No. City Country 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Rank
40 63 Vancouver Canada * * 41 68 Seattle US * * 42 71 Auckland N. Zealand * * * 43 72 Oslo Norway * *
44 76 Beta Manchester UK * * 45 79 Cape Town S. Africa * * 46 82 Minneapolis US * * 47 87 Brisbane Australia * * 48 88 Geneva Switzerland * * * 49 90 beta- Detroit US * *
final dataset for ethnic segregation analysis, by census year of the data No. City Country 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Rank 50 91 Denver US * * 51 93 Bratislava Slovakia * * 52 97 Stuttgart Germany * * 53 99 Cologne Germany * * 54 100 beta- St Louis US * * 55 103 San Diego US * * 56 105 Perth Australia * * 57 107 Cleveland US * * 58 109 Calgary Canada * * * 59 116 Cincinnati US * * 60 117 Charlotte US * * 61 118 Antwerp Belgium * * * 62 121 Baltimore US * *
63 127 gamma+ Adelaide Australia * * 64 129 Portland US * * 65 133 San Jose us US * * 66 135 Valencia Spain * * 67 136 Kansas city US * * 68 137 Phoenix US * * 69 140 Lyon France * * 70 147 Rotterdam Netherlands * * * 71 148 Tampa US * * gamma 72 149 Columbus US * * 73 150 Indianapolis US * * 74 151 Pittsburgh US * * 75 152 Edmonton Canada * * 76 153 Tallinn Estonia * * 77 155 Porto Portugal * * 78 157 Orlando US * * 79 158 Gothenburg Sweden * * 80 159 Marseille France * * 81 160 Ottawa Canada * *
82 164 gamma- Richmond US * * 83 167 Durban S. Africa * * 84 168 Austin US * * 85 174 Milwaukee US * * 86 175 Wellington N. Zealand * * *
Data and method | 119
4.7 Interpretation of the results
One of the main objectives of this research is to evaluate the validity of several widely accepted assumptions in the global city literature. These assumptions include; firstly, global cities tend to witness intensification in spatial segregation level based on socioeconomic status, and ethnic background of population as a result of the new economic functions of cities in the global economy. Secondly, the intensity of spatial change is expected to be proportional to cities’ position in the global network of cities, as the more advanced producer service firms the city host, the more stark the changes in the city’s occupational structure, income structure, pace and intensity of gentrification, and eventually intensified socioeconomic and ethnic segregation. In order to put these assumptions to test, the study adopts several statistical hypothesis-testing approaches to determine whether to reject or accept each of the aforementioned assumptions. After calculating the change in ‘spatial multi-group dissimilarity index SD(m)’ over time for all global cities of the dataset, both the direction and the value of these changes are utilized to explore whether there is a correlation between cities’ global status and their change in socioeconomic and ethnic segregation or not. The interpretation of the results takes place over several stages: In the first stage, the study presents the general observations and attempts to find a statistically significant tendency among global cities to share similar directions of change towards intensified spatial segregation. In the second stage, the study examines the correlation between the magnitude of change in spatial segregation level in cities and their global network connectivity. The third stage aims to incorporate other variables to the analysis in a way that captures the influence of region, history, or global status on the resultant change in segregation level in global cities. Finally, stage four is dedicated to discuss in more detail the situation in several individual cities in an attempt to explain any unexpected results or
120 | Data and Method discrepancies appeared during the first three stages of the analysis. These stages are discussed in more detail as follows.
• Stage one: do global cities share patterns of intensified segregation?
In this stage, the researcher inspects – both visually and statistically - the common trends of change in segregation levels among global cities, through a series of tables and graphs that compare the value of the spatial segregation index calculated for each global city. (see also appendix I and II) It is worth mentioning that before proceeding with the following stages of the analysis; the standardized z-scores of the actual SD(m) values are calculated for all cities of the dataset. The aim of calculating the z-scores is to facilitate the comparison between different samples distributions (i.e. compare levels of change in socioeconomic segregation to those in ethnic segregation). Briefly, a z-score is an indication of how many standard deviations each sample deviates from the mean. For example, a z-score equal to 1 represents a sample that is 1 standard deviation greater than the mean; similarly, a z-score equal to -2, represents a sample that is 2 standard deviations less than the mean, and so on (Cohen et al., 2013). It can be calculated from the following: z = (X - μ) / σ where z is the z-score, X is the value of the element, μ is the population mean, and σ is the standard deviation.
• Stage two: examining the correlation between intensity of spatial change and cities’ global network connectivity.
In this stage, we aim to test the correlation between the position of a city in the global network and its corresponding change in spatial segregation. In order to examine this correlation, a bivariate linear regression analysis is performed to test the dependence of the ‘response’ (that is the change in spatial segregation level) on the ‘predictor’ (that is the global
Data and method | 121 network connectivity of the city) (Weisberg, 2014). Note that the response and the predictor are referred to in the text as the dependent and the independent variables respectively. Before going on with the linear regression analysis, the distribution of the observed data has to fulfill several assumptions. Including the absence of significant outliers (Freund et al., 2006), and a confirmation that the independent variable x and the dependent variable y have a linear relationship (Hoffmann, 2010). As explained the potential outliers are excluded from the analysis based on their calculated z-scores (cities scored more than 2 standard deviations from the sample mean are considered outliers). Also, in order to check whether a linear relationship exists between the two variables, a summary scatterplot is created to visually inspect the correlation between variables for linearity. According to Kahane (2008), when dealing with a large collection of data, the correlation between x and y may not be clear. Therefore, it is possible to group the data point on the x-axis (in our case the x-axis represents the cities’ global network connectivity, where cities can be grouped in alpha, beta, and gamma categories). And for each group, the mean value of y is calculated and represented on the y-axis (in our case the y-axis represents the standardized change in spatial segregation index). Given this graph, the dots represent the mean value of SD(m) values for each of the alpha, beta, and gamma categories. And by drawing the trendline, we can verify the following assumption: the mean values of y, for given values of x, are a linear function of x. y = a + bx, where x is the independent variable and y is the dependent variable. The slope of the line is b, and a is the intercept (the value of y when x = 0).
1221 | DataData andand MethodMethod
Ethnic segregation - means - means segregation Ethnic
Socioeconomic segregation - means means - segregation Socioeconomic gamma beta alpha gamma beta alpha
Figure 4-4 check for a linear correlation between the dependent and the independent variables
It should be noted that in case that the relationship between the mean values of y and x is nonlinear, then fitting a linear regression model to the data probably will not provide a useful model, and other alternative models are to be considered. Figure (4-4) confirms that the dependent and independent variables in our case have a linear relationship and therefore the linear regression model is fit for the data. The output of the bivariate regression analysis provides several indications on the strength of the association between the two variables, which can be used as tool for hypothesis testing: For example, the analysis provides the R and R2 values. The R value represents the simple correlation between the dependent and the independent variables, where the value of R ranges from –1 and +1, with positive numbers indicating a positive association and negative numbers indicating a negative association, also, a correlation of zero means there is no statistical association (Hoffmann, 2010). While the R2 value indicates how much of the total variation in the dependent variable – segregation index - can be explained by the independent variable – cities’ global status. And by obtaining these values, the aforementioned assertion is accepted or rejected depending on the strength of the association between the dependent and independent variables.
Data and method | 123
• Stage three – examining the influence of location, history, and other factors
In case that there are more than one independent variable that have influence on the variation in the dependent variables, a multiple linear regression is used to obtain the multiple correlations R (Weinberg and Abramowitz, 2002), and the relative contribution of each of the independent variables to the total variance explained. In our case, there are several variables that may affect the level of spatial segregation in global cities. For instance, cities located in Western Europe may show different trends in those in the US or Australia. Also, cities that recently joined the classification of world cities may show different trends than those appeared in previous classifications, as it can be argued that cities with longer history as global cities will show distinct spatial change proliferated over long period of time compared to other newly added global cities. Similarly, variables such as the different urban policies, the pace of gentrification, the condition of the public housing and so on, can be added to the model. However, not all the aforementioned variables are quantitatively measured. While the conversion of qualitative data into quantitative data is possible (Loehnert, 2010), still it is proved problematic to convert nominal non-scalar or non-ordinal data into numeric values without compromising the accuracy of the results. In more detail, a city’s status as a recent global city or a city with a longer history as a global city can be given the numerical values of 0 and 1, where 0 indicates that the city didn’t appear in global city classifications prior to 2010, and 1 indicates that the city already appeared in the 2008 classification. In this case the city’s global history is a dummy variable that can have an indicative numeric value without compromising the analysis results. Similarly, if the location of a city in a certain region can affect its level of segregation, and we are faced with 6 different regions (North America, South America, Western Europe, Australia, Southeast Asia, and North Africa/ Middle East), then giving each region a code from 1 to 6 is not a proper way for quantizing the given data.
124 | Data and Method
Because the values from 1 to 6 are only indicative if the given categories are scalar or in a certain order (such as the satisfaction scale in survey data which can range from 0: no satisfaction to 5: 100% satisfaction), yet in our case, in what order can the different regions be assigned to the different values on the scale from 1 to 6. In other words, which region deserves to take the numeric value of 6, which is a higher value than 1 and on what ground. In order to avoid the quantizing of nominal data, each region can be presented in the model as a separate dummy variable. If the city is located in the region then the variable takes the value of 1, if not, the variable takes the value of 0. The problem with this approach is the possible multicollinearity between two or more dummy variable, which leads to problems with understanding which variable contributes to the variance explained and technical issues in calculating a multiple regression model. Similarly, other variables such as urban policies and public housing are not dummy variables with 0 or 1 values; rather they incorporate higher level of complexity. The case of public housing for instance requires taking into account not only the presence or absence of public housing projects, but also other factors such as the allocation of the housing units, the rate of public housing privatization, the quality of housing units, the dominant ethnic groups in the public housing projects and so on. For that matter, in order to avoid the dummy variable trap and the miscalculations associated with the inaccurate quantizing of qualitative data. Another statistical approach is deployed to compare the influence of different locations, histories, and other variables on the resultant change in spatial segregation without the need to assign numerical values for each group of cities. Instead, the one-way analysis of variance ANOVA compares means that are drawn from different groups that are part of a larger sample. Using multiple comparison procedures, we may also determine if one of the means is significantly different from one of the others (Hoffmann, 2010). For example, the one-way ANOVA can give an indication if cities located in North America are significantly different than cities of Western Europe in
Data and method | 125 terms of changes in spatial segregation. Similarly, it can indicate whether alpha cities are behaving differently from beta and gamma cities. In order to use the one-way ANOVA in testing the hypotheses postulated in the literature, a p-value is calculated and compared to significance level of 0.05 (or 0.01) to determine whether to accept or reject the hypothesis under the test. For example, the hypothesis states that the rank of a city as alpha, beta, or gamma city will have a significant effect on the changes in spatial segregation within the city. In this case, we have three groups of cities (alpha, beta, and gamma), where each group contains 20 global cities. According to Kirkman, (1996), each group probably include different values of y and different sample mean. But can we take this difference in the average of y as evidence that the groups in fact are different (and perhaps that the rank causes that difference)? The likely range of variation of the averages if there is no significant difference between groups is given by the standard deviation of the estimated means: σ/N½ where σ is the standard deviation of all the values of y and N (is the number of cities in each group: in this case N=20). Thus if we treat the collection of the 3 group means as data and find the standard deviation of those means and it is "significantly" larger than the above, we have evidence that groups are significantly different. This is to say that if some (or several) group's average change in segregation is "unusually" large or small, it is unlikely to be just "chance" (ibid.). The comparison between the actual variation of the group averages and that expected from the above formula is expressed in terms of the F ratio:
F=(found variation of the group averages)/(expected variation of the group averages)
After obtaining the F ratio, and by using the degree of freedom (in our case, it is the number of groups - 1, and the number of cities in each group - 1: (2,9)); a value x is obtained from the ‘table of critical values for the F
126 | Data and Method distribution’, which indicates how large the F needs to be for the groups to be significantly different. In other words, if the calculated F is greater than x, then the groups are significantly different and vice versa. The same principle can be applied on cities located in different regions or cities with other shared characteristics.
• Stage four – the individuality of cities: a discussion Both the regression analysis and the one-way ANOVA are expected to give robust indications on how the cities’ global status and their spatial changes may or may not be associated. Still, the detailed discussion of individual cases is required for several reasons: Firstly, the discussion of the outliers, which is useful in explaining why these particular cities have witnessed above (or below) average change in their spatial segregation level. The discussion aims to uncover the variables – other than global network connectivity – that contributed to the observed spatial change. Secondly, based on the results of the one-way ANOVA, any significant influence of one of the examined factors on the resultant segregation, e.g. if the analysis revealed that the location of a city in a certain geographic region is an influential factor, then a detailed discussion of cities of the region can assist in reaching a systemic explanation for the observed phenomenon. Also, discussion of individual cases aims to address any inconsistencies within the results. An example for such inconsistency is the case of two cities that share several characteristics including the global status or the national contexts, yet both cities show inexplicable variations in their spatial changes. In this case, the discussion of the situations in the two cities can point out certain contextual particularities to be as influential on the process of spatial restructuring as other macro developments propagated in the global city literature. The subsequent chapters present the findings of the analytical study. Chapter five deals with the change in socioeconomic segregation in global
Data and method | 127 cities. Chapter six provide the same analysis for the change in ethnic segregation. And in chapter seven, the study attempts to investigate the correlation between the change in both socioeconomic and ethnic segregation within the same global city.
128 | Data and Method
Notes
1 http://www.lboro.ac.uk/gawc/gawcworlds.html 2 Since the calculations of the spatial index incorporate both area and length of shared boundaries among division, then large rural tracts will have long shared boundaries (implying a high possibility of interaction between population groups in different divisions), which can falsely decrease the value of segregation index. While at the same time, these areas have very low population density compared to neighborhoods in inner city and suburbs, then excluding these areas from the analysis can give a better perception of the situation in the urban area with no miscalculations added by incorporating the adjacent rural areas within the metropolitan region. 3 Earlier data for the 1970s and 80s are not available in digital form on the official census bureau websites and could not be obtained by the researcher.
Chapter 5
5 Socioeconomic segregation in global cities: Analysis and findings
This chapter presents the findings and general observations that links the level of change in socioeconomic segregation in any city over time, to its position as a global city based on the GaWC classification of world cities. The collective results show if there is an association between the agglomeration of producer services firms in one city (the city’s service value) and the intensified socioeconomic segregation within this city. Also, individual cases are discussed to evaluate the influence of the contextual particularities of each city on its corresponding change in spatial segregation. As noted in the preceding chapter, the final dataset contains a total of 66 cities categorized into 20 alpha, 23 beta, and 23 gamma cities. This section presents the preliminary results and the general observations drawn from the calculations of the ‘spatial multi-group dissimilarity index SD(m)’ over time for the 66 global cities of the dataset, followed by the interpretation of the results, which takes place over the three stages of the statistical hypothesis testing discussed in chapter four. Figure (5-1) is a graphical representation of the average yearly change in SD(m) value of all cities of the dataset (see appendix i for more details about the calculated SD(m) values for all cities of the dataset). As a general observation, the figure shows that the three global ranks (alpha – beta – gamma) contain cities with both increase and decrease in levels of
130 | Analysis and findings socioeconomic segregation. Also, cities do not only differ in the direction of change towards more or less segregation, but also the intensity of the change varies greatly across cities of the dataset. In total, 36 cities (54.5% of the dataset) have witnessed an average increase in socioeconomic segregation over the past years, compared to 30 cities (45.5%) with decrease in socioeconomic segregation. The values range from the maximum increase in segregation, that is 3.0% of average yearly change in SD(m) value scored by the gamma city of Guadalajara (the city scored total of 30.3% increase in socioeconomic segregation from 2000 to 2010), to the other end of the spectrum, the maximum decrease in segregation, that is -3.7% of average yearly change in SD(m) value scored by the alpha city of Brussels (a total of -26.4% of change from 2001 to 2008). As a general observation from figure (5-1), there is no statistically significant tendency among global cities to share an increase in socioeconomic segregation levels (see appendix ii for more details regarding the significance probability). Similarly, there is no obvious correlation between the spatial change in one city and its position as an alpha, beta, or gamma city. Therefore, these results are further explored as follows:
5.1 Examining the correlation between the intensity of spatial change and cities’ global network connectivity.
This section presents the results of the linear regression analysis introduced in chapter four. The calculations are generated using IBM – SPSS statistics software, and the aim of the analysis is to test the strength of the association between the cities rank in the global network of cities and their corresponding change in socioeconomic segregation. The expected result - if the null hypothesis is correct – is a strong correlation between the two variables indicated by a value of R that is greater than 0.5. Also, the
Socioeconomic segregation in global cities | 131
New York Hong Kong Singapour Chicago Sydney Toronto Sao Paulo Los Angeles Mexico city Amsterdam Brussels San Fransisco alpha Washington Miami Melbourne Boston Dallas Atlanta Philadelphia Johannesburg Stockholm Montreal Houston Berlin Copenhagen Bogota Vancouver Seattle Auckland Oslo Cape Town Minneapolis Rio de Janeiro beta Brisbane Detroit Denver Monterrey St Louis Panama city San Diego Perth Cleveland Calgary Cincinnati Charlotte Baltimore Adelaide Portland San Jose us Kansas city Phoenix Guadalajara Rotterdam Tampa Columbus indianapolis Pittsburgh gamma Edmonton Orlando Gothenburg Ottawa Richmond Durban Austin Milwakee Wellington 4.00 3.00 2.00 1.00 0.00 -1.00 -2.00 -3.00 -4.00
FigureFi 5-15 1 alphalh – betabt – gamma citiesiti and d their th i yearly l percentage t of f change h in i socioeconomic i i segregation ti index segregation socioeconomic in change of percentage yearly their and cities – gamma – beta 5-1 alpha Figure
132 | Analysis and findings
y = -3E-06x + 0.281 R² = 0.00197
Standardized change in SD(m) index value
Global Network Connectivity
Figure 5-2 Summary scatterplot for the correlation between the standardized change in segregation index and the cities’ global network connectivity summary scatterplot in figure (5-2) represents the correlation between the dependent variable, which is the standardized change in the spatial dissimilarity index SDI on the y-axis, and the independent variable, which is the cities’ global network connectivity GNC on the x-axis. The SPSS generates several tables for the linear regression. In this section, we show the three main tables required to understand the results of the linear regression procedure: the model summary table, the ANOVA table, and the coefficients table.
Socioeconomic segregation in global cities | 133
Table 5-1 Regression output – model summary, ANOVA, and Coefficient tables
Model Summary Adjusted R Std. Error of the Model R R Square Square Estimate 1 .044a .002 -.014 1.00679900 a. Predictors: (Constant), AbsoluteGNC
ANOVAa Sum of Model Squares df Mean Square F Sig. 1 Regression .127 1 .127 .125 .725b
Residual 64.873 64 1.014 Total 65.000 65 a. Dependent Variable: zscoreINCOME b. Predictors: (Constant), AbsoluteGNC
Coefficientsa Unstandardized Standardized Coefficients Coefficients
Model B Std. Error Beta t Sig. 1 (Constant) .109 .333 .328 .744
AbsoluteGNC .000 .000 -.044 -.354 .725 a. Dependent Variable: zscoreINCOME
Firstly, the regression output table (5-3) provides the information needed to determine how well the regression model fits the data. In the model summary section, the column ‘R’ is the absolute value of the Pearson correlation coefficient between the dependent and the independent variables.
134 | Analysis and findings
It simply indicates the strength of the association between the two variables. In our case, R = 0.044 < 0.5, which indicates a weak correlation. The R2 value in the R2 column represents the proportion of variance in the dependent variable that can be explained by the independent variable. In our case, R2 = 0.002, which means that the independent variable, global network connectivity, explains only 0.2% of the variability of the dependent variable, change in segregation index. Please note that the R2 is calculated for the sample (66 global cities), SPSS generates another value called ‘adjusted R2 that represents the proportion of variance if the analysis is performed over the entire population (i.e. the full list of global cities). Table (5-3) shows that the adjusted R2 = -0.014, which indicates a very low dependence between variables. The next table is the ANOVA table, which reports how well the regression equation fits the data (i.e., how accurately the model can predicts the dependent variable SDI based on the changes in the independent variable GNC). The number in the "Regression" row and the "Sig." column indicates the statistical significance of the regression model that was run. Here, the regression model is statistically insignificant, F (1, 64) = 0.125, p = 0.725, which is greater than 0.05, and indicates that, overall, the regression model can not statistically predicts the outcome variable, which conform with the low dependence between variables observed from the model summary table. The last table is the coefficients table, which provides the coefficients of the regression equation that is used to predict the dependent variable from the independent variable. y = b0 + (b1 x) where y is the change in segregation, x is global network connectivity, b0 is the intercept and b1 is the coefficient, both can be found under the B column in the table, b0 = 0.109, and b1 = 0.000. In our case, it is already established that the model poorly fits the data due to the weak dependence between the variables. Recapitulating, the linear regression established that global network connectivity for a city could not statistically significantly predict the change
Socioeconomic segregation in global cities | 135 in spatial segregation level, as the two variables have a weak correlation. The global network connectivity of cities is accounted for only 0.2% of the explained variability in spatial segregation index.
5.2 Examining the influence of location, history, and other factors
So far, the analysis confirmed that cities’ position in the global network of cities is not associated with the direction or the intensity of change in socioeconomic segregation within these cities, which also suggests that there are other independent variables that can explain the variation in the dependent variable. This section aims to assess the influence of several factors that are assumed to be relevant to the outcomes of spatial restructuring in global cities. As explained in chapter four, the influence of these factors is evaluated by the one-way analysis of variance ANOVA. For example, the sample of cities – upon which the regression findings are drawn – contains two distinguishable groups of cities; cities that were already ranked as global cities in the GaWC classification prior to 20101, and cities that are recently added to the classification in 2010. This distinction is assumed to be relevant to the explanation of the variation in spatial segregation, as the longer the period of time the city is considered a global city, the more visible the spatial outcomes of restructuring are expected to be. One can argue that this factor (referred to as global history during data entry phase) is causing the regression results to be inaccurate as the sample is not homogeneous. Therefore, to examine the influence of a city’s global history on the resultant change in spatial segregation, the sample is divided into two groups: cities newly added to 2010 classification, and cities in the classification since 2008. Table (5- 4) reports the output of the one-way ANOVA generated by IBM- SPSS.
136 | Analysis and findings
Table 5-2 ANOVA output when cities are grouped according to their global history ANOVA Zscore: INCOME Sum of Mean Squares df Square F Sig. Between Groups .082 1 .082 .169 .683 Within Groups 28.246 58 .487 Total 28.328 59
ONEWAY zscore SD(m) BY global history
The table provides the F ratio, as explained in chapter four, the calculated value of F can be compared with x that is the critical value below which there is no statistically significantly difference between groups’ means. In other words, when FSocioeconomic segregation in global cities | 137 respective regions but they are ranked below alpha cities due to their less significant ‘service value’. In this sense, if the global/divided city model is valid, then alpha cities are expected to show higher tendency for intensified socioeconomic segregation than beta and gamma cities2. Therefore, the sample is once again divided into three groups; table (5-5) reports the output of the one-way ANOVA when cities are categorized in alpha, beta, and gamma groups.
Table 5-3 ANOVA output when cities are grouped according to their global rank ANOVA Zscore: INCOME Sum of Mean Squares df Square F Sig. Between Groups .181 2 .090 .183 .833 Within Groups 28.147 57 .494 Total 28.328 59 ONEWAY zscore SD(m) BY global rank
The calculated F ratio from the table equals 0.183 which is once again less than the critical value of x = 3.16. Accordingly, it can be concluded that changes in segregation witnessed by alpha cities is not statistically significantly different from those witnessed by beta or gamma cities. The p- value from the table also confirm this conclusion, where p= 0.833, which is greater than a significance level of 0.05. Another factor to be tested in this section is the location of a global city in a certain geographic region. The cities of the sample are divided according to their broad geographic region into 6 groups: Western Europe, North America, Latin America, Australia, Asia, and Africa. Table (5-6 – above) presents the ANOVA output, which reveals that – unlike the previous factors of global history and rank – the F ratio equals 3.025, which is greater than x= 2.32. In this case, the analysis results suggest that cities located in
138 | Analysis and findings certain region do share trends of change in spatial segregation, and these trends are significantly different from cities in other regions. This conclusion is confirmed by the p-value = 0.018 which is less than 0.05.
Table 5-4 ANOVA output when cities are grouped according to their geographic region ANOVA Zscore: INCOME Sum of Mean Squares df Square F Sig. Between Groups 6.198 5 1.240 3.025 .018 Within Groups 22.130 54 .410 Total 28.328 59 ONEWAY zscore SD(m) BY region
ANOVA Zscore: INCOME Sum of Mean Squares df Square F Sig. Between Groups 5.849 3 1.950 4.634 .006 Within Groups 21.454 51 .421 Total 27.302 54 ONEWAY zscore SD(m) BY region except Asia and Africa
In order to enhance the accuracy of the analysis and to avoid any misread conclusions, table (5-6 - below) shows a re-run of the ANOVA test when cities of Asia and Africa are excluded from the sample. Groups are excluded due to the small number of cities in each, Asia contained only 2 cities and Africa contained 3 South African cities. Accordingly the analysis is performed again among 55 cities grouped into four regions: Western Europe, North America, Latin America, and Australia. Noticeably, the F ratio = 4.63 is still greater than x = 2.79, and the p-value has improved to
Socioeconomic segregation in global cities | 139 reach 0.006, which indicates a high significance level of the influence of the geographic region on the changes of segregation in global cities. Similarly, cities can be grouped according to other characteristics to ultimately determine the influence of these characteristics on the outcomes of spatial restructuring in global cities. For example, cities can be divided into capital cities and non-capital city, which is an indication of the concentration of political power in the city and the associated change in segregation. Other factors concerning the pace of gentrification, the status of public housing, and other urban policies in general can be examined. Yet, since the collection of this data proved problematic (for instance, there is no system for ranking cities according to their gentrification level), while other factors inhibit a high level of complexity that problematize the quantizing of qualitative data (as in the case of the public housing – see chapter four for details), then the analysis will go on to stage 4 and start the discussion of the situation in individual global cities in the light of the regression and ANOVA findings.
5.3 The individuality of cities: a discussion
Table 5-7 presents the top 10 cities with increased socioeconomic segregation and the top 10 cities with decreased socioeconomic segregation. An observation drawn from table (5-7) may explain why the global/divided city model is widely accepted despite the lack of extensive empirical evidence supporting it. In the top 10 list of cities with increased segregation, the 1st, 2nd, and 5th places are filled by all three Mexican cities in the global ranking. Oddly, Mexican cities are sharing similar spatial transformations with Swedish cities that fill the 6th and 9th places on the same list. The oddness here is derived from the huge variations between the two countries in terms of quality of life (see Prescott-Allen, 2001), welfare system, and different ranking categories and their levels of integration into global economy are not the same. While, an in-depth discussion of the situation in
140 | Analysis and findings the two contexts will reveal that they might share similar changes in socioeconomic segregation, yet the changes are a result of entirely different reasons.
Top ten cities with maximum increase and decrease in socioeconomic segregation and their respective ranks and regions Top 10 cities with increased segregation City Rank Region 1 Guadalajara gamma L. America 2 Monterrey beta L. America 3 Amsterdam alpha Europe 4 Denver beta N. America 5 Mexico city alpha L. America 6 Gothenburg gamma Europe 7 Oslo beta Europe 8 Sydney alpha Australia 9 Stockholm beta Europe 10 Seattle beta N. America Top 10 cities with decreased segregation City Rank Region 1 Brussels alpha Europe 2 Rotterdam gamma Europe 3 Berlin beta Europe 4 Rio de Janeiro beta L. America 5 Austin gamma N. America 6 San Francisco alpha N. America 7 Copenhagen beta Europe 8 Atlanta alpha N. America 9 Toronto alpha N. America 10 Johannesburg alpha Africa
Socioeconomic segregation in global cities | 141 demographic changes. And still, both Mexican and Swedish cities are heading toward more segregated configuration of cities. Nevertheless, similarities between the two different contexts cannot be pinned on economic globalization, because all five cities are falling into different global categories. Overall, the absence of a collective pattern of spatial change in global cities proves that socio-spatial transformations in these cities cannot be fully understood away from their local circumstances and contexts. Therefore, the subsequent sections are dedicated to discuss in more detail the motivations and factors involved in the production of diverse socio-spatial transformations in individual global cities. And according to the results of the ANOVA, the sections are organized based on the broad geographic regions of the world, which was already proven to have a significant influence on the outcomes of spatial restructuring in global cities. For easier reading of the results presented in figures (5-1), figure (5-3) presents global cities categorized according to their location in the broad geographic regions of the world instead of their global rank. Clearly, cities in a certain region are not expected to show similar changes in their socioeconomic segregation level, due to the huge variations exist among cities within the same region in terms of national policies, history, geography, subcultural differences… etc. Yet, the observations drawn from figure (5-3) suggest that: On one hand, although the changes in income inequality and socio- spatial polarization in European cities are recognized to be ‘moderate’ when compared to the change in American cities over the last 20 years (Van Kempen and Murie, 2009), still, the results in figure (5-3) shows a similarity between the two contexts, where only 50% of global cities in both regions scored an increase in their socioeconomic segregation level, while the higher tendency towards intensified socioeconomic segregation is observed among global cities of Latin America and Australia, where 71.4% of cities of each
142 | Analysis and findings
Brussels
Rotterdam e Berlin p Copenhagen Stockholm Oslo n index index n Gothenburg Euro W. Amsterdam Austin San Fransisco Atlanta Toronto Calgary Montreal Miami Chicago Columbus Phoenix San Diego Washington St Louis Los Angeles Minneapolis indianapolis Vancouver Charlotte Houston Dallas
Edmonton lo America
Baltimore g San Jose us Philadelphia An Cleveland Kansas city Tampa New York Pittsburgh Milwakee Boston Orlando Cincinnati Portland Richmond Detroit Ottawa Seattle Denver Rio de Janeiro Bogota Panama city Sao Paulo Mexico city Monterrey Guadalajara L. America Singapour
Hong Kong Asia Adelaide Wellington Brisbane Perth Auckland Melbourne Oceania Sydney Johannesburg Cape Town
Durban Africa 4.00 3.00 2.00 1.00 0.00 -1.00 -2.00 -3.00 -4.00 Figure 5-3 global cities categorized according to their geographical region and their yearly change in socioeconomic segregatio socioeconomic in change yearly their and region geographical to their according categorized cities 5-3 global Figure Figure 5-3 global cities categorized according to their geographical region and their yearly change in socioeconomic segregation index
Socioeconomic segregation in global cities | 143 of the two regions showed an increase in their SD(m) value (note that Africa and Southeast Asia are underrepresented in this analysis due to lack of data). On the other hand, the results also show that even among cities within the same national border, similar patterns of change may not exist. For example, Brazilian cities of Sao Paulo and Rio de Janeiro show a stark contradiction in both intensity and direction of change in socioeconomic segregation, same for Johannesburg and Cape Town, Rotterdam and Amsterdam, and Austin and Denver. These cases are discussed in more detail as follows:
5.3.1 Latin American cities
The academic literature on urban development of Latin American cities underlines distinct characteristics of the region’s urban areas. On one hand, the unprecedented rapid urbanization in the second half of the twentieth century, where roughly, the share of the population living in cities grew from 40% in 1950 to more than 75% in 2000 (Lattes et al., 2002). As a result, mass numbers of rural migrants – in their search for accommodation in cities of destination - have built informal settlements at the periphery of cities (Rolnik, 2011; Monkkonen, 2011). Which led the urban structure of Latin American cities to be characterized by a center/periphery dichotomy, where high-income families are concentrated in wedge-shape areas, pointed at the historical center and expanded in a single geographical direction towards the periphery. These central and affluent urban wedges are embedded in a peripheral zone of poor and working-class families, living mostly in informal settlement (Telles, 1995; Sabatini et al., 2001; Sabatini, 2003; Feitosa et al., 2007; Audirac et al., 2012). On the other hand, uneven industrialization and growing social inequality have always been of the basic characteristics of the region (Telles, 1995). Unlike Western Europe where deindustrialization has been accompanied by a shift to services and spatial shift of manufacturing to countries of the global South such as Brazil and Mexico. Latin American
144 | Analysis and findings global cities such as São Paulo and Guadalajara faced a complex process of economic restructuring. Where metropolitan cores are deindustrializing, while at the same time, urban peripheries and hinterland are expanding and receiving off-shore manufacturing industries from Western Europe and North America (Audirac et al., 2012). Later in 2000s, industrial peripheries lost manufacturing competitiveness to Asia. Arguably, the global processes of economic, social, and spatial restructuring have impacted the longstanding history of socio-spatial inequality in Latin American cities (Telles, 1995; Audirac et al., 2012). One example is the emergence of “gated communities” for medium and high- income families within the urban periphery including poor neighborhoods (Feitosa et al., 2007; Monkkonen, 2011). The ‘islands of wealth’ as Janoschka (2002) calls them are gradually replacing the previous large-scale segregation patterns of poor peripheral areas and wealthy central cities by fragmented and more complicated urban structure (Monkkonen, 2011). In fact, Gated communities are a form of gentrification as they result from increasing demand for urban space to accommodate high-income groups, and the low land costs in poor peripheries (ibid.). Yet the difference between gentrified areas in Latin American cities and those in Western Europe and North America is the ‘walling out’3, where the gentrified areas are totally isolated from its surrounding that are still dominated by housing for low- income groups, therefore, real estate developers favored the exclusive, closed, and controlled housing project out of fear of crime and violence in the surrounding neighborhoods (Caldeira, 2000; Coy and Pohler, 2002). In general, the distinct characteristics of Latin American cities explain the historically similar patterns of socioeconomic segregation shared by cities across the region. As shown in figure (5-3), Latin American cities may still be sharing patterns of change in segregation in the past decade, where 71.4% of cities of the region scored an increase in income segregation (except for Rio de Janeiro and Bogota). The recent and collective increase in segregation is usually explained in the specialized literature by the ongoing
Socioeconomic segregation in global cities | 145 urbanization along with relocation of manufacturing industries from cities’ peripheries to Asia (Audirac et al., 2012). This relocation is accompanied by plant closures, job losses, and eventually skyrocketing unemployment rates. For example, the main industrial center of Guadalajara (Mexico’s Silicon Valley in the 1990s) has lost more than 45,000 jobs in electronics industry to China between 2001 and 2003 (Dussel, 2005). To sum up, the influx of rural – mostly poor - immigrants, increasing unemployment, and the rise of gated communities contributed to the intensification of socioeconomic segregation in a number of global Latin American cities. However, this assumption may explain why all three Mexican cities are showing sharp increase in their segregation level, but it fails to explain the inconsistency in spatial change between Sao Paulo and Rio de Janeiro. Moreover, a closer look to the situation in Mexican cities reveals that intensified segregation is generated by variables other than the historical urban development of the region. In particular, Mexico’s urban development is marked by distinct government actions, which contributed to the sharp increase in segregation in Mexico City, Guadalajara, and Monterrey. According to Monkkonen (2011), Mexico’s new federal housing finance system - initiated in the 1990s - is a government intervention that has contributed to greater levels of segregation. As, it triggered the construction of large tracts of middle-class single-family houses in the poor peripheries of cities across the country, by providing mortgages exclusively to registered salaried employees in the formal sector. Roughly, 220,000 mortgages were issued in Mexico during the 1990s, this figure jumped to 400,000 per year between 2000 and 2005 (ibid.). And since loans have been given only to registered employees, thousands of poor informal workers were excluded from the new suburban development. While at the same time, local governments were unable to effectively manage the growth of informal settlements of poor families (Audirac et al., 2012). The drawbacks of housing reform in Mexico are the creation of homogeneous large areas for residents of a similar socioeconomic status, along with the expansion of informal settlements for
146 | Analysis and findings the poor in cities’ peripheries, leading to rapid increase in socioeconomic segregation levels. The results in figure (5-3) supports Monkkonen criticism of the federal housing reform in Mexico. However, the assumption that poor population tends to expand in peripheral informal settlements could not be verified. Figure (5-4) and (5-5) show the distribution of population earning up to the minimum wage or less in Mexico City municipalities in 2000 and 2010. Unfortunately, census data does not differentiate between informal sector workers and registered employees. Nevertheless, the 2010 map shows that low- paid workers are not expanding in peripheral municipalities. Instead, poor residents were displaced and concentrated in higher densities in less number of municipalities in the center. Also, unlike ethnic segregation, socioeconomic segregation is not entirely dependent on residential mobility; part of the changes in socioeconomic compositions of municipalities is due to the changes in employment status caused by lay-offs. Therefore, the recent significant presence of poor workers in municipalities of the southeast and northeast is not necessarily due to the displacement of the poor. Instead, it can be a result of plants closure and the consequence income decline in those neighborhoods. Similar tendencies showed in Guadalajara and Monterrey, regardless the different global status of each city. Unlike the unified trend in Mexican cities, Brazilian cities showed different outcomes. Sao Paulo and Rio de Janeiro varied in both intensity and direction of change in income segregation, as shown in figure (5-3), Sao Paulo scored 4.97% increase in segregation between 2000 and 2010, while Rio de Janeiro scored 24.76% decrease in segregation over the same period of time. The noticeable difference between the two cities is explained by the uneven industrialization in previous decades and different urban development approaches in recent years. Firstly, uneven industrialization is a common characteristic of urban development in Latin America in general, and Brazil is one of the best examples of this imbalance, where metropolitan areas range from highly
Socioeconomic segregation in global cities | 147
N
10 mi
10 km Figure 5-4 distribution of workers earning the minimum wage or less in Mexico City municipalities in 2000
N
10 mi
10 km
Figure 5-5 distribution of workers earning the minimum wage or less in Mexico City municipalities in 2010
Figure 5-4 distribution of workers earning the minimum wage or less in Mexico City municipalities in 2000 Figure 5-5 distribution of workers earning the minimum wage or less in Mexico City municipalities in 2010
148 | Analysis and findings industrialized ones like Sao Paulo to others “that have grown without the benefit of industrialization” (Tells, 1995: 1200) and are consequently have higher poverty and unemployment rates. Since “Rio had never been able to attract the consumer goods sectors which led to the dynamic industrialization and strong local markets of Sao Paulo, which overtook Rio de Janeiro as Brazil’s largest city in the 1950s. Rather, the economy of Rio had been dependent on the local tertiary sector, comprised of small modern segment (finance, computers, commerce) and a large personal services sector.” (Ribeiro and Telles, 2000: 80). Therefore, Rio de Janeiro has always had higher levels of segregation between the poor and the middle class than in Sao Paulo with a clear north and south division (Tells, 1995) (see figure 5-7). While at the same time, the impact of deindustrialization in the 1980s and 90s was more severe on Sao Paulo, it followed the Mexican scenario where manufacturing industries left first to the hinterlands then to Asia in later decades (Audirac et al., 2012). Accordingly, it can be argued that uneven industrialization in previous decades has impacted the level of vulnerability of cities toward the impacts of current economic transformations on both global and local levels of urban development. Secondly, urban governance in Brazil on federal, regional, and municipal levels contributed to the production of different spatial outcomes in cities. Decentralization of the administration of urban services started in the 1960s, materialized in the form of regional offices created to manage urban growth (Levy, 2001). Then, in 1988, the new constitution gave municipal government more political, administrative, financial, and legislative independence, and made local governments responsible for all services including land use regulations (ibid.). Also, the constitution reflects the rising public awareness of cities governance, as the constitution contains a chapter dedicated to urban policies, revolving around three concepts; the social function of cities, the importance of direct participation of citizens in urban policy decision-making processes, and recognition of the land rights
Socioeconomic segregation in global cities | 149 of millions of inhabitants of informal settlements on the periphery of cities in Brazil (Rolnik, 2011). The multi-level decentralized urban governance in Brazil resulted in several urban development initiatives, on both federal and local level. Such as, Favela-Bairro project by the local government of Rio de Janeiro in 1994 (UN-Habitat, 2003), public-private partnerships projects in Sao Paulo’s master plan in early 1990s (Siqueira, 2012), and the housing program ‘minha casa, minha vida’ initiated by the state in 2008 (Barbosa, 2010). The different approaches adopted by each initiative have led to different spatial outcomes in cities across the country. As shown in figures (5-6) and (5-7), the north-south division in Rio de Janeiro in 2000 appeared to fade in 2010 despite Telles’s (1995) prediction of the continuation and increase of socioeconomic segregation in globalizing Rio. Feitosa et al. (2007) and Monkkonen (2011) explain the decrease in segregation in Rio and other Brazilian cities by the changing scale of segregation. In more detail, when a gated community appears in a poor peripheral area, the segregation level on the municipality level decline because the municipality hosts different income groups, but on the micro level, wealth is still unevenly distributed in pockets of self-segregation. Arguably, the case of Rio de Janeiro cannot be explained by this scale issue, since it depends essentially on the emergence of affluent gated communities in the poor north. Figure (5-7) shows that poor population over 10 years of age with monthly income less and up to half the minimum wage started to concentrate in the affluent south, causing income segregation levels to fall significantly. Changes in spatial distribution of the poor can be partially explained by their recent access to middle-class housing unites facilitated by the local government. The case of Sao Paulo differs significantly from Rio de Janeiro. As shown in figures (5-8) and (5-9), the concentration of poor population with monthly income less and up to half the minimum wage in 2010 increased in
150 | Analysis and findings
N
10 mi
10 km Figure 5-6 distribution of population with monthly income less and up to half the minimum wage in Rio de Janeiro in 2000
N
10 mi
10 km Figure 5-7 distribution of population with monthly income less and up to half the minimum wage in Rio de Janeiro in 2010
Figure 5-6 distribution of population with monthly income less and up to half the minimum wage in Rio de Janeiro in 2000 Figure 5-7 distribution of population with monthly income less and up to half the minimum wage in Rio de Janeiro in 2010
Socioeconomic segregation in global cities | 151
N 10 mi 10 km 9 distribution of population with monthly - Figure 5 income less and up to half the minimum wage in Sao Paulo in 2010 N 10 mi 10 km f 8 distribution of population with monthly - Figure 5 income less and up to half the minimum wage in Sao Paulo in 2000
Figure 5-8 distribution of population with monthly income less and up to half the minimum wage in Sao Paulo in 2000 Figure 5-9 distribution of population with monthly income less and up to half the minimum wage in Sao Paulo in 2010
152 | Analysis and findings neighborhoods that were already mostly poor in 2000, causing higher level of income segregation. Unlike Rio de Janeiro, the poor population did not gain access to central affluent neighborhoods. Which matches Siqueira’s (2012) description of the duality of the urban structure of Sao Paulo, where large investments are dedicated for development of strategic locations in order to “construct[ing] a skyline that now resembles other global cities around the world.” (p: 391), while poor areas lack basic services. The variations among Sao Paulo, Rio de Janeiro, and Mexican cities reflect the role of the local governments in promoting higher levels of integration or segregation among different socioeconomic groups. Latin American cities in general share historical characteristics, where their shared spatial structure is derived from colonial and pre-colonial times (Telles, 1995), they also share similar impacts of industrialization and premature deindustrialization (Audirac et al., 2012). Yet clearly, the ability of central and local governments in managing urban development is an important contingency that affects the spatial outcomes of Latin American cities under globalization.
5.3.2 Australian cities
Figure (5-3) indicates that Australian cities show parallel changes in socioeconomic segregation to that of Latin American cities, where 71.4% of Australian cities witnessed an increase in socioeconomic segregation. Generally, patterns of spatial segregation in Australian cities are discussed expansively in the academic literature. Theoretically, factors contributing to the intensified spatial segregation in Australian cities are gentrification, increased international migration, and the consequent changes in housing market including home ownership rates, housing densities, and rent costs. The common transformations in Australian cities (Baum, 2008) are basically due to similar characteristics shared by the large urban areas across the country throughout their relatively short histories (Forster, 2006). They were all highly suburbanized with high levels of home ownership. Their
Socioeconomic segregation in global cities | 153 social structure was reflected spatially in ‘doughnut’ shaped cities, where the declined inner cities are circled by growing suburban development (ibid.). Since the 1970s, it is argued that the socioeconomic segregation has worsened generally in Australian cities (Hunter, 2003), as the post-Fordist economic and social restructuring resulted in two contrasting outcomes; the increasing suburbanization of poverty and significant gentrification in inner cities (Badcock, 1997; Forster, 2006; Baum, 2008). Accordingly, the demand for housing units has undergone noticeable changes since the early 1990s. On one hand, the concentration of advanced producer services – especially in Sydney and Melbourne – predominately in the CBD and few adjacent inner suburbs has generated unprecedented demand for high-quality housing close to the CBD to serve the affluent highly-skilled professionals (Stimson, 2001; Baum, 2008). On the other hand, shifts in immigration policies have supported large-scale in-migration of international students and skilled workers (Weller and Van Hulten, 2012). Which have raised the demand for cheap residence, mostly in the outer suburbs (Stimson, 2001). The supply side has been affected by the planning policies adopted in the late 1980s, which aimed to restrain suburban expansion by promoting the construction of new dwellings within the existing built-up areas of cities (Forster, 2006). As a result, residential densities have increased and medium density housing such as apartments, townhouses, and duplexes started to gradually replace the prevalent single-family suburban houses (ibid.). Another incentive for the changes in housing supply is the decline of public housing sector in Australian cities since 1991 (Badcock, 1999), where the federal government decreased funding to public housing authorities, causing shrinkage in new public housing built in recent years, along with the demolition of deteriorated public housing built in the 1950s and 1960s to be replaced by privately owned dwellings (Arthurson, 1998). Consequently, the restricted suburban expansion, the changing residential landscape, and the declining public housing caused the growth rate of housing stock to fall
154 | Analysis and findings gradually since the early 1990s to reach the lowest rate in the last 50 years in 2009 and 2010 (Kulish et al., 2011). In what seems to be a typical example for the global/divided city model, the limited housing supply and increased demand have led to increased competition over limited residential space, which is responsible for housing prices to rise significantly in absolute terms or relative to household incomes (Forster, 2006; Kulish et al., 2011) leading to differential access to the housing market based on households’ ability to pay. Moreover, with the traditionally high rates of residential mobility in Australia (Badcock, 1997), differential access has contributed to the increased ‘neighborhood sorting’ where people with similar preferences (and the ability to pay) tend to live with others with similar characteristics (see chapter two) (Hunter, 2003). While at the same time, differential access limits the opportunities for low-income households to find affordable housing, and confine them to low-quality dwellings in private rental sector (Randolph and Holloway, 2005; Forster, 2006). Ultimately leading to a further segregation between the affluent and the disadvantaged. Figures (5- 10) and (5-11) show the distribution of poor population (with weekly income less than 160$) over Melbourne urbanized area in 2001 and 2006. Clearly, the inner-city poor in 2001 have decreased significantly in 2006, which conform to the gentrification process described in the literature. In fact, Inner Melbourne has witnessed the most extensive gentrification process across Australia (Weller and Van Hulten, 2012), according to Raskall (1995); the extent of the inner-city gentrification in Melbourne exceeded even that observed in the inner-west of Sydney. So far, Australian cities appear to be textbook cases of how globalization can affect the local housing market and generate visible changes on neighborhood level. Where, retreating policy actions and deregulated housing market have compromised cities’ resistance to gentrification and poverty concentration, leading eventually to a higher level of intensified socioeconomic segregation.
Socioeconomic segregation in global cities | 155
N 10 mi 10 km 11 Distribution of population with income less 11 - Figure 5 than 160$ per week in urbanized Melbourne 2006 N 10 mi 10 km 10 Distribution of population with income less - Figure 5 than 160$ per week in urbanized Melbourne 2001
FigureFi 5-105 10 distributiondi ib i of f population l i with ihi income less l than h 16$ per week ki in urbanized b i dM Melbournelb iin 2001 Figure 5-11 distribution of population with income less than 16$ per week in urbanized Melbourne in 2006
156 | Analysis and findings
However, as shown in other cases, government intervention is an influential contingency that can either lead to intensified socioeconomic segregation (as the case of Mexico City) or promote higher levels of integration (as the case of Rio de Janeiro). The Australian city of Adelaide is also an example of such influence of local government’s decision on the outcomes of spatial restructuring on neighborhoods level. Originally, Adelaide had a reputation to be economically disadvantaged compared to other Australian cities, where some areas of Adelaide are among the nation’s poorest suburbs (Baum, 2008). Then, in late 1990s, local government focused on issues of inequality and responded by several urban regeneration projects such as ‘The Parks neighborhood renewal’ in Adelaide’s west (Forster, 2006) and ‘Salisbury North urban improvement project’ in Adelaide’s north (Arthurson, 2012). Those projects aimed to improve housing and infrastructure of deteriorated neighborhoods, as well as implementing community development programs that reduce welfare dependence and promote social inclusion of poor residents (Forster, 2006; Arthurson, 2012). As a result, despite the economic restructuring process shared by Australian cities, Adelaide’s relatively poor outer suburbs do not show the same stark contrast to the inner city witnessed in Melbourne or Sydney. Which once again highlights the influential role of local governments on the outcomes of economic restructuring in individual cities.
5.3.3 African cities
As noted, African cities are underrepresented in the final dataset due to data availability, and the only three African cities in the dataset are all situated in South Africa. Yet the discussion here is still central to the understanding of the role of local contingencies in the complex process of cities restructuring. The case of Johannesburg shows an alpha city that does not follow the global/divided city model (since the city scored a decrease in socioeconomic segregation by 0.7% by year), while the explanation of its
Socioeconomic segregation in global cities | 157 spatial transformation is strongly linked to historical and social factors that existed on the local level in previous decades. In more detail, South African social classes have always been defined according to color line; “[wealth] was largely in the hands of white South Africans” (Beall, 2002: 47). After the first non-racial, democratic elections in 1994, which coincided with the rapid growth of service sector and changes in employment patterns, race and class association began to erode, and new black middle class started to grow to support the service industries that were predominated by whites (ibid.). For that matter, the decrease in socioeconomic segregation is explained by changes in the labor market that are caused, not only by economic changes in the global context, but also caused by changes in racial segregation in post-apartheid Johannesburg. Moreover, Beall et al. (2000) acknowledge the role of urban governance in decreasing segregation, through reaching a balance between achieving global competitiveness and tackling spatial segregation by executing programs of black empowerment and poverty reduction. Despite the fact that Cape Town and Durban share the same national, political, and historical context of Johannesburg, both cities showed different direction of spatial transformation than their fellow city. The reason behind this inconsistency is individual situation of each city. Where unlike Johannesburg, ‘shack-dwellers' movement in both Cape Town and Durban have resisted the forced eviction of blacks from slum areas in the center of these cities to peripheral townships (Huchzermeyer, 2011), which caused both racial and socioeconomic segregation to be harder to tackle. Accordingly, South African cities highlight the importance of recognizing cities as individual cases. Where generalized characteristics of cities on global or regional levels are eventually reshaped by distinct situation in each city.
158 | Analysis and findings
5.3.4 North American cities
In total, 51.2% of cities of the Anglo American region have witnessed an increase in socioeconomic segregation levels since the late 1990s, yet regional differences appear between cities in Canada and USA, and even on sub-country level especially within the USA. In case of the Canadian cities, only 33% of Canadian cities showed an increase in socioeconomic segregation level, Fong and Shibuya (2000) claim that socioeconomic segregation in these cities is influenced by racial and ethnic segregation rather than the socioeconomic status of the segregated poor. Where racial and ethnic groups have differential access to housing choices. Therefore, governmental actions that promote ethnic integration eventually reduce the associated socioeconomic segregation. Which explains the tendency for Canadian cities to be less segregated by income. This argument is discussed in more detail in the subsequent chapters. In case of American cities, issues of socioeconomic segregation have a longer history than those in Canada. Moreover, in spite of the strong historical association between socioeconomic status and race in the US (Massey and Denton, 1993), race alone cannot explain the increased poverty concentration in inner cities (Wilson, 1987), especially when considering the macro economic changes impacting western societies such as deindustrialization. Income inequalities in the US have emerged as a result of the uneven economic development on the national level since the late 1940s (Brenner, 2002). Originally, the 1950s were the golden age of manufacturing industries in central cities of the Northeast and Midwest, then the increasingly footloose industries shifted slowly - during the 1950s to the 1990s – to the Southeast then the Southwest/West then finally to Asia in the 1990s to present (Ross, 2011). With the massive impacts of deindustrialization on cities of the Northeast and Midwest, they were labeled as Rustbelt cities, a term that summarizes the process of fiscal crisis and the
Socioeconomic segregation in global cities | 159 consequent joblessness, tax revenue loss, and concentrated poverty (ibid.). As a result, the decline of Rustbelt cities accelerated the shift in population and employment to the South into the fast growing Sunbelt cities (Chapple and Lester, 2010). The North and South division in the USA was intensified in the 1980s due to neoliberal federal policies in Reagan’s era, where “state and municipalities began to adopt entrepreneurial strategies in order to attract external capital investment to their territorial jurisdictions” (Brenner, 2002: 8) including infrastructure investments and locational decisions of industries based on the supply of cheaper labor (ibid.). (See chapter six – Sunbelt cities receive sizable in- migration surge from Central and South America who are mostly low-pay workers). Apparently, as shown in figure (5-12), the socio-spatial impacts of the historically uneven development between Rustbelt cities and Sunbelt cities - since 1940s to 1980s - are not yet reversed. The map shows the 33 American cities, of which 14 cities are Sunbelt cities, the rings indicates the value and direction of change in socioeconomic segregation from 2000 to 2009. In the North, 14 out of 19 (73.7%) cities have witnessed an increase in socioeconomic segregation, compared to only 4 out of 14 (28.6%) cities in the Sunbelt. Brenner (2002) explains the continuation of the north-south division in the 21st century to be a result of the ineffective reform of federal policies under Bush and Clinton administrations, both modified the pure neoliberal policies of the preceding decades. However, urban policies are still market- based which sustained the regional competitiveness over investment. In addition, the growth of Sunbelt cities in terms of territorial size, population and economic output have widened the gap between the north and the south. According to Chapple and Lester (2010), the growth of the south is associated with the booming information technology economy of the 1990s - for example, average earnings per worker in Austin increased by 42% due to the fast growing high-tech industries in Austin metro area. Therefore, the changes in socioeconomic segregation showed in figure (5-3) conform to
160 | Analysis and findings
Figure 5-12 increase/decrease in socioeconomic segregation in both Sunbelt and Rustbelt cities in the USA USA the in cities Rustbelt and Sunbelt both in segregation socioeconomic in increase/decrease 5-12 Figure Figure 5-12 increase/decrease in socioeconomic segregation in both Sunbelt and Rustbelt cities in the USA
Socioeconomic segregation in global cities | 161
Chapple and Lester interpretation of economic development of Sunbelt cities. The increasing average earnings per worker in Austin is reflected spatially, where Austin scored the highest degree of socioeconomic integration (18.4% decrease in segregation from 2000 to 2009 which equals -2% of yearly change). While, the formerly leading industrialized cities of Detroit and Denver are amongst the top cities with increased socioeconomic segregation. Arguably, neoliberal globalization has clear impacts on American cities, exhibited in deindustrialization, increased competitiveness, and uneven development of the north and the south. However, the generalized association between globalization and socio-spatial divide does not apply to American cities as well as North American cities in general. Where, first, only half of North American cities had a significant increase in socioeconomic segregation. Second, inconsistencies in cities’ direction and value of spatial changes suggest that spatial transformations need to be explained on sub-regional and local level of individual cities. For example, the north-south division of winner and loser cities cannot explain the spatial transformations in Chicago and Dallas; both are alpha cities where command and control centers of global economy are located (Sassen, 1991), but Chicago scored less segregation despite its location in the Midwest (the losers region), while Dallas scored more segregation despite its location in the South (the winners region). Accordingly, finer grain analysis is required to comprehend the contingent factors that influence spatial transformations in individual American cities.
5.3.5 Western European cities
As shown in chapter two, several qualities have distinguished the socio-spatial transformations of Western European cities from those of American cities or any other region in general. According to Van Kempen and Murie (2009), ‘the corporatist and social democratic welfare states in Europe’ has contributed to the mitigation of the inequalities associated with
162 | Analysis and findings industrial restructuring, through strong employment protection and benefit systems. Accordingly, unemployment rates and loss of income are relatively low compared to those produced under the liberal welfare state of the USA. However, in terms of quantity, figure (5-3) shows that patterns of spatial changes in Europe are not so different than those in USA. 50% of European cities of the dataset have witnessed a significant increase in income segregation over the past decade, compared to 51.2% in North America. Also, the intensity of spatial changes appear to be higher in European cities in either directions toward more segregation or integration. The region hosts the top three cities – of the entire dataset - with less socioeconomic segregation, Brussels, Berlin, and Copenhagen, along with Amsterdam which scored the highest increase in socioeconomic segregation after Guadalajara and Monterrey. On the national level, Swedish cities of Stockholm and Gothenburg showed a consistent and strong tendency for increased socioeconomic segregation, while patterns of segregation in Amsterdam and Rotterdam are inconsistent in terms of direction and intensity of the change. Noticeably, the results do not indicate that absolute levels of spatial segregation in Europe are higher than in USA. Instead, results only refer to the pace and intensity of spatial changes over a certain period of time. The academic literature on spatial segregation in Western European cities focuses on urban policies and public housing as key aspects in defining residential segregation patterns in European cities (Musterd and Andersson, 2005; Van Kempen and Murie, 2009). Since low-income households are (in most cases) overrepresented in public housing dwellings. Then, the spatial concentration of public housing in certain areas of the city can lead to the spatial concentration of low-income households (Van Kempen and Murie, 2009). Therefore, the difference in location, quantity, and quality of public housing from city to city can explain the variations in segregation levels among European cities. However, the dynamic process of change in socioeconomic segregation patterns requires the preexisting
Socioeconomic segregation in global cities | 163 density and location of public housing to be altered, through either introducing additional public housing projects, privatization of existing ones (e.g. Musterd and Fullaondo, 2008), or demolition of old projects as part of the renewal and diversification projects adopted by several European states (e.g. Kleinhans, 2004; Musterd, 2005). Evidently, urban policies vary significantly among Western European countries (Van kempen and Murie, 2009). The next section briefly discusses housing policies and its consequent change in segregation patterns in three states, the Netherlands, Sweden, and Belgium. In the Netherlands, The national Urban Renewal Policy - since 1997 – aimed to reduce spatial concentration of low-income households and ethnic minorities to finally promote social mix in residential areas (Kleinhans, 2004; Bolt et al., 2009). The renewal processes seemed to follow a typical route of gentrification, as neighborhood restructuring resulted in the demolition of old social rented dwellings and the construction of owner- occupied expensive dwellings (ibid.). Unfortunately, that supposedly successful strategy for improving social cohesion has resulted in a “zero-sum outcome” (Musterd and Andersson, 2005: 766). As, diversification of social structure by introducing higher income households to the target area is accompanied by the displacement of low-income households who used to reside in the demolished dwellings. According to Bolt et al. (2009: 515), low-income household “tend to move to neighborhoods with a similar population composition as the areas they are leaving behind. In other words; maybe concentrations of low-income households are broken in the targeted areas, but there is a strong suggestion that new concentrations emerge elsewhere.” Causing the average segregation on city level to increase or - at least – remain the same. This theorization of residential restructuring in Dutch cities may explain the increased socioeconomic segregation in Amsterdam along with other factors discussed in chapter six. However, Rotterdam has a different story, where diversification projects are accompanied by strict housing
164 | Analysis and findings allocation rules set by local government. Those rules were defined in 2003 by the ‘Rotterdam zet door’ action program (Kleinhans, 2004), which aimed to control the spatial distribution of low-income households and prevent their re-concentration in disadvantaged neighborhoods. Apparently, Rotterdam plans succeeded in achieving social mix – at least statistically, as table 5-1 shows that Rotterdam is in the second place among the top 10 cities with decreased socioeconomic segregation. Again, the cases of Amsterdam and Rotterdam are examples of how existing local situations and governmental initiatives can alter or even reverse spatial outcomes. In Sweden, the country has always had a reputation for having a well- developed welfare system, progressive housing policy, and good quality housing (Musterd and Andersson, 2005; Andersson 2007). In 1965, the government initiated ‘The Million Homes’ program to overcome housing shortage by constructing one million new dwellings in ten years (Bråmå, 2008). In later decades – as in most western societies – neoliberal housing policies were adopted by the state. As a result, the housing sector in Sweden became deregulated and market-oriented (Hedin et al., 2012), and existing public rental dwellings are increasingly being privatized. Moreover, public housing companies became more selective in their choice of tenants, and the very poor were excluded from affordable housing (Sahlin, 2008), as a result, between 1999 and 2005, homelessness rates has doubled (Hedin et al., 2012). The overview on public housing in Sweden gives a broad idea on residential shifts expected to take place in both Stockholm and Gothenburg. Also, the absence of a counter action from both central and local governments – as in the case of Rotterdam - suggests that privatization of public housing and potential displacement of low-income households will eventually lead to higher level of concentration of disadvantaged population, which conform to the increase in segregation index in both cities shown in the results. Furthermore, growing inequalities in Sweden is perceived to be drastic, according to Musterd (2005), Swedish society is known to be more
Socioeconomic segregation in global cities | 165 egalitarian than any other society in Europe. Therefore, the recent inequalities “have led to a relatively strong labeling and stigmatization reaction in the Swedish context.” (Musterd, 2005:344). Following the same logic, the intense increase in socioeconomic segregation shown in figure 5-3 is not simply a result of the neoliberal reform of welfare system, but also because the Swedish welfare system – at least until the 1990s - was considered to be ideal and comprehensive. Thus, socio-spatial consequences of neoliberal reform are magnified, especially with the absence of governmental actions that can resist, modify, or reverse the increasingly segregated spatial outcomes. In fact, the beta and gamma cities of Stockholm and Gothenburg appear to be ideal examples of the global/divided city model discussed in the literature. In Belgium, the alpha city of Brussels offers totally different model from those discussed in the Netherlands and Sweden. Where the role of the state in housing provision has never been very large (Kesteloot and Cortie 1998), and when compared to other European cities, Belgian cities in general have relatively small shares of social housing (Kesteloot et al. 1997; Musterd, 2005). Accordingly, the changes in residential patterns in Brussels are not occurring due to demolition or privatization of public housing dwellings as in Amsterdam or Stockholm. Instead, among other factors, the expansion of the European Union institutions in Brussels as a capital of Europe has led to increased demand for additional office space and other supporting services, which eventually led to gentrification (Kesteloot, 2000; Swyngedouw et al., 2002). However, despite the fact that gentrification and its associated displacement theoretically contribute to higher levels of spatial segregation (see chapter three), still, Brussels managed to come in the first place among the top 10 cities with highest level of spatial change towards more integration. This spatial outcome is mainly due to two distinct characteristics that mark the process of gentrification in Brussels. Firstly, Kesteloot (2000) describes gentrification in Brussels to be “modest” (p: 203), compared to level of ongoing restructuring of the city.
166 | Analysis and findings
Also, the gentrification is contained in certain areas such as Leopold Quarter (Swyngedouw et al., 2002) and in the eastern edge of the 19th-century belt in general (Kesteloot, 2000), where EU institutions are located. In fact, residential blocks in the area are systematically bought by property developers and eventually demolished and replaced with offices. Yet, there is no evidence that the displaced households are re-concentrating in other neighborhoods. Secondly, according to Van Criekingen and Decroly (2003), the comparative analytical study they performed over Brussels and Montreal shows that scale (size) of gentrification projects in Brussels are small, as they put it “smaller than census tract” (p: 2466). Accordingly, the decreased segregation on city level can be explained by the scale of gentrification, where municipalities that host a small scale gentrification project will have higher social mix, which cause the average level of segregation on the municipalities level to drop significantly. Nevertheless, higher levels of income segregation may appear on smaller scale such as the neighborhood4 level, where due to ‘small scale gentrification’, adjacent blocks in the same municipality can show clear contrast in their income structure and housing quality, still, this assumption requires another detailed analytical study to be confirmed.
Socioeconomic segregation in global cities | 167
Summary
In the light of the global city literature, global cities are expected to have an increasingly fragmented socio-spatial structure due to economic restructuring and its associated growing inequality. However, the analytical study presented in this chapter concludes that generalized patterns of spatial transformations on the global level are not detected; the 66 cities of the dataset have shown diverse directions of change in segregation level, where only 54.5% of global cities have witnessed an increase in socioeconomic segregation, which challenges the validity of the generalized assumptions in the literature. This conclusion is also supported by a regression analysis, which confirmed that the global network connectivity of cities explains only 0.2% of the variability in spatial segregation index. Moreover, the one-way analysis of variance ANOVA revealed that cities’ position in the global network represented in alpha, beta, and gamma categorization are not associated with certain changes in socioeconomic segregation. While the location of global cities in certain geographic region is proved influential on the resultant levels of segregation Cases discussed in this chapter show that local situations in individual cities such as housing regulations, the existence/condition of public housing, and gentrification levels define the changes in the residential canvas of each city. And since local situations on cities level vary greatly within the same region and sometime within the same nation-state (as in cases of Johannesburg - Cape Town and Amsterdam - Rotterdam), then, generalized patterns of spatial changes in global cities are to be questioned. In fact, the impacts of globalization-related phenomena (such as gentrification and privatization of public housing) on the spatial transformation within cities are not uniform. Gentrification in Brussels and Melbourne has led to two distinct spatial outcomes due to the scale and intensity of the gentrification process. Similarly, the state provision of public housing in Mexico has contributed to higher levels of socioeconomic
168 | Analysis and findings segregation, while at he same time, the privatization of public housing in Sweden has also led to more segregated outcome. Another example for local situations that define spatial outcome is the cases of Adelaide, Rio de Janeiro, and Rotterdam. The three cities are situated in different global, regional, and sub-regional contexts. Yet, local government of each city could reach higher levels of integration through deploying different strategies such as urban development projects in Adelaide and Rio, and housing allocation regulations in Rotterdam. As a result, the three cities showed a significant decrease in socioeconomic segregation. The individuality in cities also explain the significant variations in direction and intensity of spatial change in American cities, where the factor of race is included in defining patterns of socioeconomic segregation along with other issues of affordability and access to the housing market. Race and ethnicity also show relevance to socioeconomic segregation in other contexts such as South Africa and Canada. The next chapters shed more light on patterns of ethnic segregation and the intersection of ethnicity and socioeconomic segregation in global cities.
Socioeconomic segregation in global cities | 169
Notes
1 2010 classification is compared only to 2008 classification due to the different methodology used by the GaWC in previous classifications, because the network connectivity became the main measure of importance of cities only from 2008 onwards (Taylor et al., 2011), 2 The more advanced producer service firms the city host the more stark the changes in the city’s occupational structure, income structure, pace and intensity of gentrification, and eventually intensified socioeconomic segregation. 3 By definition, ‘walling out’ is the extreme physical form of voluntary and deliberate separation of a socially and economically dominant group. Walling out may be involved in the formation of ‘exclusionary enclave’ and is also involved in the formation of a ‘citadel’ (Marcuse, 2005). 4 The 19 municipalities of Brussels are divided into 724 neighborhoods (quartiers), 606 of them have more than 200 inhabitants (Brussels Instituut voor STATISTIEK en Analyse, 2005). Statistical data of income groups on neighborhood level are not available before 2008 (only median income is available for every year since 1993). For that matter, the analysis is performed on the larger divisions of municipalities due to the availability of comparable data.
170 | Analysis and findings
Chapter 6
6 Ethnic segregation in global cities: Analysis and findings
As noted in chapter four, type of data collected by each census vary greatly from country to country. Yet, since ethnic composition within cities is related to issues of migration flows and labor mobility, which are accelerated in the age of globalization. Therefore, data collected for different cities meant to reflect the impacts of those factors, such as1, population by citizenship status (foreigners): in countries of Spain, France, and Switzerland, population by nationality by region: in countries of Belgium and the Netherlands, and foreign born population by region: in USA, Canada, and Australia. Noticeably, developing countries that are sending migration to other parts of the world have no data related to in-migration flows to global cities in these regions2, and hence are excluded from the final dataset. Accordingly, the final list contains 81 cities, to be analyzed within the period from 1995 to 2012, divided into 29 alpha, 26 beta, and 26 gamma cities. In the light of the discussion in chapter two and three, levels of ethnic segregation are expected to increase in global cities due to one of two reasons. On one hand, the increased demand for both skilled professionals and low-skilled migrants - with a relatively easier labor mobility - lead to a higher complexity of the ethnic composition of global cities, this complexity eventually contributes to higher level of spatial segregation, especially when
172 | Analysis and findings considering the preferences and behaviors of individual new-comers to the city, as most migrants follow ‘beaten paths’ and go where their fellow nationals have already established a community, making it easier to find work and accommodations (Castles, 2002). On the other hand, regardless of the in-migration flows to the city, levels of segregation among historically coexisting ethnic groups can still rise due to the socioeconomic status of ethnic groups and its related issues of access to education, labor market, and eventually, housing opportunities. This argument is discussed in more detail in chapter seven. Despite the expected increase in ethnic segregation levels in global cities, as shown in figure (6-1), global cities of the dataset showed the same lack of shared patterns of change in ethnic segregation as in the case of socioeconomic segregation. Where 45 cities (55.6% of the dataset) show a tendency for decreased levels of ethnic segregation compared to only 36 cities (44.4% of the dataset) with an increase in ethnic segregation levels. As a general observation, The maximum increase in ethnic segregation is scored by the alpha city-state of Singapore, with a 4.7% of average yearly change in SD(m) value (47% increase in ethnic segregation from 2000 to 2010), followed by the alpha city of Sydney with 2.5% by year. On the other end of the spectrum, the maximum decrease in ethnic segregation is also scored by an alpha city, that is Milan with a -5.8% of average yearly change in SD(m) value (-52.27% decrease in segregation between 2001 and 2010), followed by the beta city of Manchester with -4.5%. See subsequent sections for the in-depth interpretation of the observation.
6.1 Examining the correlation between the intensity of spatial change and cities’ global network connectivity.
This section presents the results of the linear regression analysis introduced in chapter four. The calculations are generated using IBM - SPSS
Ethnic segregation in global cities | 173
London New York Hong Kong Paris Singapore Tokyo Shanghai Chicago Sydney Milan Beijing Toronto Madrid Los Angeles ha
Frankfurt p
Amsterdam al Seoul Brussels San Fransisco Washington Miami Melbourne Zurich Munich Boston Dallas Atlanta Barcelona Philadelphia Dusseldorf Stockholm Montreal Rome Hamburg Houston Berlin Copenhagen Vancouver Seattle Auckland Oslo Manchester
Minneapolis beta Brisbane Geneva Detroit Denver Bratislava Stuttgart Cologne St Louis San Diego Perth Cleveland Calgary Cincinnati Charlotte Antwerp Baltimore Adelaide Portland San Jose us Valencia Kansas city Phoenix Lyon Rotterdam Tampa Columbus amma
indianapolis g Pittsburgh Edmonton Tallin Orlando Gothenburg Marseille Ottawa Richmond Austin Milwakee Wellington 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 Figure 6-1 alpha – beta – gamma cities and their yearly percentage of change in ethnic segregation index index segregation ethnic in of change percentage yearly their and cities – gamma – beta 6-1 alpha Figure Figure 6-1 alpha – beta – gamma cities and their yearly percentage of change in ethnic segregation index
174 | Analysis and findings statistics software, and the aim of the analysis is to test the strength of the association between the cities rank in the global network of cities and their corresponding change in ethnic segregation. The expected result - if the null hypothesis is correct – is a strong correlation between the two variables indicated by the value of R that is greater than 0.5. Also, the summary scatterplot in figure (6-2) represents the correlation between the dependent variable, which is the standardized change in the spatial dissimilarity index SDI on the y-axis, and the independent variable, which is the cities’ global network connectivity GNC on the x-axis.
Standardized change in SD(m) index value
Global Network Connectivity
Figure 6-2 Summary scatterplot for the correlation between the standardized change in segregation index and the cities’ global network connectivity
Ethnic segregation in global cities | 175
The SPSS also generates several tables of for the linear regression. In this section, we show the three main tables required to understand the results of the of the linear regression procedure: the model summary table, the ANOVA table, and the coefficients table.
Table 6-1 Regression output – model summary, ANOVA, and Coefficient tables Model Summary Adjusted Std. Error of the Model R R Square R Square Estimate 1 .048a .002 -.010 1.00516005675 a. Predictors: (Constant), AbsoluteGNC
ANOVAa Sum of Mean Model Squares df Square F Sig. 1 Regression .183 1 .183 .181 .672b Residual 79.817 79 1.010 Total 80.000 80 a. Dependent Variable: zscoreETHNICITY b. Predictors: (Constant), AbsoluteGNC
Coefficientsa Unstandardized Standardized Coefficients Coefficients
Model B Std. Error Beta t Sig. 1 (Constant) .112 .287 .392 .696
AbsoluteGNC .000 .000 -.048 -.425 .672 a. Dependent Variable: zscoreETHNICITY
176 | Analysis and findings
Firstly, the regression output table (6-2) provides the information needed to determine how well the regression model fits the data. In the model summary section, the column ‘R’ is the absolute value of the Pearson correlation coefficient between the dependent and the independent variables. It simply indicates the strength of the association between the two variables. In our case, R = 0.048 < 0.5, which indicates a weak correlation. The R2 value in the R2 column represents the proportion of variance in the dependent variable that can be explained by the independent variable. In our case, R2 = 0.002, which means that the independent variable, global network connectivity, explains only 0.2% of the variability of the dependent variable, change in segregation index. Please note that the R2 is calculated for the sample (81 global cities). SPSS generates another value called adjusted R2 that represent the proportion of variance if the analysis is performed over the entire population (i.e. the full list of global cities). Table (6-2) shows that the adjusted R2 = -0.010, which indicates a very low dependence between variables. The next table is the ANOVA table, which reports how well the regression equation fits the data (i.e., how accurately the model can predicts the dependent variable SDI based on the changes in the independent variable GNC). The number in the "Regression" row and the "Sig." column indicates the statistical significance of the regression model that was run. Here, the regression model is statistically insignificant, F (1, 79) = 0.181, p = 0.672, which is greater than 0.05, and indicates that, overall, the regression model cannot statistically predict the outcome variable, which conform with the low dependence between variables observed from the model summary table. The last table is the coefficients table, which provides the coefficients of the regression equation that is used to predict the dependent variable from the independent variable. y = b0 + (b1 x) where y is the change in
Ethnic segregation in global cities | 177 segregation, x is global network connectivity, b0 is the intercept and b1 is the coefficient, both can be found under the B column in the table, b0 = 0.112, and b1 = 0.000. In our case, it is already established that the model poorly fits the data due to the weak dependence between the variables. Recapitulating, the linear regression established that global network connectivity for a city could not statistically significantly predict the change in ethnic segregation level, as the two variables have a weak correlation. The global network connectivity of cities accounted for only 0.2% of the explained variability in spatial segregation index.
6.2 Examining the influence of location, history, and other factors
Parallel to the findings in the case of socioeconomic segregation, the analysis confirmed that cities’ position in the global network of cities is not associated with the direction or the intensity of change in ethnic segregation within these cities, which also suggests that there are other independent variables that can explain the variation in the dependent variable. This section aims to assess the influence of several factors that are assumed to be relevant to the outcomes of spatial restructuring in global cities. As explained in chapter four, the influence of these factors is evaluated by the one-way analysis of variance ANOVA. Firstly, to examine the influence of a city’s global history on the resultant change in spatial segregation, the sample is divided into two groups: cities newly added to 2010 classification, and cities in the classification since 2008. Table (6-3) reports the output of the one-way ANOVA generated by IBM- SPSS. The table provides the F ratio, as explained in chapter four, the calculated value of F can be compared with x that is the critical value below which there is no statistically significantly difference between groups’ means. In other words, when F178 | Analysis and findings under test has no significant influence on the observed variation in segregation level and vice versa. In our case, F=2.017 < x=3.97 which
Table 6-2 ANOVA output when cities are grouped according to their global history
ANOVA Zscore: ETHNICITY Sum of Mean Squares df Square F Sig. Between Groups 1.106 1 1.106 2.017 .160 Within Groups 41.114 75 .548 Total 42.220 76 ONEWAY Zscores SD(m) BY global history indicates insignificant difference between cities that are already global cities for a longer period of time than cities that are recently added to the classification. This conclusion is also confirmed with the p-value under the column ‘Sig.’ in table, where p=0.16, which is greater than a significance level of 0.05. The p-value here is another indication on the absence of statistically significantly difference in groups’ means.
Table 6-3 ANOVA output when cities are grouped according to their global rank ANOVA Zscore: ETHNICITY Sum of Mean Squares df Square F Sig. Between Groups 1.074 2 .537 .966 .386 Within Groups 41.146 74 .556 Total 42.220 76 ONEWAY Zscores SD(m) BY global rank
Similarly, the sample is once again divided into three groups; table (6- 5) reports the output of the one-way ANOVA when cities are categorized in alpha, beta, and gamma groups. The calculated F ratio from the table equals
Ethnic segregation in global cities | 179
0.966 which is once again less than the critical value of x = 3.12. Accordingly, it can be concluded that changes in ethnic segregation witnessed by alpha cities is not statistically significantly different from those witnessed by beta or gamma cities. The p-value from the table also confirms this conclusion, where p= 0.386, which is greater than a significance level of 0.05. Table 6-4 ANOVA output when cities are grouped according to their geographic region ANOVA Zscore: ETHNICITY Sum of Mean Squares df Square F Sig. Between Groups 6.034 4 1.508 3.001 .024 Within Groups 36.186 72 .503 Total 42.220 76 ONEWAY Zscores SD(m) BY region
ANOVA Zscore: ETHNICITY Sum of Mean Squares df Square F Sig. Between Groups 5.682 3 1.894 3.792 .014 Within Groups 35.464 71 .499 Total 41.146 74 ONEWAY Zscores SD(m) BY region except Eastern Europe
Another factor to be tested in this section is the location of a global city in a certain geographic region. The cities of the sample are divided according to their broad geographic region into 5 groups: Western Europe, Eastern Europe, North America, Australia, and Asia. Table (6-5 – above) presents the ANOVA output, which reveals that – unlike the previous factors of global history and rank – the F ratio equals 3.001, which is greater than
180 | Analysis and findings x= 2.5. In this case, the analysis results suggest that cities located in certain region do share trends of change in spatial segregation, and these trends are significantly different from cities in other regions. This conclusion is confirmed by the p-value = 0.024 which is less than 0.05. As in the case of socioeconomic segregation, the table (6-5 - below) shows a re-run of the ANOVA test when cities of Eastern Europe are excluded from the sample, as the region contained only 2 cities and their exclusion aims to enhance the accuracy of the analysis and to avoid any misread conclusions. Accordingly the analysis is performed again among 75 cities grouped into four regions: Western Europe, North America, Asia, and Australia. Noticeably, the F ratio = 3.792 is still greater than x = 2.73, and the p-value has improved to reach 0.014, which indicates a high significance level of the influence of the geographic region on the changes of segregation in global cities.
6.3 The individuality of cities: a discussion
Table 6-7 presents the top 10 cities with increased socioeconomic segregation and the top 10 cities with decreased socioeconomic segregation. As shown in table 6-7, the list of top 10 cities with decreased ethnic segregation is dominated by alpha cities, where 5 out of the 10 cities are ranked as alpha global cities. Yet another observation suggests that regional and national contexts have a considerable effect on the final results; where the top 8 (out of 10) cities with a decrease in ethnic segregation level are all located in Western Europe, while cities of Australia occupy the 2nd, 3rd, and 4th places of cities with the highest increase in segregation. Accordingly, the absence of clear pattern of spatial change among global cities - as well as potential similarities among cities of the same region - indicates that the understanding of the complex process of spatial restructuring of global cities requires a closer look on the local situations and particularities of individual cities. For that matter, the subsequent sections
Ethnic segregation in global cities | 181 are dedicated to discuss in more detail the motivations and factors involved in the production of diverse patterns of ethnic segregation in individual
Top ten cities with maximum increase and decrease in socioeconomic segregation and their respective ranks Top 10 cities with increased segregation City Rank Region 1 Singapore alpha Asia 2 Sydney alpha Australia 3 Perth beta Australia 4 Brisbane beta Australia 5 Valencia gamma Europe 6 Cincinnati gamma N. America 7 Tallinn gamma Europe 8 Richmond gamma N. America 9 Denver beta N. America 10 Seattle beta N. America Top 10 cities with decreased segregation City Rank Region 1 Milan alpha Europe 2 Manchester beta Europe 3 Zurich alpha Europe 4 Barcelona alpha Europe 5 Geneva beta Europe 6 Stuttgart beta Europe 7 Rome beta Europe 8 Munich alpha Europe 9 Indianapolis gamma N. America 10 Beijing alpha Asia
182 | Analysis and findings
Beijing Hong Kong Shanghai Tokyo Seoul Singapore S.E. Asia Milan Manchester Zurich Barcelona Geneva Stuttgart Rome Munich Oslo Antwerp Copenhagen e
Madrid p Brussels Rotterdam Dusseldorf Gothenburg Berlin Euro W. London Lyon Stockholm Cologne Paris Hamburg Frankfurt Amsterdam
Marseille e Valencia p Bratislava Tallin indianapolis Orlando E. Euro Charlotte Miami Minneapolis Ottawa Montreal Atlanta Houston Washington San Fransisco Dallas Chicago Phoenix San Jose us San Diego Boston New York Los Angeles St Louis Baltimore lo America
Columbus g Cleveland Tampa An Kansas city Calgary Portland Milwakee Vancouver Philadelphia Austin Toronto Pittsburgh Detroit Edmonton Seattle Denver Richmond Cincinnati Adelaide Melbourne Wellington Auckland Brisbane
Perth Oceania Sydney 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7
Figure 6-3 global cities categorized according to their geographical region and their yearly change in ethnic segregation index index segregation ethnic in change yearly their and region geographical to their according categorized cities 6-3 global Figure
Ethnic segregation in global cities | 183 global cities. And according to the results of the ANOVA, the sections are organized based on the broad geographic regions of the world, which was already proven to have a significant influence on the outcomes of spatial restructuring in global cities. Also, for easier reading of the results presented in figures (6-1), figure (6-3) presents global cities categorized according to their location in the broad geographic regions of the world instead of their global rank. Clearly, figure (6-3) shows that the dataset is geographically imbalanced, where 66 out of the 81 cities of the dataset are located in migration-receiving countries of Western Europe and North America, while migration-sending countries of Latin America and Africa have no representative cities in the dataset. As a general observation, cities of Australia and New Zealand showed higher tendency for increased ethnic segregation, where 71.4% of cities within the region have witnessed an increase in the value of SD(m), compared to 46.1% of cities in North America, and only 33.3% of cities in Western Europe. The following sections discuss in more detail the factors contributing to the changes in the spatial arrangement of different ethnic groups in individual global cities.
6.3.1 Australian cities
Spatial segregation of ethnic groups in Australian cities was initially caused, then reinforced, by several factors including location of job opportunities, location of housing opportunities, size of ethnic groups, language, and marginalization (Hugo, 1995). During the 1950s and 1960s, available cheap rental housing was located in the inner suburban areas of major cities. Accordingly, newly arrived immigrants were attracted to those areas. However, as shown in chapter five, with the flourishing service industries in the past two decades, a new middle class arose (Badcock, 2000) and inner cities have consequently been affected by gentrification (Hugo, 1995; Badcock, 2000). As a result, the inflow of higher-income residents - who do not necessarily belong to the same ethnic group residing in inner city
184 | Analysis and findings areas – contributed to changes in patterns of geographical concentration of certain ethnic groups. For example, the close proximity of the gentry group to the existing ethnic enclave can lead to the overall decrease in ethnic segregation level in the city. However, in case that the process of gentrification is accompanied with large-scale displacement and reconcentration of disadvantaged population who are – in most cases – also immigrants, then gentrification may lead to the exacerbation of ethnic segregation by dismantling the existing enclave in inner city and the creation of another enclave elsewhere in the city. This gentrification and displacement scenario may explain the growing ethnic segregation in Sydney, Perth, and Brisbane. In figures (6-4) and (6-5), areas on the waterfront of Brisbane River such as Highgate Hill, Indooroopilly, and West End Brisbane have witnessed noticeable change in their ethnic composition. Where Middle Eastern immigrants were pushed south away from the valuable waterfront area. Yet still, gentrification and displacement scenario fails to explain the decrease in ethnic segregation in Melbourne and Adelaide. Note that Melbourne has witnessed the most extensive gentrification process across Australia (Weller and Van Hulten, 2012), but the large-scale displacement of inner-city immigrants did not contribute to the intensification of ethnic segregation. The explanation here is that the displacement of immigrant is not followed by the creation of another enclave by the displaced group, instead, the displacement caused their dispersal in other poor neighborhoods that are already inhabited by other ethnic groups, leading overall level of ethnic segregation to drop. The case of Adelaide is also a good example of such dispersal of ethnic groups due to gentrification. As shown in figures (6-6) and (6-7), in 2001, Middle eastern immigrants for example were concentrated in inner city, with insignificant presence in inner suburbs. However, their displacement in 2006 from inner city areas such as, West Torrens, Marion, and Onkaparinga Woodcroft, did not only lead to their concentration somewhere else away from gentrified areas, but more specifically, they are displaced to
Ethnic segregation in global cities | 185
N
2 mi
2 km
Figure 6-4 Brisbane in 2001– distribution of North African and Middle Eastern migrants
N
2 mi
2 km
Figure 6-5 Brisbane in 2006– distribution of North African and Middle Eastern migrants
Figure 6-4 Brisbane in 2001– distribution of North African and Middle Eastern migrants Figure 6-5 Brisbane in 2006– distribution of North African and Middle Eastern migrants
186 | Analysis and findings
N 10 mi 10 km 7 urbanized area of Adelaide in 2006, distribution 7 urbanized area of - of Middle Eastern migrants Figure 6 N 10 mi 10 km 6 urbanized area of Adelaide in 2001, distribution 6 urbanized area of - Figure 6 of Middle Eastern migrants
Figure 6-6 urbanized area of Adelaide in 2001, distribution of Middle Eastern migrants Figure 6-7 urbanized area of Adelaide in 2006, distribution of Middle Eastern migrants
Ethnic segregation in global cities | 187 neighborhoods that are originally concentrated by immigrants from South and East Asia. As a result, residential displacement in the case of Adelaide promoted higher level of integration between different ethnic groups. Finally, another explanation for the decreased ethnic segregation in Adelaide is by taking into consideration that it is the only city in Australia that managed to achieve a decrease in socioeconomic segregation through efficient local urban policies – see chapter five. In this sense, based on the assumption that certain ethnic groups are also socioeconomically disadvantaged compared to nationals; then the tackling of socioeconomic segregation can lead in the process to the creation of housing opportunities for poor immigrants outside of the enclaves, and eventually contributes to decreased ethnic segregation. Further details about the association of ethnic and socioeconomic segregation in Australian cities are explained later in chapter seven.
6.3.2 Southeast and East Asian cities
As the case of Western European cities, countries of Southeast and East Asia vary greatly in their local situations which in turn affect levels of ethnic segregation within cities of the region. The dataset contains 6 Asian cities; Tokyo, Singapore, Beijing, Shanghai, Hong Kong, and Seoul. In Beijing, Shanghai and Chinese cities in general, the phenomenon of ethnic segregation is negligible, because according to Li and Wu (2008), there are 55 main ethnic minorities in China; however, their quantity is not large enough to constitute social areas. Therefore, socioeconomic status is the major dimension of residential segregation for Chinese cities. Similar situation is detected in Tokyo, where despite the doubling of foreign residents in Tokyo since the mid-1980s, still, foreign residents constitute only 3 percent of total residents of Tokyo (Matsumoto, 2009). Unlike American and European cities, Asian cities in general tend to receive migrants from within the region; the largest three foreign population groups in Tokyo are Chinese, Koreans, and Filipinos, where Chinese are the
188 | Analysis and findings largest group representing only 1% of the residents in Tokyo (ibid.). Moreover, the spatial allocation of foreigners is highly dependent on their socioeconomic status, where Americans concentrate heavily in the central area because they are most likely to be highly paid workers. While, Chinese, Koreans, and Filipinos tend to concentrate in eastern Tokyo because they are most likely to be poor students, small shop owners, and low- paid service and industrial workers (ibid.). As shown in figure (6-3), foreigner’s concentrations patterns have not changed dramatically, as Tokyo scored insignificant increase in ethnic segregation over the past decade (0.7% in the period from 2001 to 2009). The case of Tokyo as alpha city shows that global status of any city and its increasing immigrant’s inflows do not necessarily lead to spatial segregation, rather it is the local situations such as demography or population density that can affect the spatial outcomes within global cities. Another Asian alpha city is Singapore; it shares the same global status with Tokyo. However, unlike Tokyo, Singapore witnessed the highest increase in ethnic segregation of all cities of the dataset. As a leading global city, Singapore seems to fit the model of typical global/divided city characterized in the literature, where spatial segregation is the axiomatic outcome of increased in-flows of guest workers. Therefore, the next section discusses in detail the case of Singapore, in order to determine whether the increase in segregation is mainly due to the city’s global status, contextual particularities of the region, or due to local circumstances intersecting on city level. The individuality of Singapore is derived from two main conditions. First, Singapore is a historically multiracial society since its establishment in 1819, comprised of three main ethnic groups: Chinese, Malay, and Indian (Sin, 2003). Where, Chinese are the largest group with 79% of total population followed by Malay with 14%, Indian with 6%, and the remaining 1% - labeled as ‘others’ - are mainly European and other nationalities (Singapore Department of Statistics, 1996). Second, the state intervention in
Ethnic segregation in global cities | 189 housing sector is exceptionally high. In 1998, 86% of the total population resided in public housing constructed by the Housing and Development Board (HDB) (Sin, 2003). State level of control over housing market in a multiethnic society, along with government interest in achieving socio-spatial integration of ethnic minorities have led to several policy actions and restrictions imposed by the government – through HDB - over the public sector. During the 1960s, large numbers of public housing projects have been developed, offering a wide range of dwelling units for different income groups (Van Grunsven, 2000). In addition, the allocation of households over housing projects aimed to disperse ethnic groups in space (Van Grunsven, 2000; Sin, 2003). As a result, evidence from the 1970s and early 80s showed that the provision and allocation of public housing have contributed to the desegregation of ethnic groups (Van Grunsven, 2000). However, in mid 1980s, voluntary residential mobility caused re-grouping of ethnic minorities (Van Grunsven, 2000; Sin, 2003). To combat minorities’ re-grouping, a system of ethnic quotas was imposed in March 1989, where each neighborhood had to replicate the ethnic mix of Singapore. In other words, the HDB set a “maximum limit on the percentages of Chinese, Malay and Indian/Others living in each neighborhood” (Sin, 2003: 530), and when a neighborhood reached or exceeded quota limits, owners can resell their flats only to buyers of the same ethnic group. For example, in a neighborhood that is mostly inhabited by Malays, Malays can sell to Malays, but a Chinese, Indian, or others living in that neighborhood cannot sell to a potential Malay buyer (ibid.). However, the ethnic quota system did not succeed in re-achieving ethnic integration for two reasons. First, the quota was not comprehensively imposed over all neighborhoods, additionally, “household were not forced to move from or to existing neighborhoods and/or blocks which did not conform to the ethnic limit” (Van Grunsven, 2000: 119). Which implies that households of the same ethnicity could move and cluster in those
190 | Analysis and findings neighborhoods where ethnic quota is not considered. Second, defining, implementing, and following up ethnic quotas of every neighborhood require the government to impose full control over housing market transactions, where full control is ‘extreme intervention’ that the government abstained to adopt. As figures (6-8) and (6-9) show, despite the housing regulations, patterns of clustering of the Indian population have changed visibly from 2000 to 2010. As a result, the increased ethnic segregation showed in figure (6-9) reflects the spatial outcome of residential mobility of different ethnic groups based on their preferences to re-cluster with co- ethnics, under incapacitated housing regulations. The case of Singapore as presented above does not support the assumption that the in-flow of guest workers into the city is the main cause for intensified ethnic segregation. In fact, foreign workers’ regulations compel employers to provide suitable accommodation for their foreign workers by their arrival in Singapore (Cho, 2011). Also, foreign workers – with exception of highly skilled professionals – are legally banned from renting any residential properties on the market. Therefore, it is nearly impossible for low-skilled foreign workers to form their own ethnic enclave. For highly skilled professional, some ethnic groups show a tendency for clustering such as Japanese and French, however, their relatively low proportion to total population made their segregation insignificant (ibid.) especially on the macro level when compared to segregation of Malays or Indians. Comparing ethnic segregation levels of Singapore to those of Hong Kong sheds more light on the individuality of cities as the main factor in explaining spatial transformations within global cities. Both Singapore and Hong Kong are leading alpha cities located in the economically booming Asian region. However, Singapore witnessed increase in ethnic segregation due to its historical ethnic composition and local housing policies. While Hong Kong, witnessed a decrease in ethnic segregation, where its ethnic
Ethnic segregation in global cities | 191
N
5 mi
5 km
Figure 6-8 Singapore in 2000, the distribution of Indian population
N
5 mi
5 km
Figure 6-9 Singapore in 2010, the distribution of Indian population
Figure 6-8 Singapore in 2000, the distribution of Indian population Figure 6-9 Singapore in 2010, the distribution of Indian population
192 | Analysis and findings composition is mainly impacted by handover of Hong Kong to the Chinese authority in 1997, and the consequent changes in entry visa and work permits regulations especially for Mainland Chinese (Ullah, 2012). Where, 95% of total population of Hong Kong is Chinese, most of them are either descendants of migrants from China or were themselves born in China (ibid.). The handover of Hong Kong was followed by the outbreak of the H5N1 avian influenza, and then in 2003, Hong Kong and China suffered the epidemic of severe acute respiratory syndrome (SARS), where “[T]he epidemic not only severely affected the health of the people but also produced many related social, economic, and humanitarian problems including tourism, international travel and trade.” (Lee, 2006:11). Consequently, the in-flows of foreign (non-Chinese) workers to Hong Kong slowed down compared to other alpha cities in the same period of time. As a result, decreased level of ethnic segregation in Hong Kong is caused by the in-flows of Mainland Chinese in a basically Chinese community, along with relative retreat of foreign immigration. In general, the discussion of Asian cities showed that history and local circumstances are the main factors impacting socio-spatial transformations within cities. Alpha cities of Tokyo, Singapore, and Hong Kong proved that patterns of segregation are not related to global status, and the shifts on global levels are eventually defined by local contexts of each city.
6.3.3 North American cities
Despite the fact that it is a major destination of international immigrants, issue of residential segregation in the U.S. has always been orbiting around segregation of African Americans3 (Massey and Denton, 1993), and later the concerns included the segregation of Hispanic population (Cross, 1992). Only in recent decades, more focus was given to the impacts of the enormous international migration received by American cities. Between 1980 and 1989, the U.S. received third of all the world’s
Ethnic segregation in global cities | 193 international migrants (Zlotnik, 1993), where the state of California alone accommodates more permanent immigrants than any other country in the world (Massey et al., 2009). The same migration surge is observed in Canada, where in the 1990s, 16% of Canada’s total population was foreign- born residents (ibid.). According to figure (6-3), despite the increase in immigration flows received by the region, yet only 46.1% of North American cities have witnessed an increase in ethnic segregation levels. This observation matches Cutler et al. (2008) conclusion that changing pattern of immigrants’ segregation is not related to the share of immigrants of total population. Where in the U.S. – for example – the in-flows of immigrants are not distributed evenly among cities; instead, Sunbelt4 cities receive significantly higher percentage of immigrants, especially from Mexico (Thompson, 2011), compared to Rustbelt cities (Frey, 2002). Yet still, 11 out of 14 Sunbelt cities have witnessed a decrease in ethnic segregation levels over the past decade. This inverse correlation between the in-flows of immigrants and levels of ethnic segregation is explained by, on one hand, when the flows of immigrants are large enough to exceed the capacity of the existing ethnic community to accommodate the new comers, immigrants start to search for affordable housing in other neighborhoods, hence, they become dispersed in different areas of the city, which lead to higher ethnic mix and eventually the overall ethnic segregation level will drop accordingly. Note that the dispersal of immigrants in this scenario is similar to the situation in Australian cities; the only difference is that in Australia the dispersal is caused by gentrification and displacement of disadvantaged minorities, while in the U.S. the dispersal is a result of the growing in-flows of immigrants. On the other hand, instead of the dispersal of new comers to the city outside the ethnic enclave, the case of Houston (see figures 6-10 and 6- 11) shows that previous generations - who are in a better socioeconomic status then new comers - tend to move to the suburbs, while t he existing
194 | Analysis and findings
N 10 mi 10 km 10 Houston urbanized area in 2000, Distribution of Mexican, Caribbean, and Central American 10 Houston urbanized area in 2000, Distribution of Mexican, Caribbean, and Central - Figure 6
Figure 6-10 Houston urbanized area in 2000, Distribution of Mexican, Caribbean, and Central American
Ethnic segregation in global cities | 195
N 10 mi 10 km 11 Houston urbanized area in 2009, Distribution of Mexican, Caribbean, and Central American Houston urbanized area in 2009, Distribution of Mexican, Caribbean, and Central 11 - Figure 6
Figure 6-11 Houston urbanized area in 2009, Distribution of Mexican, Caribbean, and Central American
196 | Analysis and findings enclaves are sustained by new comers. Either way, ethnic segregation level decreases due to the presence of large portion of an ethnic group outside their enclave, yet in the first case, dispersal of poor immigrants can lead to the decrease in socioeconomic segregation as well, while in the second case, well-off settlers moving to the suburb can lead to the exacerbation of socioeconomic segregation, as they leave behind a concentration of disadvantaged immigrants isolated in poor neighborhoods (McGeary and Lynn, 1988). The next chapter discusses in more detail how ethnic and socioeconomic segregation are important contingencies in the understanding of one another. These explanations for the decreased ethnic segregation in Sunbelt cities clarify how cities like Miami, Phoenix, and San Francisco, where the foreign born population have increased by 15.2%, 13.2%, and 10.4% respectively 1990 to 2000 (Frey, 2002). Yet at the same time, all three cities scored significant decrease in ethnic segregation by 12.0%, 5.9%, and 8.7% respectively. While on the contrary, Cincinnati – which is a Rustbelt city - scored the highest degree of increase in ethnic segregation, although the growth of foreign-born population in Cincinnati was only 1.1% from 1990 to 2000, which strongly suggests that increased ethnic segregation is not necessarily a result of the growing in-flows of immigrants into global cities. Instead, changes in ethnic segregation patterns within cities depend highly on other factors such as characteristics of each ethnic group (Logan, 2000), their background, culture, education, and socioeconomic status. Another observation drawn from figure (6-3) suggests that there are inconsistent pattern of change in ethnic segregation in a number of individual cities. One example for such inconsistency is the cases of Austin, Houston, and Dallas; all three cities are Sunbelt cities and sharing similar characteristics of the region, firstly, they are all located in Texas. Secondly, they witnessed parallel growth in the share of immigrant workers especially from Mexico (Thompson, 2011), where the foreign born population have increased in Austin, Houston, and Dallas by 11.4%, 11.5% and 11.7%
Ethnic segregation in global cities | 197 respectively from 1990 to 2000 (Frey, 2002). And finally, the region is known for its thriving service industries, as Texas is one of the top 5 states with growth of service and high-tech industries in 2005 (The State of Texas, 2011). However, all those shared characteristics did not have equal effect on the outcomes of spatial transformation of the three cities. According to figure (6-3), Austin scored an increase in ethnic segregation by 4.8% from 2000 to 2009. While on the contrary, both Houston and Dallas scored a decrease in ethnic segregation by 9.0% and 8.5% respectively over the same period of time. The reason behind this unlikely increase in ethnic segregation in Austin is a very particular characteristic of the city, which is the interstate highway (I-35). The highway is built in the 1950s (Skop, 2009) and it is passing through the city while geographically isolating the historically persistent ‘East Austin enclave’ that is basically a Mexican community. In this case the interstate highway is acting like a physical barrier that prevents the enclave from dissolving into the city.
6.3.4 Western European cities
During the post World War II period, Western Europe witnessed sizable flows of guest workers pouring in cities of the region. Initially, guest workers were recruited from Southern European countries such as Spain and Italy, then later, workers were recruited from other countries such as Morocco, Turkey, Algeria, India, …etc. (Van Kempen, 2005). Then, with a large number of temporary guest workers turning into settlers in cities of destination – along with opening the door for family reunion – the result was the formation of ethnic communities (Castles, 2002), which became a common attribute of Western European cities The rise, then the persistence, of ethnic enclaves in Europe – as shown in chapter 3 - are perceived as an outcome of two parallel processes. First, self-segregation of minorities who choose willingly to live with co-ethnics for reasons of safety, enhancing sense of belonging, and maintaining customs and traditional values of their particular culture (Fischer, 1976).
198 | Analysis and findings
Second, in case that immigrants’ preferences are not leading to self- segregation, their low socioeconomic status comparing to nationals (Van Kempen, 2005) eventually limits their housing choices outside the enclave. Yet apparently, despite these mechanisms of voluntary and involuntary segregation of ethnic minorities, according to figure (6-3), levels of ethnic segregation have fallen in 66% of Western European cities of the dataset. Furthermore, a closer look on the situation in individual Western European cities reveals that changes in level of ethnic segregation occur for different reasons that vary greatly from country to country or even within the same country. Undoubtedly, European cities share distinctive characteristics, which differentiate them from other cities of the developed world such as North American cities (Kazepov, 2005). Yet, within Europe, dissimilarities exist on different territorial levels from states to cities (ibid.). In the process of explaining ethnic segregation, variations in the housing supply, policies, and practices on the local level can have different impacts on changes of ethnic segregation in individual cities. The role of housing policies in particular gained a special focus in segregation studies. On one hand, Van Kempen (2005) explained extensively how ethnic minorities’ choices in European cities could be affected greatly from city to city according to the type, quality, and affordability of housing supply on each market. He stated, “a lack of social rented housing in Belgium largely explains the concentrations of Turks in the city of Brussels, and the location and accessibility of social housing in the Netherlands, could be the principle explanatory factor in Dutch cities. Discriminatory regulations have a very important role in Austria, while patterns of ‘choice’ might be more important for different groups in the UK.” (Van Kempen, 2005:205). Therefore, cross-nation variations in housing policies on centralized national level or on local governments level will highly influence spatial transformation within cities. Results from figure (6-3) partially support this conclusion, on the national level, countries of France, Switzerland, and Italy
Ethnic segregation in global cities | 199 show consistent nation-based changes on their cities level, French cities of Marseille, Paris, and Lyon show parallel increase in level of ethnic segregation, while Italian cities of Rome and Milan as well as Swiss cities of Zurich and Geneva show significant decreased in ethnic segregation. On the other hand, German, Dutch, and Spanish cities show noticeable divergence in results between cities within each nation, as the cases of Frankfurt vs. Stuttgart, Amsterdam vs. Rotterdam, and Valencia vs. Barcelona. Those cases either reflect the influential role of local governance on sub-national level. Or as Musterd and Fullaondo concluded, that “more social housing and stronger welfare state neither relate to lower, nor to higher levels of segregation compared to contexts where the welfare state and the housing market are clearly different.” (Musterd and Fullaondo, 2008: 112). Their conclusion is based on comparative analytical study of ethnic segregation and the housing market in Amsterdam and Barcelona. Where the large social housing stock in Amsterdam and the relatively strong welfare state of Dutch cities are encountered by large number of migrants, causing segregation levels to be similar to that of Barcelona. Despite the fact that housing stock in Barcelona is much more private, and the welfare state is weaker compared to the Dutch, but the lower share of migrants in Barcelona – compared to Amsterdam - neutralizes the impacts of absent public housing - and the role of welfare state in general - on the process of ethnic segregation (ibid.). Resulting similar levels of segregation in two totally different situations, which imply that ethnic segregation is a complex process and it is not solely related to housing policies and issues of affordability or choice. In fact, the seemingly opposing perspectives of Van Kempen (2005) and Musterd and Fullaondo (2008) regarding the relevance of housing policies to ethnic segregation are supported by individual cases shown in figure (6-3). On one hand, the increased segregation in Amsterdam is coinciding with the retreating role of the state and the consequent privatization of socially rented housing in recent years. According to
200 | Analysis and findings
Ministry of VROM5 (2003), socially rented housing units in 2001 comprised 57% of total housing units in Amsterdam; this percentage has dropped to reach 47.4% in 2011 (Ministry of BZK6, 2012). Which backs up Van Kempen assessment of the important role of the state policies. On the other hand, Berlin has witnessed a similar and even more dramatic reduction of public housing stock than Amsterdam (from 30% of total housing units in 1990 to only 15% in 2008), along with increasing rent prices in privatized housing stock (Aalbers and Holm, 2008). Given those two conditions, the expected spatial outcome for Berlin would be increased segregation as Amsterdam. However, Berlin showed insignificant change in ethnic segregation over the years, which implies – as Musterd and Fullaondo (2008) suggested - that factors involved in producing and sustaining ethnic segregation cannot be reduced to housing policies or welfare weaknesses. That discussion of ethnic segregation in Western European cities leads to the following conclusions. First, tracing patterns on regional level are proved to be as difficult as on the global level, despite the similar characteristics shared by nations and cities of the region. Second, spatial outcomes of each city is distinctive, where individuality of cities is derived from their local situations that vary greatly in terms of housing supply on one side, and size, background, and socioeconomic status of minorities on the other side.
Ethnic segregation in global cities | 201
Summary
Spatial segregation of ethnic minorities is a complex phenomenon, the cases discussed in this chapter show that changes in ethnic segregation are highly dependent on wide range of variables, including history, housing policies, and socio-economic status of the segregated minority. Generalized patterns of spatial transformations on the global level are not detected; the 81 cities of the dataset have showed diverse directions of change in segregation level, where only 44.4% of global cities have witnessed an increase in ethnic segregation, which challenges the validity of the generalized assumptions in the literature. This conclusion is also supported by a regression analysis, which confirmed that cities’ global network connectivity explains only 0.2% of the variability in spatial segregation index. Moreover, the one-way analysis of variance ANOVA revealed that the cities’ position in the global network represented in alpha, beta, and gamma categorization are not associated with certain changes in socioeconomic segregation, while the location of global cities in certain geographic region is proved influential on the resultant levels of segregation. The regional comparisons of cities of Australia, North America, Western Europe, and Southeast Asia have showed that local situations on the level of cities vary greatly within the same region and sometime within the same nation-state (as the case of Texan cities or Amsterdam and Rotterdam). The impacts of global economic and political shifts on the spatial transformations within cities are not uniform. For example, the retreating role of the welfare state does not have the same effect on segregation levels in cities such as Singapore when compared to Amsterdam. Also, the in- flows of immigrant workers do not necessarily cause increase in ethnic segregation as the case of Californian cities. Recapitulating, the discussion over the situations in individual cities revealed a number of factors contributing to changes in ethnic segregation levels in global cities; firstly, gentrification and its associated displacement
202 | Analysis and findings and reconcentration (or dispersal) of disadvantaged minorities (as the case in Australian cities). Secondly, the growing in-flows of immigrants to the city, which can lead to the decreased ethnic segregation due to the dispersal of ethnic groups outside the enclave (as the case in American cities). Thirdly, the efficiency of urban policies and the characteristics of public housing provided by the state (in terms of distribution, access, …etc.) (As in the case of Singapore and Western European cities), can have varying effects on the resultant ethnic segregation patterns. Fourthly, a specific characteristic of the city – such as a physical barrier in Austin or the handover of Hong Kong - that can contribute to either higher or lower levels of ethnic segregation. Finally, since changes in ethnic segregation is tied to residential mobility, then segregation levels are related to the financial ability of households to move out/in the ethnic enclaves. The next chapter discusses the spatial outcome resulting from the intersection of patterns of ethnic segregation with patterns of socioeconomic segregation within the same city.
Ethnic segregation in global cities | 203
Notes
1 The variations are based on the official census methodology of each country 2 Note that alpha global cities of Mumbai (India), Johannesburg (South Africa), Jakarta (Indonesia), and beta global cities of Cairo (Egypt), Bogota (Colombia), Bucharest (Romania), and Karachi (Pakistan) are all receiving rural immigrants from within their respective regions (immigrants with the same ethnicity and therefore are irrelevant to the analysis), while at the same time, the regions are sending immigration to the usual destinations of North America, Western Europe, and Australia, and any amount of international in-migration to those global cities is outweighed by the enormous outer-migration 3 Early European migration to the U.S was also a concern for research. However, European enclaves of Irish, Germans, then later for Poles, Italians, and Czechs were perceived – for Chicago school - as a transitional phenomenon after which assimilation is inevitable (Park, 1926; Wirth, 1928; Burgess and Newcomb, 1933). 4 Sunbelt cities are located in the Southern states, while Rustbelt cities are the deindustrialized cities of the Northeastern and the East North Central States. 5 The Ministry of Housing, Spatial Planning and the Environment (Dutch: Ministerie van Volkshuisvesting, Ruimtelijke Ordening en Milieu or VROM. It was merged with the Ministry of Transport, Public Works and Water Management into the new Ministry of Infrastructure and the Environment on 14 October 2010. 6 The Ministry of the Interior and Kingdom Relations (Dutch: Ministerie van Binnenlandse Zaken en Koninkrijksrelaties; BZK) is the Dutch Ministry responsible for Home Affairs, Civil service, Intelligence and the relations with the other countries in the Kingdom of the Netherlands.
204 | Analysis and findings
Chapter 7
7 The color of poverty: the correlation between socioeconomic and ethnic segregation
The aim of this chapter is to further discuss the influential role of local contingencies on the changes in spatial segregation levels in individual global cities. The chapter follows up on the conclusions drawn from the findings of the analysis presented in chapters five and six, with a particular focus on the potential correlation between the socioeconomic status of a certain population group and their ethnic background. In other words, a number of cases presented in the previous chapters revealed that ethnic segregation patterns are changing due to the change in the socioeconomic status of the segregated ethnic groups (as in the case of Houston, where the well-off Mexican group could afford to move to the suburbs and leave the ethnic enclave). This observation matches what Timberlake et al. (2012: 74) found in their study of income polarization in 57 US cities; they concluded that cities with higher immigration in-flows are more likely to show an increase in income polarization due to the agglomeration of global economic activities in these cities compared to cities with less in-flows of immigrants. In this case, socioeconomic segregation is itself a contingency that affects patterns of ethnic segregation, and vice versa, where socioeconomic segregation patterns might be determined by patterns of clustering or regrouping of a certain ethnic group (as the case of Singapore, where the
206 | Analysis and findings voluntary regrouping of Indian population – who are mostly less affluent than the other groups - contributed to the concentration of poverty). In this case, pattern of ethnic segregation in individual city can be considered an important contingency that affects the changes in socioeconomic segregation. Investigating this correlation between socioeconomic and ethnic segregation is particularly relevant to the purpose of this research. In more detail, the three school of thoughts discussed in chapter two acknowledge the influence of macro economic changes on the micro socio-spatial settings of cities, but in case that the resultant socio-spatial change is caused (or at least modified) by other local contingent factors. Then, it can be argued that the acknowledged role of economic globalization as a main generator of socio-spatial division is overstretched. Accordingly, the subsequent sections investigate the mutual influence of socioeconomic and ethnic segregation on one another; the analysis findings are not conclusive. However, it raises several questions on the generalization tendencies in the global city literature.
7.1 Intersection of socioeconomic and ethnic segregation: a discussion
“Ethnic and social spatial inequalities cannot be understood in a one-dimensional way. Ethnic segregation has socio-economic components” (Musterd, 2005: 339)
According to Burgers and Musterd’s (2002) study of Amsterdam and Rotterdam, the ‘subcultural’ variable emerged as an important factor that affects the socioeconomic status of Turks and Moroccans, as well as their spatial distribution in both cities. Similarly, Wilson’s mismatch theory (1987) paid a particular attention to the racial/socioeconomic segregation of African Americans, as the residential segregation of African Americans is
The color of poverty | 207 not only historically persistent, but it is also institutionalized, and connected to other racial discriminatory processes; such as, limited opportunities for African Americans in the labor market, their chances of getting decent education, exclusion from the housing market, and their exposure to higher poverty rates. Accordingly, a wide range of segregation studies – especially on U.S urban areas - is dedicated to explore the connection between race/ethnicity on one hand, and poverty/socioeconomic status on the other hand. Massey and Denton (1993) explained in detail that the association between race and poverty is a key issue in explaining the dynamics of residential segregation in urban areas of the U.S. In fact, both Wilson (1987) and Massey (1990) dealt with race/poverty connection from different angles; they used the terms “minority underclass” and “black poverty” respectively, both terms are combined of two sections, the first refers to race/ethnicity status (minority – black), and the second refers to the socioeconomic status (underclass – poverty) of the segregated population, linking a certain race to a certain social and economic conditions. However, Wilson and Massey disagreed on the degree of importance of race in this equation. In Wilson’s view (1987), race alone cannot explain the sharp increase in inner city poverty in the 1970s, especially when considering the macro economic changes impacting western societies (such as deindustrialization). While, Massey (1990) and Massey and Denton (1993) insist that in a racially segregated environment, any increase of poverty will automatically lead to geographic concentration of poverty, because the higher poverty rates are absorbed by few number of racially segregated neighborhoods. In order to explain how racial segregation contributes to the concentration of poverty, Massey (1990) constructed a hypothetical city of 128,000 people, distributed evenly among 16 equal-sized neighborhoods, 8000 persons each. Massey tried to distribute the population race groups based on the real situation in many American cities in the 1970s. Therefore, the distribution of the population in the hypothetical city was determined
208 | Analysis and findings based on a previous study by Massey and Eggers (1990), in which they examined the poverty rates for different races in 60 metropolitan areas in the U.S. In average, African Americans were 25% of the population, with poverty rate of 20%, compared to 75% whites, with poverty rate of 10%. Massey started by distributing blacks and whites by those ratios evenly over the 16 neighborhood. Accordingly, average poverty rates were equal for all neighborhoods and there is no racial segregation. Then to create a racially segregated environment, he redistributed black population to be concentrated in 12 then 8 then 4 neighborhoods of the 16. Each time he raise the level of black segregation, he finds that poverty rates among black neighborhoods steadily rise while poverty rates among white neighborhoods steadily fall. Because in average, blacks tend to have higher poverty rates then whites, and their concentration in few neighborhoods lead to the geographic concentration of poverty (Massey, 1990). Since racial segregation and poverty concentration are theoretically contributing to the persistence of one another, then it is logical to assume that any external factor (macro changes) affecting either race or class composition of the city will automatically affect the other. For instance, if deindustrialization and privatization are causing higher levels of unemployment, welfare dependence, and eventually increasing poverty (Marcuse and Van Kempen, 2000; Mingione, 2005), and under the condition that the city is already highly segregated by race, then – hypothetically – increased poverty will be concentrated and the city will show higher levels of socioeconomic segregation over time – note that poverty concentration in this case is partially caused by the existing racial segregation and not only due to the rising levels of unemployment or welfare dependence . Similarly, increased poverty will contribute to the persistence of the existing racial segregation. Massey and Denton (1993) support this hypothesis by describing the situation of Blacks during the economic hardship of the 1930s and 1970s; “During the 1930s, […], the Great Depression brought a wave of factory closings, bankruptcies, bank failures, and very high rates of
The color of poverty | 209 unemployment in the black community. During the 1970s, successive recessions, bursts of inflation, and increased foreign competition eliminated many high-paying jobs in manufacturing, lowered wages, and decreased the real value of welfare payments. These dislocations took a heavy toll on the distribution of black income, especially among families in the industrial cities of the northeast and midwest. As a result of the downward shift in black incomes, poverty rates increased substantially in both decades.”(Massey and Denton, 1993:125). Moreover, as noted in chapter one, the high poverty rates among the segregated ethnic minority create a vicious cycle of social exclusion and spatial segregation, which causes spatial segregation patterns to persist. As, poverty creates very difficult conditions for Blacks or other segregated minorities to escape the segregated neighborhoods (Massey et al., 1994). Because poverty is associated with other social problems such as violence, crime, unwed child bearing, divorces, single parenting, low educational achievement, poor health, drug use ... etc. (Wilson 1987; Soja, 1989; Massey and Denton, 1993; Andersen, 2002; Wassmer, 2005; Varady, 2005). These problems affect the ability of the segregated population to find well-paying jobs, decent education, and better housing, which consequently limit their opportunities for social upward mobility (Wilson, 1987; Kazepov, 2005; Boal, 2005; Massey and Denton, 1993; Madanipour et al., 1998). Eventually, the socioeconomic status of the segregated minority is the key factor contributing to the persistence of the ethnic enclave and vice versa.
7.2 Socioeconomic and ethnic segregation: parallel or divergent changes?
In total, out of the 81 cities with available data on the spatial distribution of different ethnic groups, only 56 cities have comparable data about the distribution of socioeconomic groups. The 56 cities are categorized into 17
210 | Analysis and findings
New York Hong Kong
Singapour Chicago Sydney Toronto Los Angeles Amsterdam Brussels
San Fransisco alpha Washington Miami Melbourne Boston Dallas Atlanta Philadelphia Stockholm Montreal Houston Berlin Copenhagen Vancouver Seattle Auckland Oslo
Minneapolis beta Brisbane Detroit Denver St Louis San Diego Perth Cleveland Calgary Cincinnati Charlotte Baltimore Adelaide Portland San Jose us
Kansas city Phoenix Rotterdam Tampa Columbus
indianapolis gamma Pittsburgh Edmonton Orlando Gothenburg Ottawa Richmond Austin Milwakee Wellington Figure 7-1 alpha – beta – gamma cities and their yearly percentage of change in both socioeconomic and ethnic segregation index segregation ethnic and socioeconomic both in change of percentage yearly their and cities gamma – beta – alpha 7-1 Figure
Figure 7-1 alpha – beta – gamma cities and their yearly percentage of change in both socioeconomic and ethnic segregation index
The color of poverty | 211 alpha, 18 beta, and 21 gamma cities. Figure (7-1) presents the changes in both types of segregation in cities of each global level. For a better understanding of results in figure (7-1), figure (7-2) shows, cities categorized into four types based on the possible directions of change in both socioeconomic and ethnic segregation:
1. Cities with increase in both types of segregation 2. Cities with increased socioeconomic segregation and decreased ethnic segregation 3. Cities with decreased socioeconomic segregation and increased ethnic segregation 4. Cities with decrease in both types of segregation
The aim from this categorization is to detect if there is a tendency among cities to show parallel changes in both socioeconomic and ethnic segregation levels. According to figure (7-2) 19 cities follow the first case (increase in both types), 10 cities follow the second case (increased socioeconomic with decreased ethnic segregation), 7 cities follow the third case (decreased socioeconomic with increased ethnic segregation), and 20 cities follow the fourth case (decrease in both types). In total, the results show that 39 out of 56 cities (almost 70 percent) showed a parallel increase (or decrease) in both socioeconomic and ethnic segregation, compared to only 17 cities showed contrasting directions of change between both types of segregation. As noted, the parallel changes in both types of segregation are explained theoretically based on the credible assumption that immigrants are more likely to be in a lower socioeconomic status than nationals (Van Kempen, 2000; Massey and Denton, 1993). Consequently, any factor that contributes to more concentration (or dispersion) of ethnic minorities will also lead to socioeconomic segregation (or integration). While, the less prevalent outcome of divergent changes in both types of segregation is explained by the particular situations in individual cities. For example, the
212 | Analysis and findings
Edmonton Brisbane Philadelphia Perth
Cleveland segregation Kansas city Tampa Pittsburgh Milwakee Cincinnati Portland (+,+) Richmond Detroit Auckland Seattle Stockholm Sydney Denver Amsterdam Dallas Baltimore San Jose us New York Boston
Orlando (+,-) Ottawa Melbourne Oslo Gothenburg Austin Toronto Calgary Singapour Columbus (-,+) Wellington Vancouver Brussels Rotterdam Berlin San Fransisco Copenhagen
Atlanta Adelaide Montreal Miami Chicago
Phoenix (-,-) San Diego Washington St Louis Los Angeles Minneapolis indianapolis Hong Kong Charlotte Houston Figure 7-2 cities of the dataset categorized into four groups of possible directions of change in both socioeconomic and ethnic and socioeconomic both in of change directions of possible groups four into categorized dataset the of 7-2 cities Figure
Figure 7-2 cities of the dataset categorized into four groups of possible directions of change in both socioeconomic and ethnic segregation
The color of poverty | 213
A matrix shows cities of the dataset classified according to their global status and their direction of change in both socioeconomic and ethnic segregation: alpha beta gamma (+) Increased Sydney Stockholm Cincinnati socioeconomic Amsterdam Seattle Portland segregation Philadelphia Auckland Kansas City Brisbane Tampa (+) Increased Detroit Pittsburg ethnic segregation Denver Edmonton Perth Richmond Cleveland Milwaukee (-) Decreased Hong Kong Montreal Charlotte socioeconomic Chicago Houston Adelaide of segregation segregation Los Angeles Berlin Phoenix Brussels Copenhagen Rotterdam Parallel changes in both types (-) Decreased San Francisco Minneapolis Indianapolis ethnic segregation Washington St. Louis Miami San Diego Atlanta (+) Increased New York Oslo Baltimore socioeconomic Melbourne San Jose segregation Boston Orlando Dallas Gothenburg
(-) Decreased Ottawa ethnic segregation (-) Decreased Singapore Vancouver Austin
segregation socioeconomic Toronto Calgary Wellington segregation Columbus
(+) Increased Divergent changes in both types of ethnic segregation displacement of minorities from central areas - as a result of gentrification – to adjacent neighborhood dominated by another minority group, causing increase in socioeconomic segregation while at the same time higher level of ethnic integration (see the case of Adelaide). Another aspect affects spatial changes of socioeconomic and ethnic segregation is self-segregation, where
214 | Analysis and findings affluent minorities prefer to live with co-ethnics regardless their socioeconomic status, causing socioeconomic segregation to decrease while maintaining the ethnic divsion (see Varady, 2005). The discussion in chapters two and three suggests that both socioeconomic and ethnic segregation are expected to intensify in global cities as a result of the macro economic changes impacting these cities. However, according to table (7-1) - which presents a matrix of cities of the dataset classified according to their global status and their direction of change in both socioeconomic and ethnic segregation - only 33.9% of global cities of the dataset followed the global/divided city model and showed parallel increase in both socioeconomic and ethnic segregation. While, 35.7% of cities showed a parallel decrease in both types of segregation. Another remark drawn from table 7-1 is that – statistically speaking - alpha cities are not necessarily more prone to increase in spatial segregation, where only 3 out of 17 alpha cities (17.6%) showed parallel increase in both types of segregation. Instead, alpha cities are more likely to have a parallel decrease in both types compared to beta and gamma cities.
7.2.1 Examining the strength of association between socioeconomic and ethnic segregation
In terms of direction of change, the observations in figure (7-2) and table (7-1) suggest that the change in one type of segregation tends to follow the change in the other type, and a chi-squared test already confirmed this observation. While, in terms of the intensity of change, a linear regression analysis is performed to assess the strength of the association between the changes in socioeconomic and ethnic segregation in global cities. Figure (7- 3) presents a summary scatterplot for the correlation between the two variables. Also, The SPSS generates several tables for the linear regression analysis. In this section, we show the three main tables required to
The color of poverty | 215 understand the results of the of the linear regression procedure: the model summary table, the ANOVA table, and the coefficients table.
6
5
4
3
2
1
0 -4 -3 -2 -1 0 1 2 3 -1 Standardized change in ethnic segregation y = 0.2332x - 0.0633 -2
-3 Standardized change in socioeconomic segregation
Figure 7-3 Summary scatterplot for the correlation between the standardized change in socioeconomic segregation and ethnic segregation in global cities
Firstly, the regression output table (7-2) provides the information needed to determine how well the regression model fits the data. In the model summary section, the column ‘R’ is the absolute value of the Pearson correlation coefficient between the dependent and the independent variables. It simply indicates the strength of the association between the two variables. In our case, R = 0.226, and since the R value ranges from 0 to 1, then the observations indicate low to moderate correlation between the changes in socioeconomic and ethnic segregation in global cities. The R2 value in the R2 column represents the proportion of variance in the dependent variable
216 | Analysis and findings that can be explained by the independent variable. In our case, R2 = 0.051, which means that the independent variable, change in socioeconomic segregation, explains 5.1% of the variability of the dependent variable, change in ethnic segregation. Please note that the R2 is calculated for the sample (56 global cities), SPSS generates another value called adjusted R2 that represents the proportion of variance if the analysis is performed over the entire population (i.e. the full list of global cities). Table (7-2) shows that the adjusted R2 = 0.033.
Table 7-2 Regression output – model summary, ANOVA, and Coefficient tables Model Summary Adjusted Std. Error of the Model R R Square R Square Estimate 1 .226a .051 .033 .75379511236 a. Predictors: (Constant), zscoreINCOME
ANOVAa Sum of Mean Model Squares df Square F Sig. 1 Regression 1.650 1 1.650 2.904 .094b Residual 30.683 54 .568 Total 32.334 55 a. Dependent Variable: zscoreETHNICITY b. Predictors: (Constant), zscoreINCOME
Coefficientsa Unstandardized Standardized Coefficients Coefficients Model B Std. Error Beta t Sig. 1 (Constant) .239 .101 2.362 .022 zscoreINCOME .188 .110 .226 1.704 .094 a. Dependent Variable: zscoreETHNICITY
The color of poverty | 217
The next table is the ANOVA table, which reports how well the regression equation fits the data (i.e., how accurately the model can predicts the dependent variable – ethnic segregation - based on the changes in the independent variable socioeconomic segregation). The number in the "Regression" row and the "Sig." column indicates the statistical significance of the regression model that was run. Here, the regression model is statistically significant, F (1, 54) = 2.904, p = 0.094, which is greater than 0.05, and indicates that, overall, the regression model cannot statistically significantly predict the dependent variable. The last table is the coefficients table, which provides the coefficients of the regression equation that is used to predict the dependent variable from the independent variable. y = b0 + (b1 x) where y is the change in ethnic segregation, x is change in socioeconomic segregation, b0 is the intercept and b1 is the coefficient, both can be found under the B column in the table, b0 = 0.239, and b1 = 0.188. In our case, it is already established that the model poorly fits the data due to the weak dependence between the variables. Recapitulating, the linear regression established that the change in both types of segregation is moderately correlated, as the change in socioeconomic segregation explains 5.1% of the variability in ethnic segregation. Also, the regression model cannot statistically significantly predict the change in one type of segregation based on the change in the other. A comparison of these observations with the findings of the previous chapters reveals that change in any type of segregation is weakly correlated to the position of the city in the global network, while moderately correlated to the change in the other type of segregation within the same city, as only 0.2% of socioeconomic segregation and 0.2% of ethnic segregation are explained by cities’ global network connectivity, while this percentage jumps to 5.1% when segregation types are explained by one another. In other words, a considerable part of the increased (or decreased) spatial
218 | Analysis and findings segregation in global cities is a result of the long-standing socioeconomic and ethnic structure, rather than the cities’ level of integration in the global economy. Accordingly, a closer look on the local situation of individual cities is required, especially for cities showing divergent changes in both types of segregation, and cities excluded from the analysis for being outliers.
7.3 Regional and local patterns of association between socioeconomic and ethnic segregation
As clarified in chapter four, the sample is regionally unbalanced due to data availability limitations. Therefore, this section focuses on patterns of spatial changes in three regions: Australia and New Zealand by 7 cities, North America by 39 cities, and Western Europe by 8 cities, along with only 2 cities representing Southeast Asia that are discussed individually in the end of this chapter. Figure (7-4) shows that the increase in both types of segregation is most likely to occur in Australia and New Zealand followed by North America then Western Europe. Where, 4 out of 7 cities (57.1%) of all Australia and New Zealand cities in the dataset show parallel increase in socioeconomic and ethnic segregation, compared to 33.3% of North American cities, and 25% of Western European cities.
7.3.1 Australian cities
Healy and Birrell (2003) and Gwyther (2005) claim that ethnic and socioeconomic segregation in Australian cities are reinforcing one another, where Australian-born and English-speaking residents move to gentrified areas, while residents from non-English-speaking origins are concentrated in low-income neighborhoods. The strong and clear association between low socioeconomic status and immigrants resulted in simultaneous increase in both socioeconomic and ethnic segregation in almost 60% of Australian global cities. Also, the same logic explains the parallel decrease in both types of segregation in the case of Adelaide, where despite the inner-city
The color of poverty | 219
6
5
4
3
2
1
.2
.3
.4
.5 !!' " )'( # ((! "% $'($# (&$ ( (!#( !&- #*& ((+ $ ( $) ' &!#$ $&$#($ $# , $)'($# $&(!# $#(&! &!$(( !+ # $ !*!# $!)")' +$& !( "$& # ##( "$# "$#($# ((')& #$)*& # $')' #'' (- $'#!' ##%$! ' ' #($# # #%$! ' !!% #&#' '$ /70/70 /.0/.0 /70/.0 /.0/70 a - North America 6 6 6 5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
.2 .2 .2
.3 .3 .3
.4 .4 .4
.5 .5 .5 '!$ &( &! # -#- ! &)''!' & '# ) !# #%$)& ($ $!" !$) $# $# $((&" !! #($# "'(&" $(#)& $%## /.0/70/.0/.0 /70/70 /.0/.0 /70/.0 /70/70 /.0/.0/70/.0/.0/70 b – Western Europe c - Oceania d - Asia
Figure 7-4 Global cities of the dataset categorized according to broad regions and the four modes of income/ethnicity association
220 | Analysis and findings gentrification, the urban regeneration projects promoted by the local government contributed to the mitigation of income inequalities (Forster, 2006; Arthurson, 2012) socially as well as spatially, and consequently reduced the associated ethnic segregation. However, as argued earlier, the predictable spatial outcome of that simple and abstract correlation between both types of segregation can be altered in the presence of other contingent factors involved in the complex process of spatial changes. As in the case of Melbourne, it is a typical Australian city, but there are two distinct characteristics caused ethnic segregation to drop despite of the increased socioeconomic segregation. Firstly, Melbourne is marked by intense gentrification that surpassed any other Australian city including Sydney (Weller and Van Hulten, 2012). Consequently, poor ethnic groups are displaced from inner-city neighborhoods (see chapter five) to other poor neighborhoods that are ethnically mixed or dominated by another ethnic group. Secondly, the rapid growth of transnational students in Melbourne has shifted neighborhood housing and services structure (Fincher and Shaw, 2009), since students are attracted to affordable and relatively small apartments in accessible locations (Forster, 2006) regardless their ethnic background or preferences. Which contributes to the decreased ethnic segregation.
7.3.2 North American cities
In the United States, as suggested by Massey (1990) and Massey and Denton (1993), racial segregation and poverty concentration are historically interrelated, which is clear in the popular term of ‘black poverty’. However, segregation of foreign-born immigrants and their socioeconomic status is less acknowledged in the literature, to the extent that several studies classify foreign-born population according to color lines instead of national or regional origins (i.e. Iceland and Scopilliti, 2008). Figure (7-4) shows that only 12 out of 33 (36.4%) American cities showed increase in both types of segregation, all of them are located in the north except for Tampa.
The color of poverty | 221
Despite the fact that Rustbelt cities receive significantly lower percentage of immigrants compared to Sunbelt cities (Frey, 2002), increased ethnic segregation is more common in the north where socioeconomic segregation prevails (see chapter five). Iceland and Scopilliti (2008) explain the segregation patterns of ethnic groups by their median income, where the higher the ratio of median income of minorities to that of whites, the lower the level of income inequality, and accordingly the lower the level of ethnic segregation and vice versa. For that matter, income inequalities in Rustbelt cities contribute to a parallel increase in ethnic segregation in those cities. Noticeably, other Rustbelt cities such as Boston and Baltimore did not follow Iceland and Scopilliti (2008) explanation of parallel increase in both types of segregation. Figure (7-4) shows that only 8 American cities witnessed divergent directions of change in both types of segregation. Clearly, since divergent changes are outcome of individual situations of housing and social structure on the local level, then no unified rationalization can explain spatial changes in those 8 cities at once, especially that cities do not only show variation in directions of spatial change toward more segregated or integrated urban structure, but they also show considerable variations in the intensity of spatial change. For example, the Texan cities of Austin and Dallas, both share Sunbelt cities’ qualities. However, the decreased socioeconomic segregation in Austin is accompanied by sizable increase in ethnic segregation. While, Dallas showed insignificant change in socioeconomic segregation accompanied by decrease in ethnic segregation. In fact, Austin is the only 1 American city where socioeconomic integration failed to promote parallel ethnic integration, which is uncommon in Sunbelt cities. As noted earlier, Austin’s distinct spatial outcome is due to the historically persistent segregation of the ‘East Austin enclave’, which is physically isolated by the interstate highway (I-35) built in the 1950s (Skop, 2009), the highway is passing through the city and limits the chance of the enclave to dissolve into the city. Furthermore, despite the fact that the
222 | Analysis and findings enclave receives large numbers of poor immigrants every year (especially Mexican), socioeconomic segregation did not increase accordingly, because the fast growing Mexican community has developed its own resources and networks, which support business of immigrants entrepreneurs to flourish (ibid.) and consequently lower levels of socioeconomic segregation. Similarly, spatial changes in other US cities can only be rationalized individually. In Canada, observations from table (7-4) indicate that 4 out of 6 Canadian cities show divergent directions of change in both types of segregation. This result opposes Fong and Shibuya’s (2000) claims discussed in chapter five, as they explain the reduced socioeconomic segregation in Canadian cities as a result of governmental actions that promote ethnic integration, and eventually reduce the associated socioeconomic segregation. In fact, Montreal is the only city that matches Fong and Shibuya’s assumption where the two types of segregation decreased simultaneously. While divergent segregation patterns in Toronto, Vancouver, Ottawa, and Calgary cannot be explained accordingly. The two neighboring and largest cities in Ontario, Toronto and Ottawa show the same paradox of Austin and Dallas. However, increased ethnic segregation in Toronto – compared to Ottawa – has always been perceived as a direct outcome of the exceptional growth rate of immigrants in Toronto. According to Statistics Canada (2006), Toronto greater area alone hosts 40% of all immigrants entered Canada from 2001 to 2006, compared to only 3% of immigrants who settled in Ottawa greater area. In addition to the elevated growth rate of immigrants in Toronto, other factors - such as median income of immigrants and their residential preferences - have shaped the complex segregation patterns in the city. Haan (2005) has traced homeownership preferences of 12 ethnic groups in Toronto; his analytical study concluded that Chinese and Italians have above average tendency to self-segregate through buying their homes near or inside their ethnic enclave. Which explains the decrease in socioeconomic
The color of poverty | 223 segregation associated with increased ethnic segregation, especially that Italians in particular have higher median income than that of Toronto in general.
7.3.3 Western European cities
In Europe, the association of socioeconomic and ethnic segregation is highly acknowledged in the literature. The simultaneous changes in both types of segregation in each city are based on one (or more) of three assumptions. Firstly, ethnic groups are more likely to belong to low-income groups (Bolt et al., 2009; Van Kempen, 2000; Massey and Denton, 1993). This assumption can explain the parallel increase/decrease in both types of segregation in several European cities, because any factor contributes to the change in one type of segregation will eventually lead to similar changes in the other. Such as the case of Rotterdam, where housing allocation rules are controlling levels of concentration of low-income households, who mostly belong to ethnic minorities. The second assumption suggests that poor ethnic groups tend to concentrate in public housing – as the case of Amsterdam (Musterd and Fullaondo, 2008), and Stockholm (Bråmå, 2008). This assumption can lead to different spatial outcomes of segregation overtime. Initially, if public housing is unevenly distributed in certain neighborhoods, ethnic minorities will also be concentrated according to public housing allocation. But over time, the recent privatization of public housing may (or may not) lead to the displacement and re-concentration of poor minorities, based on the extent and intensity of the privatization process, which explains the different spatial outcomes in Swedish cities. In Gothenburg, ethnic segregation decreased despite the increased socioeconomic segregation. As most minority enclave areas are dominated by public rental housing (Bråmå, 2008). Therefore, privatization of public housing may reduce affordable housing opportunities for the poor minorities, but at the same time it introduces high-income households to the enclave, and change the ethnic composition of the targeted
224 | Analysis and findings area. Also, in case of large-scale displacement of poor minorities, they relocate to other neighborhoods with affordable housing regardless their cultural preferences to live with co-ethnics, which can push them to look for housing in neighborhoods populated by another ethnic group. Eventually, limited opportunities of affordable housing may contribute to ethnic integration accompanied by higher level of poverty concentration. Thirdly, affluent ethnic groups tend to self-segregate. This assumption explains patterns of change in cities with decreased socioeconomic segregation associated with increased ethnic segregation. This trend of spatial change is common in British cities. According to Peach (2007), Indian population in British cities has maintained high level of segregation in spite of their high socioeconomic status. Unfortunately, the analysis cannot confirm this assumption due to lack of comparable data for British cities until the analysis process was performed.
7.3.4 Southeast Asian cities
Singapore and Hong Kong are the only two global cities with available data (on both types of segregation) in the East Asian region. As shown in chapter six, both cities have their special histories and qualities. On one hand, in Singapore, the enormous increase in ethnic segregation and the associated decrease in socioeconomic segregation can be simply explained by self-segregation of different ethnic groups (see chapter six for more details). However, the decreased socioeconomic segregation is relatively low (in terms of the intensity of the change) compared to the change in ethnic segregation, which indicates that other factors are involved in the complex process of spatial change beyond the issue of self- segregation. In fact, regrouping patterns of different ethnic groups in Singapore are supposed to produce higher level of socioeconomic segregation, because Malays – for example – have higher socioeconomic status and they tend to occupy higher-quality public housing units (Sin, 2002). While, Indians are more isolated in lower-quality public housing
The color of poverty | 225 units (ibid.). Accordingly, self-segregation of Malays and Indians can lead to a wider gap between income groups. However, allocation system of public housing units overcame the socioeconomic segregation resulting from ethnic regrouping. In fact, the HDB reformed its public housing allocation system in the late 1980s, by the aim of making it “more efficient and fair” (Tu, 1999: 104). According to Sin, “[T]he entire public-housing landscape in Singapore has been engineered in such a way as to yield low segregation scores.” (2002:434). Where each planning area has fair shares of all four types2 of apartments vary in sizes and prices (Singapore Department of Statistics, 2009). And since public housing hosts 86% of the population, therefore, the housing allocation system can have an enormous effect on segregation levels of the city as a whole. On the other hand, the case of Hong Kong shows no clear association between levels of ethnic segregation and the change in socioeconomic segregation. The reason behind this special case of disassociation is the absence of clear ethnicity lines between Hong Kong Chinese, and Mainland Chinese, who are the largest and only significant immigration group in Hong Kong. According to Cheung and Leung (2012: 2), “[B]ecause the ethnicity is the same, mainland migrants to Hong Kong seem to be indistinguishable by appearance, and they can permeate everywhere in Hong Kong without encountering segregation and exclusion.” Cheung and Leung also highlighted that the only visible differences between Hong Kong Chinese and mainland migrants are some cultural practices and slightly different Cantonese dialect. Notably, younger migrants who attend school in Hong Kong are more likely to overcome the language and culture differences and blend in totally in the host society. Another evidence for the absence of differentiation between Hong Kong natives and Mainland migrants is the census data collected by ‘Census and Statistics Department of Hong Kong”, where their classification of resident population is confined to only three categories: Chinese (including natives and Mainland migrant – 95% of total population), other Asian migrants, and other non-Asian migrants.
226 | Analysis and findings
Accordingly, any changes in socioeconomic segregation cannot be explained by the non-evident segregation of Mainland immigrants. Instead, socioeconomic segregation patterns in Hong Kong are highly dependent on other characteristics of the distinctive urban landscape of the city. Hong Kong is an island with a mountainous geography, it has very high density of population, high income inequalities, and its skyline is marked by high-rise buildings that reach up to 50 stories even in areas away from the central districts (Monkkonen and Zhang, 2011). Accordingly, the high density causes the issue of scale to be particularly relevant in defining levels of segregation, where pockets of high-income/high-rise gentrification emerging in scattered locations around the city will lead to decrease in socioeconomic segregation at least on the neighborhood scale, while the fragmented urban fabric is maintained on smaller scales. This assumption conforms to the results shown in figure (7-4), as the change in socioeconomic segregation index is measured on relatively large areas (district council/constituency areas level). Furthermore, other analytical studies of Hong Kong performed by Forrest et al. (2004) and Monkkonen and Zhang (2011) have yield similar results to the findings of this research. Forrest et al. concluded that high levels of income inequality is not reflected spatially in form of residential segregation, while Monkkonen and Zhang found that levels of socioeconomic segregation have actually fallen over the period from 1991 to 2006. Among other reasons, Monkkonen and Zhang (2011) explained lower income segregation by the fragmented urban redevelopment of older areas of the city by the private sector, which have brought the new high-rise buildings in close proximity to older residential stock inhabited by low-income households. Evidently, both cases of Singapore and Hong Kong show different models of income and ethnicity association than those found in Western Europe, North America, and Oceania. Where in Singapore, the association of Malay and Indian groups to distinct socioeconomic status had no impacts
The color of poverty | 227 on segregation levels due to the spatial distribution of public housing. While in Hong Kong, there is no evidence that this association does even exist.
228 | Analysis and findings
Summary
Cases discussed in this chapter suggest that part of the change in either socioeconomic or ethnic segregation in global cities is actually caused by the change in the other. Results reveal that 70% of global cities tend to have parallel increase or decrease in both types of segregation. The regression analysis showed a moderate correlation between the change in both socioeconomic and ethnic segregation, where 9.7% of change in ethnic segregation is explained by the change in socioeconomic segregation. Moreover, the generalized assumption that alpha cities are more prone to higher levels of segregation is not confirmed. Instead, beta and gamma cities showed higher tendency to have parallel increase in both types of segregation, while a higher ratio of alpha cities showed a parallel decrease in both types. Noticeably, divergent changes in socioeconomic and ethnic segregation could not be linked to certain global status. On the regional level, results show that Australian cities are more likely to have parallel increase in both types of segregation, followed by North American cities, then Western European cities. The level of spatial change produced by the association of an ethnic group to certain socioeconomic status varies significantly from region to region, and from city to city according to the local situations in each city. For example, intense gentrification in Australia, privatization of public housing in Europe, and lower median income of immigrants in Rustbelt American cities are factors contributing to the increase of at least one type of segregation, and consequently the increase/decrease in the other. In fact, this conclusion undermines the role of macro economic change as a main generator of divided socio-spatial outcomes within global cities. Where in several cases, the local contexts proved powerful enough to modify, intensify, or reverse spatial outcomes produced on higher levels of the urban hierarchy.
The color of poverty | 229
Finally, detailed studies of individual cities reveal distinct models of urban change that are inconsistent with the global city’s characteristics discussed in the literature. Austin, Brussels, and Hong Kong are just few examples for this controversy.
230 | Analysis and findings
Notes
1 The other American city showing similar spatial change is Columbus, however, the change in ethnic segregation is insignificant compared to Austin. 2 Price of public housing units is often defined in terms of the number of rooms. Types of housing units are: 1-/2- room flats, 3-room flats, 4-room flats, 5- room/Executive flats (Sin, 2002).
Chapter 8
8 The hybrid outcomes of urban change: conclusions and discussion
Throughout the chapters, the discussion of the socio-spatial transformations in global cities revealed certain generalization tendencies in the global city literature. The generalization includes; firstly, the changes in cities’ economic function in the global economy are decisive for other social and spatial changes occurring within these cities (Friedmann, 1986), where the agglomeration of advanced producer service firms in these cities is considered the key motivation behind urban change (Sassen, 1991; Castells; 2000; Taylor, 2000; Haussermann and Haila, 2005). Secondly, as a result of economic restructuring, occupational and income structures of global cities are altered towards more polarization (Friedmann and Wolff, 1982; Sassen, 1991), dualization (Mollenkopf and Castells, 1991), and inequality (Fainstein et al., 1992; Hamnett, 2003). Consequently, the intensified social division between different population groups within the city is spatially manifested in the form of high-income gentrification in inner city, spatial concentration of poverty, and spatial segregation based on race or ethnicity (Sassen, 2001; Massey, 1996; Goldsmith, 2000; Pacione, 2009). Thirdly, despite acknowledging the relevance of local situations in individual global cities to the understanding of the complex process of urban change (Burgers and Musterd, 2002; Van Kempen, 2007; Marcuse and Van
232 | Conclusions and discussion
Kempen, 2000), yet Friedmann and Wolff (1982), Sassen (2006), along with a wide range of studies performed over individual cities (i.e. Badcock, 1997), strongly suggest that economic restructuring and its consequent socio-spatial changes are evident in different cities around the world as a result of their integration into the global economy.
8.1 The assessment of the ‘global/divided’ city model
The key findings of the analysis process presented in chapters five, six and seven answer several questions about the validity of these three widely accepted assumptions in the global city literature. Briefly, the study was designed to capture the intensity and magnitude of change in spatial segregation levels in a large number of global cities, then, by correlating these changes to the city’s level of integration into the global economy. The propagated association between the processes of economic globalization on one side, and the increasingly fragmented socio-spatial patterns within global cities on the other side, is evaluated. Also, during the analysis process, the specific local situations that affect the spatial change in individual cities were discussed. The general observations have led to the belief that the general applicability of the ‘global city/divided city’ model on global cities around the world is overstretched. The analysis was performed over three stages; stage one confirmed that global cities do not share statistically significant tendency to be more segregated based on the socioeconomic status and ethnic background of their population. The regression analysis also confirmed that; the global network connectivity of cities could explain only 0.2% of the change in socioeconomic segregation and 0.2% of the change in ethnic segregation. While this percentage jumps to 5.1% when segregation types are explained by one another. In other words, part of the increased (or decreased) spatial segregation in global cities is more likely a result of the long-standing
Conclusions and discussion | 233 socioeconomic and ethnic structure, rather than the cities’ level of integration in the global economy. The second stage of the analysis revealed that certain characteristics of individual global cities could have an influence on the resultant change in spatial segregation in these cities. A one-way analysis of variance ANOVA was performed over different groups within the sample; the analysis results confirmed that the changes in segregation differ significantly among cities located in different geographic regions of the world. On the contrary, cities divided into alpha, beta, and gamma ranks didn’t show a significant difference in their changing segregation levels. Finally, in stage three, the discussion of the situation in individual global cities aimed to find an explanation for the lack of shared pattern of spatial change among global cities. The key findings are as follows: Cities of Amsterdam, Mexico City, and Sydney are seemingly supporting the validity of the general global/divided city model. Where the three cities are sharing the same global status as alpha cities, they are situated in totally different contexts, and all of them witnessed an increase in socioeconomic segregation (see chapter five) over the past decade. However, when compared to other alpha cities of Singapore, Miami, and Brussels, - which have actually witnessed more integration between socioeconomic groups over the same period of time – it can be argued that socio-spatial segregation is not necessarily the spatial outcome associated with economic globalization as promoted in the literature. Furthermore, it is to be noted that this conclusions is not based only on these 6 cities. Instead, this conclusion is supported by the lack of common spatial change shared by the large number of cities of the dataset (total of 91 cities to avoid misleading patterns suggested by smaller datasets). As mentioned, the results of the regression analysis deny any possibility that economic restructuring is associated with intensifying socio-spatial divisions within global cities. Even among alpha cities - which have higher ‘service value’ than beta and gamma cities, the results show no tendency for intensified
234 | Conclusions and discussion socioeconomic and ethnic segregation within these cities, while on the contrary, alpha cities – statistically speaking - were less likely to witness increase in segregation levels compared to gamma cities. Beside the generalized spatial outcomes, the generalization of a mechanism of the production of urban change was also challenged. The examples of Amsterdam, Mexico City, and Sydney reveal that even if these cities share similar spatial changes. Yet, a closer look on each case clarifies that these changes occur for different reasons and through different mechanisms. In more detail, the global/divided city model suggests that the declining role of the state will lead to privatization of services and subsidies cutbacks in general, including the housing sector (see chapter two for details). This might be true for Amsterdam, where the increased socioeconomic segregation is coinciding with the privatization of socially rented housing, which dropped from 57% of total housing units in 2001 to 47.4% in 2011 (Ministry of BZK, 2012). However, the same tendencies for privatization and declined provision of public housing are not evident in Mexican cities. On the contrary, the federal government initiated a housing finance system in the 1990s, and provided around 400,000 mortgages every year between 2000 and 2005 (Monkkonen, 2011). Yet, the allocation and size of the newly introduced housing projects, along with the inequitable selection criteria for those whom eligible for mortgage contributed to higher levels of socioeconomic segregation after all. Similarly, the generalized model of the global city also suggests that labor mobility contributed to the massive increase in flows of immigrants, and consequently contributed to the intensifying of ethnic segregation - due to the concentration of new comers in existing ethnic enclaves or affordable housing in poor neighborhoods in general. Yet, alpha cities of Singapore, San Francisco, and Melbourne suggest otherwise. Despite being magnets for immigrants, changes in ethnic segregation in those cities cannot be linked to labor mobility. In Singapore, foreign workers are legally banned from renting any residential properties on the market (Cho, 2011). Therefore, any
Conclusions and discussion | 235 increase in immigrants flow will not lead to the formation of any ethnic enclaves. Instead, ethnic segregation in Singapore is historically inherited, then intensified due to inefficient housing policies (see chapter six for details). In San Francisco, as in most Sunbelt cities in the USA, the huge in- flows of immigrants to the region (especially from Mexico) (Frey, 2002; Thompson, 2011) are not followed by intensified ethnic segregation. On the contrary, most Sunbelt cities witnessed significant decrease in ethnic segregation, especially when compared to Rustbelt cities that have higher tendency for increased ethnic segregation and much less flows of immigrants, this can be explained by the ‘super-diversity’1 (Vertovec, 2007) generated in Sunbelt cities by the enormous flows of immigrants. As, the existing enclaves are no longer able to accommodate all new comers to the city, which encourage them to look for housing opportunities elsewhere in the city. While in Melbourne, the in-flows of immigrants were also accompanied by decrease in ethnic segregation, due to the demographic characteristics of immigrants. As, Melbourne attracts particularly large numbers of transnational students (Fincher and Shaw, 2009), who their preferences are based on affordability and accessibility rather than the ethnic composition of the desired neighborhood. Besides, intense gentrification in inner Melbourne contributed to the decreased ethnic segregation, due to the displacement of poor immigrants, and their concentration in other neighborhoods that are already dominated by other ethnic groups. Accordingly, labor mobility is not necessarily responsible for the complicated patterns of ethnic segregation in global cities. Increased segregation in Singapore is not even remotely related to labor mobility, while despite the evident flows of immigrants to San Francisco and Melbourne, levels of ethnic segregation dropped significantly in both cities.
236 | Conclusions and discussion
8.2 The production of spatial change in global cities
In view of the above, the lack of a general trend of change in spatial segregation levels among global cities of the dataset does not only challenge the acknowledged association between economic restructuring and the intensified socio-spatial divisions within cities, but also, the findings support a point of view that criticizes the global city literature for the oversimplification of the global/local interplay through which the complex process of urban change is taking place. In more detail, the only explanation for how cities are expected to share common trends of urban change despite their local differences is because local situations in individual cities are rendered irrelevant to the outcomes of urban change. For example, the polarization thesis acknowledges that global impacts on local settings are channeled through different scales of the urban hierarchy. It is believed that transnational institutions (global level) are downscaling central governments2 (nation-state level), which will impact urban policies (on both national and metropolitan level), to finally contribute to increased spatial segregation (on neighborhood level). For that matter, the problem with this logic is not the overlooking of the intermediate scales between the global and the local. Instead, the problem is the overlooking of the imprint each level can leave on the process as a whole. In other words, the thesis simplifies the restructuring process through the standardization of the influence of global forces and neutralizing any possible influence of national, metropolitan, or local particularities. Therefore, the lack of a shared pattern of spatial change among global cities challenges the assumptions that local situations are irrelevant, as the divergent changes in individual cities are only explained through acknowledging the important and relevant role of local situations on the outcomes of urban changes, and consequently, the findings challenges the over-simplistic view of the global/local interplay, which renders powerful global processes that modify local settings within cities in an unproblematic
Conclusions and discussion | 237 manner. Accordingly, the complex interaction between macro processes of globalization and micro situations in individual cities needs to be approached from another perspective that includes the different factors involved in the process of urban change in general, and involved in the production of distinct and non-generalizable spatial outcomes for individual global cities in particular. I don’t claim that this research offers such well-formulated alternative perspective for understanding the complex dynamics of urban change in global cities. Instead, this chapter highlights the theoretical and empirical issues that need to be addressed in future research to finally reach a framework for the systemic study of urban change in a global context.
8.2.1 Urban change and the concept of ‘hybridity’
One of the possible explanations for the observed divergent outcomes of spatial restructuring in global cities is that the potential impacts of the new functions of global cities are exaggerated, as Pretecille (1994) argues, the services industry may represent only a small part of urban employment, and accordingly, the occupational and income structures of global cities are not expected to change dramatically as a result of the cities’ integration into the global economy. In this case, a considerable part of the recent social and spatial changes in global cities is triggered by several factors that are not necessarily economic or global in reach. As in the case of Mexican cities, Guadalajara is a gamma city, Monterrey is a beta city, and Mexico City is an alpha city, yet in all three cases, the drastic intensification of socioeconomic segregation has nothing to do with these cities’ position as global cities. Instead, the change in segregation is explained by the characteristics of Mexico’s federal housing finance system (see chapter five for details). However, this perspective - which considers the influence of global economic changes on cities to be negligible - fails to fully explain the more complex situations in cities like Amsterdam, Sydney and other cases where the impacts of global economic changes on other social and spatial facets of
238 | Conclusions and discussion restructuring is more pronounced. Therefore, another systemic explanation is required in order to capture the complexity of the process of restructuring of global cities. It is already established that the generalization of a certain outcome of spatial restructuring to be applicable in global cities around the world is not valid, neither is the total elimination of the possibly influential role of global economic functions of cities on the outcomes of restructuring. Accordingly, the aim of this section is to reach ‘the middle ground’, and to introduce a potential systemic explanation that avoids any generalized assumptions by capturing the complexity of the restructuring process based on the findings of the analytical study. Briefly, the cases presented in chapters five, six and seven reveal that the spatial outcomes in global cities are a product of the ‘interaction’ between global economic functions with the existing contextual and historical particularities that varies from city to city. The nature and main characteristics of this interaction are what our proposed theoretical construct is attempting to capture. As noted in chapter two, one way to describe this ‘interaction’ is Burgers and Musterd’s (2002) three-layered model that refers to the contextual variables as the layer through which the global changes are ‘mediated’ to the local level. However, the term ‘mediate’ here implies that the direction of the change is from a top layer (the global economic change) to the bottom layer (the level of the locality). Accordingly, despite the fact that Burgers and Musterd’s model effectively explains different mechanisms of growing inequalities in post-industrial labor markets, yet the same idea of the top-down layers of change may not be suitable to describe the spatial change in global cities in general, as this model refers to the local level as a receptor for the change impinged on it by macro process that take place on higher scales of the urban hierarchy instead of acknowledging the mutual interaction between the global and the local. Other concepts in the academic literature – such as glocalization and hybridity – are one step closer to describe the nature of this global/local interaction and its outcome.
Conclusions and discussion | 239
In another publication (see Ismail, 2013), I argued that the concept of hybridity – which is widely used in cultural studies – could be extended to describe the process (and outcomes) of spatial changes taking place in urban areas as a result of globalization. The notion of hybridity originated in biology with the development of Mendelian genetics in the 1870s, which refers to the process of ‘mixing’ of genetic traits, as the resulting hybrids typically have intermediate traits of both geneses involved in the hybridization process (McCarthy, 2006). Then, the term was extended to describe the mixing of ‘cultural’ forms as a result of increasing global interconnectedness (Eriksen, 2007). Hybridity is also used to explain other types of ‘mixing’ of institutional (de Ruijter, 1996), organizational (Oliver and Montgomery, 2000), and structural forms of social organization (Nederveen Pieterse, 2001). Since global processes and local contexts are not exactly ‘idealtypical’ (Brandsen et al., 2005) entities that can be tangibly mixed to produce a hybrid outcome. Therefore, approaching such global/local interplay - from hybridity perspective - has to capture the in-between structural practices that influence one another on different hierarchal scales, then by verifying that the spatial outcome of this interplay is actually marked by traits inherited from different practices (including those on the local level), only then can the concept of hybridity be broaden to describe geographical features produced from the interaction of non-hegemonic global forces and non- transparent local contexts. The analytical study has already eliminated the possibility that spatial outcomes are defined solely by cities’ integration in the global economy, while at the same time, the detailed discussion of individual cases reveals a number of regional, national, metropolitan, and local particularities that are clearly influencing the intensity and direction of spatial changes in global cities. To finally assert that spatial changes are outcome of complex processes of structural and chronological hybridity
240 | Conclusions and discussion
a. Structural hybridity
Nederveen Pieterse explains that globalization in structural terms means “the increase in the available modes of organization: transnational, international, macro-regional, national, micro-regional, municipal, local” (1995: 50). In this sense, the global process of structural hybridization refers to the formation of a hybrid outcome as a result of the interaction of social, institutional, and economic process across the different scales of the urban hierarchy. He adds “hybrid formations constituted by the interpenetration of diverse logics manifest themselves in hybrid sites and spaces.” (p. 51). Global cities themselves are an example of such hybrid spaces, as they surpass the national scale in their importance to the global economy (Sassen, 2006), while at the same time, they are still a place of the local, where the global spaces of firms and gentrified neighborhoods are juxtaposed by the deprived neighborhoods and ethnic enclaves. In order to describe the recent spatial changes in global cities as a form of structural hybridization, the global city literature acknowledges the fact that the global impacts on local settings are channeled through different scales of the urban hierarchy. As mentioned earlier, it is believed that transnational institutions (global level) are downscaling central governments3 (nation-state level), which will impact urban policies (on both national and metropolitan level), to finally contribute to increased spatial segregation (on neighborhood level). In the light of the findings of the analytical study, all these processes on the different scales are affecting the resultant patterns of spatial segregation in terms of the direction and the magnitude of change. For that matter it is no longer possible to overlook the influence of the intermediate processes on the multitude of scales between the global and the local, through standardizing the impact of global forces and neutralizing any possible influence of national, metropolitan, or local particularities. Cases discussed in previous chapters show that, for example, public housing is
Conclusions and discussion | 241 being privatized in some cities, while new public housing provision programs are being implemented in other cities despite their high integration into the global economy. In this case, the geographical distribution of public housing units and the accessibility of the program (by disadvantaged population groups) will outline the maps of spatial segregation in unpredictable manner. Similarly, urban regeneration projects are not necessarily leading to gentrification (and its consequent displacement/ reconcentration of the poor), few examples show that urban regeneration projects are still able to achieve improvements in residents’ well-being despite the profit-based atmosphere clouding the decision-making process. Furthermore, on neighborhood level, homeownership rates, housing market flexibility, and residents’ preferences are strong modifiers of residential mobility and hence spatial segregation. One example is the preferences of affluent entrepreneurs who belong to an ethnic minority to reside in the (usually poor) ethnic enclave near to their ethnic business; their choice will probably contribute to higher level of ethnic segregation, yet socioeconomic segregation can decrease accordingly. Accordingly, it can be argued that the spatial change in global cities is a structural hybrid, formed by the mélange of the global, national, municipal, and local processes and particularities in individual cities.
b. Chronological hybridity
The metaphor of the ‘layered city’ (Massey, 1984; Kesteloot, 2000; 2005) portrays how the socio-spatial structure of any city is a product of historical processes occurring in successive rounds of capital accumulation, each round has its own arrangement of economic sites and residential areas deposited in layers one upon another across geographical space. While, each layer represents distinct spatial arrangement of economic and residential functions that belong to a certain round of accumulation, still, spatial features from previous layers may still be present in recent layers of spatial development.
242 | Conclusions and discussion
Similarly, the findings of the analytical study reveal that the patterns of spatial segregation in recent years are defined by both the historical patterns of spatial segregation, as well as the current economic, political, and social particularities in individual cities. In this sense, it can be argued that spatial outcomes in global cities are a ‘chronological hybrid’. As in the case of structural hybridity, the chronological hybridity refers to the formation of hybrid from the interaction of different processes, but instead of the interaction of these processes across different scales, the processes are taking place across the historical rounds of capital accumulation. In other words, the recent spatial changes in global cities are a product of the interaction between the historically inherited spatial forms and present-day dynamics (Soja, 1985; Kesteloot, 2000). For example, patterns of ethnic segregation in global cities are not defined solely by recent flows of international migration and the increased labor mobility. Instead, concentrations of new comers to the city are highly dependent on already existing ethnic enclaves that represented nuclei for the new comers to settle around. Therefore, labor mobility alone cannot explain present concentrations of Turkish population in German cities nor the Indian population in British cities, because the first is formed by historical migration waves since WWII, while the later is an outcome of historical ties formed in the colonial era (Peach, 2002; Massey et al., 2008). If it has any influence, labor mobility merely contributes to the altering of existing conditions that are accumulated over the years to produce a hybrid outcome, instead of imposing a non-rooted spatial outcome. To sum up, since the process of economic restructuring alone is proved inadequate to explain the complex process of urban change in global cities, then it can be argued that by acknowledging the influential role of local contexts, and through perceiving the global/local interplay as processes of structural and chronological hybridization, we can reach a better understanding of the mechanisms through which the divergent patterns of spatial change in individual cities are produced. The historically uneven
Conclusions and discussion | 243 development in Brazil, social inclusion programs in Johannesburg and Adelaide, housing allocation restrictions in Rotterdam, geography of Hong Kong, and ethnic composition of Singapore are just few examples of how the spatial transformations in global cities are a product of both global processes mélanges with historical, political, cultural, and geographical particularities across the national, metropolitan, and local contexts of each city. To finally produce a ‘hybrid’ spatial outcome; hybrid in terms that the outcome carries distinct qualities that can be traced back to all the modifiers exist – structurally - on different levels of the urban hierarchy from global to local, and – chronologically - on different layers of previous developments. One example of this ‘hybridization process’ is the gentrification of poor neighborhoods in both contexts of Latin American cities, and the city of Hong Kong. In both cases, as well as in many other European, Australian, and American cities, the demand for attractive housing conditions has increased. However, instead of the large-scale gentrification process that took place in Melbourne or Montreal for example, the gentrification process in Latin American cities was affected by the already existing informal settlements that surround urban areas due to the massive urbanization of the region since the 1950s. The mélange of gentrification (as a global process (Atkinson and Bridge, 2005)) and historical patterns of urbanization, produced what is known as the ‘gated communities’ (Feitosa et al., 2007; Monkkonen, 2011), which is technically gentrified areas4, but they are isolated by walls and secured by monitored entrances due to fear of crime and violence in the surrounding neighborhoods (Caldeira, 2000; Coy and Pohler, 2002). Similarly, the demand for attractive housing conditions in Hong Kong was met by the gentrification of deteriorated neighborhoods. Yet, due to the mountainous topography of the island of Hong Kong, its very high density of population, and its limited urban space (Monkkonen and Zhang, 2011). The gentrification process took the shape of high- income/high-rise buildings - in scattered locations in the city, and in a close proximity to older residential stock inhabited by low-income households. In
244 | Conclusions and discussion this sense, despite the fact that gentrification is a global process that occurs in cities around the world. Yet, it can be argued that, local situations – such as historical layers of urban development in Latin America, or the topography of Hong Kong – are able to bend the course of the gentrification process and produce a unique and hybrid type of gentrified areas that reflects the interaction of the global process of gentrification with the specific characteristics of every context.
8.3 So what? A research agenda for the systemic study of spatial segregation in global cities
This research aimed to answer several research questions, the main question as described in chapter one; “is the global city a divided city?” According to the theoretical discussion and the findings of the analytical study, it is argued that global cities are not necessarily divided cities; instead, it is evident that global cities vary greatly in the direction and the magnitude of change in their spatial segregation levels. Accordingly, the findings shed the light on another research problem that needs to be addressed in the future. Concisely, it can be argued that the available research lacks the ability to theoretically and empirically explain the complex dynamics of the spatial restructuring of global cities. The problem is manifested in the generalization of a number of assumptions to be applicable on global cities around the world, which indicates an oversimplification of the global/local interplay in the specialized literature. Also, the problem emerges from the evident overlooking of the possibly influential particularities of different local contexts on the process and outcomes of spatial restructuring. Briefly, in the light of this research problem, the objectives of the future research should include; the development of an all-encompassing research approach, or an alternative hypothesis that overcomes the shortcomings of the available research, as well as, the development of a comprehensive
Conclusions and discussion | 245 research methodology for the systemic study of spatial segregation in global/local context. An example of such framework for the cross-city comparison of spatial segregation patterns in global cities is proposed later in this section. Unfortunately, due to several constrains, this study could not offer such alternative hypothesis. Instead, the study examined the validity of the assumptions of the global city literature and reached the conclusion that these assumptions require to be revisited. For that matter, the future research should overcome the shortcomings of this study through: (1) enhance and update the dataset for the empirical study, (2) an in-depth investigation of the most influential contextual particularities, as not all local factors are proved to be relevant to the situation in individual cities, and (3) answer these pressing research questions; how to approach the restructuring process of global cities in a way that capture its complexity? What are the local particularities that need to be included in the study of cities’ restructuring? And how to develop the concept of hybridity to assist in the understanding of the evolving patterns of spatial segregation in global context?
8.3.1 Towards a heuristic model for the comparative studies of spatial segregation
This sub-section presents briefly an attempt to develop a framework for the comparative studies of spatial segregation in global cities. The aim here is to highlight a number of local factors that affect the course of change in spatial segregation, while at the same time, spot the possible variations in these factors from context to context. Briefly, the discussion in chapter two highlighted a number of ‘aspects’ (Burgers and Musterd, 2002) or ‘local contingencies’ (Marcuse and Van Kempen, 2000), by which the impacts of global economic change on different localities are explained. These aspects include historical and subcultural differences, as well as other variations in national policies, geography, and existing levels of inequality. Yet, another conclusion drawn
246 | Conclusions and discussion from the analytical study suggests that these aspects are too general to describe and to compare patterns of spatial segregation in individual global cities. In other words, the comparison of socioeconomic segregation patterns that are a result of – for example – the change in public housing provision in two different cities, cannot be based solely on the number or ratio of public housing units provided by the state in each city. Instead, the comparison should include other factors, such as the allocation of public housing units (are they concentrated in low-income neighborhood, or distributed across the city in a close proximity to middle and high-income households?), who has access to the service? (low-income groups, poor immigrants, the unemployed), and patterns, as well as, the pace of privatization of public housing units (if any). All these variations of a single contingency, that is public housing provision, can lead to remarkably divergent changes in segregation levels between two cities, even if both cities are witnessing a parallel decline in shares of public housing stock. Still, issues of allocation and access are able to define the resultant segregation patterns in the two cities. Similarly, based on the analysis findings, the possible variations of other relevant local contingencies can be fruitfully developed into an explanatory model for analyzing the changing patterns of spatial segregation, especially from a comparative perspective. The model compares cities along four main axes, while each axis includes a set of variables, which represent the range of contextual differences evident from city to city that are most likely related to change in the segregation level. The four axes are: (a) Gentrification, (b) Urban policies, including governmental actions, regulations, and public housing provision, (c) The mutual correlation between socioeconomic and ethnic segregation, and (d) Other distinctive characteristics of the cases under study.
Conclusions and discussion | 247
a. Gentrification
Gentrification as a global phenomenon (Atkinson and Bridge, 2005) is considered a direct outcome of the rise of service industries, while at the same time, it contributes to higher level of socioeconomic segregation by displacing the poor and creating citadels for the rich. This standardized and flat perception of gentrification has different meanings when situated in the context of specific cases. Gentrification patterns in the alpha cities of Hong Kong, Brussels, Amsterdam, and Sydney suggest that the ‘scale’ of gentrification is an important variable that is linked to certain changes in segregation levels. In more detail, the small-scale gentrification projects in both Brussels (Van Criekingen and Decroly, 2003), and Hong Kong (Monkkonen and Zhang, 2011) have contributed to the overall decrease in socioeconomic segregation levels in both cities, and vice versa, where the large-scale gentrification in Sydney and Amsterdam are contributing to the overall increase in segregation levels. Another variable related to the ‘scale’ of gentrification, is the pace and intensity of displacement of the poor and immigrants from inner city, where the larger the scale of gentrification, the larger the number of households who have to relocate. Also, the greater the difference in socioeconomic status between the poor residents and the gentry group, the starkest the upgrade in the built environment (Clark, 2005), and eventually, the higher the pace of displacement of the poor from gentrified areas. Similarly, the process of reconcentration of poor and immigrants households - that follows their displacement from gentrified areas – varies from city to city according to the distribution of affordable housing on the city level, as well as the presence and the efficiency of housing regulations that prevent the concentration of poverty and counteract the drawbacks of gentrification. As in the case of Rotterdam, the strict housing allocation rules set by local government in 2003 by the ‘Rotterdam zet door’ action program (Kleinhans, 2004), have successfully controlled the spatial distribution of low-income households and prevent their reconcentration in disadvantaged
248 | Conclusions and discussion neighborhoods. While in Amsterdam, the lack of such effective regulations have led urban development projects to contribute to the reconcentration of displaced households in other poor areas elsewhere in the city (Bolt et al., 2009). In view of the above, the gentrification process in general does not necessarily lead to higher levels of socioeconomic and ethnic segregation. Instead, the different scales of gentrification projects, the pace of the associated displacement of the poor and immigrant households, and the regulations that control the reconcentration patterns of the displaced households are the variables that need to be taken into account, especially when discussing the implications of gentrification process on spatial segregation level in cities in different contexts.
b. Urban policies, regulations, and public housing provision
It is acknowledged in the specialized literature that state regulations and housing policies can promote either higher social mix or higher degree of segregation on neighborhood levels (Kazepov, 2005; Roberts and Wilson, 2009). As, state regulations have the ability to prevent/permit both poverty and affluence concentration. Yet, it is also acknowledged that economic globalization undermined the ability of local government to impose regulative decisions on local markets (Hancher and Moran, 1989), through different trends such as deregulation, privatization, diminished public spending ...etc. (Beall, 2002; Held, 2004). As a result, the local housing market became increasingly deregulated, social housing provision declined, and existing social housing are being privatized or demolished to clear room for profitable housing projects. Accordingly, changes in local housing markets in global cities are expected to contribute to higher level of segregation. However, it is already established that generalization of certain consequences associated with economic globalization is sort of an exaggeration, because local situations in individual cities is highly influential on the outcomes of cities’ restructuring. In this sense, the
Conclusions and discussion | 249 presence of an efficient government intervention in individual cities can significantly resist segregation tendencies, or even alter existing patterns of segregation. In order to evaluate the influence of local urban policies as an important contingency affecting the course of urban change, the analytical study aimed to correlate levels of segregation to three main aspects of urban policy: housing regulations, urban development initiatives, and public housing provision. The results suggest that presence of effective housing regulations, as well as urban development initiatives has a noticeable mitigating impact on the spatial segregation level. While, the change in public housing stock – as discussed above – affect the change in segregation levels differently from city to city, based on several variables such as the allocation of public housing units, and the access of the disadvantaged groups to the service. Cases of Singapore and Rotterdam clearly reflect the influence of housing regulations on segregation levels. In Singapore, the households’ allocation rules - imposed in the 1960s - have effectively contributed to the decreased ethnic segregation in the 1970s and early 80s (Van Grunsven, 2000). Then, in 1989, the government imposed the ethnic quota system (see chapter four) to control the voluntary residential mobility of minorities (Sin, 2003). However, the government ineffective pursuance of the quota system has led eventually to the re-grouping of ethnic groups. In this sense, effective and properly implemented housing regulations are expected to contribute to lower levels of segregation. This conclusion is also supported by the case of Rotterdam as discussed above. Similarly, Johannesburg, Adelaide, and Rio de Janeiro showed special cases of socioeconomic integration in highly segregated regions. The decreased socioeconomic segregation in these cities is explained by the presence of governmental intervention aiming for improving housing and infrastructure of deteriorated neighborhoods, as well as promoting social inclusion of the poor and minorities. Apparently, black empowerment and
250 | Conclusions and discussion poverty reduction programs in Johannesburg, urban development initiatives in Rio de Janeiro and urban regeneration projects in Adelaide (see chapter five) succeeded in reducing socioeconomic segregation significantly despite the process of economic restructuring witnessed in all three cities. In this sense, it can be argued that the absence of the same governmental actions in other cities has contributed to their rising levels of residential segregation due to the unchallenged social exclusion. Finally, the debate over the relevance of public housing (in terms of size and distribution) to levels of residential segregation can be summarized as follows: On one hand, it is assumed that quality, size, and distribution of public housing are highly relevant to levels of segregation (Van Kempen, 2005). As, public housing provides affordable housing opportunities for the deprived population. So, the less the share of public housing units in the market, the less housing opportunities for the poor. Also, uneven distribution of public housing units in certain neighborhoods can lead to concentration of poverty. The analytical study revealed that Swedish, Australian, and Mexican cities support these assumptions. Where the declined provision of public housing in Sweden (Hedin et al., 2012) and Australia (Arthurson, 1998; Badcock, 1999), (as well as the demolition of deteriorated public housing dwellings) are coinciding with the increase in both ethnic and socioeconomic segregation in Swedish cities and most of Australian cities. While in Mexico, public housing provision remained steady. Yet, the distribution of public housing units contributed to segregation (Monkkonen, 2011), due to the concentration of people with similar socioeconomic attributes in certain neighborhoods. While on the other hand, it is also assumed that housing policies are not necessarily related to lower or to higher levels of segregation (Musterd and Fullaondo, 2008). This assumption is supported by cases of Brussels, Barcelona, Berlin, and Singapore. Where, ethnic integration in Barcelona and socioeconomic integration in Brussels occurred anyway, despite the
Conclusions and discussion | 251 documented small shares of public housing in both cities. In Berlin, the reduction of public housing stock - from 30% of total housing units in 1990 to 15% in 2008 (Aalbers and Holm, 2008) - is not associated with increased socioeconomic segregation as the case of Swedish cities. Instead, Berlin showed insignificant change in levels of socioeconomic segregation. While in Singapore, 86% of the population resides in public housing dwellings (Sin, 2003). Yet, the increase in ethnic segregation is exceptionally high despite the fair distribution and high quality of public housing in Singapore. To sum up, contingency of local urban policies is proved to be – to some extent - relevant and influential on the course of spatial changes in global cities. As, housing regulations and inclusion policies can have a great impact on the accessibility of low-income households to decent housing opportunities, the prevention of poverty concentration, and the improvement of social cohesion. It is also concluded that size and distribution of public housing stock may or may not be related to a certain change in spatial segregation levels in individual cities.
c. The mutual correlation between socioeconomic and ethnic segregation
As shown in chapter seven, the intersection between maps of socioeconomic and ethnic segregation in the same city suggests that both type of segregation are mutually affecting one another. Where, 70% of cities of the dataset support this statement by showing a parallel increase (or decrease) in both socioeconomic and ethnic segregation. For example, in the case of Toronto, the decreased socioeconomic segregation is explained and only understood by the tendency among affluent Italian immigrants to cluster near or inside their ethnic enclave. As a result, the ethnic background and preferences of a certain population group resulted in the change in socioeconomic segregation. Similarly, increased ethnic segregation in Rustbelt cities in the USA is explained by the low median income of minorities compared to that of whites. As a result, the limited opportunities
252 | Conclusions and discussion of affordable housing for the poor minorities have contributed to their concentration in poor and deteriorated neighborhoods, while leading levels of ethnic segregation to rise accordingly. In this sense, comparing ethnic segregation levels in two cities needs to incorporate the socioeconomic status of the segregated ethnic minority as an important contingency and vice versa.
d. Other distinctive characteristics of the cases under study
Other factors that need to be considered in comparative studies of spatial segregation may include special and distinctive characteristic of a certain city. On the regional level, historically uneven economic development between cities of the same geographic region may affect the recent changes in spatial segregation levels. Where, historically uneven economic development between Rustbelt and Sunbelt cities in the USA, or between Sao Paulo and Rio de Janeiro in Brazil (see chapter five for details) is reflected in the tendency for Rustbelt cities, as well as Sao Paulo, to be more segregated. Apparently, impacts of uneven industrialization - since the 1940s - are still visible in the two contexts. On level of individual cities, distinctive characteristics of Hong Kong and Austin have noticeably impacted levels of spatial segregation in both cities. Despite being an alpha city, Hong Kong witnessed a relative retreat in foreign immigration in 2000s with the outbreak of H5N1 avian influenza, followed by SARS (severe acute respiratory syndrome). Also, the handover of Hong Kong to the Chinese authority in 1997 sparked the flows of Mainland Chinese into Hong Kong. As a result, local circumstances of Hong Kong contributed to less ethnic segregation in the city despite its global status. Likewise, gamma city of Austin receives fewer immigrants than other Sunbelt alpha cities such as Miami and San Francisco (Frey, 2002). Yet, Austin suffered an increase in ethnic segregation due to the historically persistent ‘East Austin enclave’, which is physically isolated by the interstate highway (I-35) built in the 1950s (Skop, 2009). Accordingly, it is
Conclusions and discussion | 253 clear that patterns of segregation in Austin and Hong Kong are strongly linked to the distinct characteristics of each city rather than its global status. For that matter, paying enough attention to the distinctive characteristics of different cities can assist in explaining the seemingly inconsistent changes in segregation levels showed by individual global cities. Finally, this sub-section underlines the remaining local contingencies that were not included in the comparative model, because they are either proved to be irrelevant to the process of spatial segregation, or its relevance is not confirmed based on the findings of the analytical study. On one hand, Marcuse and Van Kempen (2000) highlight the contingency of globalization to be influential on the spatial transformations in contemporary cities. In their view, the contingency of globalization is defined as; the city’s “position in the processes of globalization” (p: 268), where the processes of globalization are not limited to the concentration of advanced producer service firms in the city, but the processes also include other aspects such as the importance of international trade and extent of technological development. Accordingly, Marcuse and Van Kempen perception of globalization as a multi-dimensional process has led to their conclusion that spatial transformation in global cities are dependent on cities’ position in the processes of globalization. Which is not necessarily the case when a city’s global status is primarily determined based on the size of producer service firms hosted by the city. In fact, the results show no association between levels of segregation (both socioeconomic and ethnicity) and cities’ global status. On the contrary, each of the alpha, beta, and gamma categories had its share of cities with various directions and magnitudes of change in residential segregation. Furthermore, the discussion of the spatial transformations in individual cities did not pinpoint cities’ global status as a contingency that affects levels of segregation in anyway. Instead, there are numerous examples of alpha cities that show contrasting transformations in terms of the direction and magnitude of change in segregation level despite the fact that these cities share the same global
254 | Conclusions and discussion status as alpha cities such as Amsterdam, Brussels, Johannesburg, and Los Angeles to name a few. On the other hand, existing patterns of racial and ethnic segregation is another local contingency stated by Marcuse and Van Kempen. The discussion in chapter seven - and in the previous sub-sections - reveals that the existing patterns of ethnic segregation are highly relevant to the production of distinct patterns of socioeconomic segregation in individual cities. While in case of ethnic segregation, the historically persistent patterns of ethnic segregation are not associated with a recent intensification of ethnic segregation. For example, almost 60% of Western European cities have witnessed a significant decrease in ethnic segregation despite the persistent post-war ethnic enclaves in those cities. Accordingly, the relevance of existing patterns of ethnic segregation as an influential contingency is not confirmed. In more detail, the only case where the size and location of the existing ethnic enclave is proved relevant to recent patterns of ethnic segregation is when preferences of self-segregation are particularly high, as in the case of Indian population in Manchester, or the voluntary ethnic regrouping in Singapore. Other than that, the persistence or the dismantling of the existing enclave is determined by other factors such as the process of gentrification, where the displacement of poor immigrants from inner city is followed by either their concentration elsewhere, hence, ethnic segregation level is expected to rise, or their dispersal in other low- income and ethnically-mixed neighborhoods, which leads ethnic segregation level to drop. Similarly, the size of in-flows of immigrants to the city determine whether existing patterns of ethnic segregation is sustained or not, because when the flows of immigrants are large enough to surpass the capacity of the existing enclave to accommodate the new comers, immigrants start to search for affordable housing in other neighborhoods, and the resultant patterns of ethnic segregation in this case is also determined by either the concentration or the dispersal of immigrants outside of the existing enclaves, not by the presence of the enclave itself. Note that
Conclusions and discussion | 255 patterns of concentration of immigrants are also dependent on other aspects such as housing regulations and allocation schemes that vary from city to city. To sum up
The main contributions of this research are the in-depth investigation of the complex process of cities’ multifaceted restructuring, its motivations, outcomes, and dynamics. Also, the research managed to overcome a number of shortcomings of the available research by providing an empirical study - performed over a large dataset - for investigating the presumed socially and spatially divided nature of global cities. The research concluded that the widely acknowledged global/divided city model is insufficient to capture the potential influence of local contexts of individual cities on the outcomes of spatial changes, as the model oversimplify the global/local interplay by marginalizing the influence of the distinctive characteristics of individual cities on the local level. Therefore, the research offered a research agenda for future research; its main objective is developing an alternative theoretical construct for interpreting the socio-spatial transformations in global cities. The proposed construct should acknowledge the complexity of the global/local interplay through perceiving the process of urban change as a form of structural and chronological hybridity (see Ismail, 2013). Where, the spatial transformations in global cities are a product of both global processes mélanges with historical, political, cultural, and geographical particularities across the national, metropolitan, and local contexts of each city. To finally produce a ‘hybrid’ and unique spatial outcome; hybrid in terms that the outcome carries distinct qualities that can be traced back to all the modifiers exist – structurally - on different levels of the urban hierarchy from global to local, and – chronologically - on different layers of previous developments. A clear example for such hybridity is the case of Hong Kong, where segregation patterns are an outcome of complex interaction of (1) the city’s economic functions in the global economy, (2) the political milestone of the
256 | Conclusions and discussion handover of Hong Kong to the Chinese authority in 19975, (3) the natural outbreak of H5N1 avian influenza, then the epidemic of severe acute respiratory syndrome (SARS) in 20036, (4) the city’s demography and its high population density, and finally (5) the mountainous topography of the island. All these factors have contributed to the distinctive urban characteristics of Hong Kong including its skyline, social structure, ethnic composition, and its spatially fragmented urban fabric. Similarly, each global city has its own sets of ‘modifiers’ that cannot be overlooked, because these modifiers are proved to be an essential part of the complex global/local dynamics. The absence of standardized spatial outcome for global cities affirms that it is no longer possible to perceive the global/local interplay as a unidirectional cause-effect relation, in which the global processes are expected to reshape local settings of cities. Instead, context matters, history matters, and both are reflected on the outcomes of cities’ restructuring, even if this restructuring is taking place under the seemingly prevalent conditions of globalization. Moreover, the study also developed a scheme for studying the changes in socioeconomic and ethnic segregation within global cities, especially from a comparative perspective. The proposed scheme is based on a number of relevant and highly influential local contingent factors that are drawn from the discussion of a large number of individual cases throughout the chapters. The scheme is an attempt to facilitate the cross-city comparisons of the motivation behind spatial change in different local contexts. Finally, although the study could not provide a detailed discussion of the situation in all individual cities of the dataset (due to either lack of data or other time constrains), still, unexplained patterns of spatial change in cities and regions are to be the focus of an upcoming research that aims to develop the argument presented in this study. Overall, the study acknowledges the complexity of the urban changes taking place in contemporary cities. Generalized or not, the transformations in any city are an outcome of a dynamic process that can never come to halt. Yet, the nature
Conclusions and discussion | 257 of these transformations, as well as their motivations, signifies that each city is unique.
258 | Conclusions and discussion
Notes
1 Vertovec uses the term super-diversity to describe the unprecedented level of diversity witnessed in Britain’s demographic and social patterns. He explains, “Britain can now be characterized by ‘super-diversity,’ a notion intended to underline a level and kind of complexity surpassing anything the country has previously experienced. Such a condition is distinguished by a dynamic interplay of variables among an increased number of new, small and scattered, multiple-origin, transnationally connected, socio-economically differentiated and legally stratified immi- grants who have arrived over the last decade.” (p. 1024) 2 One example is the pressure of transnational institutions on central government to alter migration restriction to promote easier labor mobility, the result is the increased cross-border migratory flows on nation-state level (Castles, 2002), and with concentration of immigrants in metropolitan areas, ethnic segregation is expected to increase on the local level due to the deregulation of housing markets in global cities. 3 One example is the pressure of transnational institutions on central government to alter migration restriction to promote easier labor mobility, the result is the increased cross-border migratory flows on nation-state level (Castles, 2002), and with concentration of immigrants in metropolitan areas, ethnic segregation is expected to increase on the local level due to the deregulation of housing markets in global cities. 4 Gated communities are a from of gentrification, as – by definition – gentrification involves the gradual up-scaling of land-users where the new users are of a higher socioeconomic status than the previous users, together with an associated change in the built environment through a reinvestment in fixed capital (Clark, 2005) 5 Which altered the entry visa and work permits regulations especially for mainland Chinese (Ullah, 2012) causing unprecedented influx of workers to the city 6 Consequently, in-flows of foreign (non-Chinese) workers to Hong Kong slowed down compared to other alpha cities in the same period of time. Which altered the ethnic composition of the city due to the relative retreat of foreign immigration and the increased inflows of Mainland Chinese in a basically Chinese community
Conclusions and discussion | 259
Appendices
Appendix I
i. Tables of results
ii. iii. iv. This Appendix presents the values of the ‘spatial multi-group dissimilarity index’ SD(m) calculated for all cities of the dataset. As well as the total percentage and average yearly percentage of change in SD(m) value over time. v. vi. vii. viii. ix. x. xi. xii. xiii. xiv. xv. xvi.
Tables of results | 261
Table i-1: Socio-economic segregation index value for global cities, census year, and total/yearly change in index value
Table i-1 Socio-economic segregation index value for global cities, census year, and total/yearly change in index value in index change and total/yearly year, census cities, global for value index segregation Socio-economic i-1 Table
262 | Appendix I
xvii. e in index index in e g chan y earl y and total/ and , ear y census census , lobal cities g ation index value for value index ation g re g Cont. Table i-1: Socio-economic se Socio-economic i-1: Table Cont.
Tables of results | 263 xviii.
index in change total/yearly and year, census cities, for global value index segregation Socio-economic i-1: Table Cont.
264 | Appendix I
Table i-2 Ethnic segregation index value for global cities, census year, and total/yearly change in index value index in change total/yearly and year, census cities, global for value index segregation Ethnic i-2 Table
Tables of results | 265
Cont. Table i-2: Ethnic segregation index value for global cities, census year, and total/yearly change in index value in index change total/yearly and year, census cities, for global value index segregation i-2: Ethnic Table Cont.
Table i-2 Ethnic segregation index value for global cities, census year, and total/yearly change in index value
266 | Appendix I
xix.
Cont. Table i-2: Ethnic segregation index value for global cities, census year, and total/yearly change in index value index in change and total/yearly year, census cities, for global value index segregation i-2: Ethnic Table Cont.
Tables of results | 267
xx. xxi.
xxii. xxiii. xxiv. xxv. xxvi. xxvii. xxviii. xxix. xxx. xxxi. xxxii. xxxiii. xxxiv. xxxv. xxxvi. xxxvii. xxxviii. xxxix. xl. xli. xlii. xliii. xliv. xlv.
Cont. Table i-2: Ethnic segregation index value for global cities, census year, and total/yearly change in index value in index change total/yearly and year, census cities, for global value index segregation i-2: Ethnic Table Cont.
Appendix II
ii. Statistical hypothesis testing
Basically, statistical hypothesis tests aim to determine whether the observations drawn from the dataset are near the expected results defined by the hypothesis under examination (H0), or the observations are significantly various to the extent that there is high probability the ‘null’ hypothesis (H0) is rejected in favor of an alternative hypothesis (H1) (Gentle, 2002; Gibbons and Chakraborti, 2003). Simply put, the null hypothesis (H0) in our case is the association between economic restructuring and intensified spatial segregation within global cities. And according to (H0), the expected result is a tendency among global cities to have parallel changes towards increased segregation. However, in case that a number of global cities in the dataset shown a change in the opposite direction (towards more integration instead of more segregation); then, how to determine whether this number of cities is significant enough to rule out the null hypothesis or not? In other words, do cities with divergent behavior provide enough evidence against the global/divided city model? In order to decide whether to reject the null hypothesis or not based on the study observations, a statistical measure called the p-value (also known as probability value or the significance probability (Gibbons and Chakraborti, 2003)) is computed. By definition, the p-value is the probability of obtaining observations from the study that are within the normal range defined by the null hypothesis (Goodman, 1999). If this probability is high (higher than a pre-determined significance level of – for instance - 1% or 5%) (Stigler, 2008; Gentle, 2002), then the null hypothesis cannot be rejected as there is high probability that the observations are
Data catalogue | 269 within normal range and the null hypothesis might be true. While in contrast, if the p-value is below 5%, then the null hypothesis is rejected, because the observations are significantly outside the normal range defined by the null hypothesis. For example, if the global/divided city model requires – as an assumption - 85% of cities of the dataset to show increase in their spatial segregation, yet the observations confirmed that only 70% of cities show such increase, then the calculation of the p-value will determine whether that 70% is significant enough to oppose the global/divided city model, or it is within the normal range of the expected observations. One of the ways1 to determine the p-value is through performing the Pearson's Chi-squared test (χ2); which basically compares the expected counts of particular cases to their observed counts (Kirkman, 1996) where:
Where
Oi = the observed count of cases
Ei = the expected count of cases asserted by the null hypothesis n = the number of possible categories within the cases - for example, a coin flipping test has two possible categories, heads and tails. In our case cities are also categorized into two: cities with increase in segregation, and cities with decrease in segregation. After calculating the χ2 value, statistics textbooks - such as Gibbons and Chakraborti (2003) - already provide Chi-squared distribution tables that determine the equivalent p-value for the calculated χ2 for different values of n (see the subsequent chapters for more details about how the test is performed over the final dataset). In case that the p-value is less than the defined significance level (usually 5%), then the generalized global/divided city model is opposed, and the alternative hypothesis (H1) is supported.
270 | Appendix II
Where, (H1) suggests that local contextual differences are producing significantly divergent outcomes of restructuring for global cities in every context. The results in chapter 5 showed inconclusive patterns of change in socioeconomic segregation in global cities, more than half of cities of the dataset (54.5%) showed a tendency towards increased segregation, yet the only way to assert that the remaining (45.5%) of cities is significant enough to rule out any association between cities’ integration in the global network and their change in spatial segregation is through determining the ‘significance probability’ p-value of this null hypothesis. The process of determining the p-value is explained in the following steps:
1. Confirm that the ‘null hypothesis’ H0 is stated clearly, according to
the discussion in the preceding chapters, H0 states that ‘there is a tendency among global cities to show intensified patterns of spatial segregation due to the growing income inequality and social polarization within these cities as a direct outcome of their economic restructuring’. 2. Identify the expected results of the analysis in case the null hypothesis is true. In view of the above, the tendency among global cities to show increase in segregation level suggests that a ‘high percentage’ of cities of the dataset must show such increase. However, this ‘high percentage’ is not clearly defined, does it mean 95% or 80% or even 70% of cities of the dataset with increased segregation constitute a tendency? Therefore, since it is not an easy task to translate this ‘tendency’ into a definitive numerical value, then the χ2 is calculated several times based on the range of several possible expected observations suggested by the null hypothesis. See table 5-1 for more details 3. Calculate the χ2 value. 4. Determine the DF (degree of freedom), where DF = n -1, (n= the number of possible categories within the cases under study – as noted
Data catalogue | 271
in chapter 4, cities are categorized into two: cities with increase in segregation, and cities with decrease in segregation. Accordingly, DF = 2 - 1 = 1 5. Choose a significance level for p-value. By convention, scientists usually set the significance value for their experiments at 0.05, or 5 percent (Vaughan, 2001). If the P value is less than or equal to 0.05,
the decision is to reject H0; otherwise, the decision is not to reject H0 (Gibbons and Chakraborti, 2003). 6. Finally, in a Chi-squared distribution table (see table ii-1), each row in the table represents the corresponding χ2 values for different degrees of freedom, while the columns represents intervals of the p-value for each χ2 value. Given that the DF = 1, then by reading the row where DF equals 1 across from the left to the right until the first cell with a value bigger than the calculated χ2 value appears, the corresponding p-value appears in the top of this column.
Chi-squared distribution table As noted in step 2, the expected number of cities with increased segregation varies based on the definition of the word ‘tendency’ in the null hypothesis. Accordingly, χ2 is calculated when 95% of the dataset is 272 | Appendix II expected to show an increase in their socioeconomic segregation, then the χ2 is calculated again if this ration is expected to be 90%, 85%, 80%, 75%, or 70%. With each value for χ2 the p-value is determined and a decision about rejecting the null hypothesis or not is made. Table ii-2 shows the different χ2 value and their corresponding’s p-value. Note that the dataset contains total of 66 cities, the observations confirmed that only 36 cities showed an increase in their socioeconomic segregation index compared to 30 cities showed a decrease in their segregation index. The (36 to 30) ratio is compared to other expected ratios as follows:
Chi-squared value and their corresponding p-value χ2 value when expected Expected ratio ratio is Cities (of the of cities with compared dataset) with increased to to the tendency for cities with observation Null increased decreased (36 to 30) hypothesis segregation segregation cities p-value (H0) 95% 62 to 4 180 0.000 < 0.05 Rejected 90% 59 to 7 84.5 0.000 < 0.05 Rejected 85% 56 to 10 47.1 0.000 < 0.05 Rejected 80% 52 to 14 23.2 0.000 < 0.05 Rejected 75% 49 to 17 13.4 0.000 < 0.05 Rejected 70% 46 to 20 7.17 0.005 < 0.05 Rejected Noticeably, based on a significance value 0.05, the null hypothesis is rejected as the p-value is founded less than the significance value in every case shown in table ii-2 (whether the null hypothesis require 95% of cities of the dataset to show an increase in segregation or even only 70%). In other
Data catalogue | 273 words, the 45.5 % of cities with decreased segregation are significant enough to rule out any suggested association between cities’ integration in the global network of cities and the intensified socioeconomic segregation within them. The case of socioeconomic segregation required performing the Chi- squared test, as the observations were not conclusive on their own. However, in the case of ethnic segregation, the test is not necessary as it confirms the obvious, because only less than half of cities of the dataset showed an increased in ethnic segregation index, which suggests that there is no tendency among global cities to show an increase in their ethnic segregation as propagated in the world/global city literature.
Notes
1 The Pearson’s chi-squared test is recommended by Dr. Manal Nassar - Professor of statistics in Ain Shams University to be suitable for the requirement of this research, compared to other statistical hypothesis tests including: t-test, student’s test, and The Kolmogorov-Smirnov test (KS-test).
Appendix III
iii. Data catalogue
This appendix presents all cities of the dataset with respect to their available data, data type, data source, as well as a guide basemap showing the boundaries and the internal sub-city divisions of each case. Cities are listed in alphabetical order.
Data catalogue | 275
City: Adelaide Country: Australia Global status: gamma (+) Rank: 127 Types of divisions: statistical local areas Number of divisions: 54 Income based segregation: Type of data: Gross weekly individual income of population over 15 years of age Population categories: 0$, 1$ to 159$, 160$ to 399$, 400$ to 599$, 600$ to 799$, 800$ to 999$, 1000$ to 1549$, 1550$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: People born in UK, surrounding territories, Anglo America, Western Europe, Eastern Europe and Former SU, South/East Asia, Middle East, Africa, and elsewhere. Data source: 2001 Census of Population and Housing 2006 Census of Population and Housing Australian Bureau of Statistics
City: Amsterdam Country: The Netherlands Global status: alpha Rank: 21 Types of divisions: neighborhoods Number of divisions: 94 Income based segregation: Type of data: Households in top middle and bottom of the income structure Population categories: Bottom 40% Middle 40% Top 20% Ethnic based segregation: Type of data: Population ethnic origin Population categories: Surinamese, from Antil and Aruba, Turkish, Moroccan, other not European, other European, and natives Data source: Gemeente Amsterdam Dienst Onderzoek en Statistiek: Stadsdelen in cijfers 2000, 2002, 2003, 2011
276 | Appendix II
City: Antwerp Country: Belgium Global status: gamma(+) Rank: 118 Types of divisions: Wijken (quarters) Number of divisions: 62
Income based segregation: Type of data: No detailed data available, 2001 and 2008 income data are available on city level
Ethnic based segregation: Type of data: Population by nationality Population categories: Belgian, EU 15 countries, EU 12 countries, other European, Turkish, Moroccan, Africa, Asia, America/Oceania, and Refugees Data source: Stad Antwerpen Stad Antwerpen in Cijfers, 2000, 2006, 2012
City: Atlanta Country: USA Global status: Alpha (-) Rank: 40 Types of divisions: census tracts Number of divisions: 601 Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/Australia, Caribbean/Central America, South America, North America, and not foreign born population. Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 277
City: Auckland Country: New Zealand Global status: beta Rank: 71 Types of divisions: area units Number of divisions: 292 Income based segregation: Type of data: Total Personal Income, for the Census Usually Resident Population Count Aged 15 Years and Over Population categories: 5000$ or less, 5001$ to 10000, 10001$ to 20000$, 20001$ to 30000$, 30001$ to 50000$, 50001$ or more Ethnic based segregation: Type of data: Population by ethnic group Population categories: European, Maori, Pacific, Asian, MELLA, and other Data source: 1996 Census, 2000 Census, 2006 Census, Statistics New Zealand
City: Austin Country: USA Global status: gamma(-) Rank: 168 Types of divisions: census tracts Number of divisions: 204 Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/Australia, Caribbean/Central America, South America, North America, and not foreign born population. Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
278 | Appendix II
City: Baltimore Country: USA Global status: gamma(+) Rank: 121 Types of divisions: census tracts Number of divisions: 543 Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/Australia, Caribbean/Central America, South America, North America, and not foreign born population. Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Barcelona Country: Spain Global status: alpha (-) Rank: 41 Types of divisions: zones de recca Number of divisions: 248 Income based segregation: Type of data: No comparable data on income is available
Ethnic based segregation: Type of data: Population by nationality Population categories: People from Spain, other European countries, Africa, North America, Central America, South America, Asia, and Oceania Data source:
Ajuntament de Barcelona Xifres oficials de població a 1 de gener de 2003 y 2009
Data catalogue | 279
City: Beijing Country: China Global status: alpha (+) Rank: 12 Types of divisions: Districts Number of divisions: 12 Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status
Population categories: Foreigners Not foreigners Data source:
2010 Beijing sub-counties major statistical data directory.
City: Berlin Country: Germany Global status: beta(+) Rank: 56 Types of divisions: districts Number of divisions: 12 Income based segregation: Type of data: Personal monthly net income Population categories: Under 500€, 500€ to 899€, 900€ to 1299€, 1300€ to 1999€, 2000€ to 2599€, 2600€ and more Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners not Foreigners Data source:
Statistik Berlin: Results of the micro-census In Berlin in 2006, and 2010
280 | Appendix II
City: Bogota Country: Colombia Global status: beta(+) Rank: 62 Types of divisions: districts Number of divisions: 19 Income based segregation: Type of data: Population by income covers basic needs Population categories: People with income not enough to cover minimum needs People with income just enough to cover minimum needs people with income cover more than minimum needs Ethnic based segregation: Type of data: No data available
Data source:
DANE - DAPD, Quality of Life Survey 2003 Bogotá DANE - SDP, Quality of Life Survey 2007 Bogotá
City: Boston Country: USA Global status: alpha (-) Rank: 36 Types of divisions: census tracts Number of divisions: 895 Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/Australia, Caribbean/Central America, South America, North America, and not foreign born population. Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 281
City: Bratislava Country: Slovakia Global status: beta(-) Rank: 93 Types of divisions: districts Number of divisions: 5
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by nationality by continent. Population categories: Slovak, Russian, Hungarian, Gipsy, Czech, Ruthenian, Ukrainian, German, Polish, and others.
Data source: Statistical Office of the Slovak Republic – Regional Office of Bratislava
City: Brisbane Country: Australia Global status: beta(-) Rank: 87 Types of divisions: Statistical Local Areas Number of divisions: 66 Income based segregation: Type of data: Gross weekly individual income of population over 15 years of age Population categories: 0$, 1$ to 159$, 160$ to 399$, 400$ to 599$, 600$ to 799$, 800$ to 999$, 1000$ to 1549$, 1550$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: People born in UK, surrounding territories, Anglo America, Western Europe, Eastern Europe and Former SU, South/East Asia, Middle East, Africa, and elsewhere. Data source: 2001 Census of Population and Housing 2006 Census of Population and Housing Australian Bureau of Statistics
282 | Appendix II
City: Brussels Country: Belgium Global status: alpha Rank: 25 Types of divisions: municipalities Number of divisions: 19
Income based segregation: Type of data: Population by yearly taxable income
Population categories: Less than 10000€, 10001€ to 20000€ 20001€ to 30000€, 30001€ to 40000€ 40001€ to 50000€, 50001€ and more
Ethnic based segregation: Type of data: Population by nationality (broad regions)
Population categories: European, Turkish, North Africa, Sub-Sahara and West Africa, and others.
Data source: Nationaal Instituut voor de Statistiek. Algemene Directie Statistiek en Economische Informatie.
City: Calgary Country: Canada Global status: beta(-) Rank: 109 Types of divisions: neighborhoods Number of divisions: 145
Income based segregation: Type of data: Yearly income of population over 15 years of age
Population categories: No income, under 1000$, 1000$ to 2999$, 3000$ to 4999$, 5000$ to 6999$, 7000$ to 9999$, 10000$ to 11999$, 12000$ to 14999$, 15000$ to 19999$, 20000$ to 24999, 25000$ to 29999$, 30000$ to 34999$, 35000$ to 39999$, 40000$ to 44999$, 45000$ to 49999$, 50000$ to 59999$, 60000 and over.
Ethnic based segregation: Type of data: Population by visible minority group
Population categories: Not minority, South Asian, Chinese, Japanese, Black, Latino, Arab and West Asian, multiple race, and not classified
Data source: 2001 Census – 2006 Census: Statistics Canada.
Data catalogue | 283
City: Cape Town Country: South Africa Global status: beta Rank: 79 Types of divisions: wards Number of divisions: 100
Income based segregation: Type of data: Population by their individual monthly income Population categories: No income, 1$ to 6400$, 6401$ to 12800$, 12801 to 51200$, 51201$ to 102400$, 102401$ to 204800$, Over 204800
Ethnic based segregation: Type of data: Population by race Population categories: Black, White, Colored, and Indian
Data source: Statistics South Africa: 1996 census - 2001 census.
City: Charlotte Country: USA Global status: gamma(+) Rank: 117 Types of divisions: census tracts Number of divisions: 161
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
284 | Appendix II
City: Chicago Country: USA Global status: alpha (+) Rank: 8 Types of divisions: census tracts Number of divisions: 1912
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Cincinnati Country: USA Global status: gamma(+) Rank: 116 Types of divisions: census tracts Number of divisions: 392 Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/Australia, Caribbean/Central America, South America, North America, and not foreign born population. Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 285
City: Cleveland Country: USA Global status: beta(-) Rank: 107 Types of divisions: census tracts Number of divisions: 615 Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/Australia, Caribbean/Central America, South America, North America, and not foreign born population. Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Cologne Country: Germany Global status: beta(-) Rank: 99 Types of divisions: district Number of divisions: 85
Income based segregation: Type of data: No data available.
Ethnic based segregation: Type of data: Population by nationality by continent Population categories: German, African, Asian, other European, North American, Australian, and Turkish
Data source: City of Cologne - Department of City Development and Statistics (Statistical Information System). Kölner Stadtteilinformationen .
286 | Appendix II
City: Columbus Country: USA Global status: gamma Rank: 149 Types of divisions: census tracts Number of divisions: 292
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Copenhagen Country: Denmark Global status: beta(+) Rank: 60 Types of divisions: district Number of divisions: 15
Income based segregation: Type of data:
Households by taxable income
Population categories: Less than 49.999kr, 50000kr to 99999kr, 100000kr to 149999kr, 150000kr to 199999kr, 200000kr to 299999kr, 300000kr to 399999kr, 400000kr to 499999kr, 500000kr to 599999kr, 600000kr to 699999kr, 700000kr and more.
Ethnic based segregation: Type of data: Population by ethnic origin
Population categories: Denmark, other Europe, Africa North America , South and central America, Asia, Oceania, stateless, and unknown
Data source: Statistics Copenhagen, Copenhagen City - www.kk.dk/statistik
Data catalogue | 287
City: Dallas Country: USA Global status: alpha(-) Rank: 38 Types of divisions: census tracts Number of divisions: 901
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Denver Country: USA Global status: beta(-) Rank: 91 Types of divisions: census tracts Number of divisions: 484
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
288 | Appendix II
City: Detroit Country: USA Global status: beta(-) Rank: 90 Types of divisions: census tracts Number of divisions: 1148
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Durban Country: South Africa Global status: gamma(-) Rank: 167 Types of divisions: wards Number of divisions: 103
Income based segregation: Type of data: Population by their individual monthly income Population categories: No income, 1$ to 6400$, 6401$ to 12800$, 12801 to 51200$, 51201$ to 102400$, 102401$ to 204800$, Over 204800
Ethnic based segregation: Type of data: Population by race Population categories: Black, White, Colored, and Indian
Data source: Statistics South Africa: 1996 census - 2001 census.
Data catalogue | 289
City: Dusseldorf Country: Germany Global status: beta(+) Rank: 48 Types of divisions: districts Number of divisions: 49
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status Population categories: German, and foreigners.
Data source: Landeshauptstadt Düsseldorf. (city of Dusseldorf). Landesamt für Datenverarbeitung und Statistik NRW. (State Office for Data Processing and Statistics NRW).
City: Edmonton Country: Canada Global status: gamma Rank: 152 Types of divisions: neighborhoods Number of divisions: 210
Income based segregation: Type of data: Yearly income of population over 15 years of age
Population categories: No income, under 1000$, 1000$ to 2999$, 3000$ to 4999$, 5000$ to 6999$, 7000$ to 9999$, 10000$ to 11999$, 12000$ to 14999$, 15000$ to 19999$, 20000$ to 24999, 25000$ to 29999$, 30000$ to 34999$, 35000$ to 39999$, 40000$ to 44999$, 45000$ to 49999$, 50000$ to 59999$, 60000 and over.
Ethnic based segregation: Type of data: Population by visible minority group
Population categories: Not minority, South Asian, Chinese, Japanese, Black, Latino, Arab and West Asian, multiple race, and not classified
Data source: 2001 Census – 2006 Census: Statistics Canada.
290 | Appendix II
City: Frankfurt Country: Germany Global status: alpha Rank: 19 Types of divisions: neighborhoods Number of divisions: 45
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status Population categories: German, and foreigners.
Data source: Statistisches Jahrbuch Frankfurt am Main: 2007, and 2011. Stadt Frankfurt am Main Bürgeramt, Statistik und Wahlen.
City: Geneva Country: Switzerland Global status: beta(-) Rank: 88 Types of divisions: municipality Number of divisions: 16
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners, and not foreigners.
Data source: Office cantonal de la statistique – OCSTAT: Statistique Genève.
Data catalogue | 291
City: Gothenburg Country: Sweden Global status: gamma(-) Rank: 158 Types of divisions: district Number of divisions: 94
Income based segregation: Type of data: Individuals disposable income in thousands kr.
Population categories: No income, 0.1 kr to 79 kr, 80 kr to 159 kr, 160 kr to 239 kr, 240 kr, 359 kr and more.
Ethnic based segregation: Type of data: Foreign-born population
Population categories: Swedish, other Scandinavian, other European, Asia, Africa, North America, South American, and others.
Data source: Statistiska centralbyrån (SCB). INK3 Göteborg.
City: Guadalajara Country: Mexico Global status: gamma Rank: 139 Types of divisions: municipalities Number of divisions: 8
Income based segregation: Type of data: Employed population by their income Population categories: Individuals earning less income than the minimum wage, individual earning the minimum wage to double the minimum wages, individual earning more than double the minimum wage, and individual whose income is not specified
Ethnic based segregation: Type of data: No data available
Data source: INEGI: XII Censo general de poblacion y vivienda 2000. INEGI: Censo de Población y Vivienda 2010.
292 | Appendix II
City: Johannesburg Country: South Africa Global status: alpha(-) Rank: 47 Types of divisions: wards Number of divisions: 109
Income based segregation: Type of data: Population by their individual monthly income Population categories: No income, 1$ to 6400$, 6401$ to 12800$, 12801 to 51200$, 51201$ to 102400$, 102401$ to 204800$, Over 204800
Ethnic based segregation: Type of data: Population by race Population categories: Black, White, Colored, and Indian
Data source: Statistics South Africa: 1996 census - 2001 census.
City: Hamburg Country: Germany Global status: beta(+) Rank: 53 Types of divisions: neighborhoods Number of divisions: 98
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status Population categories: German, and foreigners.
Data source: Statistisches Amt für Hamburg und Schleswig- Holstein (Statistical Office of Hamburg and Schleswig-Holstein).
Data catalogue | 293
City: Hong Kong Country: China Global status: alpha(+) Rank: 3 Types of divisions: statistical divisions Number of divisions: 128
Income based segregation: Type of data: Monthly income of individuals over 15 years of age
Population categories: No income, under 2000$, 2000$ to 3999$, 4000 to 5999$, 6000$ to 7999$, 8000$ to 9999$, 10000$ to 14999$, 15000$ to 19999$, 20000$ to 24999$, 25000 to 39999$, 40000$ and more.
Ethnic based segregation: Type of data: Population by origin Population categories: Chinese, other Asian, and others.
Data source: Census and statistics department: The Government of the Hong Kong Special Administrative Region.
City: Houston Country: USA Global status: beta(+) Rank: 55 Types of divisions: census tracts Number of divisions: 731
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
294 | Appendix II
City: Indianapolis Country: USA Global status: gamma Rank: 150 Types of divisions: census tracts Number of divisions: 275
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Kansas City Country: USA Global status: gamma Rank: 136 Types of divisions: census tracts Number of divisions: 423
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 295
City: London Country: UK Global status: alpha(++) Rank: 1 Types of divisions: boroughs Number of divisions: 32
Income based segregation: Type of data: No comparable data available.
Ethnic based segregation: Type of data: Population by ethnic group Population categories: Whites, Black African, Black Caribbean, other Blacks, Indians, Pakistanis, Bangladeshis, Chinese, other Asians, and others (including mixed ethnicity).
Data source: London Borough Profiles: Greater London Authority.
City: Los Angeles Country: USA Global status: alpha Rank: 17 Types of divisions: census tracts Number of divisions: 2295
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
296 | Appendix II
City: Lyon Country: France Global status: gamma Rank: 140 Types of divisions: commune Number of divisions: 138
Income based segregation: Type of data: No data available.
Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners , and not foreigners.
Data source: Recensement de la population de 1982, 1990, 1999 et 2008: Insee, Institut national de la statistique et des etudes economiques.
City: Madrid Country: Spain Global status: alpha Rank: 15 Types of divisions: barrios Number of divisions: 128
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners, and not foreigners.
Data source: Madrid Datos: Direccion General de Estadistica (Madrid info: General Bureau of Statistics)
Data catalogue | 297
City: Manchester Country: UK Global status: beta Rank: 76 Types of divisions: wards Number of divisions: 32
Income based segregation: Type of data: No comparable data available
Ethnic based segregation: Type of data: Population by ethnic groups
Population categories: Whites, Black African, Black Caribbean, other Blacks, Indians, Pakistanis, Bangladeshis, Chinese, other Asians, and others (including mixed ethnicity).
Data source: Manchester City Council. Office of National Statistics: Neighborhood profile.
City: Marseille Country: France Global status: gamma(-) Rank: 159 Types of divisions: commune Number of divisions: 67
Income based segregation: Type of data: No data available.
Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners , and not foreigners.
Data source: Recensement de la population de 1982, 1990, 1999 et 2008: Insee, Institut national de la statistique et des etudes economiques.
298 | Appendix II
City: Melbourne Country: Australia Global status: alpha(-) Rank: 31 Types of divisions: statistical local areas Number of divisions: 57
Income based segregation: Type of data: Gross weekly individual income of population over 15 years of age
Population categories: 0$, 1$ to 159$, 160$ to 399$, 400$ to 599$, 600$ to 799$, 800$ to 999$, 1000$ to 1549$, 1550$ or more
Ethnic based segregation: Type of data: Population by birthplace
Population categories: Individuals born in the UK, surrounding territories, Anglo America, Western Europe, Eastern Europe and former SU, South/East Asia, Middle East, Africa, and elsewhere
Data source: 2001 Census of Population and Housing 2006 Census of Population and Housing Australian Bureau of Statistics
City: Mexico city Country: Mexico Global status: alpha Rank: 20 Types of divisions: municipalities Number of divisions: 50
Income based segregation: Type of data: Employed population by their income Population categories: Individuals earning less income than the minimum wage, individual earning the minimum wage to double the minimum wages, individual earning more than double the minimum wage, and individual whose income is not specified
Ethnic based segregation: Type of data: No data available
Data source: INEGI: XII Censo general de poblacion y vivienda 2000. INEGI: Censo de Población y Vivienda 2010.
Data catalogue | 299
City: Miami Country: USA Global status: alpha (-) Rank: 29 Types of divisions: census tracts Number of divisions: 878
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Milan Country: Italy Global status: alpha(+) Rank: 11 Types of divisions: Nuclei di Identità Locale Number of divisions: 88
Income based segregation: Type of data: No comparable data available
Ethnic based segregation: Type of data: Population by citizenship status
Population categories: Italian, and not Italian
Data source: Comune di Milano - Settore Statistica e S.I.T. - Servizio Statistica. (City of Milan - Statistics Service)
300 | Appendix II
City: Milwaukee Country: USA Global status: gamma(-) Rank: 174 Types of divisions: census tracts Number of divisions: 396
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Minneapolis Country: USA Global status: beta Rank: 82 Types of divisions: census tracts Number of divisions: 643
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 301
City: Monterrey Country: Mexico Global status: beta(-) Rank: 92 Types of divisions: municipalities Number of divisions: 11
Income based segregation: Type of data: Employed population by their income Population categories: Individuals earning less income than the minimum wage, individual earning the minimum wage to double the minimum wages, individual earning more than double the minimum wage, and individual whose income is not specified
Ethnic based segregation: Type of data: No data available
Data source: INEGI: XII Censo general de poblacion y vivienda 2000. INEGI: Censo de Población y Vivienda 2010.
City: Montréal Country: Canada Global status: beta(+) Rank: 51 Types of divisions: neighborhoods Number of divisions: 81
Income based segregation: Type of data: Yearly income of population over 15 years of age
Population categories: No income, under 1000$, 1000$ to 2999$, 3000$ to 4999$, 5000$ to 6999$, 7000$ to 9999$, 10000$ to 11999$, 12000$ to 14999$, 15000$ to 19999$, 20000$ to 24999, 25000$ to 29999$, 30000$ to 34999$, 35000$ to 39999$, 40000$ to 44999$, 45000$ to 49999$, 50000$ to 59999$, 60000 and over.
Ethnic based segregation: Type of data: Population by visible minority group
Population categories: Not minority, South Asian, Chinese, Japanese, Black, Latino, Arab and West Asian, multiple race, and not classified
Data source: 2001 Census – 2006 Census: Statistics Canada.
302 | Appendix II
City: Munich Country: Germany Global status: alpha(-) Rank: 34 Types of divisions: districts Number of divisions: 25
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by nationality Population categories: German, Turkish, other EU nationals, East European, and other nationalities
Data source: Statistisches Amt Munchen
City: New York Country: USA Global status: alpha (++) Rank: 2 Types of divisions: census tracts Number of divisions: 4429
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 303
City: Orlando Country: USA Global status: gamma(-) Rank: 157 Types of divisions: census tracts Number of divisions: 262
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Oslo Country: Norway Global status: beta Rank: 72 Types of divisions: commune Number of divisions: 16
Income based segregation: Type of data: population by gross income of tax payers 14 years of age and older. Population categories: No income, 100NOK to 49900NOK, 50000 to 99900, 100000 to 149900, 150000 to 199900, 200000 to 249900, 250000 to 299900, 300000 to 349900, 400000 to 449900, 450000 to 499900, 500000NOK or more.
Ethnic based segregation: Type of data: Population by continent of origin Population categories: Nordic, Western European, Eastern European, North America and Oceania, and others including: Asia, Africa, and Latin America.
Data source: Statistisk sentralbyrå/statistisk kontor (Statistics Norway - statistical office)
304 | Appendix II
City: Ottawa Country: Canada Global status: gamma(-) Rank: 160 Types of divisions: neighborhoods Number of divisions: 235
Income based segregation: Type of data: Yearly income of population over 15 years of age
Population categories: No income, under 1000$, 1000$ to 2999$, 3000$ to 4999$, 5000$ to 6999$, 7000$ to 9999$, 10000$ to 11999$, 12000$ to 14999$, 15000$ to 19999$, 20000$ to 24999, 25000$ to 29999$, 30000$ to 34999$, 35000$ to 39999$, 40000$ to 44999$, 45000$ to 49999$, 50000$ to 59999$, 60000 and over.
Ethnic based segregation: Type of data: Population by visible minority group
Population categories: Not minority, South Asian, Chinese, Japanese, Black, Latino, Arab and West Asian, multiple race, and not classified
Data source: 2001 Census – 2006 Census: Statistics Canada.
City: Panama City Country: Panama Global status: beta(-) Rank: 102 Types of divisions: districts Number of divisions: 21
Income based segregation: Type of data: Employed population by their income Population categories: No income, less than 100, 100 to 124, 125 to 174, 175 to 249, 250 to 399, 400 to 599, 600 to 799, 800 to 999, 1000 to1499, 1500 to 1999, 2000 to 2499, 2500 to 2999, 3000 to 3999, 4000 to 4999, 5000 and more, and not stated.
Ethnic based segregation: Type of data: No data available
Data source: Contraloria General de la Republica de Panama: Instituto Nacional de Estadistica y Censo. INEC.
Data catalogue | 305
City: Paris Country: France Global status: alpha(+) Rank: 4 Types of divisions: commune Number of divisions: 134
Income based segregation: Type of data: No data available.
Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners , and not foreigners.
Data source: Recensement de la population de 1982, 1990, 1999 et 2008: Insee, Institut national de la statistique et des etudes economiques.
City: Perth Country: Australia Global status: beta(-) Rank: 105 Types of divisions: statistical local areas Number of divisions: 28
Income based segregation: Type of data: Gross weekly individual income of population over 15 years of age
Population categories: 0$, 1$ to 159$, 160$ to 399$, 400$ to 599$, 600$ to 799$, 800$ to 999$, 1000$ to 1549$, 1550$ or more
Ethnic based segregation: Type of data: Population by birthplace
Population categories: Individuals born in the UK, surrounding territories, Anglo America, Western Europe, Eastern Europe and former SU, South/East Asia, Middle East, Africa, and elsewhere
Data source: 2001 Census of Population and Housing 2006 Census of Population and Housing Australian Bureau of Statistics
306 | Appendix II
City: Philadelphia Country: USA Global status: alpha(-) Rank: 46 Types of divisions: census tracts Number of divisions: 1360
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Phoenix Country: USA Global status: gamma Rank: 137 Types of divisions: census tracts Number of divisions: 630
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 307
City: Pittsburgh Country: USA Global status: gamma Rank: 151 Types of divisions: census tracts Number of divisions: 599
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Portland Country: USA Global status: gamma(+) Rank: 129 Types of divisions: census tracts Number of divisions: 369
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
308 | Appendix II
City: Richmond Country: USA Global status: gamma(-) Rank: 164 Types of divisions: census tracts Number of divisions: 228
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Rio de Janiero Country: Brazil Global status: beta(-) Rank: 86 Types of divisions: municipal district Number of divisions: 30
Income based segregation: Type of data: Employed population by their income Population categories: Individuals with no income, individuals earning less income than half the minimum wage, individuals earning 1 to 2 the minimum wage, individual earning 2 to 3 the minimum wage, individual earning 3 to 5 the minimum wage, individual earning 5 to 10 the minimum wage, , individual earning 10 to 15 the minimum wage, individual earning 15 to 20 the minimum wage, individual earning more than 20 the minimum wage.
Ethnic based segregation: Type of data: No data available
Data source: IBGE: Instituto Brasileiro de Geografia e Estatistica
Data catalogue | 309
City: Rome Country: Italy Global status: beta(+) Rank: 52 Types of divisions: municipalities Number of divisions: 20
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status
Population categories: Italian, and not Italian
Data source: Ufficio di Statistica e Censimento del Comune di Roma. Bureau of Statistics and Census of the City of Roma
City: Rotterdam Country: The Netherlands Global status: gamma Rank: 147 Types of divisions: neighborhoods Number of divisions: 92
Income based segregation: Type of data: Households by income strata
Population categories: Household in the bottom 40% of the income distribution, household in the middle 40%, and household in the top 20% of the income distribution.
Ethnic based segregation: Type of data: Population by ethnic origin
Population categories: Natives, other European, Surinamese, Turkish, Moroccan, Anthill & Aruba, Cape Verde, other Western, and other non-Western.
Data source: Gemeente Rotterdam, Centrum voor Onderzoek en Statistiek.
310 | Appendix II
City: San Diego Country: USA Global status: beta(-) Rank: 103 Types of divisions: census tracts Number of divisions: 487
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: San Francisco Country: USA Global status: alpha Rank: 27 Types of divisions: census tracts Number of divisions: 724
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 311
City: San Jose Country: USA Global status: gamma(+) Rank: 133 Types of divisions: census tracts Number of divisions: 315
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Sao Paulo Country: Brazil Global status: alpha Rank: 14 Types of divisions: neighborhood Number of divisions: 8
Income based segregation: Type of data: Employed population by their income Population categories: Individuals earning less income than 3 the minimum wage, individual earning 3 to 5 the minimum wage, individual earning 5 to 10 the minimum wage, individual earning 10 to 20 the minimum wage, individual earning more than 20 the minimum wage,
Ethnic based segregation: Type of data: No data available
Data source: IBGE: Instituto Brasileiro de Geografia e Estatistica
312 | Appendix II
City: Seattle Country: USA Global status: beta Rank: 68 Types of divisions: census tracts Number of divisions: 608
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Seoul Country: South Korea Global status: alpha Rank: 24 Types of divisions: statistical divisions Number of divisions: 25
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by ethnic origin Population categories: China and Taiwan, North America, Europe and Russia, Central and South Asia, East and Southeast Asia, and others.
Data source: Seoul Metropolitan Government: Seoul Statistical Yearbook.
Data catalogue | 313
City: Shanghai Country: China Global status: alpha(+) Rank: 7 Types of divisions: districts Number of divisions: 12
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population immigration status Population categories: Foreign Not foreign
Data source: Shanghai Statistics Bureau
City: Singapore Country: Republic of Singapore Global status: alpha(+) Rank: 5 Types of divisions: planning zones Number of divisions: 35
Income based segregation: Type of data: Monthly income of population over 15 years of age
Population categories: Under 1000$, 1000$ to 1499$, 1500$ to 1999$, 2000$ to 2499, 2500$ to 2999$, 3000$ to 3999$, 40004 to 4999$, 5000$ to 5999$, 6000$ and over.
Ethnic based segregation: Type of data: Population by ethnic group
Population categories: Chinese, Malay, Indian, and others.
Data source: Census of Population 2000 - 2010 Singapore Department of Statistics.
314 | Appendix II
City: St Louis Country: USA Global status: beta(-) Rank: 100 Types of divisions: census tracts Number of divisions: 446
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Stockholm Country: Sweden Global status: beta(+) Rank: 49 Types of divisions: district Number of divisions: 128
Income based segregation: Type of data: Individuals disposable income in thousands kr.
Population categories: No income, 0.1 kr to 120 kr, 120 kr to 240 kr, 240 kr to 360 kr, 360 kr and more.
Ethnic based segregation: Type of data: Foreign-born population
Population categories: Swedish, other Scandinavian, other European, Asia, Africa, North America, South American, and others.
Data source: Stockholms stads Utrednings- och statistikkontor. (Stockholm office for Research and Statistics).
Data catalogue | 315
City: Stuttgart Country: Germany Global status: beta(-) Rank: 97 Types of divisions: statistical zones Number of divisions: 23
Income based segregation: Type of data: No data available.
Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners , and not foreigners.
Data source: Landeshauptstadt Stuttgart, Statistisches Amt (City of Stuttgart, Statistical Office).
City: Sydney Country: Australia Global status: alpha(+) Rank: 10 Types of divisions: statistical local areas Number of divisions: 24
Income based segregation: Type of data: Gross weekly individual income of population over 15 years of age
Population categories: 0$, 1$ to 159$, 160$ to 399$, 400$ to 599$, 600$ to 799$, 800$ to 999$, 1000$ to 1549$, 1550$ or more
Ethnic based segregation: Type of data: Population by birthplace
Population categories: Individuals born in the UK, surrounding territories, Anglo America, Western Europe, Eastern Europe and former SU, South/East Asia, Middle East, Africa, and elsewhere
Data source: 2001 Census of Population and Housing 2006 Census of Population and Housing Australian Bureau of Statistics
316 | Appendix II
City: Tallinn Country: Estonia Global status: gamma(-) Rank: 153 Types of divisions: districts Number of divisions: 8
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by ethnic groups Population categories: Natives, Russian, Ukrainian, Belarus, Finnish, Jews, Tatars, and others
Data source: Tallinna Linnavalitsus (Tallinn City Government). Estonian Ministry of the Interior: Population Register.
City: Tampa Country: USA Global status: gamma Rank: 148 Types of divisions: census tracts Number of divisions: 491
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
Data catalogue | 317
City: Tokyo Country: Japan Global status: alpha(+) Rank: 6 Types of divisions: wards Number of divisions: 23
Income based segregation: Type of data: No data available on ward level.
Ethnic based segregation: Type of data: Population by ethnic minority Population categories: population belongs to minority group, and native population (data does not include undocumented migrants)
Data source: Statistics Bureau, Ministry of Internal Affairs and Communications. Bureau of General Affairs: Tokyo Metropolitan Government.
City: Toronto Country: Canada Global status: alpha Rank: 13 Types of divisions: neighborhoods Number of divisions: 874
Income based segregation: Type of data: Yearly income of population over 15 years of age
Population categories: No income, under 1000$, 1000$ to 2999$, 3000$ to 4999$, 5000$ to 6999$, 7000$ to 9999$, 10000$ to 11999$, 12000$ to 14999$, 15000$ to 19999$, 20000$ to 24999, 25000$ to 29999$, 30000$ to 34999$, 35000$ to 39999$, 40000$ to 44999$, 45000$ to 49999$, 50000$ to 59999$, 60000 and over.
Ethnic based segregation: Type of data: Population by visible minority group
Population categories: Not minority, South Asian, Chinese, Japanese, Black, Latino, Arab and West Asian, multiple race, and not classified
Data source: 2001 Census – 2006 Census: Statistics Canada.
318 | Appendix II
City: Valencia Country: Spain Global status: gamma Rank: 135 Types of divisions: districts Number of divisions: 19
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by nationality by continent. Population categories: Spanish, other European, African, North and Central America, South America, Asia, and others.
Data source: Padró Municipal d’Habitants 2001 y 2010: Oficina d’Estadística. Ajuntament de València. (Municipal Population Census 2001 and 2010: Statistical Office. City of Valencia.)
City: Vancouver Country: Canada Global status: beta(+) Rank: 63 Types of divisions: neighborhoods Number of divisions: 375
Income based segregation: Type of data: Yearly income of population over 15 years of age
Population categories: No income, under 1000$, 1000$ to 2999$, 3000$ to 4999$, 5000$ to 6999$, 7000$ to 9999$, 10000$ to 11999$, 12000$ to 14999$, 15000$ to 19999$, 20000$ to 24999, 25000$ to 29999$, 30000$ to 34999$, 35000$ to 39999$, 40000$ to 44999$, 45000$ to 49999$, 50000$ to 59999$, 60000 and over.
Ethnic based segregation: Type of data: Population by visible minority group
Population categories: Not minority, South Asian, Chinese, Japanese, Black, Latino, Arab and West Asian, multiple race, and not classified
Data source: 2001 Census – 2006 Census: Statistics Canada.
Data catalogue | 319
City: Washington Country: USA Global status: alpha Rank: 28 Types of divisions: census tracts Number of divisions: 865
Income based segregation: Type of data: yearly earnings for population 16 years old and over Population categories: 1$ to 9999$, 10000$ to 19999$, 20000$ to 29999$, 30000$ to 39999$, 40000$ to 49999$, 50000$ to 64999$, 65000$ to 74999$, 75000$ to 99999$, 100000$ or more
Ethnic based segregation: Type of data: Population by birthplace Population categories: North/West Europe, South/East Europe, East Asia, South/Central Asia, South/East Asia, Middle East, East/Middle Africa, South/West Africa, Oceania/ Australia, Caribbean/Central America, South America, North America, and not foreign born.
Data source: 2000 US census - 2009 American Community Survey: US Census Bureau
City: Wellington Country: New Zealand Global status: gamma(-) Rank: 175 Types of divisions: wards Number of divisions: 158
Income based segregation: Type of data: Total Personal Income, for the Census Usually Resident Population Count Aged 15 Years and over Population categories: 5000$ or less, 5001$ to 10000, 10001$ to 20000$, 20001$ to 30000$, 30001$ to 50000$, 50001$ or more
Ethnic based segregation: Type of data: Population by ethnic group Population categories: European, Maori, Pacific, Asian, MELLA, and other
Data source: 1996 Census, 2000 Census, 2006 Census, Statistics New Zealand.
320 | Appendix II
City: Zurich Country: Switzerland Global status: alpha(-) Rank: 32 Types of divisions: neighborhoods Number of divisions: 34
Income based segregation: Type of data: No data available
Ethnic based segregation: Type of data: Population by citizenship status Population categories: Foreigners, and not foreigners.
Data source: Statistik Stadt Zürich.
References and bibliography
Aalbers, M. B., & Holm, A. (2008). Privatising social housing in Europe: The cases of Amsterdam and Berlin. In K. Adelhof, B. Glock, J. Lossau, & M. Schulz (Eds.), Urban trends in Berlin and Amsterdam (pp. 12-23). Berlin: Berliner Geographische Arbeiten, Humboldt Universität zu Berlin. Abrahamson, M. (2004). Global cities. New York: Oxford University Press. Andersen, H. T. (2002). Globalization, spatial polarization, and the housing market. Geografisk Tidsskrift: Danish Journal of Geography, 102, 93-102. Anderson, J. (1996). The shifting stage of politics: new medieval and postmodern territorialities?. Environment and Planning D, 14, 133-154. Andersson, R. (2007). Ethnic residential segregation and integration processes in Sweden. In K. Schönwälder (Ed.), Residential segregation and the integration of immigrants Britain, the Netherlands and Sweden (pp. 61-90). Berlin: WZB. Arthurson, K. (1998). Redevelopment of public housing estates: The Australian experience. Urban Policy and Research, 16(1), 35-46. Arthurson, K. (2012). Social mix and the city: Challenging the mixed communities consensus in housing and urban planning policies. Collingwood, Vic.: CSIRO Publishing. Atkinson, A. (1970). On the measurement of inequality. Journal of Economic Theory, 2(3), 244-263. doi: 10.1016/0022-0531(70)90039-6 Atkinson, R., & Bridge, G. (2005). Gentrification in a global context: The new urban colonialism. London: Routledge. Audirac, I., Cunningham-Sabot, E., Fol, S., & Moraes, S. T. (2012). Declining suburbs in Europe and Latin America. International Journal of Urban and Regional Research, 36(2), 226-244. doi: 10.1111/j.1468-2427.2011.01093.x Badcock, B. (1997). Recently observed polarising tendencies and Australian cities. Australian Geographical Studies, 35(3), 243-259. doi: 10.1111/1467- 8470.00025 Badcock, B., (1999): Doing more with less: public housing in the 1990s. In O’Connor, K. (ed.) Houses and Jobs in Cities and Regions: Research in Honour of Chris Maher. University of Queensland Press, Brisbane, 81–95. Badcock, B. (2000). The imprint of the post-fordist transition on Australian cities. In P. Marcuse & R. V. Kempen (Eds.), Globalizing cities: A new spatial order? (pp. 211-227). Oxford: Blackwell.
322 | References and bibliography
Barbosa, N. (2010). Latin America: Counter-cyclical policy in Brazil: 2008-09. Journal of Globalization and Development, 1(1). doi: 10.2202/1948- 1837.1052 Baum, S. (2008). Suburban scars: Australian cities and socio-economic deprivation (Research Paper 15). Urban Research Program, Griffith University. Retrieved November 9, 2012, from http://www.griffith.edu.au/__data/assets/pdf_file/0017/53009/urp-rp15- baum-2008.pdf Bauman, Z. (1998). On glocalization: or globalization for some, localization for some others. Thesis Eleven, 54(1), 37-49. Beall, J. (2002). Globalization and social exclusion in cities: Framing the debate with lessons from Africa and Asia. Environment and Urbanization, 14(1), 41-51. Beall, J., Crankshaw, O., & Parnell, S. (2000). Victims, villains and fixers: The urban environment and Johannesburg’s poor. Journal of Southern African Studies, 26(4), 803-855. Beauregard, R. A., & Deitrick, S. (1995). From front runner to also-run: The transformation of once-dominant region, Pennsylvania, USA. In P. Cooke (Ed.), The rise of the rustbelt (pp. 52-71). New York: St. Martin's Press. Beauregard, R. A., & Haila, A. (2000). The unavoidable continuities of the city. In P. Marcuse & R. Van Kempen (Eds.), Globalizing cities: A new spatial order? (pp. 22-36). Oxford: Blackwell. Beaverstock, J. V., Smith, R. G., & Taylor, P. J. (1999). A roster of world cities. cities, 16(6), 445-458. Beaverstock, J. V., Smith, R. G., & Taylor, P. J. (2000). World-city network: A new metageography? Annals of the Association of American Geographers, 90(1), 123-134. doi: 10.1111/0004-5608.00188 Beaverstock, J. V., Faulconbridge, J. R., & Hoyler, M. (2011). Globalization and the city. The SAGE Handbook of Economic Geography, 189-201. Beaverstock, J. V. (2011). Review of global urban analysis: a survey of cities in globalization, edited by Peter J. Taylor: Pengfei Ni, Ben Derudder, Michael Hoyler, Jin Huang and Frank Witlox, London, Earthscan, 2011, Urban
Research & Practice, 4(2), 219-221. Beck, U. (2000). The brave new world of work. Cambridge: Polity Press. Behr, J. G. (2004). Race, ethnicity, and the politics of city redistricting: Minority- opportunity districts and the election of Hispanics and Blacks to city councils. Albany: State University of New York Press. Bell, W. (1954). A probability model for the measurement of ecological segregation. Social Forces, 32, 337-364.
References and bibliography | 323
Bluestone, B., & Harrison, B. (1982). The deindustrialization of America: Plant closings, community abandonment, and the dismantling of basic industry. New York: Basic Books. Boal, F. W. (2005). Urban ethnic segregation and scenarios spectrum. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 62-78). Albany, NY: State University of New York Press. Bolt, G., Van Kempen, R., & Van Weesep, J. (2009). After urban restructuring: Relocations and segregation In Dutch cities. Tijdschrift Voor Economische En Sociale Geografie, 100(4), 502-518. doi: 10.1111/j.1467- 9663.2009.00555.x Bornschier, V. (2008, June). Income Inequality in the World–Looking Back and Ahead. In conference on “Inequality Beyond Globalisation” organised by the World Society Foundation and the RC02 of the International Sociological Association, University of Neuchatel (Vol. 28). Boter, I., Comabella, J., Lles, C. & Tobio,C. (1988). Socially transversal processes in a context of an increasingly differential social structure, Madrid, 1975- 1986. Paper presented at the Conference of the Research Committee on the Sociology of Urban and Regional Development, of the International Sociological Association, Rio de Janeiro, September. Brenner, N. (1998). Global cities, glocal states: Global city formation and state territorial restructuring in contemporary Europe. Review of International Political Economy, 5(1), 1-37. doi: 10.1080/096922998347633 Brenner, N. (2002). Decoding the Newest “metropolitan regionalism” in the USA: A critical overview. Cities, 19(1), 3-21. Brenner, N. (2004). New state spaces: Urban governance and the rescaling of statehood. Oxford: Oxford University Press. Briggs, X. (2005). Conclusion: Desegregating the city. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 233-257). Albany, NY: State university of New York press. Bråmå, Å. (2008). Dynamics of ethnic residential segregation in Göteborg, Sweden, 1995–2000. Population, Space and Place, 14(2), 101-117. doi: 10.1002/psp.479 Brandsen, T., Van de Donk, W., & Putters, K. (2005). Griffins or chameleons? Hybridity as a permanent and inevitable characteristic of the third sector. International Journal of Public Administration, 28(9-10), 749-765. doi: 10.1081/PAD-200067320 Bridge, G., & Watson, S. (2003). City differences. In G. Bridge & S. Watson (Eds.), A companion to the city (pp. 251-260). Oxford, UK: Blackwell.
324 | References and bibliography
Brussels Instituut voor STATISTIEK en Analyse. (2005). Buurtatlas van de bevolking van het Brussels Hoofdstedelijk Gewest bij de aanvang van de 21 eeuw (Nr.42) (Belgium, Ministerie van het Brussels Hoofdstedelijk Gewest). Burgers, J., & Engbersen, G. (1996). Globalization, migration, and undocumented migrants. New Community, 22(4), 619-635. Burgers, J., & Musterd, S. (2002). Understanding urban inequality: A model based on existing theories and an empirical illustration. International Journal of Urban and Regional Research, 26(2), 403-413. doi: 10.1111/1468- 2427.00387 Burgers, J. (2002). De gefragmenteerde stad. Amsterdam: Boom. Burgess, E. W., & Newcomb, C. S. (1933). Census data of the city of Chicago, 1930,. Chicago, IL: The University of Chicago press. Burke, R. J., & Cooper, C. L. (2000). The organization in crisis: Downsizing, restructuring, and privatization. Oxford, UK: Blackwell. Butler, T. (2005). Globalization and gentrification: The emergence of a middle range theory. Cities and Territories, 14, working paper series. Cairncross, F. (1997). The Death of distance: how the communications revolution will change our lives. Cambridge, MA: Harvard Business School. Caldeira, T.P.R. (2000). City of walls: crime, segregation, and citizenship in São Paulo. University of California Press, Berkeley. Carr, S., Francis, M., Rivlin, L. G., & Stone, A. M. (1992). Public space. Cambridge, Angleterre: Cambridge University Press. Castells, M., & Mollenkopf, J. H. (1991). Conclusion: Is New York a dual city? In J. H. Mollenkopf & M. Castells (Eds.), Dual city: Restructuring New York. New York: Russell Sage Foundation. Castells, M. (1989). The informational city: Information technology, economic restructuring, and the urban regional process. Oxford: Blackwell. Castells, M. (1996). The rise of the network society. Malden, MA: Blackwell. Castells, M. (1997). End of millenium. Oxford: Blackwell Scientific Publications. Castells, M. (1998). The Rise of Network Society. Malden, MA: Blackwell. Castells, M. (1999). Information technology, globalization and social development. Discussion paper-United Nations Research Institute for Social Development, (114), 1-15. Castells, M. (2000). The information age: Economy, society and culture. Cambridge, MA: Blackwell. Castles, S. (2002). Migration and community formation under conditions of globalization. International Migration Review, 36, 1143-1168.
References and bibliography | 325
Chapple, K., & Lester, T. W. (2010). The resilient regional labour market? The US case. Cambridge Journal of Regions, Economy and Society, 3(1), 85-104. doi: 10.1093/cjres/rsp031 Cheung, C., & Leung, K. (2012). Chinese migrants’ class mobility in Hong Kong. International Migration IOM. doi: 10.1111/j.1468-2435.2012.00778.x Cho, H. (2011, April). Immigration policy and settlement patterns of migrants in South Korea and Singapore: Understanding ethnic and socio-economic class settlement behaviors in Asia. In PSA Conference 2011. Retrieved October 15, 2012, from http://www.psa.ac.uk/journals/pdf/5/2011/1258_669.pdf Clark, E. (2005). The order and simplicity of gentrification - a political challenge. In R. Atkinson & G. Bridge (Eds.), Gentrification in a global context: The new urban colonialism (pp. 256-264). London: Routledge. Clark, W. A. (1991). Residential preferences and neighborhood racial segregation: A test of the Schelling segregation model. Demography, 28(1), 1-19. Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2013). Applied multiple regression/correlation analysis for the behavioral sciences. Routledge. Coldwell Banker. (1987). National survey of international investment ownership of major office buildings in 19 largest United States. Boston: Coldwell Banker. Cortes, C., Mohri, M., Riley, M., & Rostamizadeh, A. (2008, January). Sample selection bias correction theory. In Algorithmic Learning Theory (pp. 38-53). Springer Berlin Heidelberg. Coy, M. and M. Pohler (2002) Gated communities in Latin American mega cities: case studies in Brazil and Argentina. Environment and Planning B: Planning and Design 29.3, 335–70. Cross, M. (1992). Ethnic minorities and industrial change in Europe and North America. Cambridge, England: Cambridge University Press. Cutler, D. M., Glaeser, E. L., & Vigdor, J. L. (2008). Is the melting pot still hot? explaining the resurgence of immigrant segregation. Review of Economics and Statistics, 90(3), 478-497. doi: 10.1162/rest.90.3.478 De Ruijter, A. (1996). Hybridization and governance. The Hague: ISS. Dodge, M., & Kitchin, R. (2007). Virtual places. In I. Douglas, R. J. Huggett, & C. R. Perkins (Eds.), Companion encyclopedia of geography: From local to global (pp. 519-536). London: Routledge. Duncan, G. J. (2009). New lessons from the Gautreaux and moving to opportunity residential mobility program. Symposium of housing after the fall: reassessing the future of the American dream, 19-20 Feb 2009, The University of California Irvine and The Paul Merage School of Business Center for Real Estate.
326 | References and bibliography
Duncan, O. D., & Duncan, B. (1955). A methodological analysis of segregation indexes. American Sociological Review, 20, 210-217. Dussel, E. (2005) The implications of China’s entry into the WTO for Mexico. Global Issue Papers 24, 1–41. Elander, I., & Blanc, M. (2001). Partnerships and democracy: A happy couple in urban governance? In H. T. Andersen & R. Van Kempen (Eds.), Governing European cities: Social fragmentation, social exclusion and urban governance (pp. 93-124). Aldershot: Ashgate. Eriksen, T. H. (2007). Globalization: The key concepts. Oxford: Berg. Espino, N. A. (2005). Inequality, segregation, and housing market: The US case. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 145-157). Albany, NY: State University of New York Press. Fainstein, S. S., & Campbell, S. (1996). Readings in urban theory. Cambridge, Mass., USA: Blackwell. Fainstein, S. S., & Harloe, M. (2003). Ups and downs in the global city: London and New York at the millennium. In G. Bridge & S. Watson (Eds.), A companion to the city (pp. 155-167). Malden: Blackwell. Fainstein, S. S. (2001). Inequality in global city-regions. In A. J. Scott (Ed.), Global city-regions: Trends, theory, policy (pp. 285-298). Cambridge, UK: Oxford University Press. Fainstein, S. S. (2010). The just city. Ithaca: Cornell University Press. Fainstein, S. S., Gordon, I., & Harloe, M. (1992). Divided cities: New York & London in the contemporary world. Oxford, UK: Blackwell. Farley, R., Fielding, E. L., & Krysan, M. (1997). The residential preferences of blacks and whites: A fourmetropolis analysis. Housing Policy Debate, 8(4), 763-800. Feitosa, F. F., Câmara, G., Monteiro, A. M. V., Koschitzki, T., & Silva, M. P. S. (2004). Spatial measurement of residential segregation. VI Brazilian Symposium on GeoInformatics, 70-83. In VI Brazilian Symposium on GeoInformatics, 22-24 Nov 2004, Campos do Jordão, Geneve: IFIP. Feitosa, F. F., Câmara, G., Monteiro, A. V., Koschitzki, T., & Silva, M. S. (2007). Global and local spatial indices of urban segregation. International Journal of Geographical Information Science, 21(3), 299-323. Feitosa, F. F., Le, Q. B., & Vlek, P. L. (2011). Multi-agent simulator for urban segregation (MASUS): A tool to explore alternatives for promoting inclusive cities. Computers, Environment and Urban Systems, 35(2), 104-115. doi: 10.1016/j.compenvurbsys.2010.06.001 Fernandez, M. (2007, October 15). Study finds disparities in mortgage by race. New York Times. Retrieved September 9, 2011, from
References and bibliography | 327
http://www.nytimes.com/2007/10/15/nyregion/15subprime.html?pagewanted =all Fernandez, S., Brudney, J. L., & Ryu, J. E. (2005, August). Variations in local service delivery: Examining the effects of state-level factors on local government contracting for services. In annual meeting of the American Political Science Association, Washington, DC. Fernandez, S., Smith, C. R., & Wenger, J. B. (2007). Employment, privatization, and managerial choice: Does contracting out reduce public sector employment? Journal of Policy Analysis and Management, 26(1), 57-77. doi: 10.1002/pam.20227 Fincher, R., & Shaw, K. (2009). The unintended segregation of transnational students in central Melbourne [Abstract]. Environment and Planning A, 41(8), 1884-1902. doi: 10.1068/a41126 Findlay, A. (1995). Skilled transients: The invisible phenomenon. In R. Cohen (Ed.), The Cambridge survey of world migration (pp. 515-522). Cambridge: Cambridge University Press. Fischer, C. S. (1976). The urban experience. New York: Harcourt Brace Jovanovich. Fong, E., & Shibuya, K. (2000). The spatial separation of the poor in Canadian cities. Demography, 37(4), 449. doi: 10.2307/2648071 Foreign Policy. (2010, August 16). Global cities index. Retrieved October 7, 2012, from http://www.foreignpolicy.com/articles/2010/08/11/the_global_cities_index_2 010 Forrest, R., La Grange, A., & Yip, N. M. (2004). Neighborhood in a high-rise, high- density city: Some observations on contemporary Hong Kong. The Sociological Review, 50(22), 215-240. Forster, C. (2006). The challenge of change: Australian cities and urban planning in the new millennium. Geographical Research, 44(2), 173-182. doi: 10.1111/j.1745-5871.2006.00374.x
Freund, R. J., Wilson, W. J., & Sa, P. (2006). Regression analysis. Academic Press. Frey, W. H. (2002). Census 2000 reveals new native-born and foreign-born shifts across U.S. (Rep. No. 02-520). Population Studies Centre, the Institute for Social Research, University of Michigan. Friedman, T. L. (2005). The world is flat: A brief history of the twenty-first century. New York: Farrar, Straus and Giroux. Friedmann, J., & Wolff, G. (1982). World city formation: An agenda for research and action. International Journal of Urban and Regional Research, 6(3), 309-344.
328 | References and bibliography
Friedmann, J. (1986). The world city hypothesis. Development and Change, 17(1), 69-83. doi: 10.1111/j.1467-7660.1986.tb00231.x Friedmann, J. (1995). The world city hypothesis. In P. L. Knox & P. J. Taylor (Eds.), World cities in a world system (pp. 317-331). Cambridge: Cambridge University Press. (Original work published 1986) Friedrichs, J., Galster, G., & Musterd, S. (2003). Neighbourhood effects on social opportunities: The European and American research and policy context. Housing Studies, 18(6), 797-806. Gardner, A. (2010). City of strangers: Gulf migration and the Indian community in Bahrain. Ithaca: Cornell/ILR Press. Geniş, S. (2007). Globalization of cities: Towards conceptualizing a new politics of place-making in a transnational era. University of Gaziantep Journal of Social Sciences, 6(1), 59-77.
Gentle, J. E. (2002). Elements of computational statistics. New York: Springer. Gibbons, J. D., & Chakraborti, S. (2003). Nonparametric statistical inference. New York: Marcel Dekker. Goldsmith, W. W. (2000). From the metropolis to globalization: The dialectics of race and urban form. In P. Marcuse & R. V. Kempen (Eds.), Globalizing cities: A new spatial order? (pp. 37-55). Oxford: Blackwell. Goodman, S. N. (1999). Toward evidence-based medical statistics. Annals of internal medicine, 130(12), 995-1004. Gray, J. (1996). After social democracy: Politics, capitalism and the common life. London: Demos. Gregory, D., & Urry, J. (1985). Introduction. In D. Gregory & J. Urry (Eds.), Social relations and spatial structures (pp. 1-8). New York: St. Martin's Press. Gwyther, G. (2005). Paradise planned: Community formation and the master planned estate. Urban Policy and Research, 23(1), 57-72. doi: 10.1080/0811114042000335304 Haan, M. (2005). Summary of: Are immigrants buying to get in?: The role of ethnic clustering on the homeownership propensities of 12 Toronto immigrant groups, 1996 - 2001. Ottawa, Ontario: Statistics Canada. Heckman, J. J. (1979). Sample Selection Bias as a Specification Error. Econometrica, 47, 151–161. Hackworth, J. (2002). Postrecession gentrification in New York city. Urban Affairs Review, 37(6), 815-843. Haila, A. (1991). Four types of investment in land and property. International Journal of Urban and Regional Research, 15(3), 343-365.
References and bibliography | 329
Hamnett, C., & Cross, D. (1998). Social polarisation and inequality in London: The earnings evidence, 1979 - 95. Environment and Planning C: Government and Policy, 16(6), 659-680. Hamnett, C. (1994). Social polarisation in global cities: Theory and evidence. Urban Studies, 31(3), 401-424. Hamnett, C. (1996). Social polarisation, economic restructuring and welfare state regimes. Urban studies, 33(8), 1407-1430. Hamnett, C. (2001). Social segregation and social polarization. In R. Paddison (Ed.), Handbook of urban studies (pp. 162-176). London: SAGE. Hamnett, C. (2003). Gentrification and the middle-class remaking of inner London, 1961-2001. Urban Studies, 40(12), 2401-2426. Hancher, L., & Moran, M. (1989). Introduction: Regulation and deregulation. European Journal of Political Research, 17(2), 129-136. Harloe, M., & Fainstein, S. S. (1992). Conclusions: The divided cities. In S. S. Fainstein, I. Gordon, & M. Harloe (Eds.), Divided cities: New York and London in the contemporary world. Oxford: Blackwell. Harvey, D. (1978). The urban process under capitalism: A framework for analysis. International Journal of Urban and Regional Research, 2(1-4), 101-131. Harvey, D. (1985a). The geopolitics of capitalism. In D. Gregory & J. Urry (Eds.), Social relations and spatial structures (pp. 128-163). London: MacMillan. Harvey, D. (1985b). The urbanization of capital. Oxford: Blackwell. Harvey, D. (1989). The condition of postmodernity (Vol. 14). Oxford: Blackwell. Harvey, D. (2001). Globalization and the spatial fix. Geographische revue, 2(3), 23-
31. Harvey, D. (2005). From globalization to the new imperialism. In R. P. Appelbaum & W. I. Robinson (Eds.), Critical globalization studies (pp. 91-100). New York: Routledge. Harvey, D. (2010). The enigma of capital: And the crises of capitalism. Oxford [England: Oxford University Press. Haussermann, H., & Haila, A. (2005). The European city: A conceptual framework and normative project. In Y. Kazepov (Ed.), Cities of Europe: Changing contexts, local arrangements, and the challenge to urban cohesion (pp. 43- 63). Malden, MA: Blackwell Publishing. Healey, P., Cameron, S., Davoudi, S., Graham, S., & Madanipour, A. (1995). Managing cities: The new urban context. Chichester: Wiley. Healy, E., & Birrell, B. (2003). Metropolis divided: The political dynamic of spatial inequality and migrant settlement in Sydney. People and Place, 11, 65-87.
330 | References and bibliography
Hedin, K., Clark, E., Lundholm, E., & Malmberg, G. (2012). Neoliberalization of housing in Sweden: Gentrification, filtering, and social polarization. Annals of the Association of American Geographers, 102(2), 443-463. Held, D. (2004). Global covenant: The social democratic alternative to the Washington Consensus. Cambridge: Polity. Hoffmann, J. P. (2010). Linear regression analysis: applications and assumptions. Brigham Young University, Provo. Huchzermeyer, M. (2011). Cities with 'slums': From informal settlement eradication to a right to the city in Africa. Cape Town: University of Cape Town Press. Hugo, G. J. (1995). Understanding where immigrants live. Canberra: Australian Government Publishing Service. Hunter, B. H. (2003). Trends in neighbourhood inequality of Australian, Canadian, and United States of America cities since the 1970s. Australian Economic History Review, 43(1), 22-44. doi: 10.1111/0004-8992.00039 Iceland, J., & Scopilliti, M. (2008). Immigrant residential segregation in U.S. metropolitan areas, 1990–2000. Demography, 45(1), 79-94. doi: 10.1353/dem.2008.0009 Ismail, A. (2010). Spatial segregation in global cities: Global pressures and local changes in housing market. In 22nd International Housing Research Conference. Istanbul: ENHR. Ismail, A. (2013). The hybrid outcome of urban change: Global city, polarized city? Glocalism, Journal of Culture, Politics and Innovation., (1). doi: 10.12893/gjcpi.2013.1.5 Jacobs, J. (1992). The death and life of great American cities. Random House LLC.
Originally published in (1961). Janoschka, M. (2002) Urbanizaciones privadas en Buenos Aires: ¿hacia un nuevo modelo de ciudad latinoamericana? [Gated communities in Buenos Aires: towards a new Latin American urban model?] In L.F. Cabrales Barajas (ed.), Latinoamérica: países abiertos, ciudades cerradas [Latin America: open countries, closed cities], Pandora, Guadalajara and Paris. Jargowsky, P. A. (1997). Poverty and place: Ghettos, barrios, and the American city. New York: Russell Sage Foundation. Johnston, R., Poulsen, M., & Forrest, J. (2007). The geography of ethnic residential segregation: A comparative study of five countries. Annals of the Association of American Geographers, 97(4), 713-738. doi: 10.1111/j.1467- 8306.2007.00579.x
Kahane, L. H. (2008). Regression basics. Sage. Kazepov, Y. (2005). Cities of Europe: Changing contexts, local arrangements, and the challenge to urban cohesion. Malden, MA: Blackwell Publishing.
References and bibliography | 331
Keane, D., & McGeehan, N. (2008). Enforcing Migrant Workers' Rights in the United Arab Emirates. International Journal on Minority and Group Rights, 15(1), 81-115. Kesteloot, C., & Cortie, C. (1998). Housing Turks and Moroccans in Brussels and Amsterdam: The difference between private and public markets. Urban Studies, 35(10), 1835-1853. doi: 10.1080/0042098984178 Kesteloot, C. (2000). Post-fordist polarization in a fordist spatial canvas. In P. Marcuse & R. Van Kempen (Eds.), Globalizing cities: A new spatial order? (pp. 186-210). Oxford: Blackwell. Kesteloot, C. (2005). Urban socio-spatial configurations and the future of European cities. In Y. Kazepov (Ed.), Cities of Europe: Changing contexts, local arrangements, and the challenge to urban cohesion (pp. 123-148). Malden, MA: Blackwell Publishing. Kesteloot, C., De Decker, P., & Manço, A. (1997). Turks and their housing conditions in Belgium, with special reference to Brussels, Ghent and Visé. In A. S. Özüekren & R. Van Kempen (Eds.), Turks in European cities: Housing and urban segregation (pp. 67-97). Utrecht: European Research Centre on Migration and Ethnic Relations. Kirkman, T.W. (1996) Statistics to Use. http://www.physics.csbsju.edu/stats/ Kirkwood, B. R. (1988). Essentials of medical statistics. Blackwell Scientific Publications. Kirsch, S. (1995). The incredible shrinking world? Technology and the production of space. Environment and Planning: Society and Space, 13, 529–55. Kleinhans, R. (2004). Social implications of housing diversification in urban renewal: A review of recent literature. Journal of Housing and the Built Environment, 19(4), 367-390. doi: 10.1007/s10901-004-3041-5 Kloosterman, R. C. (1996). Double dutch: Polarization trends in Amsterdam and Rotterdam after 1980. Regional Studies, 30(5), 467-476. Knox, P. L. (1991). The restless urban landscape: Economic and sociocultural change and the transformation of metropolitan Washington, DC. Annals of the Association of American Geographers, 81(2), 181-209. Kraidy, M. M. (1999). The global, the local, and the hybrid: A native ethnography of glocalization. Critical studies in media communication, 16(4), 456-476. Kulish, M., Richards, A., & Gillitzer, C. (2011). Urban structure and housing prices: Some evidence from Australian cities (Research Discussion Paper, RDP2011-03). Reserve Bank of Australia. Retrieved November 9, 2012, from http://www.rba.gov.au/publications/rdp/2011/pdf/rdp2011-03.pdf Larbi, G. A. (1999). The new public management approach and crisis states. Geneva: United Nations Research Institute for Social Development.
332 | References and bibliography
Lattes, A., J. Rodríguez and M. Villa (2002) Population dynamics and urbanization in Latin America: concepts and data limitations. Paper presented at the International Union for the Scientific Study of Population Expert Meeting, Bellagio, 11–19 March. Lee, H. (1997). The employment of foreign workers in Korea: Issues and policy suggestion. International Sociology, 12(3), 353-371. Lee, S. H. (2006). SARS in China and Hong Kong. New York: Novinka. Lefebvre, H. (1991). The production of space (Vol. 142). Blackwell: Oxford. Levy, E. (2001). Decentralization and governance in São Paulo, Brazil. In B. A. Ruble, R. E. Stren, J. S. Tulchin, & D. H. Varat (Eds.), Urban governance around the world (pp. 22-25). Washington, DC: Woodrow Wilson International Center for Scholars, Comparative Urban Studies Project. Levy, F. (1998). The new dollars and dreams: American incomes and economic change. New York: Russell Sage Foundation. Li, Z., & Wu, F. (2008). Tenure-based residential segregation in post-reform Chinese cities: A case study of Shanghai. Transactions of the Institute of British Geographers, 33(3), 404-419. doi: 10.1111/j.1475- 5661.2008.00304.x Little, R. J. A., & Rubin, D. B. (1986). Statistical analysis with missing data. New York, NY, USA: John Wiley & Sons, Inc. Loehnert, S. (2010). About Statistical Analysis of Qualitative Survey Data. Journal of Quality and Reliability Engineering. Hindawi Publishing Corporation. Logan, J. R. (2000). Still a global city: The racial and ethnic segmentation of New York. In P. Marcuse & R. V. Kempen (Eds.), Globalizing cities: A new spatial order? (pp. 158-185). Oxford: Blackwell. (Original work published 1997). MacLeod, G. (2002). From urban entrepreneurialism to a “revanchist city”? On the spatial injustices of Glasgow’s renaissance. Antipode, 34(3), 602-624. Madanipour, A. (2003). Social exclusion and space. In R. T. LeGates & F. Stout (Eds.), The city reader (pp. 181-188). London: New York. Madanipour, A. (2004). Marginal public spaces in European cities. Journal of Urban Design, 9(3), 267-286. doi: 10.1080/1357480042000283869 Madanipour, A., Cars, G., & Allen, J. (1998). Social exclusion in European cities: Processes, experiences, and responses. London: Jessica Kingsley. Maloutas, T. (2007). Segregation, social polarization and immigration in Athens during the 1990s: Theoretical expectations and contextual difference. International Journal of Urban and Regional Research, 31(4), 733-758. doi: 10.1111/j.1468-2427.2007.00760.x
References and bibliography | 333
Marceau, D. J. (1999). The scale issue in social and natural sciences. Canadian
Journal, 25(4), 347-356. Marcuse, P., & Van Kempen, R. (2000). Introduction. In P. Marcuse & R. Van Kempen (Eds.), Globalizing cities: A new spatial order? (pp. 1-21). Malden, MA: Blackwell. Marcuse, P., & Van Kempen, R. (2002). Of states and cities: The partitioning of urban space. Oxford: Oxford University Press. Marcuse, P. (1989). Dual city: A muddy metaphor for a quartered city. International Journal of Urban and Regional Research, 13(4), 697-708. Marcuse, P. (1993). What's so new about divided cities?. International Journal of Urban and Regional Research, 17(3), 355-365. Marcuse, P. (1996). Space and race in the post-Fordist city: The outcast ghetto and advanced homelessness in the United States today. Urban poverty and the underclass, 176-216. Marcuse, P. (2002). The shifting meaning of the black ghetto in the United States. In P. Marcuse & R. V. Kempen (Eds.), Of states and cities: The partitioning of urban space (pp. 109-142). Oxford: Oxford University Press. Marcuse, P. (2003). Cities in quarters. In G. Bridge & S. Watson (Eds.), A companion to the city (pp. 270-281). Malden, Mass: Blackwell. Marcuse, P. (2005). Enclaves yes, ghettos no: Segregation and the state. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 15-30). Albany, NY: State University of New York Press. Martin, P. L., & Miller, M. (2000). Employer sanctions: French, German and US experiences (International Migration Papers, Publication). Geneve: International Labour Organisation. Martinotti, G. (2005). Employer sanction: French, German and US experience. In Y. Kazepov (Ed.), Cities of Europe: Changing contexts, local arrangement and the challenge to urban cohesion (pp. 90-108). Oxford: Blackwell publishing. Massey, D. (1984). Spatial divisions of labor: Social structures and the geography of production. New York: Methuen. Massey, D. (1985). New directions in space. In D. Gregory & J. Urry (Eds.), Social relations and spatial structures (pp. 9-19). London: MacMillan. Massey, D. S., & Denton, N. A. (1993). American apartheid: Segregation and the making of underclass. Cambridge, MA: Harvard University press. Massey, D. S., & Eggers, M. L. (1990). The ecology of inequality: Minorities and the concentration of poverty, 1970-1980. American Journal of Sociology, 95(5), 1153-88. doi: 10.1086/229425
334 | References and bibliography
Massey, D. S., & Fischer, M. J. (1999). Does rising income bring integration? new results for Blacks, Hispanics, and Asians in 1990. Social Science Research, 28(3), 316-326. doi: 10.1006/ssre.1999.0660 Massey, D. S., & International Union for the Scientific Study of Population. (2008). Worlds in motion: Understanding international migration at the end of the millennium. Oxford: Clarendon Press. Massey, D. S. (1990). American Apartheid: Segregation and the Making of the Underclass. American Journal of Sociology, 96(2), 329-357. doi: 10.1086/229532 Massey, D. S. (1996). The age of extremes: Concentrated affluence and poverty in the twenty-first century. Demography, 33(4), 395-412. Massey, D. S. (2008). Globalization and inequality: Explaining American exceptionalism. European Sociological Review, 25(1), 9-23. Massey, D. S., Arango, J., Hugo, G., Kouaouci, A., Pellegrino, A., & Taylor, J. E. (2009). Worlds in motion: Understanding international migration at the end of the millennium. Oxford: Clarendon Press. Massey, D. S., Gross, A. B., & Shibuya, K. (1994). Migration, segregation, and the geographic concentration of poverty. American Sociological Review, 59(3), 425-445. Matsumoto, Y. (2009, July 12). Living together: segregation and the norm of multicultural urbanism in Tokyo. Lecture presented at Workshop on enclave urbanism as problem or solution in University of Utrecht, Utrecht, the Netherlands. Retrieved October 13, 2012, from http://www.rikkyo.ne.jp/web/ymatsumoto/LivingTogether.pdf McCarthy, E. M. (2006). Handbook of avian hybrids of the world. Oxford: Oxford University Press. McCarthy, J. (1999). Chicago: A case study of social exclusion and city regeneration. Cities, 16(5), 323-331. doi: 10.1016/S0264-2751(99)00030-X McGeary, M. G., & Lynn, L. E. (Eds.). (1988). Urban change and poverty. National Academies. McLuhan, M. (1962). The Gutenberg Galaxy. London: Routledge. McLuhan, M. (1964). Understanding Media: The Extensions of Man. New York: Signet. Mingione, E. (2005). Urban social change: A socio-historical framework of analysis. In Y. Kazepov (Ed.), Cities of Europe: Changing contexts, local arrangements, and the challenge to urban cohesion (pp. 67-89). Malden, MA: Blackwell Publishing.
References and bibliography | 335
Ministry of BZK. (2012). De starter op de Amsterdamse woningmarkt [The starter on the Amsterdam housing] (Publication). General Directorate for Housing, Communities and Integration. Ministry of Interior. Ministry of VROM. (2003). Dossier kwalitatieve woningregistratie 2000 [file qualitative housing register] (Publication). Den Haag: Ministerie van VROM. MMF. (2011). GPCI Gobal Power City Index 2011 (Rep.). Institute for Urban Strategies The Mori Memorial Foundation. Retrieved September 9, 2012, from http://www.mori-m- foundation.or.jp/english/research/project/6/pdf/GPCI2011_English.pdf Mollenkopf, J. H., & Castells, M. (1991). Dual city: Restructuring New York. New York: Russell Sage Foundation. Monkkonen, P., & Zhang, X. (2011). Socioeconomic segregation in Hong Kong: Spatial and ordinal measures in a high-density and highly unequal city (Working paper No. 2011-3). Berkeley: Institute of Urban and Regional Development IURD. Monkkonen, P. (2011). Housing finance reform and increasing socioeconomic segregation in Mexico. International Journal of Urban and Regional Research, 63(4), 757-772. doi: 10.1111/j.1468-2427.2011.01085.x Morgan, B. S. (1975). The segregation of socioeconomic groups in urban areas: A comparative analysis. Urban Studies, 12(1), 47-60. doi: 10.1080/00420987520080041 Morgan, K. (2004). The exaggerated death of geography: Learning, proximity and territorial innovation systems. Journal of Economic Geography, 4(1), 3-21. doi: 10.1093/jeg/4.1.3 Munck, R. (2005). Globalization and social exclusion: A transformationalist perspective. Bloomfield, CT: Kumarian Press. Murray, W. (2006). Geographies of globalization. Routledge. Musterd, S., & Andersson, R. (2005). Housing Mix, Social Mix, and Social Opportunities. Urban Affairs Review, 40(6), 761-790. doi: 10.1177/1078087405276006 Musterd, S., & Fullaondo, A. (2008). Ethnic segregation and the housing market in two cities in northern and southern Europe: The cases of Amsterdam and Barcelona. Architecture, City and Environment, 3(8), 93-115. Musterd, S., & Ostendorf, W. J. (1998). Urban segregation and the welfare state: Inequality and exclusion in western cities. London: Routledge. Musterd, S., & Ostendorf, W. J. (2005). Social exclusion, segregation, and neighborhood effect. In Y. Kazepov (Ed.), Cities of Europe: Changing contexts, local arrangements, and the challenge to urban cohesion (pp. 170- 189). Malden, MA: Blackwell.
336 | References and bibliography
Musterd, S. (1994). A rising European underclass?. Built Environment, 20(3), 185- 191. Musterd, S. (2005). Social and ethnic segregation in Europe: Levels, causes, and effects. Journal of Urban Affairs, 27(3), 331-348. doi: 10.1111/j.0735- 2166.2005.00239.x Nederveen Pieterse, J. (1993). Globalization as hybridization. ISS Working Paper Series/General Series, 152, 1-18. Nederveen Pieterse, J. (2001). Hybridity, so what?: The anti-hybridity backlash and the riddles of recognition. Theory, Culture & Society, 18(2-3), 219-245. doi: 10.1177/026327640101800211 Nederveen Pieterse J. (2009). Globalization and culture: Global mélange. Lanham, MD: Rowman & Littlefield. Ohmae, K. (2005). The next global stage: Challenges and opportunities in our borderless world. Upper Saddle River, NJ: Wharton School Pub. Oliver, A. L., & Montgomery, K. (2000). Creating a Hybrid Organizational Form from Parental Blueprints: The Emergence and Evolution of Knowledge Firms. Human Relations, 53(1), 33-56. doi: 10.1177/0018726700531003 Pacione, M. (2009). Urban geography: A global perspective. London: Routledge. Pahl, R.E. (1988). Some remarks on informal work, social polarization and social structure. International Journal of Urban and Regional Research, 12: 247– 267. Panitch, L. (1998). The state in a changing world: Social-democratizing global capitalism? Monthly Review, 50(5), 11-22. Parakatil, S. (2012, July 11). Defining quality of living. Retrieved October 7, 2012, from http://www.mercer.com/articles/quality-of-living-survey-report-2011 Park, R. E. (1926). The urban community as a spatial pattern and a moral order. In E. W. Burgess (Ed.), The urban community: Selected papers from the Proceedings of the American Sociological Society, 1925. Chicago: University of Chicago Press. Patel, V. R., & Mehta, R. G. (2011). Impact of Outlier Removal and Normalization Approach in Modified k-Means Clustering Algorithm. International Journal of Computer Science Issues (IJCSI), 8(5). Peace, R. (2001). Social exclusion: A concept in need of definition? Social Policy Journal of New Zealand, 16. Peach, C. (1981). Conflicting interpretations of segregation. In P. Jackson & S. Smith (Eds.), Social interaction and ethnic segregation. London: Academic Press. Peach, C. (1996). Good segregation, bad segregation. Planning Perspectives, 11(4), 379-398.
References and bibliography | 337
Peach, C. (2002). Ethnic diversity in the city. In M. Martiniello & B. Piquard (Eds.), Diversity in the city (pp. 21-42). Bilbao: University of Deusto. Peach, C. (2007). Sleepwalking into ghettoisation? the British debate over segregation. In K. Schönwälder (Ed.), Residential segregation and the integration of immigrants Britain, the Netherlands and Sweden (pp. 7-40). Berlin: WZB. Pendall, R. (2005). Does density exacerbate income segregation? evidence from US metropolitan areas. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 175-199). Albany, NY: State University of New York Press. Piketty, T., & Saez, E. (2003). Income Inequality in the United States, 1913–1998*. The Quarterly journal of economics, 118(1), 1-41. Prescott-Allen, R. (2001). The wellbeing of nations: A country-by-country index of quality of life and the environment. Washington, DC: Island Press. Preteceille, E. (1994). Cidades globais e segmentação social. In L. C. Ribeiro & O. A. Santos Junior (Eds.), Globalização, fragmentação e reforma urbana: O futuro das cidades brasileiras na crise. (pp. 65-89). Rio de Janeiro, RJ: Civilização Brasileira. Qadeer, M. A. (2005). Ethnic segregation in a multicultural city. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 49-61). Albany, NY: State University of New York Press. Randolph, B., & Holloway, D. (2005). Social disadvantage, tenure and location: An analysis of Sydney and Melbourne. Urban Policy and Research, 23(2), 173- 201. doi: 10.1080/08111470500135136 Raskall, P. (1995, October). Who gets what where? Spatial inequality between and within Australian cities. Paper presented to The Commonwealth Department of Housing and Regional Development Seminar on Aspects of Spatial Inequality, Canberra. Ray, L. J. (2007). Globalization and everyday life. London: Routledge. Reardon, S. F., & Firebaugh, G. (2002). Measures of Multigroup Segregation. Sociological Methodology, 32(1), 33-67. doi: 10.1111/1467-9531.00110 Reardon, S. F., & O'Sullivan, D. (2004). Measures of spatial segregation. Sociological Methodology, 34(1), 121-162. doi: 10.1111/j.0081- 1750.2004.00150.x Reichl, A. J. (2007). Rethinking the dual city. Urban Affairs Review, 42(5), 659- 687. Ribeiro, L. C., & Telles, E. E. (2008). In a historically unequal city. Globalizing Cities, 78.
338 | References and bibliography
Ritzer, G., & Ryan, J. M. (Eds.). (2011). The concise encyclopedia of sociology. Chichester, West Sussex, U.K.: Wiley-Blackwell. Ritzer, G. (Ed.). (2004). Handbook of social problems: A comparative, international perspective. Thousand Oaks, CA: Sage Publications. Roberts, B., & Wilson, R. (2009). Spatial differentiation, inequality and urban policy: The findings. Urban segregation and governance in the Americas, 205-222. Robinson, J. (2002). Global and world cities: A view from off the map. International Journal of Urban and Regional Research, 26(3), 531-554. doi: 10.1111/1468-2427.00397 Robinson, J. (2005). Urban geography: World cities, or a world of cities. Progress in Human Geography, 29(6), 757-765. doi: 10.1191/0309132505ph582pr Rodriguez, A., Swyngedouw, E. and Moulaert, F. (2003). Urban restructuring, social-political polarization and new urban policies, in: F. Moulaert, R. Arantxa and E. Swyngedouw (Eds) The Globalized City: Restructuring and Social Polarization in European Cities, pp. 28– 45. Oxford: Oxford University Press. Rofe, M. W. (2009). Globalisation, gentrification and spatial hierarchies in and beyond New South Wales: The local/global nexus. Geographical Research, 47(3), 292-305. doi: 10.1111/j.1745-5871.2009.00574.x Rolnik, R. (2011). Democracy on the edge: Limits and possibilities in the implementation of an Urban reform agenda in Brazil. International Journal of Urban and Regional Research, 35, 239-255. doi: 10.1111/j.1468- 2427.2010.01036.x Ross, R. J. (2011, July 6). New Orleans as a Rust Belt City? Retrieved from http://www.metropolitiques.eu/New-Orleans-as-a-Rust-Belt-City.html Rubinowitz, L. S., & Rosenbaum, J. E. (2000). Crossing the class and color lines: From public housing to white suburbia. Chicago: University of Chicago Press. Sabatini, F. (2003) The social spatial segregation in the cities of Latin America. Social development strategy document, Inter-American Development Bank, Washington, DC. Sabatini, F., Caceres, G., Cerda, J., 2001, Residential segregation pattern changes in main Chilean cities: Scale shifts and increasing malignancy. In International Seminar on Segregation in the City, 26-28 July 2001, Lincoln Institute of Land Policy, Cambridge. Available online at: www.lincolninst.edu/pubs/dl/615_sabatini_caceres_cerda.pdf Sahlin, I. (2008, June 12). ‘Social housing’ som bostadspolitiskt spoke (’Social housing’ as housing policy ghost). Alba.nu. Retrieved from http://www.alba.nu/?showArticle=19296
References and bibliography | 339
Sakoda, J. (1981). A generalized index of dissimilarity. Demography, 18, 45-50. Sassen, S. (1988). The mobility of labor and capital: A study in international investment and labor flow. Cambridge: Cambridge University Press. Sassen, S. (1990). Economic restructuring and the American city. Annual Review of Sociology, 16(1), 465-490. Sassen, S. (1991). The global city: New York, London, Tokyo. Princeton, NJ: Princeton University Press. Sassen, S. (1994). Cities in a World Economy. Pine Forge Press: London. Sassen, S. (1996). Cities and communities in the global economy: Rethinking our concepts. American Behavioral Scientist, 39(5), 629-639. Sassen, S. (1999). Guests and Aliens. The New Press: New York. Sassen, S. (2000). The state and the new geography of power. In D. Kalb (Ed.), The ends of globalization: Bringing society back in. Lanham, MD: Rowman & Littlefield. Sassen, S. (2001). The global city: New York, London, Tokyo (2nd ed., Original work published 1991). Princeton, NJ: Princeton University Press. Sassen, S. (2002). Locating cities on global circuits. Environment and Urbanization, 14(1), 13-30. doi: 10.1630/095624702101286034 Sassen, S. (2006). Cities in a world economy (3rd ed., Original work published 1994). Thousand Oaks, CA: Pine Forge Press. Sayer, A. (1985). The difference that space makes. In D. Gregory & J. Urry (Eds.), Social relations and spatial structures. New York: St. Martin's Press. Scott, A. J. (2008). Inside the city: on urbanisation, public policy and planning.
Urban Studies, 45(4), 755-772. Sennett, R. (1970). The uses of disorder: Personal identity & city life. New York: A.A. Knopf. Sheppard, E. (2002). The spaces and times of globalization: Place, scale, networks, and positionality. Economic Geography, 78(3), 307. doi: 10.2307/4140812 Short, J.R. & Kim, Y. (1999). Globalization and the city. Addison Silver, H. (1993). National conceptions of the new urban poverty: Social structural change in Britain, France and the United States. International Journal of Urban and Regional Research, 17(3), 336-354. Simon, P. (2005). Gentrification of old neighborhoods and social integration in Europe. In Y. Kazepov (Ed.), Cities of Europe: Changing contexts, local arrangements, and the challenge to urban cohesion (pp. 210-232). Oxford: Blackwell Publishing.
340 | References and bibliography
Sin, C. H. (2002). The interpretation of segregation indices in context: The case of P in Singapore. The Professional Geographer, 54(3), 422-437. doi: 10.1111/0033-0124.00340 Sin, C. H. (2003). The politics of ethnic integration in Singapore: Malay 'regrouping' as an ideological construct. International Journal of Urban and Regional Research, 27(3), 527-544. doi: 10.1111/1468-2427.00465 Singapore Department of Statistics. (1996). General household survey 1995: Socio- demographic and economic characteristics, statistical release 1 (Singapore, Department of Statistics). Siqueira, M. T. (2012). Urban operations: public–private partnerships globalizing São Paulo. Advances in Education in Diverse Communities: Research, Policy and Praxis, 8, 389-413. Skop, E. (2009). Fueling Austin’s boom: The new 21st century immigrant metropolis. In A. Singer, C. Brettell, & S. Hardwick (Eds.), America’s twenty-first century immigrant gateways: Immigrant incorporation in suburbia. Brookings Institution Press. Smith, M. P. (2005). Power in place: retheorizing the local and the global. The urban sociology reader, 241-250. Smith, N. (1996). Spaces of vulnerability: the space of flows and the politics of scale. Critique of Anthropology, 16, 63– 77. Soja, E. (1985). The spatiality of social life: Towards a transformative retheorization. In D. Gregory & J. Urry (Eds.), Social relations and spatial structures (pp. 90-127). London: MacMillan. Soja, E. W. (1989). Postmodern geographies: The reassertion of space in critical social theory. London: Verso. Squires, G. D., Friedman, S., & Saidat, C. (2005). Experiencing residential segregation: A contemporary study of Washington D.C. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 127- 144). Albany, NY: State University of New York Press. Stanley, B. (2003). "Going global" and wannabe world cities: (Re)conceptualizing regionalism in the Middle East. In W. A. Dunaway (Ed.), Emerging issues in the 21st century world-system: Crisis and resistance in the 21st century world-system (Vol. 1, pp. 151-170). Westport: Greenwood Publishing Group. Statistics Canada. (2006, May 16). Retrieved from http://www12.statcan.gc.ca/census-recensement/2006/index-eng.cfm
Stigler, S. (2008). Fisher and the 5% level. Chance, 21(4), 12-12. Stimson, R. (2001). Dividing societies: The socio-political spatial implications of restructuring in Australia. Australian Geographical Studies, 39(2), 198-216. doi: 10.1111/1467-8470.00140
References and bibliography | 341
Sullivan, H. J. (1987). Privatization of public services: A growing threat to constitutional rights. Public Administration Review, 47(6), 461-467. Swyngedouw, E. C., & Kesteloot, C. (1991). Le passage sociospatial du fordisme à la flexibilité : Une interprétation des aspects spatiaux de la crise et de son issue. Espace Et Sociétés, 54(2), 243-268. Swyngedouw, E., Moulaert, F., & Rodriguez, A. (2002). Neoliberal urbanization in Europe: Large-scale urban development projects and the new urban policy. Antipode, 34(3), 542-577. doi: 10.1111/1467-8330.00254 Swyngedouw, E. (1997). ‘Neither global nor local: “glocalization” and the politics of scale’, in K. R. Cox, (ed.) Globalization: reasserting the power of the local, New York and London: Guilford, pp. 137–66. Taylor, P. J., Catalano, G., & Walker, D. R. (2002). Exploratory analysis of the world city network. Urban Studies, 39(13), 2377-2394. Taylor, P., & Walker, D. R. (2001). World cities: A first multivariate analysis of their service complexes. Urban Studies, 38(1), 23-47. Taylor, P. J., Ni, P., Derudder, B., Hoyler, M., Huang, J., Lu, F., ... Shen, W. (2010, April 13). Measuring the World City Network: New Results and Developments. Retrieved from http://www.lboro.ac.uk/gawc/rb/rb300.html Taylor, P. J. (2000). World cities and territorial states under conditions of contemporary globalization. Political geography, 19(1), 5-32. Taylor, P. J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181-194. doi: 10.1111/j.1538-4632.2001.tb00443.x Taylor, P. J. (2004). World city network: A global urban analysis. London: Routledge. Taylor, P. J., Ni, P., Derudder, B., Hoyler, M., Huang, J., Lu, F., ... Shen, W. (2011). Global urban analysis: A survey of cities in globalization. London: Earthscan. Telles, E. E. (1995). Structural sources of socioeconomic segregation in Brazilian metropolitan areas. American Journal of Sociology, 100(5), 1199-1223. doi: 10.1086/230636 The State of Texas. (2011). Foreign investment in Texas: The industries and countries leading current growth (Rep.). Retrieved October 10, 2012, from Office of the Governor - Economic Development & Tourism website: http://www.governor.state.tx.us/files/ecodev/Foreign_Investment.pdf Thierstein, A., Lüthi, S., Kruse, C., Gabi, S., & Glanzmann, L. (2008). Changing value chain of the Swiss knowledge economy: spatial impact of intra-firm and inter-firm networks within the emerging mega-city region of Northern Switzerland. Regional studies, 42(8), 1113-1131.
342 | References and bibliography
Thompson, H. (2011). International economics: Global markets and competition (3rd ed.). Singapore: World Scientific. Tickell, A., & Peck, J. (2003). Making global rules: Globalization or neoliberalism? In J. Peck & H. W. Yeung (Eds.), Remaking the global economy: Economic- geographical perspectives (pp. 163-181). London: SAGE. Tiesdell, S., & Oc, T. (1998). Beyond 'fortress' and 'panoptic' cities -- towards a safer urban public realm. Environment and Planning B: Planning and Design, 25(5), 639-655. Timberlake, M., Sanderson, M. R., Ma, X., Derudder, B., Winitzky, J., & Witlox, F. (2012). Testing a global city hypothesis: an assessment of polarization across US cities. City & Community, 11(1), 74-93.
Tomlinson, J. (1999). Globalization and culture. University of Chicago Press. Tu, Y. (1999). Public homeownership, housing finance and socioeconomic development in Singapore. Review of Urban and Regional Development Studies, 11(2), 100-113. doi: 10.1111/1467-940X.00009 Ullah, A. (2012). Bangladeshi migrant workers in Hong Kong: Adaptation strategies in an ethnically distant destination. International Migration IOM. doi: 10.1111/j.1468-2435.2012.00779.x UN-Habitat (2003) Global Report on Human Settlements 2003, The Challenge of Slums, Earthscan, London; Part IV: 'Summary of City Case Studies', pp195- 228. Urry, J. (1985). Social relations, space and time. In D. Gregory & J. Urry (Eds.), Social relations and spatial structures. New York: St. Martin's Press. Van Criekingen, M., & Decroly, J. (2003). Revisiting the diversity of gentrification: Neighbourhood renewal processes in Brussels and Montreal. Urban Studies, 40(12), 2451-2468. doi: 10.1080/0042098032000136156 Van der Waal, J. (2010).“Unravelling the Global City Debate. Economic Inequality and Ethnocentrism in Contemporary Dutch Cities” (Doctoral dissertation, PhD thesis, Erasmus University Rotterdam). Van der Waal, J., & Burgers, J. (2009). Unravelling the global city debate on social inequality: a firm-level analysis of wage inequality in Amsterdam and Rotterdam. Urban Studies, 46(13), 2715-2729. Van Grunsven, L. (2000). Singapore: The changing residential landscape in a winner city. In P. Marcuse & R. Van Kempen (Eds.), Globalizing cities: A new spatial order? (pp. 95-126). Oxford: Blackwell. Van Kempen, R., & Marcuse, P. (1997). A new spatial order in cities? American Behavioral Scientist, 41(3), 285-298.
References and bibliography | 343
Van Kempen, R., & Murie, A. (2009). The new divided city: Changing patterns in European cities. Tijdschrift Voor Economische En Sociale Geografie, 100(4), 377-398. doi: 10.1111/j.1467-9663.2009.00548.x Van Kempen, R., & Özüekren, A. S. (1998). Ethnic segregation in cities: New forms and explanations in a dynamic world. Urban Studies, 35(10), 1631- 1656. Van Kempen, R. (2000). Big cities policy in the Netherlands. Tijdschrift voor economische en sociale geografie, 91: 197–203. doi: 10.1111/1467- 9663.00107 Van Kempen, R. (2005). Segregation and housing conditions of immigrants in Western European cities. In Y. Kazepov (Ed.), Cities of Europe: Changing contexts, local arrangements, and the challenge to urban cohesion (pp. 190- 209). Malden, MA: Blackwell Publishing. Van Kempen, R. (2007). Divided cities in the 21st century: Challenging the importance of globalisation. Journal of Housing and the Built Environment, 22(1), 13-31. Vandell, K. D. (1995). Market factors affecting spatial heterogeneity among urban neighborhoods. Housing Policy Debate, 6(1), 103-139. Varady, D. P. (2005). Desegregating the city: Ghettos, enclaves, and inequality. Albany, NY: State University of New York Press. Wade, R. (1996). Globalization and its limits: Reports of the death of the national economy are exaggerated. In S. Berger & R. P. Dore (Eds.), National diversity and global capitalism (pp. 60-88). Ithaca, NY: Cornell University Press. Walker, R. A. (1985). Class, division of labour, and employment in space. In D. Gregory & J. Urry (Eds.), Social relations and spatial structures (pp. 164- 189). London: MacMillan. Wassmer, R. A. (2005). An economic view of the causes as well as the costs and some of the benefits of urban spatial segregation. In D. P. Varady (Ed.), Desegregating the city: Ghettos, enclaves, and inequality (pp. 158-174). Albany, NY: State University of New York Press. Waters, M. (1995). Globalization. London: Routledge. Wegener, M., Gnad, F., & Vannahme, M. (1986). The time scale of urban change. In B. Hutchinson & M. Batty (Eds.), Advances in urban systems modelling (pp. 175-197). Amsterdam: North Holland. Weinberg, S. L., & Abramowitz, S. K. (2002). Data analysis for the behavioral sciences using SPSS. Cambridge University Press.
Weisberg, S. (2014). Applied linear regression. John Wiley & Sons.
344 | References and bibliography
Weller, S., & Van Hulten, A. (2012). Gentrification and displacement: The effects of a housing crisis on Melbourne's low-income residents. Urban Policy and Research, 30(1), 25-42. Wessel, T. (2000). Social Polarisation and Socioeconomic Segregation in a Welfare State: The Case of Oslo. Urban Studies, 37(11), 1947-1967. doi: 10.1080/713707228 WHO | Glossary of globalization, trade and health terms. (n.d.). Retrieved September 09, 2011, from http://www.who.int/trade/glossary/en/ Wills, J., Datta, K., Evans, Y., Herbert, J., May, J., & McIlwaine, C. (2010). Global Cities at Work. New Migrant Divisions of Labour. London: Pluto. Wilson, W. J. (1987). The truly disadvantaged: The inner city, the underclass, and public policy. Chicago: University of Chicago Press. Wilson, W. J. (1996). When work disappears: The world of the new urban poor. New York: Knopf. Wirth, L. (1928). The ghetto. Chicago: The University of Chicago Press. Wong, D. D. (2004). Changing local segregation of selected U.S. metropolitan areas between 1980 and 2000 (Rep.). Retrieved http://gesg.gmu.edu/seg/change_locseg.pdf Wong, D. W. (1993). Spatial indices of segregation. Urban Studies, 30(3), 559-572. doi: 10.1080/00420989320080551 Wong, D. W. (1998). Measuring multi-ethnic spatial segregation. Urban Geography, 19(1), 77-87. Wong, D. W. (2002). Modeling local segregation: A spatial interaction approach. Geographical and Environmental Modelling, 6(1), 81-97. doi: 10.1080/13615930220127305 Wong, D. W. (2003). Implementing spatial segregation measures in GIS. Computers, Environment and Urban Systems, 27(1), 53-70. doi: 10.1016/S0198-9715(01)00018-7 Wong, D. W. (2005). Formulating a general spatial segregation measure. The Professional Geographer, 57(2), 285-294. doi: 10.1111/j.0033- 0124.2005.00478.x Wong, D. (2009). The modifiable areal unit problem (MAUP). In A. Fotheringham, & P. Rogerson (Eds.), The SAGE handbook of spatial analysis. (pp. 105- 125). London: SAGE Publications Ltd. doi: http://dx.doi.org/10.4135/9780857020130.n7 Xu, J., Yeh, A., & Wu, F. (2009). Land commodification: Mew land development and politics in China since the late 1990s. International Journal of Urban and Regional Research, 33(4), 890-913.
References and bibliography | 345
Yeung, H. W. (2002). The limits to globalization theory: A geographic perspective on global economic change. Economic Geography, 78(3), 285. doi: 10.2307/4140811 Zlotnik, H. (1993). South-to-north migration since 1960: The view from the south. In General Population Conference, Montreal 1993 (pp. 3-14). Liege: International Union for the Scientific Study of Population.