada 2015). The visible minority data was consistent in the four census years. However, some reorganisation of data was necessary to match to the foreign-born immigrant population data. Thus, the Chinese, Filipino, Southeast Asian, Japanese, and Korean visible minority population counts were grouped together and defined as Asian visible minority, and the South Asian visible minority was used with no changes and referred as South Asian visible minority. The total visible minority population was used to calculate the two group’s respective percentages. For assessing income inequality, the analytic focus was on average family income. The family income reflects the actual income status of a family in the society. It is believed to be a better representative than individual income to assess income inequalities at a neighbourhood level based on the fact that a person can have a high income at an individual scale, but may belong to a middle-class status when normalised by the number of dependents in the individual’s family. Also, in urban planning, understanding income dynamics at a family level is beneficial for designing planning endeavours (Salvi et al., 2016). The average family income was also used with the educational attainment data to characterise the ethnic groups. Educational attainment was selected to describe the socioeconomic status of the based on the assertion that the new immigrants who are settling down in the ethnoburbs possess a high education and skill levels. Three educational attainment data were used for the analysis: the population that have completed a secondary level or equivalent level of education, the population that has certification or training below bachelor level, and the population that have bachelor's or higher degree was used to characterise the socioeconomic status of the ethnic groups. Each of the variables corresponds to the age group of 15 and above. Each of the variables was normalised by the total population in the age group 15 and above at each spatial unit to control for the population size differences. The average family income at each spatial unit was controlled for inflation by normalising by the mean of the average family income in the region under study. 2.2 Geographic Units Canada’s four largest metropolitan areas – Greater Toronto Area, Montreal, Vancouver, and Calgary – were chosen for this study. Ontario, British Columbia, Quebec, and Alberta are the major provinces that the foreign-born immigrants tend to settle down, predominantly in the urban areas, which makes the four selected metropolitan areas a preferred location for settlement. The four metropolitan areas together host 70% of the country’s immigrant population (Statistics Canada 2013). The neighbourhoods within each of the metropolitan areas are defined by the census tracts (CT). Census tracts are small and relatively stable geographic areas where the population range between 2,500 to 8,000. If the population at a census tract exceeds 8,000, the census tract is split into two or more new census tracts. CTs usually respect the census metropolitan area boundaries. Although the number of census tract increases in every census year corresponding to the population changes, based on the fact that the CTs respect metropolitan boundaries enables to make comparisons among metropolitan areas in cross-sectional studies like this. Thus, all the analysis performed in the study were at the CT level. 2.3 Methods for characterising ethnic groups The Asian and South Asian ethnic groups were characterised regarding the differences in their spatial settlement patterns and socio-economic dynamics. To identify the difference in the settlement patterns between the two ethnic groups and to determine the temporal changes, a hotspot analysis was performed using the Getis Ord statistics (Gi*) for the four metropolitan areas from 1996 to 2011 at five-year intervals. The variable used in the analysis was the percentage of Asian and South Asian visible minorities at each census tract. Gi* identify locations of clusters of high values and low values that refers to as hotspots and coldspots respectively (Harun and Ogneva-himmelberger 2013). The analysis also finds the statistical significance of the identified clusters by calculating a Z-score. Gi* is a popularly applied method to study spatial patterns using spatial autocorrelation and is used across disciplines (Hoalst- Pullen et al. 2013; Harun and Ogneva-himmelberger 2013; Jana and Sar 2016; Johnston, Poulsen, and Forrest 2011). By comparing the hotspots of the two ethnic groups in this study, the difference/similarity in the locational preferences between the two ethnic groups was identified and the changes over the four time periods were assessed. Multiple neighbourhood relationships were conceptualised during the analysis. However, the results of the hotspot analysis using the “Contiguity Edges Corners” is documented in this article. The “Contiguity Edges Corners” are often referred as “Queen Contiguity”, where the polygons that share an edge and/or corner are considered neighbours (ESRI 2017). A correlation analysis was performed to understand the differences in the social dynamics of the two ethnic groups. Correlation analysis helps to determine the direction and strength of the relationship among variables. The resulting values of a correlation analysis can range between -1 to +1. A positive correlation between two variables indicates that if the value of one variable increases, the value of the other variable will also increase and the negative correlation indicates that if the value of one variable increase, the value of the other variable decrease. The variables that were used in the correlation analysis were the percentage of the visible minority groups at a CT, mean of the average family income at a CT, and the three educational attainment variables (with secondary education, with certification and training from non-university institutions, and with bachelor's or above degree). The direction of the relationship derived from the correlation analysis was primarily considered for interpreting the results of this study. 2.4 Methods for identifying ethnoburbs This study did not focus on contributing an innovative approach for ethnoburb delineation, rather intended to criticise a few practices commonly applied in the ethnoburb delineation process. Thus, a simple threshold of 10% to 15% ethnic concentration in each census tract was used to identify the census tracts that can be potential candidates for ethnoburb delineation. Although the resulting census tracts that meet the criteria are potential candidates for ethnoburb considerations, not ethnoburb by its entity, the selected census tracts are referred as ethnoburbs in this study for simplicity. Four types of ethnoburbs were identified using the threshold on two sets of variables. Asian and South Asian visible minority based ethnoburbs were determined using the visible minority variables, and Asian and South Asian immigrant based ethnoburbs were identified using the foreign-born immigrant population born in east and southeast Asian countries, and southern Asian countries. The percentages of Asian foreign-born immigrant population and Asian visible minority population, South Asian visible minority population, and South Asian visible minority population were calculated for each of the census tracts in the four study areas, and the 10% to 35% concentration threshold was applied to identify the four types of ethnoburbs. Following the ethnoburb identification process, the efficacy of visible minority based ethnoburb and the immigrant based ethnoburbs were assessed on their abilities to capture the immigrant settlement dynamics in the study areas. Location Quotient (LQ), a traditional economic model commonly applied to identify industry clusters, was used to evaluate the efficacy. The LQ quantifies how concentrated an industry is in a location is relative to a benchmark (Miller, Gibson, and Wright 1991). The application of LQ has extended beyond economic analysis and have also been applied to study the concentration of certain ethnic groups in enclaves (Qadeer, Agrawal, and Lovell 2010). In this study, LQ was calculated using Equation (1).

�� = !"/! ………………….Equation (1) !!"/!! where, �� is the number of immigrant population of i ethnicity at a census tract; � is the total immigrant population in the census tract; ��� equals the number of immigrant population belonging to I ethnicity in a benchmark (i.e., each of the four metropolitan areas); and �� is the total immigrant population in the region. The LQ value in this study indicates the relative standing of each census tract in terms of immigrant influx relative to the benchmark. A LQ value of more than 1 indicates that the census tract is experiencing higher influx of immigrant population relative to the benchmark and a value of less than 1 indicates that the census tract is experiencing lower influx of immigrant population relative to the benchmark. A LQ value equal to 1 indicates a subsistence level performance. After the LQ was calculated for each of the census tracts, the two types of identified ethnoburbs, visible minority based and foreign-born immigrant population based, were overlaid and evaluated to determine which one of the two types of ethnoburb could efficiently capture the immigrant settlement dynamics. The kind of ethnoburb that contained more of the census tracts that overlapped with the census tracts with LQ values of 1 or above were identified to possess higher efficacy in capturing immigrant settlement dynamics. Thus, for each type of ethnoburb, the percentage of census tracts that overlaps with census tracts with LQ values 1 or above was calculated for each of the four types of ethnoburbs for the entire study areas from 1996 to 2011. The kind of ethnoburb that had the higher value was perceived efficient on capturing the immigrant settlement dynamics. 2.5 Methods for comparing income inequalities between ethnoburb and non-ethnoburbs Income inequality in ethnoburbs and non-ethnoburbs were evaluated and compared regarding income segregation, and changes in the proportions of the middle-class income groups. The census tracts that had less than 10% or above 35% of foreign-born Asian and South Asian immigrant population were defined as non-ethnoburbs. Once the ethnoburbs and non-ethnoburbs for the two ethnic groups were identified, the income segregation was calculated using Gini index. Income inequality looks at how income is distributed across the entire population and Gini coefficient is acknowledged the best measure of income inequality (Dinca-Panaitescu and Walks 2015). The Gini coefficient evaluates the distribution of income among families within an economy and how it deviates from a perfectly equal distribution (UN-HABITAT 2011). The Gini coefficient for the average family income for all the census tracts was calculated for the ethnoburbs and non-ethnoburbs of the two ethnic groups in all four metropolitan areas from 1996 to 2011 for this study. The Gini Coefficient was calculated using Equation (2).

! � = ��� [1 − � − � )(� + � ] ……………..Equation (2) !!! ! !!! ! !!! where, � equals to Gini Coefficient; �! equals to the cumulated proportion of the population in a region k; and �! equals to the cumulated proportion of the income variable in region k. The value of Gini coefficient varies between 0 and 1, where 0 indicates perfect income equality and 1 indicates perfect income inequality. The proportions of middle-income groups in the ethnoburbs and non-ethnoburbs were calculated for all four municipalities to explain the income discrepancies. The middle-class population was defined as the population whose average family income is above or below 20% of the mean of the average family income in the study area following Hulchanski (2007). A low proportion of middle class potentially indicates segregation, a division of wealth in two separate groups of the rich and the poor, and a large proportion of middle class means the more even distribution of wealth across the economy.

3.0 Results 3.1 Difference between Ethnic Groups The difference between the two ethnic groups of interests was assessed on the spatial clustering patterns demonstrating locational preferential differences and the socioeconomic dynamics attributed to the income and education levels across all four metropolitan regions from 1996 to 2011 in five-year intervals. Figure 1 to figure 4 shows the shows the spatial clustering patterns of Asian and South Asian population in all four metropolitan areas. High clusters of the Asian population in the GTA is identified in the census tracts of Richmond Hill and Markham areas, which are suburban areas located at the north and east parts of Toronto. In Vancouver, the Asian are clustered in significant concentrations at the census tracts located in Richmond, Vancouver, and Burnaby areas, which are in the western parts of Vancouver metropolitan area. In both the cases of GTA and Vancouver, the clusterization pattern remains consistent over the four time periods. In the case of Calgary, the clusterization of Asian population seems to vary over space and time. However, in all four time periods, census tracts that are located in the northern parts of the City of Calgary consistently shows statistically significant clusters of the Asian population. Unlike Calgary, Montreal shows consistency in the Asian clusterization patterns from 1996 to 2011. Statistically significant high clusters of Asian population are identified in the Island of Montreal, around Mont-Royal, Dollar-Des Ormeaux, Cote-Saint-Luc and Dorval, and to the east in Brossard and La Prairie suburbs. [ Insert Figure 1 to figure 4 about here] The spatial clustering patterns of the South Asian population is quite different from the Asian ethnic group. Significantly high clusters of South Asian population are identified in the census tracts of Brampton and Mississauga that are located on the west side of GTA. Some clusters are also noticed at the east of the City if Toronto. Expanding clusters of South Asian population can be seen in Pickering and Ajax from 2006. In Vancouver, South Asians are identified significantly clustered in the census tracts of Surrey across 1996 to 2011. Patches of the high clusters can be observed in parts of Richmond and Vancouver. Similar to the Asian counterpart in Calgary, no clear and consistent clusterization pattern can be noticed for the South Asian population also. In the case of Montreal, similar to the Asian population, the South Asian population are identified to cluster in the Island of Montreal, predominantly in parts of Mont- Royal, Dollard-Des Ormeaux, Dorval, and Pointe-Claire. The correlation analysis results as shown in Table 2 demonstrate substantial difference in the socioeconomic dynamics between Asian and South Asian visible minority populations. The Asian population in all four metropolitan areas shows a positive correlation between the population with bachelor degrees and above accredited from a University and the percentage of Asian visible . An exception is observed in the case of GTA for the year 2011, where a negative correlation is found between the population with bachelor degrees and above and concentration of Asian visible minority population. The difference can be explained by the limitations in the 2011 census data resulting from the discontinuation of the long form population census and the low response rates. A similar relationship can be witnessed between income and the concentration of Asian visible minorities. The average family income has a positive correlation with the percentage of Asian visible minority population in all the four metropolitan areas over the four time periods. The findings characterise the Asian visible minority population as highly educated and in good economic standing across Canada. [ Insert Table 2 about here] The correlation analysis results for the South Asian visible minority population varies over the locations. In GTA and Vancouver, there is a negative correlation between the population between the population with bachelor degrees and above and the concentration of South Asian visible minority population. However, the relationship between the two variables has a positive correlation in the case of Calgary and Montreal. An exception to the relationship can be noticed for the year 2001 in Calgary, but the relationship is statistically insignificant at 95% confidence interval. Regarding income, the South Asian visible minority has a negative correlation with income in GTA and Vancouver metropolis. However, a positive relationship between the income and the concentration of South Asian visible minority can be noticed in Calgary and Montreal. The findings indicate that the South Asian visible minority population residing in GTA and Vancouver are not as highly educated as the Asians visible minority groups and may not have a good economic standing, whereas the South Asians in Calgary and Montreal are highly educated and are in good financial standings.

3.2 Identifying Ethnoburbs and Assessing Efficacy to Capture Immigrant Settlement Dynamics A threshold of 10% to 35% was applied to the visible minority and foreign-born immigrant variables for both the Asian and the South Asian ethnic groups to identify the census tracts that are potential candidates for ethnoburb delineation. Table 3 shows the percentages of census tracts identified as visible minority based and immigrant based ethnoburbs relative to the total number of census tracts for each of the four metropolitan areas from 1996 to 2011. The results show that on an average, Asian visible minority based ethnoburbs in Vancouver and Montreal respectively comprises 41% and 49% of all the census tracts in the metropolitan areas from 1996 to 2011. In the case of Calgary and Vancouver, the percentage is lower, comprising about 22% and 28% of the total census tracts respectively across the four time periods. However, the number of Asian immigrant based ethnoburbs were higher than the Asian visible minority based ethnoburbs in GTA, Vancouver and Calgary from 1996 to 2011, but was lower in the case of Montreal. The density of the Asian visible minority population and Asian immigrant population in each census tract can explain the difference between the percentage composition of visible minority and immigrant based ethnoburbs as shown in Table 4. On an average, the Asian visible minority population comprises more than 35% of the total visible minority population living in GTA, Vancouver, and Calgary at each census tract level. Since the value exceeds the 35% threshold applied during ethnoburb delineation, may census tracts that did not satisfy the criteria were not selected as a potential candidate for ethnoburb. On the other hand, the Asian immigrant population constituted between 12% to 30% of the total immigrant population in the three metropolitan areas, resulting in an increased number of identified ethnoburbs compared to the Asian visible minority based ethnoburbs. [ Insert Tables 3 and four about here] The situation is a little different in the case of South Asian ethnoburbs. The number of the South Asian visible minority based ethnoburbs was higher than the South Asian immigrant based ethnoburbs in all four metropolitan areas from 1996 to 2011. The South Asian visible minority population comprises between 10% to 27% of the total visible minority population in all the locations. On the other hand, the South Asian immigrant population comprises between 6% to 17% of the total immigrant population in all four cases. Since the South Asian visible minority population composition falls within the 10% to 35% threshold applied for ethnoburb delineation, a larger number of census tracts were identified as ethnoburb. Whereas, since the South Asian immigrant population in many cases did not meet the minimum threshold of 10%, less number of census tracts was delineated as ethnoburbs. [Insert Figure 5 about here] Nevertheless, the higher proportions of census tracts described as ethnoburb does not indicate better encapsulation of immigrant settlement dynamics. The efficacy of the identified ethnoburbs to capture immigrant settlement dynamics rely on the spatial accuracy of the overlap between the identified ethnoburb to the LQ of the census tracts. The overlap will indicate better spatial accuracy to capture the locations that are experiencing a higher influx of immigrant population. Thus, between the visible minority and immigrant based ethnoburbs, the type of ethnoburb that overlapped with the census tracts with the LQ value of 1 or above in higher proportions demonstrated greater efficacy. The proportions of the two types of delineated ethnoburbs and the percentage of the census tracts overlapping with the census tracts that have LQ values equal to or greater than one is shown in figure 5. A Higher proportion of Asian immigrant minority based ethnoburbs are foound to overlap with the census tracts that demonstrates increased the influx of immigrant population in all the four metropolitan areas from 1996 to 2011 compared to the Asian visible minority based ethnoburbs. The South Asian visible minority based and immigrant based ethnoburbs showed similar trends. The South Asian immigrant based ethnoburb outperformed the South Asian visible minority based ethnoburbs in all the metropolitan areas across the four time periods. The results clearly indicate that the performance to capture immigrant settlement dynamics was better for the ethnoburbs that were delineated based on the foreign-born immigrant population variable compared to the visible minority variable.

3.3 Comparison of Income Inequality between Ethnoburbs and Non-ethnoburbs Since the immigrant based ethnoburbs demonstrated higher efficiency to capture immigrant settlement dynamics and ensured better suitability to delineate ethnoburbs, the immigrant based ethnoburbs and non-ethnoburbs were chosen to make the comparisons for income inequalities. Figure 6 shows the relative positions of income disparities in the ethnoburbs and non-ethnoburbs of GTA, Vancouver, Montreal and Calgary. Figure 6 reveals that for the Asian population in GTA and Montreal, except for a few years, Gini coefficient is estimated equal or higher in the Asian ethnoburbs compared to the non-ethnoburbs. However, in Vancouver and Calgary, the higher coefficient is observed in the Asian non- ethnoburbs compared to the ethnoburbs. In the case of South Asian population, the Gini coefficient is lower in the ethnoburbs compared to non-ethnoburbs in GTA and Calgary. But in Montreal, the coefficient is estimated higher or equal across the four time periods. Vancouver did not demonstrate any generalizable trend. The values indicate the prevalence of inequality concerns among the Asian ethnoburbs in GTA and Montreal, and among the South Asian population predominantly in Montreal. [Insert Figure 6 and 7 about here] Although income inequality is a complex issue, the proportions of the middle-class population are often used to explain inequalities. Figure 7 shows the proportions of middle-class populations in the Asian and South Asian ethnoburbs in the metropolitan areas from 1996 to 2011. The proportion of population belonging to the middle-class group in the Asian ethnoburbs of GTA, Montreal and Calgary is lower than that in the non-ethnoburb areas. In Vancouver, the proportion of middle-class Asian population has been consistently higher compared to the non- ethnoburbs. For the South Asian population, a high percentage of the middle-class population is identified in the ethnoburbs of GTA and Calgary, but a low proportion is observed in the ethnoburbs of Montreal. The proportions in Vancouver tends to vary over time, restricting from inferring any generalizable trend. 4.0 Discussion The empirical findings presented above suggests some points concerning Asian and South Asian groups residing in Canada and ethnoburb research. The results suggest substantial differences in the locational preferences between the two ethnic groups in the major Canadian metropolis. The Asian population in the GTA area are significantly concentrated in the eastern parts of the GTA whereas the South Asian population are concentrated in the western part of GTA. In Vancouver, clusters of the Asians are located in the west near Richmond, and the South Asians are concentrated further southeast around Surrey. Although no clear spatial clustering pattern was noticed in Calgary across time, Montreal showed thought-provoking clusterization patterns for both the minority groups. The Island of Montreal hosted both the Asian and South Asian population in significant concentrations around Mount Royal. The liberal use of English on the Island of Montreal may have influenced the spatial settlement patterns. An increasing suburbanization trend was observed among the Asians as significant clusters of them have located around the Brossard and La Prairie area, upscale suburbs on the south shore of Montreal. In addition to spatial settlement patterns, the Asians and South Asian populations differed in their socioeconomic dynamics too. The Asians in all the four metropolitan areas showed high educational attainment and financial levels across 1996 to 2011. The long history of Asian immigrants in Canada and their entrepreneurial capabilities can explain the attenuated success in education and finance. Similar to Asians, South Asians in Calgary and Montreal showed a positive correlation with high educational achievement and income. However, the South Asian population in the GTA and Vancouver were found to have a negative correlation with high educational and income levels. A detailed look at the type of skills and employment patterns is essential to tease out the forces causing the differences. Based on the concentration of Asian and South Asian population at each census tract, the census tracts that showed ethnoburb characteristics were identified. Two variables, visible minority and foreign-born Immigrant, were used to calculate the concentration of the two groups that resulted in two types of ethnoburbs: visible minority based ethnoburbs and immigrant based ethnoburbs for Asian and South Asian populations. The application of 10% to 35% threshold in the ethnoburb delineation process resulted in the exclusion of many census tracts. The Asian population in many census tracts resided in a high concentration exceeding 35%, and the South Asian population in many census tracts was less than 10% (see Tables 3 and 4). Since the census tracts did not meet the set criteria, they were not identified as ethnoburbs. In reality, Asians are relatively more concentrated compared to South Asians. Applying the same threshold for the two different ethnic group is not appropriate. In addition, population concentration is certainly not the sole criteria that determines ethnoburb. Other factors such as transboundary connection, the degree of integration of ethnic businesses to the mainstream economy, cultural kinship are key to ethnoburb delineation. Thus, scrutinising census tracts solely on population concentration is not justified. Apart from the threshold, the use of the type of variable to delineate ethnoburb had a substantial impact on the abilities to capture the immigrant settlement dynamics. The evolution of the ethnoburb was immigrant-based settlement phenomenon in North American cities. Thus, theoretically, the census tracts that will be recognised as potential candidates for ethnoburb delineation should capture the immigrant settlement dynamics. The two types of ethnoburbs were compared in terms of their abilities to capture locations that experience high immigrant influx. The immigrant based ethnoburbs in all four metropolitan areas consistently contained higher proportions of census tracts that showed a large influx of immigrant population compared to the visible minority based ethnoburbs (see Figure 5). The visible minority variable includes the immigrants born outside Canada and the population who were born Canada but share the same ethnicity. The population who were born in Canada may have substantially different preferences from the foreign-born immigrants. Most importantly, the ethnic population that were born in Canada are Canadians, not immigrants, and including them in studies on immigrant population is an injustice. Thus, in the ethnoburb delineation process, the foreign-born immigrant variable is a better choice over the visible minority in the ground of efficacy to capture immigrant settlement dynamics and ethics. Since the immigrant based ethnoburbs outperformed the visible minority based ethnoburbs in capturing immigrant dynamics, the immigrant based ethnoburbs were subjected to assess for income disparities. The comparison between the Asian and South Asian ethnoburb and non- ethnoburb revealed a high income inequality in the Asian ethnoburbs of GTA, but low income inequality in the South Asian ethnoburbs. The inequality issues was identified prevalent in Montreal for both the Asian and South Asian. The ethnoburbs in Calgary did not indicate any income disparity problem. The evaluation of the middle-class compositions in the ethnoburbs to explain income disparities did not perform well across locations. The small proportions of middle-class in the Asian ethnoburbs and high proportions of middle-class in the South Asian ethnoburbs can respectively explain the low and high income inequalities in GTA. However, the middle-class population was unable to explain the patterns of income disparities observed in Montreal in Calgary. In Montreal, income inequality was higher in both the Asian and South Asian ethnoburbs when compared to their ethnoburbs, but the proportions of the middle-class were larger for the ethnoburbs than the non-ethnoburbs for both the ethnic groups. In contrast, although Calgary had low income inequality in the Asian and South Asian ethnoburbs in Calgary, the ethnoburbs had low proportions of middle-class people. The income disparity analysis thus reveals the prevalence of income inequality issues in the identified ethnoburbs, which varies over space and ethnic group. The composition of the middle- class composition is inadequate to explain the income disparity concerns. An in-depth analysis of the factors influencing the income inequality for each of the ethnic groups in each of metropolitan area is essential to design an inclusive society. A few limitations to this study must be noted that may also offer future research directions. First, the study used aggregated census tract level data. The results using individual-level data may change the findings. Also, the longitudinal analysis was not possible for this study as the number of census tracts varies in each census year. Reconciling the number of census tracts to one census year could solve the problem, but may potentially introduce errors. Thus future research should base on individual data to track movement patterns of immigrants and focus on longitudinal trend analysis. Second, the variables available in the Canada census data was regrouped to match the visible minority and immigrant population variables used in this analysis. No validation of the accuracy of data was possible. Finally, the study used only ethnic concentrations for the analysis of ethnoburbs. Ethnic economy information of the four metropolitan areas was not available. However, the ethnic economy is key to ethnoburb characterization and is mandatory to consider.

5.0 Conclusion Geographers have long been interested in understanding immigrant settlement patterns. Ethnoburb is a new addition to the steam of models that have gained popularity in recent years. Ethnoburb research is recently receiving momentum in Canada, and a few studies are attempting to delineate ethnoburbs in Canadian metropolis. Based on some observations made on existing literature, this study identifies that the use of visible minority variable to understand the ethnoburb composition should be revisited. Since ethnoburb is an immigrant phenomenon, and visible minority includes both the immigrant and the first generation, who are not immigrants, the validity of the use of a visible minority in ethnoburb delineation process can be inappropriate. The results indicated that foreign-born immigrant population variable available from the Canadian population census strictly represents the immigrant population and better captures the immigrant settlement dynamics. In addition, applying a threshold value of ethnic concentration may be an inappropriate process of ethnoburb delineation. Multiple other factors are important for ethnoburbs that require equal attention in the delineation process. Also, the study identified issues with income inequality that are existing in the ethnoburbs. However, the pattern of income inequality changes depending on the ethnic group and location. Ethnoburb is a complex system and needs coordinated steps from researchers to advance scholarship. The coordination is required to agree on multiple issues. First, the researchers should agree on the appropriate scale for ethnoburb research. The aggregate level data is easily available but undergo many approximations that can potentially introduce errors in the analytical process. An individual level data can be difficult to obtain but provides better insight on immigrant settlement patterns. Such data can also help to conduct longitudinal analysis without reconciliation of census tracts that is seen in the recent literature. Second, the researchers should coordinate to operationalize the criteria that Wei Li used to delineate ethnoburbs. The lack of operationalization is causing incoherence in the conducted studies worldwide and misuse of the term ethnoburb in literature. Finally, it is about time that ethnoburb research pushes their boundaries beyond descriptive research and start investigating socioeconomic issues that are important for society and urban planning purposes. Since the number of immigrant population is growing in Canada, there is a high likelihood that ethnoburb can potentially become a dominant immigrant settlement pattern. If that happens, a better understanding of the ethnoburb dynamics is essential to understand for building an inclusive society.

References Chan, Arlene. 2012. “From to Ethnoburb: The Chinese in Toronto.” In World Conference of Institutes and Libraries in Chinese Overseas Studies 5th (WCILCOS) International Conference, 1–11. Vancouver, BC: University of British Columbia Library and Ohio University Libraries. http://wcilcos.library.ubc.ca/about/wcilcos/. Dinca-Panaitescu, Mihaela;, and Alan Walks. 2015. “Income Inequality , Income Polarization , and Poverty.” Toronto, ON. http://neighbourhoodchange.ca/documents/2015/12/inequality- polarization-poverty-definitions.pdf. Dwyer, Claire, Justin Tse, and David Ley. 2016. “‘Highway to Heaven’: The Creation of a Multicultural, Religious Landscape in Suburban Richmond, British Columbia.” Social & Cultural Geography 17 (5): 667–93. doi:10.1080/14649365.2015.1130848. ESRI. 2017. “Modeling Spatial Relationships.” Spatial Statistics Toolbox. http://pro.arcgis.com/en/pro-app/tool-reference/spatial-statistics/modeling-spatial- relationships.htm. Harun, Rafael;, and Yelena Ogneva-himmelberger. 2013. “Distribution of Industrial Farms in the United States and Socioeconomic, Health, and Environmental Characteristics of Counties.” Geography Journal 2013: 12. doi:10.1155/2013/385893. Hoalst-Pullen, Nancy;, Vanessa; Slinger-Friedman, Harold; Trendell, and Mark Patterson. 2013. “Spatial and Temporal Patterns of an Ethnic Economy in a Suburban Landscape of the Nuevo South.” Southeastern Geographer 53 (3): 310–27. doi:10.1353/sgo.2013.0019. Hong, Seong-yun;, and Hong-key Yoon. 2014. “Ethno-Economic Satellite : The Case of Korean Residential Clusters in Aukland.” Population, Space and Place 20: 277–92. Hulchanski, J. David. 2007. “The Three Cities within Toronto: Income Polarization among Toronto’s Neighbourhoods , 1970-2005.” Research Bulletin 41. Toronto, ON. http://www.urbancentre.utoronto.ca/pdfs/curp/tnrn/Three-Cities-Within-Toronto-2010- Final.pdf. Ishizawa, Hiromi;, and Dharma Arunachalam. 2014. “Ethnic Neighbourhoods in , New Zealand.” Urban Policy and Research 32 (4): 417–36. doi:10.1080/08111146.2013.877391. Jana, Mrityunjoy;, and Nityananda Sar. 2016. “Modeling of Hotspot Detection Using Cluster Outlier Analysis and Getis-Ord Gi* Statistic of Educational Development in Upper-Primary Level, India.” Modeling Earth Systems and Environment 2 (2). Springer International Publishing: 60. doi:10.1007/s40808-016-0122-x. Johnston, Ron;, Michael; Poulsen, and James Forrest. 2011. “Using Spatial Statistics to Identify and Characterise Ethnoburbs: Establishing a Methodology Using the Example of Auckland, New Zealand.” GeoJournal 76 (5): 447–67. doi:10.1007/s10708-010-9366-6. Li, Wei. 1998a. “Anatomy of a New Ethnic Settlement: The Chinese Ethnoburb in .” Urban Studies 35 (3): 479–501. doi:10.1080/0042098984871. ———. 1998b. “Ethnoburb versus Chinatown : Two Types of Urban Ethnic Communities in Los Angeles.” Cybergeo: European Journal of Geography, 8–11. http://cybergeo.revues.org/1018?lang=es&la. ———. 2006. “Spatial Transformation of an Urban Ethnic Community: From Chinatown to Ethnoburb in Los Angeles.” In From Urban Enclave to Etghnic Suburb: New Asian Communitis in Pacific Rim Countries, edited by Wei Li, 74–94. Honolulu: University of Hawaii Press. Li, Wei;, Emily; Skop, and Wan Yu. 2016. “Enclaves, Ethnoburbs, and Settlement.” In Contemporary Asian America, edited by Min; Zhou and Anthony Ocampo, 3rd ed., 193– 213. New York, NY: New York University Press. Miller, Mark;, Lay; Gibson, and Gene. Wright. 1991. “Location Quotient: A Basic Tool for Economic Development Analysis.” Economic Development Review 9 (2): 65. http://search.proquest.com/openview/d6011b83d027b7ad1dba29bb96b74a53/1.pdf?pq- origsite=gscholar&cbl=38209. Phillips, Bruce. 2016. “Not Quite White: The Emergence of Jewish Ethnoburbs in Los Angeles, 1920 - 2010.” American Jewish History 100 (1): 73–104. Pih, Kay;, Akihiko; Hirose, and KuoRay Mao. 2012. “The Invisible Unattended: Low-Wage Chinese Immigrant Workers, Health Care, and Social Capital in Southern ’s San Gabriel Valley.” Sociological Inquiry 82 (2): 236–56. doi:10.1111/j.1475- 682X.2012.00408.x. Qadeer, Mohammad, Sandeep K. Agrawal, and Alexander Lovell. 2010. “Evolution of Ethnic Enclaves in the Toronto Metropolitan Area, 2001-2006.” Journal of International Migration and Integration 11 (3): 315–39. doi:10.1007/s12134-010-0142-8. Salvi, Angelica;, Willem; Adema, Valeria; Ferraro, and Valérie Frey. 2016. “Policies to Promote Access to Good-Quality Affordable Housing in OECD Countries.” 176. OECD Social, Employment and Migration Working Papers. Paris. Statistics Canada. 2013. “Immigration and Ethnocultural Diversity in Canada. National Household Survey, 2011.” National Household Survey, 2011. Vol. 99-10-X20. doi:99-010- X2011001. ———. 2015. “Classification of Visible Minority.” Definitions, Data Sources and Methods. https://unstats.un.org/unsd/methodology/m49/. UN-HABITAT. 2011. “Cities and Climate Change.” Washington DC. United Nations Statistics Division. 2017. “Geographic Regions.” Standard Country or Area Codes for Statistical Use (M49). https://unstats.un.org/unsd/methodology/m49/. Wang, Shuguang;, and Jason Zhong. 2013. “Delineating Ethnoburbs in M Etropolitan Toronto.” 100. The CERIS Working Paper Series. Toronto, ON. Waters, Johanna. 2003. “Flexible Citizens? Transnationalism and Citizenship amongst Economic Immigrants in Vancouver.” Canadian Geographer 47 (3): 219–34. doi:10.1111/1541- 0064.00019. Wen, Ming;, Diane; Lauderdale, and Namratha Kandula. 2009. “Ethnic Neighborhoods in Multi- Ethnic America, 1990-2000: Resurgent Ethnicity in the Ethnoburbs?” Social Forces 88 (1): 425–60. doi:10.1353/sof.0.0244. Xue, Jingjing;, Wardlow; Friesen, and New Zealand. 2012. “Diversity in Chinese Auckland : Hypothesising Multiple Ethnoburbs.” Population, Space and Place 18: 579–95.