Changes in aboriginal settlement patterns in two Canadian cities: a comparison to immigrant settlement models.

Peters, Evelyn J. ; Starchenko, Oksana M.

Abstract

This paper explores the changing settlement patterns of Aboriginal peoples in Winnipeg, Manitoba and Edmonton, Alberta, between 1971 and 2001. The objective is to compare these residential patterns to the traditional model for immigrant settlement put forward by the Chicago school model and supported by other researchers. We recognize that Aboriginal people are not immigrants. Immigrant models have often been employed to interpret Aboriginal peoples settlement patterns in cities, though, and we attempt to evaluate the applicability of these models to the Aboriginal experience, and to explore variations between these two cities. Our results suggest that the classical model of immigrant settlement patterns does not describe the urban Aboriginal experience very well. They also show that patterns vary substantially between cities. Interviews with key informants in the two cities suggest that much of this variation is related to differences in the spatial distribution of low-cost housing in Edmonton and Winnipeg.

Keywords: urban Aboriginal, residential settlement patterns, inner city

Resume

Cet article explore la mobilite residentielle des autochtones a Winnipeg, (Manitoba) et Edmonton (Alberta) entre 1971 et 2001. L'objectif est de comparer ces modeles residentiels au modele traditionnel d'amenagement des immigrants de l'ecole de Chicago et soutenu par d'autres chercheurs. Nous reconnaissons que les autochtones ne sont pas des immigrants. Toutefois, les modeles d'amenagement des immigrants ont souvent ete utilises pour interpreter l'amenagement des autochtones au sein de la ville. Ainsi, notre objectif est d'evaluer l'applicabilite de ces modeles traditionnels a l'experience autochtone, et analyser les variations entre les deux villes. D'apres les resultats, le modele classique d'amenagement des immigrants ne decrit pas adequatement l'experience urbaine des autochtones. De plus, les resultats indiquent des variations considerables entre les deux villes. Les entrevues, conduites dans les deux villes, indiquent que les variations sont liees aux differentes distributions spatiales du logement a loyer modique a Winnipeg et Edmonton.

Mots cles: mobilite residentielle, autochtones, quartiers centraux

**********

The model of changing settlement patterns of immigrants to American cities, put forward by Chicago sociologists in the early decades of the 1900s, continues to influence interpretations in urban geography, sociology and planning. According to this model, immigrants first move to reception areas in the inner city, and disperse to suburban areas as they assimilate (Burgess 1967). With increasing Aboriginal urbanisation in Canada after WWII, a number of researchers expected Aboriginal migrants to urban areas to follow patterns like turn of the century immigrant groups (Kerri 1976:147; Renaud 1961, Svenson 1978). However numerous researchers have questioned the universality of early settlement models, suggesting that they reflect the social and economic structure of urban areas at the time, the unique characteristics of particular immigrants, and the nature of immigration policy and administration at a particular juncture (for example Bourne 1989, Burnley and Hiebert 2001, Yancey et al 1976). Moreover the Aboriginal population differs from the immigrant population in a number of significant ways, so it is not certain that Aboriginal people in cities would follow the same settlement strategies as immigrant groups.

The paper compares three decades of Aboriginal people's settlement patterns in two Canadian cities to the Chicago school model and to more contemporary research. Clearly, Aboriginal people are not immigrants to cities, and this comparison is not meant to suggest that they are like contemporary ethnic groups. Instead, it is important to recognize that the Aboriginal groups migrating to urban areas were often removed from those very spaces, during the development of settler cities. However, comparing settlement patterns can be a first step in understanding both the nature of the urbanization experience for Aboriginal people, and the characteristics of the contemporary cities to which they migrate.

The paper begins with a brief review of some of the main themes in the literature on immigrant settlement patterns in cities. It then summarizes existing research on urban Aboriginal settlement patterns in Canada. Following a description of methods and data, the paper describes changing Aboriginal settlement patterns in Winnipeg, Manitoba and Edmonton, Alberta, between 1971 and 2001, and compares these patterns to models put forward in the immigrant literature. Finally the paper summarizes interviews with key informants that help interpret different patterns in the two cities. The implications of the research are described by way of conclusion.

Models Of Immigrant Settlement Patterns In Cities

In 1925, Ernest Burgess (1967) published a paper on city growth that would subsequently be adopted as a model for immigrant settlement patterns by generations of future researchers. The paper described a "zone of deterioration" near the central business district as the first area of settlement for immigrant groups, with gradual movement toward more suburban areas with success and in subsequent generations. Burnley and Hiebert (2001:127-8) summarised the characteristics of immigrant settlement patterns associated with researchers who developed Burgess's model. The model assumes that the least expensive housing is in the inner city, and because immigrants generally arrive with little wealth, they settle in inner city neighbourhoods. These reception areas assist immigrants in adjusting to their new environments. With time, immigrants improve their linguistic and employment skills, and become upwardly mobile. Socioeconomic mobility is associated with acculturation, relocation into better housing, and, in the process, dispersal into suburban areas, either by the same or by subsequent generations. Lieberson's (1963) analysis of changes in the settlement patterns of immigrants in ten US cities from 1930 to 1960 found that more recent immigrants were located nearer the centre of the city, and that there was a general outward movement and dispersion over time. Lieberson (1963:15-16) concluded that, when economic differences were taken into account, "the degree of suburbanization of immigrant groups and their children generally fits into the Burgess model of centralisation." This model still influences urban studies. For example, Clark's (1998:140) examination of immigration in California maintained: "When immigrants first arrive, they cluster in central locations dominated by members of their own ethnic group, a tendency that is no different in 1998 than it was in 1908."

However a number of researchers have questioned the universal applicability of this model. Ward (1971) drew attention to the particular developments in industry and transportation that provided the context for immigrant settlement patterns in urban areas around the turn of the century. Warner and Burke (1969) noted that most immigrants to American cities did not follow this pattern, and that these settlements patterns were applicable to a limited time span (1880-1940), to the largest cities, and to southern and eastern Europeans. Yancey et al (1976:399) argued that "much of the behaviour that is commonly associated with ethnicity is largely a function of the structural situations in which groups have found themselves." Boal (1999) identified five ethnic settlement scenarios. Initial concentration of ethnic groups in one location such as inner city and subsequent dispersion due to assimilation to the "mainstream" society constituted only one of five possible scenarios, four others being pluralism, segmentation, polarization and ethnic cleansing.

Recent work on immigrant settlement patterns supports the idea that researchers should not expect the model developed to describe turn of the century immigration to predict patterns for all cities, time periods, or for groups migrating to the city. This literature is voluminous, and we refer to only a few more recent Canadian examples that suggest that the Chicago School traditional ecology model is particular rather than general, and that current research has moved beyond the Chicago model. A number of contemporary studies show that some ethnic groups move directly to suburban locations, rather than initially to inner city areas (Archembault et al 1999, Chard and Renaud 1999, Doucet 1999, Ray 1992, Ray et al 1997). Some ethnic groups remain highly concentrated even when they settle in suburban areas (Breton et al 1990, Burnley and Hiebert 2001, Driedger 1996). Bourne (1989:314) suggested that contemporary urban areas demonstrated a "more detailed and fragmented ethnocultural mosaic" than put forward in the Chicago School's traditional ecological model. This proposal is supported by Bauder and Sharpe's (2002) recent paper on residential segregation of visible minorities in Vancouver, Toronto and Montreal, in which they found evidence of fragmentation and dispersal between 1986 and 1996. Finally, the spatial structure of urban areas also seems to affect settlement patterns (Bauder 2001, Bauder and Sharpe 2002, Fong 1996). All of these examples suggest that the residential patterns of migrants to urban areas may be unique to particular groups, the structure of particular cities, and the socioeconomic circumstances surrounding urbanization. With reference to traditional models of immigrant settlement patterns, Bauder and Sharpe suggested (2002:219) that "we should not conceptualise residential patterns as 'natural' trends and spatial laws. In other words, we must avoid the trap of spatial fetishism which lures with the idea that space alone can explain residential configuration."

While there is considerable evidence that the Chicago School model is not appropriate to all contemporary immigrant settlement patterns, there is, at present, very little work that explores changing residential patterns of Aboriginal people in Canadian cities. In this context, we argue that comparing urban Aboriginal settlement patterns to the Chicago school model represents a useful place of departure. In this paper, we examine four dimensions of ethnic settlement patterns found in the Chicago school framework, and explore the degree to which the residential locations of Aboriginal people in the two Canadian cities with the largest Aboriginal populations replicate these patterns between 1971 and 2001. We assess whether Aboriginal people were most likely to be found in inner city areas in these cities in 1971, and explore the extent to which the Aboriginal population is clustered or dispersed within the inner city. Then we discuss whether or not Aboriginal settlement patterns in these cities show movement toward suburbanization by 2001, and ask whether Aboriginal people living in suburban areas are clustered or dispersed. (1) Finally, we report on interviews with key informants who gave their interpretations of Aboriginal settlement patterns in each of the cities.

Aboriginal Urbanization in Canada

The 1951 Census of Canada showed that approximately seven percent of the Aboriginal population lived in cities. The Aboriginal population in large cities counted in the 1951 Census varied between 48 individuals in Saskatoon to 805 individuals in Toronto. By 2001, the proportion of the Aboriginal population living in cities had had increased to 49 percent (Statistics Canada 2003). Aboriginal populations varied from 8,200 in Thunder Bay to 55,755 in Winnipeg. (2) As a proportion of total Census Metropolitan Area (CMA) populations in 2001, Aboriginal populations varied between 0.3 percent of the Montreal population to 9.1 percent of the Saskatoon population. Aboriginal people differ from immigrants because they are migrating to cities within their homelands. However their relatively lower levels of education and income compared to most urban populations mean that they face some of the same challenges earlier immigrants faced in moving to North American cities.

There is relatively little research available that analyses urban Aboriginal people's settlement patterns. The work that does exist demonstrates some contradictions. On the one hand, concerns about Aboriginal urbanization at mid-century emphasized the possibility of the emergence of slums in inner cities (summarized in Peters 2000). Other research suggests Aboriginal people were scattered throughout urban areas even in the 1960s and 1970s, and did not form ethnic neighbourhoods thought to be common to the experience of immigrant groups (Davis 1965:361-2, Dosman 1972:183-4, Johnston 1979, Krotz 1980, Nagler 1970:65). In fact, the lack of Aboriginal concentrations in the city and assumptions about the role of neighbourhoods in facilitating adaptation stimulated several suggestions to encourage urban Aboriginal people to follow the traditional pattern of ethnic groups (Damas & Smith 1975, Dosman 1972:183). For example, in 1978 a consultant describing the situation of urban Aboriginal people in Saskatchewan suggested that they should be encouraged to cluster and build urban support networks (Svenson, 1978:22).</p> <pre> Housing programs need to be adjusted to facilitate the development of 'Indian [sic] Neighbourhoods' in urban centres. Policies of intentionally scattering Indian housing throughout non-Indian neighbourhoods should be abandoned. This policy isolates Indian people from any social support and restricts the possibility of successful adjustment to urban living. </pre> <p>There is relatively little work that explores contemporary urban Aboriginal settlement patterns. Some of what is available, though, suggests that urban Aboriginal people are not currently concentrated in particular areas of the city, nor invariably located in the inner city. Work using indices of dissimilarity (3) to describe urban Aboriginal settlement patterns uniformly concludes that segregation is low to moderate. Values up to 0.3 are considered low, 0.4-0.5 are moderate and 0.6 and over are considered high. Researchers using 1981, 1991 and 1996 single origin ancestry census data for CMA's have found moderate dissimilarity indices for Aboriginal people, ranging from about 0.2 to about 0.4 (Bauder and Sharpe 2002, Bourne et al 1986, Clatworthy 1994:256, Maxim et al 2000). Clatworthy (1994:259) found that migrants (people who had moved to urban areas in the last five years) were not more residentially segregated than non-migrants, suggesting that a model that associates concentration with arrival in the city may not accurately describe the urban Aboriginal experience. Maxim et al (2000) found that, while Aboriginal people were more likely to be located in central areas in some cities, this was not the only or even the main settlement pattern.

At the same time, other analyses drew conclusions about urban Aboriginal settlement patterns that assumed segregation and an inner city location (Darden and Kamel 2002, Krauter and Davis 1978:20). Drost (1995:47-48), in his multivariate analysis of determinants of unemployment in 1991 CMA's, suggested that high unemployment rates of urban Aboriginal people were partially related to residential segregation in the core areas of western CMAs. Richards' (2001) study of western cities, Toronto and Montreal, used the model of US black inner city ghetto concentrations to interpret his results.

The existing work on Aboriginal settlement patterns, then, seems to raise a number of questions. Early observers worried about Aboriginal concentrations in cities, but available studies suggested that there were no urban Aboriginal neighbourhoods. More contemporary work using the index of dissimilarity indicated low to moderate levels of segregation with inconsistent centralization, but some other work suggested that inner city concentration was a feature of urban Aboriginal settlement patterns. In this paper we compare settlement patterns in two cities over three decades in order to provide a basis for assessing similarities between urban Aboriginal settlement patterns and classical models of immigrant residential settlement patterns.

Data and Methods

The following section describes the decisions about definition and measures that were used to assess aspects of Aboriginal settlement patterns. It also outlines the protocols the guided key informant interviews in Edmonton and Winnipeg.

Definitions and Boundaries

There are a variety of definitions of 'Aboriginal' using census data, and the definition chosen affects on the socio-economic characteristics of the population (Siggner 2003). We chose to use Aboriginal single origin ancestry data for this study. This definition limits the number of people in the group to those with particular "direct" ancestry, excluding people with multiple origins (Bauder and Sharpe 2002:208). Other possible definitions were the multiple origin ancestry data (people who indicated that they had Aboriginal and other group ancestry) and identity data (people who identified as Aboriginal). However these data were not available in 1971, making it impossible to compare patterns over several decades. Moreover, the majority of the Canadian researchers examining ethnic residential segregation employ single origin ancestry data in their studies. We argue that this measure is most appropriate for the analysis that follows.

We chose the cities of Winnipeg and Edmonton because they have the largest and the second largest single origin ancestry Aboriginal populations respectively, of large Canadian cities. Our analysis focuses on Aboriginal settlement patterns within the city because both cities have First Nations reserves within their CMA boundaries. Because these are areas of concentration of First Nations people, they represent different settlement patterns from those of the urban core, and they would distort the analysis. We focus on census tracts because this is the unit of geography used in the majority of studies of ethnic immigrant settlement within cities (Deacon 2002), and because census tracts are considered the best and closest practical approximation to the concept of neighbourhood within the census (White 1987). To ensure the accuracy of comparison, the boundaries of both cities were kept at their 1971 limits. (4)

Inner city and suburban areas were identified on the basis of the age of the housing stock in 2001, and those boundaries were used for the 1971 analysis. Arguably, the age of housing construction identifies areas of the city that came into being during periods with distinct social and economic characteristics (Filion and Bunting 1991), and also corresponds to general perceptions of city regions. The inner city was defined as those census tracts where the proportion of housing built before 1946 was twice the metropolitan average in 2001. Census tracts that did not meet the criteria for the inner city, but were surrounded on three sides by inner city tracts were classified as 'inner city' to allow us to include areas that have been redeveloped. All other tracts were identified as suburban.

Comparing Residential Patterns

While index-based approaches have been employed in numerous studies on ethnic residential patterns in Canadian and the US cities (for example Massey and Denton 1989, 1993; Maxim et al 2000), they are not conducive to comparative studies because they are dependent on a group's absolute and relative size within the city (Voas and Williamson 2000; Poulsen, Forrest and Johnston 2002). Some researchers have introduced a measure based on calculating the percentage of an minority group's population found in certain areas (usually census tracts) where they comprise different proportions of the population in that area--for example, the proportion of the group found in census tracts where they comprise 20 percent of the population, 30 percent of the population ... and so on (Johnston et al 2001; Peach 1996; Poulsen and Johnston 2000). The main advantage of this method over the traditional index-based approach is that it employs absolute rather than relative measures of residential segregation, thus "providing a valid and robust measure for comparative studies over time and space" (Johnston et al. 2001:151). In this study, we rely mainly on a this approach, using census tracts and inner city and suburban areas as a basis, and documenting changes in this distribution over time. However, we retain the dissimilarity index because it has been widely used in ethnic segregation studies and thus provides a convenient "reference point."

Key Informant Perspectives

In the summer of 2002, interviews were conducted with Aboriginal service providers and individuals who were knowledgeable about the Aboriginal population in these cities. There were 13 interviews in Edmonton, and 21 interviews in Winnipeg. Interviewers showed participants maps (5) of the distribution of Aboriginal people in Edmonton and Winnipeg, and began with an open ended question that asked respondents to indicate what they thought affected Aboriginal people's decisions to locate in particular areas. Then interviewers probed for the effects of a number of factors: the location of services, housing costs, tenure and size, the location of employment opportunities, neighbourhood improvement initiatives, by-law enforcement and zoning regulations, the decisions of Native housing organizations, general attitudes towards an area, and historic Aboriginal attachments to particular areas. Interviews were taped, transcribed, and analysed using Atlas-ti software. The analysis looked for themes associated with the questions, and also for any elements that had not been expected. The interviews were exploratory, because there has been virtually no work that addresses factors affecting Aboriginal settlement patterns. Despite their exploratory nature, however, the interviews suggest some directions for interpreting settlement patterns in these two cities.

Analysis

The cities of both Winnipeg and Edmonton experienced substantial increases in Aboriginal populations between 1971 and 2001. In 1971, in Edmonton, the Aboriginal single origin ancestry population was 4,330 or slightly less than 1 percent of the city population. By 2001, the Aboriginal population had increased to 12, 015, or 2 percent of the city population. Winnipeg's Aboriginal single origin ancestry population in 1971 was 6,340, or 1.2 percent of the total city population. By 2001, it had increased to 22,265 or 3.6 of the total city population.

Our analysis of Aboriginal settlement patterns addresses four main questions that explore whether Aboriginal settlement patterns reflect dimensions of Burgess's model of change in immigrants settlement patterns over time. We begin by asking whether most Aboriginal people found in inner city areas at the beginning of the time period (1971), whether they are clustered or evenly distributed in these areas, initially, and whether this distribution changes. Then we examine Aboriginal settlement patterns at the end of the period (2001) to see whether Aboriginal people are more likely to be found in suburban areas at the time, and whether suburban populations are dispersed or clustered. Finally, we turn to an explanation of differences between the cities, based on key informant interviews.

Are Most Aboriginal People Found In Inner City Areas At The Beginning Of The Time Period (1971)?

In 1971 the majority of the Aboriginal population was found within the inner city boundaries of both Edmonton and Winnipeg (Table 1). This represented 4,780 people in Winnipeg and 2,530 people in Edmonton. However, there were considerable differences between the cities. In Winnipeg, over three quarters (75.4 percent) of the Aboriginal population was found in the inner city. In Edmonton, the proportion of the population was only slightly over half (58.4 percent). While Aboriginal people in Winnipeg could be seen to replicate the immigrant model of locating in the inner city, in Edmonton almost half live outside these boundaries in 1971.

If Aboriginal People Are Found In Inner Cities, Are They Clustered Or Evenly Distributed In These Areas, Initially? Does The Distribution Change Over Time?

Figure 1 describes the residential patterns of Aboriginal people in the inner city areas of Winnipeg and Edmonton. It shows the percentage of the total inner city Aboriginal population in census tracts where Aboriginal people comprise 0-4.9 percent of the tract population, 5-9.9 percent of the tract population, and so on. In 1971 in Edmonton, about half of the inner city Aboriginal population lived in census tracts where they comprised less than 5 percent of the total tract population. Tracts with this level of concentration comprised 25 of 26 inner city census tracts. In one census tract, Aboriginal people comprised between 5 and 10 percent of the population. There were no census tracts where Aboriginal people comprised more than ten percent of the total population. The dissimilarity index for Aboriginal people in the inner city of Edmonton was .239, which is low. In 1971 in Winnipeg, almost 40 percent of the Aboriginal population lived in census tracts where they comprised less than 5 percent of the tract population. Tracts with this level of concentration comprised 38 of the 45 inner city tracts. In six census tracts, 5 to 9.9 percent of the tract population was Aboriginal, and in one census tract, 10 to 19.9 percent of the population was Aboriginal. Less than 10 percent of the Aboriginal population lived in tracts where they comprised up to one fifth of the tract population. The dissimilarity index for this region of Winnipeg was .409, which shows a moderate level of concentration. In both cities, Aboriginal people were dispersed throughout the inner city.

[FIGURE 1 OMITTED]

While the distributions had changed by 2001, there was still no evidence that Aboriginal people were concentrated in a few inner city tracts. By 2001, the Aboriginal population in inner city Edmonton had increased to only 3,830 people. The dissimilarity index had increased to .377, but all of the Aboriginal people living in the inner city lived in census tracts where they comprised less than 10 percent of the census tract population. By 2001, inner city Winnipeg had 14,290 Aboriginal people, showing a much greater increase than Edmonton's inner city population. However, the dissimilarity index was .411, only slightly higher than the 1971 index (Table 1). There were fewer census tracts where Aboriginal people comprised less than 5 percent of the population, but the majority of Aboriginal people in the inner city lived in census tracts where they comprised one fifth or less of the population (Figure 2). There were 6 census tracts where Aboriginal people made up between 20 and 29.9 percent of the population, and two where they made up between 30 and 39.9 percent of the population. In summary, Aboriginal people were relatively dispersed in both inner cities, but the degree of dispersion was slightly higher in Edmonton. At this scale of analysis, neither city appeared to have densely concentrated, inner city, Aboriginal neighbourhoods.

[FIGURE 2 OMITTED]

Are Aboriginal people more likely to be found in suburban areas at the end of the time period (2001)?

While the proportion of the Aboriginal population living outside the inner city increased in both cities, the change was more dramatic in Edmonton than in Winnipeg. In Edmonton, most of the growth of the Aboriginal population occurred in suburban areas. The total Aboriginal population grew from 4,330 in 1971 to 12, 015 in 2001, but the suburban population grew from 1,800 in 1971 to 8,185 in 2001. In 2001, more than two thirds (68.1 percent) of the Aboriginal

population in Edmonton lived outside the inner city. By 2001, the Aboriginal population of Winnipeg had grown from 6.340 to 22, 265 (Table 2). Only about one third (35.8 percent) of that population lived in suburban areas. While the suburban Aboriginal population grew from 1,560 to 7,975, the inner city Aboriginal population grew even more rapidly (Figure 1). The aspect of the Chicago school model that predicts suburbanization over time describes the Edmonton situation, then, but not the Winnipeg situation.

How Are Aboriginal People Living In Suburban Locations Distributed, And Does This Distribution Change Over Time?

In 1971, 1,560 Aboriginal people lived in suburban census tracts in Winnipeg, and all of them lived in census tracts where they comprised less than 5 percent of the population. The dissimilarity index for Aboriginal people in suburban Winnipeg was .315. The situation was virtually identical for Edmonton, where 1,800 Aboriginal people lived in suburban locations, all in census tracts where they comprised less than 5 percent of the population, and with a dissimilarity index of .346. In both cities, all suburban Aboriginal people lived in census tracts where they comprised less than 5 percent of the tract population (Figure 3). Although the number of Aboriginal people living in suburban areas had increased by 2001, their settlement patterns remained very similar to the situation in 1971. In 2001, in Edmonton, the suburban Aboriginal population had grown to 8,185, with a dissimilarity index of .389. No Aboriginal people lived in census tracts where they comprised more than 10 percent of the tract population. In Winnipeg, 7,975 Aboriginal people lived in suburban areas in Winnipeg. The dissimilarity index was .360, and they comprised between 10 and 19.9 percent of the tract population in only 2 census tracts. At the scale of census tracts, then, there appears to be relatively little concentration of Aboriginal people in suburban areas in either of these cities.

Perspectives on Aboriginal Settlement Patterns in Winnipeg and Edmonton

Aboriginal settlement patterns in Edmonton and Winnipeg were different from each other in 1971, and by 2001 they had become even more differentiated. While neither the Edmonton nor the Winnipeg Aboriginal population was highly segregated, the Edmonton Aboriginal population was even more dispersed than the Winnipeg Aboriginal population. Most of the Aboriginal population in Edmonton lived in suburban areas in 2001, while most of the Aboriginal population in Winnipeg lived in inner city areas. Interviews in 2002 with key informants in Edmonton and Winnipeg provided some interpretation of these patterns.

In a series of interviews, key informants from Aboriginal service providers and individuals who were knowledgeable about urban Aboriginal populations in these cities identified a number of common factors contributing to urban Aboriginal settlement patterns in both cities. The ones mentioned most often were the location of low cost housing, the location of other Aboriginal people in decisions about location, and the location of services. However, participants also identified some important differences between the two cities. A brief summary of the results follows. Quotations are identified only by interview and page number for reasons of confidentiality. While there is only one quotation for each theme, each quote represents elements also identified by other participants.

In Winnipeg, participants most often mentioned the cost of housing as a factor contributing to Aboriginal people's settlement patterns. The inner city area was one that had low-cost housing historically, and at present.</p> <pre> There's concentration of people in inner city areas, which are the low income, lower cost neighbourhoods in terms of housing. Given the poverty situation, poverty statistics of Aboriginal people, particularly those that are newcomers to Winnipeg, it's not surprising to see the concentration of higher numbers in

the areas where the housing is less expensive. That would probably be one of the major factors (Q2:10). </pre> <p>When we started buying homes back in the seventies, homes that were affordable were in this area. So that dictated a large part of that. You buy where you can afford to buy. Houses were very low at that time, even with repairs. Your market was very restricted. I would say that nothing has changed over the years. The prices that we were allowed went up, but the market didn't change much. That's why you see the population here, here, and here [points to map]. The maximum prices went up, but it didn't give us free range of the city. Our intention was to purchase these older homes and renovate them. At that time, we would provide not only housing opportunities, but employment and training as well in doing the renovations. One of the things we looked at was the long term effects of having houses scattered too far. If we were able to centralize, our maintenance people didn't have to drive all over the city to each house. They would be going from house to house in smaller areas. That decision had an effect on how this map is, for sure (Q14:6).

The next most often cited factor influencing location decisions was the location of other Aboriginal people.</p> <pre> I know people who come to the city ... with modest resources often come from northern communities, reserves, or whatever. They end up gravitating to the areas in which others that they know live. It's not uncommon to have people staying with others for a point in time, but moving very little beyond that (Q9:3). </pre> <p>One respondent argued that the decision to locate near acquaintances was linked to the decision to live in areas with lower cost housing, but that the decision-making process was not straightforward.</p> <pre> Well, the most important one is ... when I first came into the city, I'll take my own case, I came to where people were. They happened to be in a specific location at that time and that's how I ended up there. It was only after I got a bit climatized and involved with what was going on in the city ... all of a sudden options opened up.... And housing and cost of living and all that other stuff ...more or less you bought into

that whole package when you moved into that particular area. So,

it wasn't really a conscious choice. I don't think that most Aboriginal people, when they come look at all these different things, they don't say "is the cost of housing going to be this, is the cost of living going to be" ... I don't think they look at it like that at all. I think they look at it in terms of relationships they have with people. They say, "we're here because my cousin's here," or there's some connection with people that are in the area. All this other stuff is just secondary (Q13:7). </pre> <p>In Winnipeg, respondents indicated that the services generally located where Aboriginal people were, rather than Aboriginal people locating close to services.</p>

<pre> Basically, what happens is a lot of the people come who already have relatives that are living in the city. I think that more determines where they're going to live, than the services that

are available to them. They don't know where the services are, when they first get here. If you've come from a rural area and, if you're on social assistance, you're going to either go to some relatives you know. They learn about the services after they get here. I don't think that would determine where they're going to live. Services locate where the people are, rather than the people locating where the services are.... Ma Ma Wi was one of the agencies that really looked at where their clientele were ... to decide where to operate their services. Because of the various needs and what not, they accommodated that. Other groups, they're already there for the people. Even ourselves, we're right in the midst of the Aboriginal community ... in the Point Douglas area. We just, naturally, would be here. That's where the Aboriginal Centre is, as well. It's right in the heart of the Point Douglas area, where there's a lot of people (Q16:14). </pre> <p>Like the Winnipeg key informants, Edmonton respondents identified the location of low-cost and Native-run public housing as important in affecting location decisions. However, in Edmonton, this meant that Aboriginal people were dispersed rather than clustered. Part of this was due to the history of development in Edmonton. The slow development of residential areas allowed for a dispersal of low-income/Aboriginal populations across the city rather than a concentration in one or several particular areas.</p> <pre>

One of the things that is characteristic of Edmonton early development pattern, is that the development was scattered. And it took a long time for the neighbourhoods to develop--and it goes back to the pattern where I showed you on this map, because of excessive subdivision of land. Even close to the center of the city, you have a pattern of comparatively few houses that are in some empty lots. Like, um, 30, 40, 50 years. They were laying along side of the university (Q6:15). </pre> <p>In addition, there were explicit municipal policies that encouraged developers to provide a diverse range of housing types (e.g. apartments, single family dwellings, affordable housing) in an area. As a result, there was affordable housing throughout the city.</p>

<pre> Now the specific change that [former mayor] made was a

land use development policy that required new neighbourhoods to be designed with housing appropriate to all, if you will, financial classes of people, and so if all the housing and all that, the only housing (more correctly) that was available to people that was living in poverty, was, prior to that, was in the inner city, any new neighbourhoods that were built prior to that had been single family, complete with middle class and upper class housing. There was no mixed housing in the neighbourhoods. You know there was following [former mayor's] position of this by-law and policy, every new neighbourhood had single family, had multi-residential, that is to say apartment building of various scales and low income supported housing. Every new neighbourhood. And that remained the case for about 20 years, until about 1991 actually. So, as a result of that, the various social services agencies,

support agencies for people in poverty had available to them, income supported housing in other areas in the city (Q5:14). </pre> <p>Edmonton participants noted that several Native non-profit housing organizations located their housing units outside of the inner city, because of the dispersal of low-cost housing. As a result, Aboriginal populations in Edmonton were also relatively dispersed.

Like key informants in Winnipeg, Edmonton participants indicated that Aboriginal people chose to locate close to friends and family already in the city in order to receive additional support (e.g. how to get around town, what services are available, provide social activities) and to be around people of the same culture. The word "comfortable" was used in both the Winnipeg and Edmonton interviews to describe the importance of initially being located close to other Aboriginal people. Being located close to people of the same culture could ease some of the pressures and stresses of adjusting to an urban way of life. However, as the following quote shows, this meant that in Edmonton, Aboriginal people were found in a variety of locations.</p> <pre> I guess for the reason why they would go inner city, unfortunately, sometimes they're, a lot of their relatives will have moved there, because rent was cheaper. Often the housing is substandard, but you know people there, and usually you see you know what's available for rent and things and there is some low income housing also in that area. The same with Abbottsfield you have a lot of ... lower income housing, subsidized housing, and also the same in the south east and in the west end, yes, there would be more, a lot more people out there. And again what happens with First Nations is just the

comfort level. Like, you know your people are around there and everything, there is not that stigma, you know. And they tend to move together as well. If one family moves in and have found a house, another family will move in to that area because there, pretty soon you've got enough people there (Q10:7). </pre> <p>To summarize then, key informants in Edmonton and Winnipeg identified very similar objectives underlying Aboriginal people's choices of location in the city. Aboriginal people located in areas with less expensive housing, and they located near other Aboriginal people. However, because the location of this type of housing varied between the two cities, the settlement patterns of Aboriginal people also varied. These views support arguments that the spatial structure of local housing markets affects minority group settlement patterns (Bauder and Sharpe 2002, Yancey et al 1976).

Conclusion

Moreover, settlement patterns varied substantially between the two cities studied, suggesting that the particular characteristics of urban areas need to be taken into account. Interviews with key informants provided some interpretation of these patterns. While interviewees identified similar factors affecting the location decisions of Aboriginal migrants to cities--the location of low-cost housing, the location of other Aboriginal people, and the location of services--the spatial distribution of low-cost housing in these cities varied substantially, with the result that settlement patterns also varied. Because low-cost housing was more dispersed in Edmonton, the Aboriginal population and services that targeted that group were also more dispersed than in Winnipeg. These results support researchers who have suggested that the spatial structure of housing affects settlement patterns (Bauder 2001, Bauder and Sharpe 2002, Fong 1996, Yancey, 1976).

References

Archembault, J., B. Ray, D. Rose and A-M Seguin. 1999. Atlas Immigration et Metropoles. http:/im.metropolis.net/research-policy/research_content/atlas/ index.html

Bauder, H. 2001. Visible minorities and urban analysis. Canadian Journal of Urban Research 10 (1):69-90.

Bauder, H. and B. Sharpe. 2002. Residential segregation of visible minorities in Canada's gateway cities. The Canadian Geographer 46 (3):204-222.

Boal, F.W. 1999. From Undivided Cities to Undivided Cities: Assimilation to Ethnic Cleansing. Housing Studies 14 (5):585-600.

Bourne, L.S. 1989. Are new urban forms emerging? Empirical tests for Canadian urban areas. The Canadian Geographer 33:312-328.

Bourne, L.S., A.M. Baker, W. Kalbach, R. Cressman, and D. Green. 1986. Canada's Ethnic Mosaic: Characteristics and Patterns of Ethnic Origin Groups in Urban Areas. Toronto: Centre for Urban and Community Studies, University of Toronto.

Breton, R., W. W. I. Warren, E. Kalbach, and J. G. Reitz. 1990. Ethnic Identity and Equality. Varieties of Experience in a Canadian City. Toronto: University of Toronto Press.

Burgess, E. W. 1967. The Growth of the City: An Introduction to a Research Project. In The City, ed. R. E. Park, E. W. Burgess and R. D. McKenzie, 4762. First published in 1925. Chicago: University of Chicago Press.

Burnley, I. and D. Hiebert. 2001. Emerging patterns of immigrant settlement at the metropolitan scale. In Progress in Planning 55, ed. David Ley and R Murphy, 127-140.

Chard, J. and V. Renaud. 1999. Visible minorities in Toronto, Vancouver and Montreal. Canadian Social Trends 54:20-25.

Clark, W.A.V. 1998. The California Cauldron: Immigration and the Fortunes of Local Communities. New York and London: Guilford Press.

Clatworthy, S. 1994. The Migration and Mobility Patterns of Canada's Aboriginal Population. Prepared for the Royal Commission on Aboriginal Peoples. Winnipeg, Manitoba: Four Directions Consulting Group.

Damas & Smith Ltd. 1975. A Report on the Feasibility Study prepared for Neeginan (Manitoba) Incorporated. Winnipeg: Damas & Smith Ltd.

Darden, J.T. and S. M. Kamel. 2002. The Spatial and Socioeconomic Analysis of First Nation People in Toronto CMA. The Canadian Journal of Native Studies 22 (2):239-268.

Davis, A.K. 1965. Edging Into Mainstream. Urban Indians in Saskatchewan. Bellingham: Western Washington State College, Bellingham.

Deacon, P. 2002. Urban Aboriginal Population Distribution- Winnipeg and Thompson, CHMC (unpublished).

Dosman, E.J. 1972. Indians: The Urban Dilemma. Toronto: McClelland and Stewart.

Doucet, M. 1999. Toronto in Transition: Demographic Change in the Late Twentieth Century. Virtual Libra,, Joint Centre of Excellence for Research on Immigration and Settlement - Toronto. http://ceris.metropolis.net/vl/other/ doucet2.html

Driedger, L. 1996. Multi-ethnic Canada: Identities and Inequalities. Toronto: Oxford University Press.

Drost, H. 1995. The Aboriginal-White Unemployment Gap in Canada's Urban Labor Market. In Market Solutions for Native Poverty: Social Policy for the Third Solitude, ed. H. Drost, B. L. Crowley, R. Schwindt. Toronto: C.D. Howe Institute.

Filion, E and T. Bunting. 1991. Introduction: Perspectives on the City. In Canadian Cities in Transition, ed. T. Bunting and P. Filion, 2-22. Toronto: Oxford University Press.

Fong, Eric. 1996. A Comparative Perspective on Radical Residential Segregation: American and Canadian Experiences. The Sociological Quarterly 37 (2): 199226.

Johnston, F. 1979. Researcher, Institute of Urban Studies, University of Winnipeg, Winnipeg, Manitoba, personal communication.

Johnston, R., J. Forrest, and M. Poulsen. 2001. Sydney's Ethnic Geography: new approach to analysing patterns of residential concentration. Australian Geographer 32(2): 149-162

Kerri, J.N. 1976. Indians in a Canadian City: Analysis of Social Adaptive Strategies. Urban Anthropology 5(2): 143-156.

Krauter, J.F. and M. Davis. 1978. Minority Canadians: Ethnic Groups. Toronto: Methuen.

Krotz, L. 1980. Urban Indians: the Strangers in Canada's Cities. Edmonton: Hurtig Publishers Ltd.

Leiberson, S. 1963. Ethnic Patterns in American Cities. New York: The Free Press of Glencoe.

Massey, D. S. and N.A. Denton. 1993. American Apartheid." Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press.

Maxim, P.S., J.P. White, P.C. Whitehead, D. Beavon. 2000. Patterns of Residential Settlement among Canada's First Nations Communities. London, Ontario: Population Studies Centre, University of Western Ontario.

Nagler, M. 1970. Indians in the City. Ottawa: Canadian Research Centre for Anthropology, St. Paul University.

Norris, M. J. and S. Clatworthy. 2003. Aboriginal Mobility and Migration within Urban Canada: Outcomes, Factors and Implications. In Not Strangers in These Parts: Urban Aboriginal People, eds. D. Newhouse and E. Peters, 51-78. Ottawa: PCRI.

Peach, C. 1996. Does Britain have ghettoes? Trans. Inst. Br. Geogr., NS 21:216-235.

Peters, E. J. 2000. Aboriginal People in Urban Areas. In Visions of the Heart: Canadian Aboriginal Issues, 2nd edition, ed. D. Long and O. P. Dickason. Harcourt Canada, Toronto.

Poulsen, M.F., J. Forrest, and R. Johnston. 2002. Contemporary ethnic residential segregation in four US metropolitan areas. Cities 19(3): 161-72.

Poulsen, M. F. and R. J. Johnston. 2000. The Ghetto Model and Ethnic Concentration in Australian Cities. Urban Geography 21 (1):26-44.

Renaud, Father, Director-General of the Oblate Fathers' Indian and Eskimo Welfare Commission, Ottawa, Ontario. 1961 Address to the Third Annual Short Course on Northern Community Development. Centre for Community Studies, University of Saskatchewan, Saskatoon, Saskatchewan, April 14.

Richards, J. 2001. Neighbours Matter. Poor Neighbourhoods and Urban Aboriginal Policy. Toronto: C.D. Howe Institute.

Svenson, K. 1978. The Explosive Years: Indian and Metis Issues in Saskatchewan to 2001. Prepared for the Department of Indian Affairs and Northern Development, Regina, Saskatchewan.

Ray, Brian K. 1992. Immigrants in a 'multicultural' Toronto: exploring the contested social and housing geographies of post-war Italian and Caribbean immigrants. Unpublished PhD Dissertation, Queen's University, Kingston, Ontario.

Ray, Brian K., Greg Halseth and Benjamin Johnson. 1997. The Changing 'Face' of the Suburbs: Issues of Ethnicity and Residential Change in Suburban Vancouver. International Journal of Urban and Regional Research. 21:75-99.

Siggner, Andrew J. 2003. The Challenge of Measuring the Demographic and Socio-Economic conditions of the Urban Aboriginal Population. In Not Strangers in These Parts: Urban Aboriginal People, ed. David Newhouse and Evelyn J. Peters, 119-130. Ottawa: Policy Research Institute.

Statistics Canada. 2001. Aboriginal peoples of Canada, 2001 Census Analysis series, http://www12.statcan.ca/english/census01/Products/Analytic/ companion/abor/canada.cfm (Accessed January 21, 2003.)

Voas, D. and P. Williamson. 2000. The scale of dissimilarity, measurement and an application to socio-economic variation across England and Wales. Trans Inst Br Geog NS 25:465-81.

Ward, David. 1971. Cities and Immigrants. A Geography of Change in Nineteenth Century America. New York: Oxford University Press.

Warner, Sam Bass and Colin Burke. 1969 Cultural change and the ghetto. Journal of Contemporary History 4:173-188.

White, M. J. 1987. American Neighborhoods and Residential Differentiation. New York: Russell Sage.

Yancey, William L., Eugene P. Ericksen and Richard N. Juliani. 1976. Emergent Ethnicity: A Review and Reformulation. American Sociological Review 41(3):391-403.

Notes

(1) The data do not support an analysis of whether or not particular Aboriginal populations are moving to different urban locations within cities. In other words, we cannot tell whether individuals living in suburban areas moved from inner city areas, or whether they moved directly from rural and reserve communities to particular locations in the city. Urban Aboriginal people exhibit rates of movement that are higher than non-Aboriginal populations (Norris and Clatworthy 2003). Some of this movement involves relocation within urban areas, and some of it involves migration between urban areas and between urban, and rural and reserve locations. The mobility patterns of individuals and households remain an important area of research that is beyond the scope of this paper.

(2) The 1951 population figure is based on a question that asks about Aboriginal ancestry and does not include Metis populations. The 2001 figure is based on a question that asks about Aboriginal identity.

(3) The index of dissimilarity measures whether a minority population is distributed more or less evenly than the population (usually the host population) to which it is being described.

(4) In case of Winnipeg, this was relatively easy to do, as the city boundary remained constant between 1971 and 1996. In 1996, the city actually became smaller when parts of two suburban census tracts were included into the rural municipality of Headingley. Consequently, we included these dissemination areas into the analysis of 2001 data for Winnipeg. On the other hand, the city of Edmonton experienced considerable growth between 1971 and 2001, which was reflected in the continuous boundary changes during this period. However, when accounting for subdivision within the tracts, the boundaries of the majority of census tracts falling within 1971 city limits remained fairly constant. If in 1981, 1991 or 2001, the census tract boundaries were crossing the 1971 city limits, population data for these tracts were adjusted to that of enumeration areas falling within the 1971 limits, a method used by Fong (1996:207).

(5) Because the 2001 data were not available at the time, the maps were based on 1996 data. These maps show a similar patterns as the 2001 distributions documented in this paper.

Evelyn J. Peters

Oksana M. Starchenko

Department of Geography

University of Saskatchewan

Figure 3. Percent of Total Suburdan Aboriginal Population Living in
Suburban Tracts, by Percent of the Census Tract Population that is
Aboriginal, Winnepeg and Edmonton, 1971-2001

 Edmonton Winnipeg

% of the Tract Population Aboriginal

0-4.9 53 55*
5-9.9
10-19.9

% of Total Suburban Aboriginal Population

0-4.9 105 104
5-9.9 10 5
10-19.9 2
20-29.9

* Number of tracts

Note: Table made from bar graph.

Source: Statistics Canada, special cross-tabulations

Table 1: Settlement Patterns, Aboriginal Single Origin Ancestry
Population, Edmonton and Winnipeg, 1971

 % of total
 Aboriginal Aboriginal Dissimilarity
City part population population Index

Edmonton
City 4,330 100.00 0.356
Inner city 2,530 58.43 0.239
Suburbs 1,800 41.57 0.346

Winnipeg
City 6,340 100.00 0.479
Inner city 4,780 75.39 0.409
Suburbs 1,560 24.61 0.315

Source: Statistics Canada, special cross-tabulations

Table 2: Settlement Patterns, Aboriginal Single Origin Ancestry
Population, Edmonton and Winnipeg, 2001

 % of total
 Aboriginal Aboriginal Dissimilarity
City part population population Index

Edmonton
City 12,015 100.00 0.402
Inner city 3,830 31.88 0.377
Suburbs 8,185 68.12 0.389

Winnipeg
City 22,265 100.00 0.518
Inner city 14,290 64.18 0.411
Suburbs 7,975 35.82 0.360