Disentangling the Residential Clustering of New Immigrant Groups in Suburbia +

Transcription

1 人口學刊第 35 期,2007 年 12 月, 頁 Journal of Population Studies No. 35, December 2007, pp research article Disentangling the Residential Clustering of New Immigrant Groups in Suburbia + Eric Fong *,TracyMatsuo * and Rima Wilkes ** + This research was supported by the Social Sciences and Humanities Research Council of Canada. Direct to Eric Fong, Department of Sociology, 725 Spadina Avenue, University of Toronto, Toronto, Ontario, Canada M5S 2J4. fong@chass.utoronto.ca. * University of Toronto ** University of British Columbia Received: October 2, 2006; accepted: September 28,

2 Journal of Population Studies No. 35 Abstract This paper explores an emerging common phenomenon of new immigrant groups clustering in suburban areas. We argue that the households of new immigrant groups are concerned with neighborhood safety. They are attracted to suburban neighborhoods which generally have lower crime rates. Since suburban neighborhoods have become more heterogeneous, households of new immigrant groups can look for safer neighborhoods in the suburbs without necessarily paying a higher housing cost than they would reside in the central city or in other neighborhoods in the same suburban context. With such budget constraints, only a limited number of suburban neighborhoods are available. Subsequently, these households spatially cluster in suburban areas. The 1996 Toronto census data in general confirms this picture. Keywords: new immigrant, clustering. 38.

3 Eric Fong, Tracy Matsuo and Rima Wilkes I. Introduction As cities become multiethnic, a trend of ethnic neighborhoods other than black neighborhoods emerging in various suburban areas has been documented. These ethnic suburban neighborhoods are comprised mostly of new immigrant groups, such as Chinese in Monterey Park (Los Angeles), Koreans in Flushing (New York), and Mexicans in Cicero (Chicago). In Canada, the northern neighbor of the United States, the same pattern can be found in various suburban locations, such as Asian Indians in Scarborough (Toronto) and Chinese in Richmond (Vancouver). This phenomenon of ethnic suburban clustering poses major challenges to the conventional understanding of suburbanization. Suburbanization has always been considered the means by which minority groups achieve spatial contact with the majority group. This formulation is derived from the early Chicago School, based on experiences of European immigrants at the beginning of the last century. The emergence of suburban neighborhoods of new immigrant groups, however, reveals that the suburbanization of these groups, the majority from non-european countries, does not necessarily improve their spatial contact with other groups. In fact, in moving to the suburbs some may have to give up their spatial achievement of sharing neighborhoods with other groups in the central city (Fong and Shibuya 2000). Thus, this pattern directly challenges the understanding of suburbanization as an indicator of the increase in spatial contact with other groups. It also suggests that suburbanization is a multidimensional process. The movement to the suburbs may result in an amelioration of various neighborhood characteristics without a concomitant increase, in some cases a decline, in spatial contact with other groups. An. 39.

4 Journal of Population Studies No. 35 effort to gauge this emerging phenomenon will provide a good opportunity to re-examine our understanding of suburbanization within the contemporary context of the increasing racial and ethnic diversification of North American cities. In addition, the study of ethnic suburban neighborhoods provides a unique context in which to study suburbanization, especially in light of recent findings regarding immigrant adaptation. Recent studies of immigrant adaptation, such as those emphasizing segmented assimilation, envisage an important role for ethnic neighborhoods in the adaptation of immigrants and their children (Portes and Rumbaut 2001; Portes and Zhou 1993). A study of ethnic suburban neighborhoods can delineate the emergence and possibly the internal dynamics of these neighborhoods, which in turn may help to further clarify the effect of ethnic neighborhoods, especially those in suburban areas, on the adaptation of immigrants and their children. Despite the importance of the topic and the extent of suburban residential clustering among new immigrant groups, efforts to explain ethnic clustering in suburban areas have so far focused primarily on African Americans (Frey and Speare 1988) or have been descriptive in nature (Logan, Alba, and Zhang 2002). To fill this research gap, we propose a perspective that incorporates recent literature on residential patterns and immigrant adaptation in order to understand the suburban ethnic residential clustering of new immigrant groups. Using the 1996 census data for Toronto, we examine the suburban ethnic patterns of three major new immigrant groups-eastern Europeans, East and Southeast Asians, and blacks. Implications of the findings are discussed in the final section of the paper.. 40.

5 Eric Fong, Tracy Matsuo and Rima Wilkes II. Literature Review An ethnic neighborhood is usually defined as the residential clustering of ethnic group members. Although some writers further assume that ethnic neighborhoods are associated with the presence of ethnic institutional structures (Alba, Logan, and Crowder 1997), such as ethnic stores or ethnic restaurants, this article makes no such assumption. In suburban areas, where residential and commercial land use are usually separated, ethnic neighborhoods do not always include ethnic institutional structures. In fact, many suburban residents deliberately maintain a distinction between residential and leisure locations (Zhou 1992). The study of ethnic neighborhoods, from the classic work associated with the Chicago School, such as Park (1936) and Burgess (1925), to more recent discussion, has tended to concentrate on ethnic neighborhoods in the inner city. This focus largely reflects the human ecological approach and, in recent years, the spatial assimilation perspective, which postulate a connection between social mobility and spatial mobility (Massey, Condran, and Denton 1987; Massey and Denton 1985). These authors demonstrate recurring patterns whereby members of immigrant groups cluster in less desirable neighborhoods with lower housing costs in the inner city when they first arrive in the country. As these groups stay in the country longer and improve their socioeconomic resources, they move from these ethnic neighborhoods into ethnically mixed neighborhoods associated with better amenities, which are typically found in suburban areas (Fong and Wilkes 1999). Therefore, the basic tenet of this perspective argues that the suburbanization of new immigrant groups, usually associated with an increase in residential contact with other groups and improvement in. 41.

6 Journal of Population Studies No. 35 neighborhood amenities, reflects their economic resources. Consistent with this perspective, Massey and Denton (1988), using the1980 census, showed that the income and education of new immigrant groups, such as Asians and Hispanics, were positively associated with suburbanization. Based on the estimation from both aggregate-level and individual-level data, Logan and Alba (1993) found similar results. Further, Fong and Shibuya (2000), using 1990 census data, found that period of immigration, education, and income had significant effects on suburbanization even after controlling for housing tenure. Although these studies in general support the spatial assimilation model regarding the effects of socioeconomic resources on suburbanization, some studies suggest that the suburbanization of minority groups does not necessarily increase their spatial contact with whites (Logan, Alba, and Leung 1996; Alba and Logan 1993; Massey and Denton 1988). Research has shown that some new immigrant groups, such as Cubans, are less able to translate their socioeconomic resources into higher levels of contact with whites in suburban neighborhoods (Alba and Logan 1993; Massey and Denton 1988). In addition, the documentation of recently emerging ethnic suburban communities suggests that a considerable proportion of new immigrant groups in suburban communities retain their residential clustering (Fong 1994; Saito 1998). These findings of ethnic suburban clustering challenge the key assumption of the spatial assimilation perspective: that suburbanization is associated with more spatial contact with other groups. These discrepancies have led to an alternative theoretical perspective, the place stratification perspective. In a series of papers, based on the concept of a dual housing market (Stearns and Logan 1986), Logan and Alba (Alba and Logan 1993; Logan, Alba, and Leung 1996) suggest that. 42.

7 Eric Fong, Tracy Matsuo and Rima Wilkes the residential location of groups, especially blacks in the United States, reflects structural sources of inequality. Their view is that places and groups are stratified in society. Since places can be understood as a means for advantaged groups to maintain social resources and symbolic power, these groups will use various means - such as real estate agents (Yinger 1995), financial institutions (Squires and O'Connor 1998), local residential organizations, and governments (Saito 1998) - to create barriers restricting the residential mobility of disadvantaged minority groups. The place stratification perspective acknowledges the importance of socioeconomic resources on suburbanization and spatial contacts with the majority groups, but suggests that the effects vary for different groups. Although the place stratification perspective successfully explains anomalies not captured by the spatial assimilation perspective, the explanation focuses on certain racial groups, particularly blacks. The application of the perspective to explain the clustering of new immigrant groups in suburban areas is less developed, even though the phenomenon is commonly found in most major North American cities. The residential clustering of new immigrant groups in suburban areas poses unique challenges to these perspectives. Studies based on previous generations of immigrants showed that an improvement in the socioeconomic resources of immigrants usually was associated with moving away from ethnic clustering neighborhoods where intergroup interaction was limited. Why then do these new immigrant groups cluster rather than disperse to mixed neighborhoods, even when they have socioeconomic resources, contrary to the prediction of the spatial assimilation perspective? The question is particularly important as recent studies have demonstrated that immigrants in general still gradually integrate over time in terms of their economic achievement, language. 43.

8 Journal of Population Studies No. 35 ability, and friendship networks, especially those immigrants with economic resources. Similarly, how much can immigrants residential clustering in suburban areas be explained by the place stratification perspective, since their ability to translate their socioeconomic resources into spatial contact with the majority group, according to previous studies, is substantially significant, controlling for their length of stay in the country (Alba, Logan, Stults, Marzan and Zhang 1999; Fong and Wilkes 1999). In the following section, we address issues associated with the contending spatial assimilation perspective and place stratification perspective on the topic of ethnic suburban neighborhoods. First, we discuss the uniqueness of new immigrant groups and the recent developments in suburban areas. From the discussion, we conclude suburban neighborhoods are no longer homogeneous and that the socioeconomic backgrounds of recent immigrants are increasingly diverse. Considering these changes in suburban contexts and immigrant characteristics, we propose that new immigrant groups, seeking a safer environment in the suburbs but constrained by budget, limit their locational choices and eventually cluster in suburban areas. We then develop testable hypotheses to account for these arguments. The final section of the paper tests these hypotheses by examining the ethnic residential clustering patterns of three new immigrant groups in the suburban areas of Toronto, Canada. III. New Immigrant Groups and Suburbanization The new immigrant groups are unique in their socioeconomic. 44.

9 Eric Fong, Tracy Matsuo and Rima Wilkes background due to changes in North American immigration policies in the last thirty years. Most came from non-traditional sending countries with considerable language skills (Farley 1996). There is substantial variation in educational credentials among these new immigrant groups, with overrepresentation at the top and bottom ends of the educational spectrum. Since the economic achievement of immigrants is largely related to educational attainment, these new immigrant groups show a wide distribution in occupations held and income earned (Zhou 2001). This information suggests that the ability of these immigrants to gain access to desirable neighborhoods, including suburban neighborhoods, varies considerably. In the same period, the suburban environment has undergone drastic changes. The uniformity and homogeneity of suburban communities and residents has been gradually replaced by variations in the characteristics of suburban neighborhoods and diversity in the socioeconomic background of suburban residents (Fong and Shibuya 2000; Alba et al. 1999). Furthermore, with the spread of automobiles over the years and federal support of home mortgages, households with a wider spectrum of socioeconomic backgrounds can afford to live in the suburbs (Jackson 1985). Recent research has documented considerable variation in suburban neighborhood qualities that attract suburban residents with substantial variation in education, income, and even the duration of immigrants in the country (Fong and Shibuya 2000; Zelinsky and Lee 1998). Given the variation in suburban neighborhoods, households are attracted to neighborhoods congruent with their preferences. Among various neighborhood characteristics, many studies have shown that levels of crime stand out as a consistent and significant concern. For example, in their recent study Emerson, Yancy and Chai (2001) showed that a. 45.

10 Journal of Population Studies No. 35 neighborhood crime rate is strongly related to the likelihood of buying a house in the neighborhood. Similarly, Fillon, Bunting, and Warriner (1999) documented that safety is the most common concern among various neighborhood characteristics. Research has also pointed out that concern about crime usually translates into a push factor leading to residential mobility. Sampson and Wooldredge (1986) documented that the level of crime in neighborhoods facilitates out migration to the suburbs. South and Crowder (1997) showed the same effect in a recent study using the 1979 to 1985 waves of the Panel Study of Income Dynamics. The effect also applies to immigrant groups. A study by Rosenbaum and Friedman (2001), which specifically focused on immigrants, showed that immigrants are more likely to reside in desirable neighborhoods with low crime rates. They suggested that immigrants, like others, are conscious of the neighborhood characteristics where they settle. Case studies, such as Zhou's (1992) detailed study of ChinatowninNewYorkCity,showedthestrongdeterminationofChinese immigrant families to improve the neighborhood qualities where they live, including neighborhood safety, even if they have to pool resources from different family members. Similarly, Waters' study of West Indian immigrants in New York showed the strong desire of these immigrants to move to neighborhoods free of violence and fear (Waters 1999: ). Neighborhood crime is a major concern for most people. It triggers mobility because it is usually perceived as a salient indicator of the deterioration and undesirable inhabitability of a neighborhood (Taub, Taylor, and Dunham 1984). These associated perceptions of the crime level, to a large extent, can lead to neighborhood instability, including substantial population change (Morenoff and Sampson 1997). Immigrant groups are especially sensitive to the neighborhood crime level. With. 46.

11 Eric Fong, Tracy Matsuo and Rima Wilkes limited information and lack of familiarity with a broad array of neighborhoods in the city, the locational decisions of new immigrant groups are largely based on specific heuristics that serve as proxies to provide them with informational cues (Bickers and Stein 1998; Lowery and Lyons 1989). Since crime level is usually associated with the social and physical deterioration of a neighborhood, it provides a "clear" and "natural" cue to the desirability of the neighborhood for new immigrant groups. Given that neighborhood choices are constrained by household budgets and suburban neighborhoods are usually associated with lower crime levels, households that cannot afford to pay substantial housing prices but still seek safer neighborhoods will reside in suburban neighborhoods with housing values that are lower than or at least similar to their neighborhoods in the central city or neighborhoods in the same suburban context (i.e. inner or outer suburbs). Budget constraints considerably limit purchasing power, which in turn exacerbates the restriction of locational options. Drawing from the discussion, we suggest three hypotheses: Hypothesis 1: Suburban ethnic neighborhoods have lower crime rate than central city ethnic or non-ethnic neighborhoods. Hypothesis 2: Housing values of suburban ethnic neighborhoods are either lower than or similar to housing values in central city neighborhoods. Hypothesis 3: Housing values of suburban ethnic neighborhoods are either lower than or similar to housing values in non-ethnic neighborhoods in the same suburban context. However, as documented by the place stratification perspective, not all groups have equal access to locational choices. Since locations are associated with resource distribution and status, minority groups, such as blacks in the United States, may face barriers to entering certain. 47.

12 Journal of Population Studies No. 35 neighborhoods. We expect this experience also applies to blacks in Canada even though most are immigrants. This expectation is based on two lines of research. First, the differences in social values, especially the perception of race and ethnic relations, are increasingly insignificant between the two countries (Reitz and Breton 1994). This phenomenon is largely due to the impact of American media on Canadians. Second, although the socioeconomic and demographic backgrounds of blacks in Canada are different from their counterparts in the United States, the study of blacks in Canada shows that they also experience discrimination in the housing market (Fong and Gulia 2000; Fong and Wilkes 1999). Therefore, we expect that suburbanization of blacks in Canada also experiences demonstrates place stratification. Although most studies focus on the differential return rates of blacks to translate their socioeconomic resources into suburban location, it is also possible that blacks may face a limited choice of suburban neighborhoods that they would feel comfortable moving into. The housing values of these neighborhoods are pushed up because of the limited supply. This has led to the fourth hypothesis: Hypothesis 4: Suburban black neighborhoods, usually associated with lower crime levels, are higher in housing values than central city black and non-black neighborhoods. Before proceeding with the discussion of the data and results, we would like to acknowledge that we are fully aware that the locational choice of immigrants is complex. Locational choice involves the matching of socioeconomic resources and preferences over a variety of amenities. Without discounting the importance of other amenities considered by immigrant families, our study only argues that neighborhood safety is only one of the significant factors that contributes to immigrant residential clustering in suburbs.. 48.

13 Eric Fong, Tracy Matsuo and Rima Wilkes IV. Data and Methods The units of analysis in this study are 808 census tracts of the Toronto census metropolitan area. Toronto is the largest city in Canada and has historically been the most favored destination for immigrants to Canada. Early records have documented the residential enclaves of Jews, Italians, and Irish in various parts of the city (Harney 1985). With the change in immigration policies, recent immigrants in Toronto come from even more diverse places. The city has become the destination of recent immigrants from China, India, Russia, Taiwan, and Yugoslavia. With the considerable size of these ethnic groups, diverse ethnic neighborhoods have emerged in various parts of the city. Data for the study were drawn from three sources. The first source comes from the 1996 Census of Canada: Profile of Census Tracts. The data set provides detailed socioeconomic and demographic information on census tracts. The second source is a specially requested table from Statistics Canada that contains detailed census tract information on the socioeconomic and demographic background of various racial and ethnic groups. The website of a major local newspaper, the Toronto Star, was our third source of information. The website provided data on neighborhood homicide rates, including all homicides that occurred between 1994 and 1996: one hundred seventy-eight in total (Toronto Star 2001) 1. This source of data was used because crime statistics are rarely reported at the census 1 Three homicides were excluded because insufficient information was reported to accurately map their locations. Homicides that happened two years prior to the census year were included because homicides that occurred in immediately previous years would have had a greater effect on residential choices.. 49.

14 Journal of Population Studies No. 35 tract level in Canadian government publications. We include in our analysis three major new immigrant groups: Eastern Europeans, East and Southeast Asians, and blacks. They represent new immigrant groups from various parts of the world. We classify groups according to broad geographic region instead of specific ethnicity because most data are suppressed when specific ethnic groups are used. The general principle of data suppression practiced by Statistics Canada is that all information about the group in the census tract will be suppressed if the group size in the tract is less than 40. We choose to use broad geographic regions because research on residential patterns in Canada has shown that the residential patterns of ethnic groups from the same geographic region are similar (Fong and Wilkes 1999). Information on blacks is obtained directly from the census. Blacks have been included as one of the response categories since The three groups included in the analysis represent a large percentage of the recent immigrant population in Canada. About 38.9% of Eastern Europeans, 56.7% of East and Southeast Asians, and 43.3% of blacks arrived in Canada after The three groups comprise more than 50% of the immigrants in Canada and more than 45% of those who arrived after The key concept of the study is ethnic neighborhoods. There are two approaches to defining ethnic neighborhoods in the literature. One is based on an absolute cutoff percentage that applies to all groups: the other assigns a specific cutoff percentage to each group reflecting their 2 We did not include other European groups - such as Western, Northern, and Southern Europeans - because the majority of Western and Northern Europeans arrived in Canada before 1950, and most Southern Europeans arrived between 1950 and 1960 (Kalbach 1987). South Asians are not included because of their small population size.. 50.

15 Eric Fong, Tracy Matsuo and Rima Wilkes overrepresentation in the neighborhood based on their population percentage in the Toronto census metropolitan area. We used the latter approach because it takes group size into consideration. When one cutoff point is applied to all groups without considering their size in the city, the designation of neighborhoods as ethnic neighborhoods can be largely influenced by the group composition in the city. Even with high concentrations, the residential clustering of a smaller group still may not reach the cutoff percentage. In contrast, larger groups can easily reach the cutoff percentage just because of their sizes. Therefore, we defined ethnic neighborhoods as those neighborhoods with an ethnic population one standard deviation above their percentage in the city 3. Following the strategy of defining ethnic neighborhoods suggested by Alba, Logan and Crowder (1997), the decision to use one standard deviation above the average percentage in the city is based on careful ethnographic studies of these neighborhoods, knowledge of neighborhoods in the city, and consultation with ethnic members. We are interested in understanding factors related to ethnic clustering in suburban areas. Since Statistics Canada does not provide official boundaries for the central city, we defined the central city as the area within the boundaries of the former city of Toronto 4. The areas that fall in the Toronto census metropolitan area but outside the former city of Toronto are classified as suburban. We further differentiate inner- and outer-suburban areas to recognize the possible ecological differences of neighborhoods, which in turn may affect the locational choice of new immigrant groups. We 3 Percentages of Eastern Europeans, East and Southeast Asians, and blacks one standard deviation above their percentage in the Toronto census metropolitan area are 10.2%, 23.6%, and 12.3% respectively. 4 The city of Toronto and five surrounding municipalities (East York, Etobicoke, Scarborough, North York, and York) were amalgamated in

16 Journal of Population Studies No. 35 define as inner suburbs those tracts in the five former suburban municipalities (East York, Etobicoke, North York, Scarborough, and York), where the percentage of housing built before 1946 is more than one standard deviation below the mean. Tracts that do not fall into this classification are usually located on the fringes of these former municipalities and have been developed in recent decades. They should be considered outer suburbs. The outer-suburban areas therefore include areas beyond the five former suburban municipalities and a small number of census tracts within them where the percentage of housing built before 1946 is no more than one standard deviation below the mean of the five municipalities. V. Results We first present findings on the locational distribution of the three major immigrant groups in Toronto. Two sets of information are reported. The top panel in Table 1 shows the number of census tracts that are classified as ethnic and non-ethnic neighborhoods for a given group in the central city, inner-suburbs, and outer-suburbs 5. The bottom panel in Table 1 also shows the percentage of Eastern Europeans, East and Southeast Asians, and blacks residing in these neighborhoods. These results provide detailed information about the residential distribution of ethnic groups in the central city, inner suburbs and outer suburbs. The distribution of Eastern European and black neighborhoods are similar in Toronto. The two groups have considerable numbers of ethnic neighborhoods in both inner and outer suburbs, but the numbers of ethnic 5 Non-ethnic neighborhoods hereafter are defined as non-ethnic neighborhoods for the specific group under discussion.. 52.

17 Eric Fong, Tracy Matsuo and Rima Wilkes Table 1. Distribution of Various Neighborhood Types and Population in Various Neighborhood Types by Location and Major New Immigrant Group, 1996 Eastern European East and Southeast Asian Black Number of Neighborhoods Toronto Ethnic Non-Ethnic Inner Suburbs Ethnic Non-Ethnic Outer Suburbs Ethnic Non-Ethnic Total Group Proportion Toronto Ethnic Non-Ethnic Inner Suburbs Ethnic Non-Ethnic Outer Suburbs Ethnic Non-Ethnic Total neighborhoods in the inner suburbs are less than in the outer suburbs for both groups. The patterns simply reflect that there are more neighborhoods in the outer suburbs, and that there has been substantial outward expansion of the suburbs in the past decades. In fact, earlier work, such as Frey and Speare's (1988) study of metropolitan growth based on the 1980 U.S. census data, found similar patterns in U.S. cities. The finding that more ethnic neighborhoods are located in the outer suburbs than in the inner suburbs is more pronounced for East and Southeast Asians. Only one East and. 53.

18 Journal of Population Studies No. 35 Southeast Asian neighborhood is found in the inner suburbs. In short, the considerable number of ethnic neighborhoods in the suburbs with a high number of them located in the outer suburbs clearly challenges the conventional association of suburbanization with an increase in spatial contact with other groups. The second panel shows the population distribution of the three new immigrant groups in ethnic and non-ethnic neighborhoods for a given group by location. The results show that most members reside in suburban areas. Among them, most are residents of outer suburbs. Not only do a large proportion of group members live in suburbs, but the results also show that a substantial proportion of them are clustered in ethnic neighborhoods despite a large proportion of them living in non-ethnic neighborhoods. Of those living in suburbs, about 33.3% of Eastern Europeans, 43.2% of East and Southeast Asians, and 43.3% of blacks are living in ethnic neighborhoods. Among those living in suburban ethnic neighborhoods, an average of 85% lives in outer suburbs. The results again challenge the notion that suburbanization leads to an increase in spatial contact with other groups. The pattern of living in suburban ethnic neighborhoods is not uncommon. Table 2 goes one step further to identify whether any of these three new immigrant groups is over represented in ethnic neighborhoods in the inner or outer suburbs when compared to other groups. The comparison gives us a larger picture of group suburban residential clustering in relation to other groups. The table shows the odds ratios that compare the odds of a group being in an ethnic or non-ethnic neighborhood in a given area (i.e., central city, inner suburbs, outer suburbs) with the odds of its representation in the same type of neighborhood in all other areas. The advantage of an odds ratio is to avoid the sensitivity of group size based on the percentage. It. 54.

19 Eric Fong, Tracy Matsuo and Rima Wilkes Table 2. Odds Ratios of the Distribution of Major New Immigrant Groups by Neighborhood Type and Location, Eastern European East and Southeast Asian Black Toronto Ethnic Neighborhoods Non-Ethnic Neighborhoods Inner Suburbs Ethnic Neighborhoods Non-Ethnic Neighborhoods Outer Suburbs Ethnic Neighborhoods Non-Ethnic Neighborhoods provides us with a clearer picture of group representation in inner and outer suburbs. To interpret the results, an odds ratio of 1 and above suggests that a group is overrepresented in ethnic neighborhoods in a given area. Although ethnic neighborhoods are commonly found in inner and outer suburbs, the results show diverse patterns of group representation in ethnic neighborhoods. Eastern Europeans are overrepresented in innersuburban Eastern European neighborhoods but underrepresented in the outer-suburban Eastern European neighborhoods. However, the pattern for East and Southeast Asians is the reverse: they are overrepresented in the outer-suburban East and Southeast Asian neighborhoods but underrepresented in the inner-suburban East and Southeast Asian neighborhoods. These two opposite patterns suggest that the different neighborhood characteristics associated with inner and outer suburbs may attract groups with different characteristics. Blacks are overrepresented in black neighborhoods in both inner and outer suburbs. The results obviously show that the suburbanizaiton of blacks is not necessarily associated with an increase in their spatial contact with other groups. These results echo. 55.

20 Journal of Population Studies No. 35 findings on the suburban experiences of their counterparts in the United States. VI. Neighborhood Characteristics and Suburban Ethnic Clustering The results so far suggest that a sizeable percentage of new immigrant groups live in suburban ethnic neighborhoods, with varied representation in ethnic neighborhoods in inner and outer suburbs. In this section, we conduct multinomial logit regression to explore the effects of neighborhood characteristics on the patterns of suburban ethnic clustering. This analysis differentiates central city neighborhoods as ethnic and non-ethnic to take into consideration differences in neighborhoods characteristics. Table 3 presents results from the multinomial logit model that contrasts the neighborhood characteristics of the central city ethnic neighborhoods with ethnic and non-ethnic neighborhoods in inner- and outer-suburban areas. Results that contrast neighborhood characteristics of central city non-ethnic neighborhoods with suburban neighborhoods will be presented in the next table. The two major neighborhood characteristics in the analysis are housing values and crime rate. We expect that suburban ethnic neighborhoods will have lower crime rates and lower or similar housing values in comparison to central city neighborhoods. The average housing value in a neighborhood is a good indicator of the housing budget of those residing in the area. The crime rate is measured by only one indicator: the homicide rate. The occurrence of homicide usually attracts wide media attention and is the most obvious type of crime to be known to immigrant groups, who usually have limited information about neighborhoods. As. 56.

21 Eric Fong, Tracy Matsuo and Rima Wilkes Table 3. Multinomial Logistic Regression of Contrasting Selected Neighborhood Characteristics of Central-City Ethnic Neighborhoods with Various Suburban Neighborhood Types by Three New Immigrant Groups, Eastern Europeans East and Southeast Asians Blacks Inner-Suburban Ethnic Neighborhoods Average Household Income Prop. Not Knowing Official Languages * Prop. Unemployment Homicide Rate ** * Housing Value ** 0.816* Intercept * Inner-Suburban Non-Ethnic Neighborhoods Average Household Income 0.862** * Prop. Not Knowing Official Languages ** ** Prop. Unemployment 0.197* Homicide Rate * ** ** Housing Value ** ** Intercept * * Outer-Suburban Ethnic Neighborhoods Average Household Income 0.983** 2.819** 6.504** Prop. Not Knowing Official Languages * ** Prop. Unemployment ** Homicide Rate ** ** ** Housing Value ** * Intercept ** Outer-Suburban Non-Ethnic Neighborhoods Average Household Income 1.58** 2.417* 6.609** Prop. Not Knowing Official Languages ** ** Prop. Unemployment 0.37** Homicide Rate ** ** ** Housing Value ** ** Intercept * * Likelihood Ratio N Note: p** < 0.05, p* <

22 Journal of Population Studies No. 35 discussed above, the homicide rate in each neighborhood was based on information from the website of a local newspaper, the Toronto Star. We created a variable to take into consideration the filtering effect on adjacent neighborhoods. All census tracts within a 1 km of a homicide site were given a value of 2. Those beyond 1 km but within 2 km of the homicide were given a value of 1. A value for each neighborhood was created to account for the effect of homicide occurrence by summing up the scores. Though using the homicide rate does not have the same underreporting problem associated with other crime statistics, and the analysis is therefore free from problems due to underreporting bias, using only homicide rates for understanding the effects of crime on neighborhood choices among immigrant groups can have limitations. Occurrences of other types of crimes- such as robbery, aggravated assault, or property damage- can also affect neighborhood choices. However, occurrence of homicide is repeatedly shown to be highly correlated with other types of crime across neighborhoods. Based on the 1990 data, Galster and Mikelsons (1995) showed the correlation of violent crimes with drug sale/possession arrests and with property crimes in neighborhoods to be about 0.76 and 0.92, respectively. Therefore, even though only homicide is used in our analysis, the statistics highlight those neighborhoods most likely to have occurrences of other types of crimes. The neighborhoods where homicide occurred can be interpreted as a conservative indicator of neighborhoods with multiple crime occurrences. Nevertheless, our results based on a single indicator of neighborhood crime rate must be interpreted with caution. In the analysis, we also controlled three other neighborhood characteristics: average household income, proportion of unemployed residents, and proportion of residents not knowing either official language (English or French). They represent, though by no means exclusively, key. 58.

23 Eric Fong, Tracy Matsuo and Rima Wilkes socioeconomic neighborhood characteristics that concern households of new immigrant groups. We expect that new immigrant households, like other households, prefer to reside in neighborhoods with higher household income and fewer unemployed residents, which usually are characterized as high-status neighborhoods. The presence of new immigrant households with a smaller proportion of residents not knowing either official language in neighborhoods are also desirable to new immigrant groups. Residing in such neighborhoods implies that they will not limit their social networks to include only those who do not know either official language. The discussion of the results focuses on the comparison of central city ethnic neighborhoods with inner- and outer-suburban ethnic neighborhoods. In other words, we explore why the suburbanization of new immigrant groups occurs without increasing their spatial contact with other groups. The results that contrast central city ethnic neighborhoods with inner- and outer-suburban non-ethnic neighborhoods of a given group are mentioned only as reference for the discussion. Based on the discussion in the previous section, we expect that new immigrant groups, seeking a safer environment in suburban areas but constrained by budget to not pay more for housing, face limited locational choices and eventually cluster in suburban areas. Therefore, the homicide rate of both inner- and outersuburban ethnic neighborhoods should be consistently and significantly lower than that of central city ethnic neighborhoods. The housing values should be statistically insignificant or significantly lower than in the central city ethnic neighborhoods 6. 6 It is also possible that housing values may change because new immigrant groups had moved into these neighborhoods. A comparison of housing values between 1991 and 1996 in all ethnic neighborhoods in our analysis show that the change is insignificant after adjusting for inflation.. 59.

24 Journal of Population Studies No. 35 The first column contrasts the central city Eastern European (hereafter EE) ethnic neighborhoods with EE and non-ee neighborhoods in the inner and outer suburbs. The results, as expected, show only the homicide rate and the housing value to be different between central city EE neighborhoods and inner-suburban EE neighborhoods. Although there is no increase in the spatial contact with other groups, a safer environment with lower housing cost than central city EE neighborhoods definitely attracts Eastern Europeans to these inner-suburban EE neighborhoods. In contrast, the comparison between central city EE neighborhoods and inner-suburban non-ee neighborhoods suggests that the latter type of neighborhoods are associated with higher exposure to other groups and lower housing values, and that most of the included neighborhood characteristics are also improved. Then why do immigrant groups cluster in inner-suburban EE neighborhoods rather than inner-suburban non-ee neighborhoods? A comparison between inner-suburban EE and non-ee neighborhoods shows that the housing value of inner-suburban EE neighborhoods is significantly lower (see Appendix A). Taken together, the results for Eastern Europeans support the hypothesis that new immigrant groups clustering in inner-suburban ethnic neighborhoods is associated with lower housing cost. Although residing in inner-suburban ethnic neighborhoods does not increase the spatial contact with other groups, nor necessarily improve other neighborhood characteristics, these neighborhoods have lower crime levels than the central city ethnic neighborhoods. As for the contrast between central city and outer-suburban EE neighborhoods, the results demonstrate that homicide rates and average housing values are lower in the outer-suburban EE neighborhoods. However, these results have a pattern very similar to the contrast between. 60.

25 Eric Fong, Tracy Matsuo and Rima Wilkes central city EE neighborhoods and outer-suburban non-ee neighborhoods. To address the question of why new immigrant groups cluster in outersuburban EE neighborhoods, a comparison between EE and non-ee neighborhoods in the outer-suburbs demonstrates that non-ee neighborhoods have higher housing values and higher homicide rates (see Appendix B). In other words, Eastern Europeans who reside in outersuburban EE neighborhoods, like their co-ethnic members in innersuburban EE neighborhoods, may limit their exposure to other groups but are able to secure a safer environment without paying a higher housing value, in contrast to those living in outer-suburban non-ee neighborhoods. The second column shows differences in neighborhood characteristics between central city East and Southeast Asian (hereafter ESA) neighborhoods and neighborhoods in inner- and outer-suburban areas. Inner-suburban ESA neighborhoods are not included in this multinomial logit regression because only one neighborhood falls into this category. In outer-suburban areas, whether ESA or non-esa neighborhoods, more residents know an official language; and these residents have more economic resources compared to those in central city ESA neighborhoods. Both also enjoy safer neighborhoods and the housing values are not significantly different from the central city ESA neighborhoods. A comparison of outer-suburban ESA and non-esa neighborhoods suggests that, as with Eastern Europeans, the outer-suburban ESA neighborhoods are safer and have lower housing values (see Appendix B). The model for blacks reveals a different picture. Although innersuburban black neighborhoods have lower homicide rates and a higher proportion of residents who know an official language in comparison to central city black neighborhoods, these neighborhoods have higher housing values as compared to central city black neighborhoods. As well,. 61.

26 Journal of Population Studies No. 35 comparison of neighborhood characteristics of inner-suburban black and non-black neighborhoods with central city black neighborhoods yields similar results. As with other groups, it raises the same question of why blacks reside in inner-suburban black neighborhoods. Further comparison between inner-suburban black and inner-suburban non-black neighborhoods suggests the difference in the homicide rate is insignificant (see Appendix A). The inner-suburban black neighborhoods, however, have lower housing values than the inner-suburban non-black neighborhoods. Taken together, like other groups, blacks reside in innersuburban black neighborhoods to secure a safer environment and lower housing prices than in other neighborhoods in the same suburban context. However, unlike other groups, blacks in inner-suburban ethnic neighborhoods have to pay a higher housing price than those living in central city black neighborhoods. Neighborhood characteristics of outer-suburban black neighborhoods are all significantly different from those of the central city black neighborhoods. These outer-suburban black neighborhoods have lower homicide rates, higher household incomes, and lower proportions of residents not knowing an official language. Similar to inner-suburban black neighborhoods, the housing values are also higher than in central city black neighborhoods. A further comparison of outer-suburban non-black neighborhoods and central city black neighborhoods suggests that outersuburban non-black neighborhoods also enjoy almost all the desirable neighborhood characteristics. The findings again lead to the question of why blacks cluster in outer-suburban black neighborhoods. Similar to previous groups, the comparison of outer-suburban black and outersuburban non-black neighborhoods suggests that outer-suburban non-black neighborhoods have a higher housing value (see Appendix B).. 62.

27 Eric Fong, Tracy Matsuo and Rima Wilkes Table 4. Multinomial Logistic Regression of Contrasting Selected Neighborhood Characteristics of Central-City Non-Ethnic Neighborhoods with Various Suburban Neighborhood Types by Three New Immigrant Groups, Ethic Neighborhoods in Inner Suburbs Eastern Europeans East and Southeast Asians Blacks Average Household Income * Prop. Not Knowing Official Languages Prop. Unemployment ** Homicide Rate ** ** Housing Value ** Intercept Non-Ethnic Neighborhoods in Inner Suburbs Average Household Income Prop. Not Knowing Official Languages Prop. Unemployment Homicide Rate ** ** ** Housing Value ** ** Intercept Ethnic Neighborhoods in Outer Suburbs Average Household Income 0.379** 1.258** 0.805** Prop. Not Knowing Official Languages ** Prop. Unemployment ** Homicide Rate ** -0.39** ** Housing Value ** ** ** Intercept 3.097* 4.724** ** Non-Ethnic Neighborhoods in Outer Suburbs Average Household Income 0.947** 0.893** 0.948** Prop. Not Knowing Official Languages Prop. Unemployment ** 0.126** Homicide Rate ** ** ** Housing Value ** ** ** Intercept Likelihood Ratio Note: p**<0.05; p*<

28 Journal of Population Studies No. 35 Table 4 shows the results of the multinomial logit regression that contrasts the neighborhood characteristics of central city non-ethnic neighborhoods with ethnic and non-ethnic neighborhoods for a given group in inner- and outer-suburban areas. In other words, the results address the seemingly paradoxical pattern that ethnic households, who have already achieved residential proximity with other groups in the central city, give up the opportunity and reside in suburban ethnic neighborhoods. The results for Eastern European and East and Southeast Asians show similar patterns. Homicide rates in their suburban ethnic neighborhoods are lower than in their central city non-ethnic neighborhoods. While results show the housing values in both ethnic and non-ethnic neighborhoods in the same suburban context are lower than in central city non-ethnic neighborhoods, further comparison shows that ethnic neighborhoods consistently have lower housing values (see Appendixes A and B). Although inner-suburban black neighborhoods are safer and have lower housing values, they suffer from households with lower income and higher unemployment in comparison to central city non-black neighborhoods. As contrast to central city non-black neighborhoods, the neighborhood amenities of these inner-suburban black neighborhoods are definitely less desirable neighborhood amenities than inner-suburban nonblack neighborhoods. This raises the question of why blacks cluster in inner-suburban black neighborhoods. A comparison of inner-suburban black neighborhoods with inner-suburban non-black neighborhoods indicates that the former have lower housing values (see Appendix A). In outer-suburban black neighborhoods, the set of desirable neighborhood characteristics is the same as those in outer-suburban non-black neighborhoods. Further analysis, however, suggests that outer-suburban black neighborhoods have lower housing values than outer-suburban non-. 64.

29 Eric Fong, Tracy Matsuo and Rima Wilkes black neighborhoods (see Appendix B). VII. Conclusion This paper explores an emerging common phenomenon of new immigrant groups clustering in suburban areas in multiethnic cities. We incorporated recent immigrant characteristics and urban structures by suggesting that suburbanization of new immigrant groups is not necessary related to increasing spatial contact with other groups as has commonly been understood. We argued that the households of new immigrant groups are concerned with neighborhoods safety. These groups are attracted to suburban neighborhoods which generally have lower crime rates. Since suburban neighborhoods have become more heterogeneous, households of new immigrant groups can look for safer neighborhoods in the suburbs without necessarily paying a higher housing cost than they would to reside in the central city. With such budget constraints, only a limited number of suburban neighborhoods are available. Subsequently, these households spatially cluster in suburban areas. Our analysis of the Toronto data in general confirms this picture. The results also show that black suburbanization differs from that of other groups. Blacks have to pay a higher housing price even when residing in suburban ethnic neighborhoods. This is because the housing price of non-black suburban neighborhoods is generally is much higher than black suburban neighborhoods. Blacks with limited housing budgets have little choice but to settle in black suburban neighborhoods. The study offers a few implications for the understanding of suburbanization. First, the understanding of suburbanization should not. 65.